Exercise interventions for adults with cancer receiving radiation therapy alone

Abstract

Background

Radiation therapy (RT) is given to about half of all people with cancer. RT alone is used to treat various cancers at different stages. Although it is a local treatment, systemic symptoms may occur. Cancer‐ or treatment‐related side effects can lead to a reduction in physical activity, physical performance, and quality of life (QoL). The literature suggests that physical exercise can reduce the risk of various side effects of cancer and cancer treatments, cancer‐specific mortality, recurrence of cancer, and all‐cause mortality.

Objectives

To evaluate the benefits and harms of exercise plus standard care compared with standard care alone in adults with cancer receiving RT alone.

Search methods

We searched CENTRAL, MEDLINE (Ovid), Embase (Ovid), CINAHL, conference proceedings and trial registries up to 26 October 2022.

Selection criteria

We included randomised controlled trials (RCTs) that enrolled people who were receiving RT without adjuvant systemic treatment for any type or stage of cancer. We considered any type of exercise intervention, defined as a planned, structured, repetitive, objective‐oriented physical activity programme in addition to standard care. We excluded exercise interventions that involved physiotherapy alone, relaxation programmes, and multimodal approaches that combined exercise with other non‐standard interventions such as nutritional restriction.

Data collection and analysis

We used standard Cochrane methodology and the GRADE approach for assessing the certainty of the evidence. Our primary outcome was fatigue and the secondary outcomes were QoL, physical performance, psychosocial effects, overall survival, return to work, anthropometric measurements, and adverse events.

Main results

Database searching identified 5875 records, of which 430 were duplicates. We excluded 5324 records and the remaining 121 references were assessed for eligibility. We included three two‐arm RCTs with 130 participants. Cancer types were breast and prostate cancer. Both treatment groups received the same standard care, but the exercise groups also participated in supervised exercise programmes several times per week while undergoing RT. Exercise interventions included warm‐up, treadmill walking (in addition to cycling and stretching and strengthening exercises in one study), and cool‐down.

In some analysed endpoints (fatigue, physical performance, QoL), there were baseline differences between exercise and control groups.

We were unable to pool the results of the different studies owing to substantial clinical heterogeneity.

All three studies measured fatigue. Our analyses, presented below, showed that exercise may reduce fatigue (positive SMD values signify less fatigue; low certainty).

• Standardised mean difference (SMD) 0.96, 95% confidence interval (CI) 0.27 to 1.64; 37 participants (fatigue measured with Brief Fatigue Inventory (BFI))
• SMD 2.42, 95% CI 1.71 to 3.13; 54 participants (fatigue measured with BFI)
• SMD 1.44, 95% CI 0.46 to 2.42; 21 participants (fatigue measured with revised Piper Fatigue Scale)

All three studies measured QoL, although one provided insufficient data for analysis. Our analyses, presented below, showed that exercise may have little or no effect on QoL (positive SMD values signify better QoL; low certainty).

• SMD 0.40, 95% CI −0.26 to 1.05; 37 participants (QoL measured with Functional Assessment of Cancer Therapy‐Prostate)
• SMD 0.47, 95% CI −0.40 to 1.34; 21 participants (QoL measured with World Health Organization QoL questionnaire (WHOQOL‐BREF))

All three studies measured physical performance. Our analyses of two studies, presented below, showed that exercise may improve physical performance, but we are very unsure about the results (positive SMD values signify better physical performance; very low certainty)

• SMD 1.25, 95% CI 0.54 to 1.97; 37 participants (shoulder mobility and pain measured on a visual analogue scale)
• SMD⁠⁠⁠⁠⁠⁠ 3.13 (95% CI 2.32 to 3.95; 54 participants (physical performance measured with the six‐minute walk test)

Our analyses of data from the third study showed that exercise may have little or no effect on physical performance measured with the stand‐and‐sit test, but we are very unsure about the results (SMD 0.00, 95% CI −0.86 to 0.86, positive SMD values signify better physical performance; 21 participants; very low certainty).

Two studies measured psychosocial effects. Our analyses (presented below) showed that exercise may have little or no effect on psychosocial effects, but we are very unsure about the results (positive SMD values signify better psychosocial well‐being; very low certainty).

• SMD 0.48, 95% CI −0.18 to 1.13; 37 participants (psychosocial effects measured on the WHOQOL‐BREF social subscale)
• SMD 0.29, 95% CI −0.57 to 1.15; 21 participants (psychosocial effects measured with the Beck Depression Inventory)

Two studies recorded adverse events related to the exercise programmes and reported no events. We estimated the certainty of the evidence as very low. No studies reported adverse events unrelated to exercise.

No studies reported the other outcomes we intended to analyse (overall survival, anthropometric measurements, return to work).

Authors' conclusions

There is little evidence on the effects of exercise interventions in people with cancer who are receiving RT alone. While all included studies reported benefits for the exercise intervention groups in all assessed outcomes, our analyses did not consistently support this evidence. There was low‐certainty evidence that exercise improved fatigue in all three studies. Regarding physical performance, our analysis showed very low‐certainty evidence of a difference favouring exercise in two studies, and very low‐certainty evidence of no difference in one study. We found very low‐certainty evidence of little or no difference between the effects of exercise and no exercise on quality of life or psychosocial effects. We downgraded the certainty of the evidence for possible outcome reporting bias, imprecision due to small sample sizes in a small number of studies, and indirectness of outcomes.

In summary, exercise may have some beneficial outcomes in people with cancer who are receiving RT alone, but the evidence supporting this statement is of low certainty. There is a need for high‐quality research on this topic.

Plain language summary

Exercise interventions for adults with cancer who are receiving radiation therapy without additional cancer therapy

What is radiation therapy?

Radiation therapy (also called radiotherapy) is a treatment that delivers high doses of radiation to a specific part of the body to kill cancer cells. One in two people with cancer will undergo radiation therapy. Some people receive radiation therapy alone, while others receive radiation therapy combined with other cancer treatments that affect the whole body (chemotherapy, immunotherapy, or hormone therapy). The unwanted effects of radiation therapy usually affect the part of the body where the radiation is delivered, but there may also be symptoms that affect the whole body. These unwanted effects can lead to reduced physical activity, physical performance, and quality of life. There is evidence that people with cancer who perform exercise may be less likely to die from cancer or from other causes, may be less likely to have their cancer return, and may have fewer unwanted effects of cancer treatment. 

What did we want to find out?

We wanted to find out if exercise could help to improve the following outcomes in people with cancer receiving radiation therapy alone.

• Fatigue
• Quality of life
• Physical performance
• Psychosocial effects (such as depression)
• Overall survival
• Return to work
• Anthropometric measurements (such as weight)
• Unwanted effects

What did we do?

We searched electronic medical literature databases for randomised controlled trials (RCTs) that enrolled people with all types and stages of cancer who were receiving RT alone. Eligible RCTs randomly assigned some participants to receive any type of exercise intervention plus standard care and others to standard care alone. We excluded exercise interventions that involved physiotherapy alone, relaxation programmes, or combination programmes with exercise and, for example, dietary restrictions.

We compared the results of the studies and rated our confidence in the evidence, based on factors such as study methods and sizes.

What did we find?

We included three studies that enrolled 130 people with breast or prostate cancer. The exercise groups participated in a supervised exercise programme three to five times per week for five to eight weeks. The exercise interventions included warm‐up, aerobic exercise, and cool‐down.

We analysed the differences between the exercise groups and control groups in the outcome values after radiation therapy. We could not compare the differences between the groups in the change in outcome values from before to after radiotherapy because the studies did not provide enough information for this comparison. In some outcomes (fatigue, physical performance, quality of life), there were already differences between the exercise and control groups at the beginning of the studies.

Exercise may improve fatigue and may have little or no effect on quality of life. Exercise may improve physical performance, but we are very uncertain about the results. Exercise may have little or no effect on psychosocial effects, but we are very uncertain about the results. Two studies reported no unwanted effects of exercise. No studies measured our other outcomes of interest. 

Exercise programmes in people with cancer receiving RT alone may provide some benefits, but the evidence to support this is poor. Due to the lack of evidence, we could not detect and also not rule out clear differences in outcomes.

What are the limitations of the evidence?

We have little or very little confidence in the evidence because the results are based on a small number of studies that enrolled very few people, because the people in two studies knew which group they were in, and because the evidence focused on a specific population whereas the question we wanted to answer was broader. Further research is likely to change our results.

How up to date is the evidence?

The evidence is up to date to 26 October 2022. 

Summary of findings

Summary of findings 1. Exercise intervention compared to no exercise intervention for people receiving radiation therapy.

Population: adults with breast or prostate cancer receiving radiation therapy alone Settings: medical centre Intervention: exercise intervention Comparison: no exercise intervention
Outcome	Narrative synthesis	No of participants (studies)	Certainty of the evidence	Comments
Fatigue Measured at 5‐8 weeks (short term); end scores. Positive SMD values signify less fatigue.	There is evidence of a difference between groups favouring exercise:  SMD 0.96, 95% CI 0.27 to 1.64; 37 participants SMD 2.42, 95% CI 1.71 to 3.13; 54 participants SMD 1.44, 95% CI 0.46 to 2.42; 21 participants	112 (3 studies)	⊕⊕⊝⊝ Lowa,b	We did not combine data from different studies owing to clinical heterogeneity.
Quality of life Measured at 5‐8 weeks (short term); end scores. Positive SMD values signify better quality of life.	There is evidence of little or no difference between groups: SMD 0.4, 95% CI −0.26 to 1.05; 37 participants SMD 0.47, 95% CI −0.40 to 1.34; 21 participants	112 (3 studies)	⊕⊕⊝⊝ Lowa,b	We did not combine data from different studies owing to clinical heterogeneity. In Kulkarni 2013, the difference was not estimable due to lack of data.
Physical performance Measured at 5‐8 weeks (short term); end scores. Positive SMD values signify better physical performance.	Evidence from 2 studies favours exercise; 1 study shows no difference between groups:  SMD 1.25, 95% CI 0.54 to 1.97; 37 participants SMD 3.13, 95% CI 2.32 to 3.95; 54 participants SMD 0.0, 95% CI −0.86 to 0.86; 21 participants	112 (3 studies)	⊕⊝⊝⊝ Very lowa,b,c	We did not combine data from different studies owing to clinical heterogeneity. Because of the different endpoints in the studies, between‐group comparability is limited.
Psychosocial effects Measured at 5‐8 weeks (short term); end scores. Positive SMD values signify better psychosocial effects.	There is evidence of little or no difference between groups: SMD 0.48, 95% CI −0.18, to 1.13; 37 participants SMD 0.29, 95% CI −0.57 to 1.15; 21 participants	58 (2 studies)	⊕⊝⊝⊝ Very lowa,d	We did not combine data from different studies owing to clinical heterogeneity. We were unable to analyse data from Kulkarni 2013 owing to poor presentation of results.
Overall survival	See comment	See comment	See comment	No studies measured overall survival.
Return to work	See comment	See comment	See comment	No studies measured return to work.
Adverse events	2 studies reported "no exercise‐related adverse events"	45 (2 studies)	⊕⊝⊝⊝ Very lowa,d	The studies only reported exercise‐related adverse events in the intervention groups.
CI: confidence interval; SMD: standardised mean difference.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

^a Downgraded one level for risk of bias concerns (participants and personnel administering the exercise intervention were not blinded; possible outcome reporting bias due to missing study protocols).
^b Downgraded one level for serious imprecision due to small sample sizes in few studies. 
^c Downgraded one level for indirectness of outcomes.
^d Downgraded two levels for very serious imprecision due to very small sample size in few studies. 

Background

Description of the condition

Cancer is a common cause of death worldwide in both high‐ and low‐income countries. The incidence of cancer is increasing as the global population grows and ages. In low‐income countries in particular, the adoption of lifestyle behaviours that are established risk factors for cancer, such as smoking, obesity, lack of exercise, and changing reproductive patterns (lower parity and later maternal age at birth), together with urbanisation and economic development, has further increased the cancer rate (Torre 2015).

Surgery, irradiation, and chemotherapy are considered the three pillars of cancer treatment. Cancer immunotherapy (use of antibodies, small molecules, cells, and viruses that stimulate the host's immune system to attack and destroy tumour cells) is becoming more common and could be considered the fourth pillar of cancer treatment (Smyth 2017).

Half of people with cancer receive radiation therapy (RT) at some point during the course of their disease. In contrast to drug‐based systemic chemotherapy or immunotherapy, which affects the whole body, RT is a localised treatment where the tumour‐destroying effect is focused on a specific area, called the radiation field. Stereotactic RT delivers high doses of radiation to a very precise irradiation field. RT can be delivered alone or in combination with systemic therapies like chemotherapy (chemoradiation or chemoradiotherapy), immuno‐ or hormone therapy. It can also be delivered before surgery (neoadjuvant RT) or after surgery (adjuvant RT). An additional use of RT is to relieve or prevent cancer‐specific symptoms.

Conventional or stereotactic RT alone constitutes a treatment option for various cancers at different stages, including breast cancer, prostate cancer, early stages of lung cancer and Hodgkin's lymphoma, sarcomas, hepatocellular and renal cell carcinomas, endometrial cancer, cancer of salivary glands, non‐melanoma skin cancer, uveal/choroidal melanomas, and meningioma. It is also used to treat local metastases. Furthermore, RT can reduce tumour burden and relieve pain in advanced disease in the palliative setting. People with a contraindication for chemotherapy (e.g. comorbidities) or who refuse systemic treatment methods often receive RT alone.

RT destroys cancer cells through ionising radiation or particle radiation. The radiation damages the DNA of the cells so that they stop dividing or die. Because it is a local treatment, toxic effects predominantly occur in the radiation field; however, systemic symptoms such as fatigue, nausea, loss of appetite, fever, and weakness may also occur, mainly because the body is exposed to large amounts of cell debris and degradation products (Schmoll 2006; Stöver 2018). The systemic and local adverse events that arise from RT can lead to exhaustion, weight loss, and physical deconditioning. The term cancer‐related fatigue (CRF) covers the physical, emotional, and cognitive fatigue or exhaustion associated with cancer or cancer treatment (Berger 2015). Between 60% and 90% of people with cancer experience CRF during or after cancer treatment, and more than 30% in their first year after diagnosis (Wagner 2004; Weis 2015). Many people with breast cancer receive RT alone, and impairment of lung function due to RT could prevent them from performing physical activity. Up to 30% of all people with breast cancer receiving thoracic RT can develop radiation‐induced pneumonitis as a subacute treatment‐associated toxicity, and they are at high risk of developing radiation‐induced lung fibrosis as late toxicity (Käsmann 2020; Keffer 2019). Treatment‐related death is uncommon, and the incidence is strongly related to patient characteristics such as age and BMI, and treatment characteristics such as mean lung dose, irradiated volume, and radiation technique (Chao 2017; Kahán 2007; Lee 2015).

All these cancer‐ or treatment‐related side effects – including cognitive impairment, sleep disorders, depression, pain, anxiety, and physical disorders – can lead to reduced physical activity, physical performance, and quality of life (QoL) in people with cancer.

Description of the intervention

Exercise is planned, structured, repetitive, and objective‐oriented physical activity in the sense that improving or maintaining one or more components of physical fitness is an intended goal (Caspersen 1985). According to the World Health Organization (WHO), physical activity is one of nine alterable risk factors for cancer. Low physical activity levels, together with obesity, are associated with 20% to 33% of all colorectal, breast, kidney, and gastrointestinal cancers (WHO 2009).

The outdated belief that exercise during cancer therapy could be harmful and that people with cancer should rest until complete remission has no scientific basis (Andrykowski 1989; Dimeo 1996). The current literature shows increased survival in people with breast, colon, or colorectal cancer who have higher exercise levels (Barbaric 2010). In one study that enrolled people with colon and colorectal cancer, exercise improved survival rates when combined with standard cancer treatment (Meyerhardt 2006).

One systematic review examined all available evidence on the role of exercise in the management of cancer, without distinguishing between cancer site or stage/grade (Cormie 2017). It demonstrated a sustained trend towards reduced risk of cancer‐specific mortality, recurrence of cancer, and all‐cause mortality in participants with higher levels of exercise after a diagnosis of cancer. The included studies found an association between exercise levels and cancer mortality in participants with several types of cancer, including breast, colorectal, and prostate cancer.

Preliminary evidence suggests that exercise is safe and beneficial and can improve the management of various adverse effects of cancer and cancer treatments (Lipsett 2017). One narrative review evaluated the effects of an exercise programme to alleviate treatment‐related side effects in people with cancer undergoing RT among other cancer treatments, finding benefits in people with breast, prostate, rectal, lung, head, and neck cancer (Piraux 2020).

Research suggests that aerobic exercise, resistance training, and mindful forms of physical activity (e.g. yoga or tai chi) are effective for helping people with cancer manage their disease (Mustian 2012).

How the intervention might work

Physically active people compared with physically inactive people may have a 20% to 40% lower risk of developing several cancer types (Parent 2011). Cancer is a complex disease, and various mechanisms related to the site and stage of cancer, and to the type and quality of the exercise performed, can influence cancer outcomes. Animal models and human epidemiological studies suggest that exercise prior, during, and after cancer treatment provides positive outcomes, and that regular exercise might reduce the risk of developing cancer. The growing use of exercise in people undergoing cancer therapies has shown promising results (improved cardiorespiratory fitness, muscle strength, and physical functioning; Mishra 2014).

Researchers have proposed different biological and immunological mechanisms to explain the protective effect of exercise on cancer outcomes. Alterations in the secretion of sex hormones or insulin and insulin‐like growth factor have a direct effect on the tumour and its biology, and modify the immune system and the whole body composition (Li 2010; Westerlind 2003). Exercise is associated with an increased number of anti‐inflammatory, interleukin‐10 (IL‐10)‐producing T‐lymphocytes (Weinhold 2016), and could contribute to a benefit in primary prevention and relapse of cancer. Exercise may normalise the tumour microenvironment, leading to an increase in the transport of systemic therapies to the cancer cells (Pedersen 2015), and a change of the inflammatory status (Zeng 2012).

There are different hypotheses regarding how exercise can influence survival in different cancer types. In breast cancer, exercise may reduce lifetime oestrogen exposure and lower blood oestrogen levels (reducing proliferative activity in breast tissue), improve immune function, decrease insulin resistance, lower fat tissue levels, and reduce bodyweight (Abrahamson 2006; Irwin 2005). Exercise may reduce colon and colorectal cancer mortality by shortening gastrointestinal transit time, altering prostaglandin levels and tumour growth‐related hormone pathways (e.g. insulin or insulin‐like growth factors), and increasing adipokines such as adiponectin and leptin (hormones secreted by adipose tissue), all of which can affect a person's inflammatory status (Haydon 2006; Meyerhardt 2006; Sandhu 2002; Schoen 1999; Stansfield 2014; Westerlind 2003).

Exercise is associated with reduced incidence of and reduced mortality from several diseases and comorbidities, such as diabetes, hypertension, cardiovascular diseases, obesity, depression, and osteoporosis (Cormie 2017; Courneya 2007; Pedersen 2006; Warburton 2006). Moreover, a higher level of fitness has been associated with better surgical outcomes and fewer complications (Carli 2005; Moran 2016).

The positive effects of exercise during cancer treatment are demonstrated improvement in CRF, cognitive impairment, sleep problems, depression, pain, anxiety, and physical dysfunction, including muscular function, cardiopulmonary function, and bone density (Mustian 2012; Mustian 2017; Velthuis 2010). Exercise can improve people's psychological and physical tolerance of treatment regimens, leading to higher rates of treatment completion (van Waart 2015). Exercise can also improve their ability to manage activities of daily life and return to previous routines, such as work (Baumann 2012; Colemann 2012; Courneya 2009; Hwang 2008). It is an effective therapeutic intervention to prepare people for the successful completion of treatments; to reduce acute, chronic, and late side effects; and to improve QoL during and after treatment. In this way, exercise directly and indirectly benefits overall survival.

Why it is important to do this review

There are published systematic reviews about the effects of exercise in people with cancer who have undergone different treatment approaches (Baumann 2018; Galvão 2005; Lahart 2018; Mustian 2017; Piraux 2020; Thorsen 2008). There are also reviews on exercise in the context of selective treatments, including chemotherapy, surgery, and hormone therapy (Cavalheri 2019; Cave 2018; Chang 2016; Segal 2003). Most available reviews on cancer and exercise focus on one type of cancer or do not distinguish between different cancer treatments. In studies that combine people who receive different treatments (systemic, local, or both), it is difficult to determine which particular therapy benefits most from the addition of exercise. It is also difficult to identify the supportive treatments that most effectively address specific treatment‐related toxicities. Because RT is a local treatment method with its own spectrum of local and systemic side effects, it may not be directly comparable to other cancer therapies. Although most people with cancer receive multimodal treatment, there are indications for RT alone, as described in Description of the condition. To date, there have been no reviews focusing on the effects of exercise on adverse events in people with cancer who undergo RT alone.

The aim of this review was to determine whether exercise in addition to standard care, compared to standard care alone, affects CRF, anthropometric outcomes; systemic or local side effects, or both; overall survival; well‐being; return to work; and psychosocial outcomes in people receiving RT without adjuvant systemic cancer therapy. It should help us determine whether an exercise intervention is beneficial in this population, clarify the most effective type and intensity of exercise intervention, and reach conclusions on feasibility, safety, and efficacy. The results of this Cochrane Review should help to inform future guidelines of cancer and cancer therapies.

Objectives

To evaluate the benefits and harms of exercise plus standard care compared with standard care alone in adults with cancer receiving RT alone.

Methods

Criteria for considering studies for this review

Types of studies

We included randomised controlled trials (RCTs), published in any language.

Types of participants

Eligible RCTs enrolled people with cancer receiving RT of any type. Trial participants had to be at least 18 years old at the time of therapy and intervention. We placed no restrictions on sex or ethnicity. We considered all types and stages of cancer. We excluded RCTs that enrolled people who were receiving chemotherapy, hormone therapy, or immunotherapy. Surgical treatment was not an exclusion criterion. We excluded RCTs where participants received any nutrition restriction or diet intervention.

Types of interventions

We defined exercise as a medical intervention involving physical activity that is generated by skeletal muscles and leads to expenditure of energy. 

We considered any type of exercise intervention, but excluded physiotherapy alone and relaxation programmes. We also excluded multimodal approaches that combined exercise with other non‐standard interventions such as nutritional restriction.

After the diagnosis of cancer, the exercise intervention could take place before, during, or after RT. We considered exercise intervention before the start of RT as prehabilitation. Exercise programmes could be supervised or non‐supervised. We only included studies that provided exercise programmes in addition to standard care, where this refers to the accepted and routine care offered to each person for their specific medical condition in the specific study centre

Types of outcome measures

We considered all trials that met our inclusion criteria, irrespective of outcomes reported.

Primary outcomes

Fatigue, measured by a validated questionnaire such as the Multidimensional Fatigue Inventory (MFI 20; Smets 1996), Functional Assessment of Cancer Therapy‐Fatigue (FACT‐F; Cella 2002), Brief Fatigue Inventory (BFI; Mendoza 1999), or Piper Fatigue Scale (PFS; Reeve 2012).

Secondary outcomes

• QoL (cancer‐specific QoL/cancer site‐specific QoL/health‐related QoL) measured with validated instruments such as the European Organisation for Research and Treatment of Cancer (EORTC) questionnaires (Aaronson 1993)

• Physical performance (e.g. oxygen system, cardiorespiratory fitness, muscle strength, physical functioning)

• Psychosocial effects (e.g. depression, anxiety level)

• Overall survival from randomisation to death or censoring

• Return to work

• Anthropometric measurements (e.g. weight, body mass index)

• Adverse events

Search methods for identification of studies

Electronic searches

To reduce possible language bias, we applied no language restrictions to the searches. For original texts in non‐English languages (Persian and Portuguese in the case of this review), we translated the texts with the help of bilingual native speakers. We did not restrict the searches by publication date or publication status (e.g. abstract, conference proceedings, unpublished data, dissertations, etc).

We searched the following databases:

• Cochrane Central Register of Controlled Trials (CENTRAL; 2022, Issue 10) in the Cochrane Library (searched 26 October 2022; Appendix 1)

• MEDLINE Ovid (1946 to 26 October 2022; Appendix 2)

• Embase Ovid (1980 to 26 October 2022; Appendix 3)

• CINAHL (1961 to 26 October 2022; Appendix 4)

We designed the initial search strategy for MEDLINE as suggested in the Cochrane Handbook for Systematic Reviews of Interventions, then adapted it to the other databases (Higgins 2022).

Searching other resources

We searched the following trials registries.

• EU Clinical Trials Register (clinicaltrialsregister.eu/ctr-search/search)

• WHO International Clinical Trials Registry Platform (who.int/trialsearch)

• ClinicalTrials.gov (clinicaltrials.gov)

• ISRCTN Registry (isrctn.com)

• metaRegister of Controlled Trials (mRCT; isrctn.com/page/mrct)

We also searched conference proceedings of annual meetings of the following societies for abstracts, if not included in CENTRAL (2010 to October 2022).

• European Society for Radiotherapy and Oncology (estro.org)

• American Society for Radiation Oncology (astro.org)

• American Medical Society for Sports Medicine (amssm.org)

We checked the references of all identified trials and relevant review articles for further literature. In addition, we handsearched trial registries and abstract books of annual conferences of the major societies of radiology, oncology, and sports medicine. 

Data collection and analysis

Selection of studies

We imported all the records from the searches into Covidence 2021. Two review authors (TN, RR) independently screened the titles and abstracts of all records and classified them as "eligible/potentially eligible" or "ineligible". We obtained the full‐text articles of all records that at least one review author had considered eligible or potentially eligible (Higgins 2022). If the full texts were unavailable, we contacted the study authors where possible by telephone or email. Two review authors (TN, RR) independently assessed the full‐text reports against our eligibility criteria, resolving any disagreements by discussion or, if necessary, by involving a third review author (MT). We documented the selection process in sufficient detail to complete a PRISMA flow chart (Page 2021; Figure 1).

1.

PRISMA flow diagram summarising the study selection process.

Data extraction and management

Two review authors (TN, RR) independently extracted study characteristics and outcome data from the included studies using a piloted data collection form (see Table 2) in Covidence 2021 according to Cochrane guidance (Higgins 2022). We resolved disagreements by discussion or by involving a third review author (MT). One review author (MT) transferred data into the statistical software Review Manager 5 (Review Manager 2020).

1. Template data extraction form.

Study Identification

Sponsorship source

Country

Setting

Comments

Author name

Institution

Email

Address

Methods

Design

Group

Population

Inclusion criteria

Exclusion criteria

Group differences

Interventions

Instruction

Specific intervention

Intervention details

Intensity

Timing

Frequency

Outcomes

Quality of life

Overall health

Fatigue

Physical well‐being

Psychological well‐being

Social well‐being

Environmental well‐being

Emotional well‐being

Functional well‐being

Strength

Flexibility

Depression

Exercise capacity

Maximal oxygen consumption

Metabolic equivalents (METS)

Pain

Forward flexion

Abduction

Internal rotation

External rotation

Relationship with physician

Prostate cancer symptoms

We double‐checked that data had been entered correctly by comparing the data presented in the systematic review with the study reports. A fourth review author (FTB) spot‐checked the accuracy of the study characteristics against the trial reports. We extracted the following items where possible.

• General information: author, title, source, publication date, country, language, duplicate publications

• Study characteristics: trial design, aims, setting and dates, source of participants, inclusion and exclusion criteria, comparability of groups, subgroup analysis, statistical methods, power calculations, treatment cross‐overs, compliance with assigned treatment, length of follow‐up, time point of randomisation

• Participant characteristics: underlying disease; stage of disease; age; sex; ethnicity; number of participants recruited, allocated, evaluated, and lost to follow‐up; type and concept of RT

• Interventions: type, duration, and intensity of exercise intervention; standard care; duration of follow‐up

• Outcomes: overall survival, mortality, QoL, fatigue, physical performance (e.g. oxygen system, cardiorespiratory fitness, muscle strength, physical functioning), physical activity level, anthropometric measurements (e.g. weight, body mass index), adverse events, drop‐outs, return to work, psychosocial effects (e.g. depression, anxiety level)

We extracted data from trials reported in more than one publication into a single data extraction form. If these sources did not provide sufficient information, we contacted the study authors for additional details.

Assessment of risk of bias in included studies

Two review authors (TN, RR) independently assessed the risk of bias for each study in Covidence 2021 using the following criteria, as outlined in Chapter 8 of the Cochrane Handbook for Systematic Reviews of Intervention (Higgins 2011; Higgins 2017).

• Sequence generation

• Allocation concealment

• Blinding (participants, personnel, outcome assessors)

• Incomplete outcome data

• Selective outcome reporting

• Other potential sources of bias

For every criterion, we made one of the following three judgements.

• Low risk: if the study adequately fulfilled the criterion

• High risk: if the study did not fulfil the criterion

• Unclear: if the study report did not provide sufficient information to enable a judgement

We resolved any disagreements by discussion. We summarised the results of this analysis in the risk of bias graph and summary (Figure 2 and Figure 3).

2.

Review authors' judgements about each risk of bias item presented as percentages across all included studies.

3.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Measures of treatment effect

We uploaded the outcome data for each study to Review Manager 5 to calculate treatment effects. We had planned to calculate continuous outcomes as mean differences (MDs) with 95% confidence intervals (CIs) where studies measured outcomes on the same scale, or standardised mean differences (SMDs) with 95% CIs where studies reporting the same outcome used different scales.

Because no studies reported dichotomous data for any of our outcomes, we did not estimate treatment effect measures as risk ratios (RRs) with 95% CIs. As no studies reported survival data, we could not estimate treatment effects by extracting hazard ratios (HRs) of individual studies and analyse these using the methods described in Parmar 1998 and Tierney 2007.

Unit of analysis issues

As recommended in Chapter 16.5.4 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011), We could not combine arms of studies with multiple treatment groups as subtypes of the same intervention. We could not conduct pairwise meta‐analysis.

Dealing with missing data

There are many potential sources of missing data, listed in Chapter 16 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011), at a study level, outcome level, and summary data level. We contacted the study investigators to request missing numerical outcome data and further information on key study characteristics. Because we received no responses, we made explicit our assumptions for any methods used; for example, that the data were assumed to be missing at random or that missing values were assumed to have a particular value. We addressed the potential impact of missing data on the findings of the review in the Discussion. If numerical outcome data such as standard deviations (SDs) were missing, we calculated them from other available statistics. When specific numbers were missing from the text, we measured them using the scales in the figures in the study report.

Assessment of heterogeneity

Had we conducted meta‐analyses to assess heterogeneity of treatment effects between trials, we could have explored potential causes of heterogeneity through sensitivity and subgroup analyses (Higgins 2022).

Assessment of reporting biases

We assessed any potential reporting bias by investigating unpublished studies (Characteristics of studies awaiting classification). Had we included more than 10 trials in a meta‐analysis, we would have explored potential reporting bias by generating a funnel plot and performing a linear regression test (Higgins 2022).

Data synthesis

Owing to substantial clinical heterogeneity between trials (various types of disease), we did not pool results in a fixed‐effect meta‐analysis or use the random‐effects model for sensitivity analysis of the primary outcome. Instead, we analysed the data from each study without calculating an overall estimate. We performed analyses with the statistical software Review Manager 5 and according to guidance provided in Chapter 9 of the Cochrane Handbook for Systematic Reviews of Interventions (Review Manager 2020; Deeks 2011).

Subgroup analysis and investigation of heterogeneity

Meta‐analysis and subgroup analyses were not possible.

Sensitivity analysis

If the studies had been sufficiently similar, we would have pooled the results and used the random‐effects model as a sensitivity analysis for the primary outcome.

Summary of findings and assessment of the certainty of the evidence

We created a summary of findings table based on the methods described the Cochrane Handbook for Systematic Reviews of Interventions and using GRADEpro software (Higgins 2022; GRADEpro). We included the following outcomes.

• Fatigue

• QoL

• Physical performance

• Psychosocial effects

• Overall survival

• Return to work

• Adverse events

 We chose this order of priority according to the existing data on the topic of 'Physical Activity and Cancer' or 'Exercise and Cancer'. Several studies have investigated fatigue, physical performance, and adverse events in people with cancer after an exercise intervention.

We presented the overall certainty of the evidence for each outcome listed in Types of outcome measures according to the GRADE approach, which takes into account issues related to internal validity (risk of bias, inconsistency, imprecision, publication bias) and external validity (directness of results; Langendam 2013; Schünemann 2020). We referred to the GRADE checklist and GRADE Working Group certainty of evidence definitions and the GRADE Working Group grades of evidence (Meader 2014). We downgraded the certainty of the evidence from 'high' by one level for serious (or by two for very serious) concerns for each limitation, and interpreted the resulting grade as follows.

• High certainty: we are very confident that the true effect lies close to that of the estimate of the effect.

• Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.

• Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect.

• Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

Results

Description of studies

Results of the search

The literature search recovered 5875 references, of which 430 were duplicates. In the title and abstract screening, we selected 121 records for the full‐text assessment based on our eligibility criteria. After this process, we included three studies (Hwang 2008; Kulkarni 2013; Monga 2007). We found one study protocol from the UK in the ISRCTN Registry that would have met our inclusion criteria, but no results have been published to date (ISRCTN26140710). We were unable to obtain a full‐text article even after contacting an investigator. In another study, it was unclear how many people were enrolled and whether they were truly randomised (Milecki 2013). We excluded 118 studies during the full‐text review and recorded the reasons for exclusion in the Characteristics of excluded studies table. Figure 1 summarises the study selection process as recommended in the PRISMA statement (Page 2021).

Included studies

All three included studies were two‐arm RCTs (Hwang 2008; Kulkarni 2013; Monga 2007). In total, they enrolled 130 individuals with either a breast or prostate cancer diagnosis. The exercise groups participated in an exercise programme several times per week for several weeks. All three studies included treadmill walking in the exercise intervention. Table 3 provides a tabular overview of the inclusion and exclusion criteria of the included studies. Table 4 shows the characteristics of the included studies.

2. Inclusion and exclusion criteria of included studies.

Study	Inclusion criteria	Exclusion criteria
Monga 2007	• Localised prostate cancer • First time cancer diagnosis • Ambulatory • Able to complete self‐report measures	• Concurrent chemotherapy • Major health problems (uncontrolled hypertension (seated systolic blood pressure < 160 mmHg or seated diastolic blood pressure < 90 mmHg) uncontrolled insulin‐dependent diabetes mellitus, severe arthritis, or obvious cognitive dysfunction) • Recent history of sudden onset of shortness of breath on exertion or recent history of dizziness, blurred vision, or fainting spells • Recent history of unstable angina, coronary artery disease, myocardial infarction, or cardiac failure • Bone, back, or neck pain of recent origin, or inability to exercise
Hwang 2008	• Female sex • Outpatient waiting list for radiotherapy for breast cancer	• Concurrent major health problems, including: ‐ uncontrolled hypertension; ‐ cardiovascular disease; ‐ acute or chronic respiratory disease; and ‐ cognitive dysfunction.
Kulkarni 2013	• Female sex • Age 30–60 years • Clinical diagnosis of stage I or II breast cancer • Unilateral modified radical mastectomy • Radiotherapy	• Any cardiovascular abnormality • Impaired cognitive function • Musculoskeletal disorders • Neurological disorders • Emotional instability • Very low level of activity

3. Characteristics of included studies.

ID

Title

Methods

Participants

Recruitment

Interventions

Outcomes (assessment tool)

Quality assessment

Study design

Statistical methods

Follow‐up duration

Cancer type and stage

Sex (n)

Mean age (years)

Inclusion criteria

Exclusion criteria

Type and concept of RT

Intervention group

Control group

Blinding (participants, personnel, outcome assessors)

Other potential sources of bias

Monga 2007

Exercise prevents fatigue and improves quality of life in prostate cancer patients undergoing radiotherapy

Parallel‐ group, 2‐arm RCT

SASb software:  • Rank‐sum or 2‐sample t tests • Univariate or multivariate ANOVA • Paired‐difference and 2‐sample t tests (equivalent to univariate ANOVA) • Nonparametric Wilcoxon rank‐sum and signed‐rank tests P ≤ 0.05 considered significant for all comparisons.

8 weeks

Localized prostate cancer

Only male (30)

Intervention group: 68 (SD 4.2); control group: 70.6 (SD 5.3); overall: 69.24 (SD 4.82)

• Localized prostate cancer • Only first‐time cancer diagnosis • Ambulatory • Able to complete self‐report measures

• Concurrent chemotherapy  • Major health problems (uncontrolled hypertension, uncontrolled insulin‐dependent diabetes mellitus, severe arthritis, and obvious cognitive dysfunction)  • Recent history of sudden onset of shortness of breath on exertion or recent history of dizziness, blurred vision, or fainting spells • Recent history of unstable angina, coronary artery disease, myocardial infarction, or cardiac failure •Bone, back, or neck pain of recent origin, or inability to exercise

Over a 2‐year period, participants referred to the radiotherapy service at the Houston Veterans Affairs Medical Center for radiation treatment of localised prostate cancer were approached for participation

External beam radiotherapy: all participants received radiotherapy for 7–8 weeks from a Varian 2100 linear accelerator with 18‐MV photon beams. Each participant received 68 to 70Gy in 34 to 38 fractions at 1.80 to 2.0Gy per fraction.

3 times/week for 8 weeks: aerobic (cardiovascular conditioning) exercise program with 10‐minute warm‐up, 30‐minute aerobic segment (treadmill walking), and 5–10‐minute cool‐down period

Standard care that included patient education and radiotherapy without exercise prescription

• Cardiovascular fitness (Bruce treadmill test + METS + target HR) • Flexibility (modified sit‐and‐reach test) • Strength (stand‐and‐sit test) • Fatigue (PFS) • QoL (FACT‐P) • Prostate cancer symptoms • Depression (BDI)

Control group was aware of the exercise arm of the study.

Transcription error between tables (3 and 4/5) for "FACT‐P" and "Prostate cancer symptoms".

Hwang 2008

Effects of supervised exercise therapy in patients receiving radiotherapy for breast cancer

Parallel‐ group, 2‐arm RCT

SPSS version 10.0 software (SPSS Inc., Chicago, IL, USA): • Independent‐samples t tests • ANCOVA in which groups were compared according to follow‐up data with baseline data as the covariate P value < 0.05 taken as significant.

5 weeks

Breast cancer

Only female (40)

46.3 (SD 8.52)

• Women on outpatient waiting list for radiotherapy for breast cancer

• Concurrent major health problems • Uncontrolled hypertension • Cardiovascular disease • Acute or chronic respiratory disease • Cognitive dysfunction.

Consecutive unselected women on the outpatient waiting list for radiotherapy for breast cancer were approached at their first planned visit.

Participants were irradiated with a dose of 50 Gy during 5 weeks with a dose per fraction of 2 Gy.

3 times/week for 5 weeks: supervised exercise program with 10‐minute warm‐up, 30 minutes of exercise (shoulder stretching exercises, aerobic exercise, and strengthening exercise), and a 10‐minute cool‐down (relaxation period)

Standard care that included showing participants how to perform shoulder ROM exercises and encouraging them to continue with normal activities.

• QoL (overall, psychological, environmental, social, physical; WHOQOL‐BREF) • Fatigue (BFI) • ROM of shoulder (forward flexion, abduction, and internal rotation and external rotation) • Pain (VAS) • Adverse events

—

—

Kulkarni 2013

A randomized controlled trial of the effectiveness of aerobic training for patients with breast cancer undergoing radiotherapy

Parallel‐ group, 2‐arm RCT

Not reported.

0–6 weeks

Breast cancer, stage I or II

Only female (60)

45.59 (SD 7.33)

• Age 30–60 years • Clinical diagnosis of stage I or II breast cancer • Unilateral modified radical mastectomy •Radiotherapy

•Any cardiovascular abnormality •Impaired cognitive function • Musculoskeletal disorders •Neurological disorders •Emotional instability •Very low level of activity

Not reported

Not reported

5 times/week for 6 weeks: aerobic training combined with conventional physiotherapy. Aerobic training consisted of a 10‐minute warm‐up (mild stretching of large muscle groups), 10–30 minutes treadmill walking at self‐adjusted speeds, and 10‐minute cool‐down

Standard care plus conventional physiotherapy

• Fatigue (BFI) • Exercise capacity (6‐minute walk test) • VO2max • QoL (WHOQOL‐BREF) • Adverse events

Participants were blinded to the intervention that they were allocated to. Physiotherapists were not blinded because they had to schedule the intervention sessions.

Calculation error of the SD of the average age in the results section. Bar chart for fatigue does not match the numbers given in the text.

ANCOVA: analysis of covariance; ANOVA: analysis of variance; BDI: Beck Depression Inventory; BFI: Brief Fatigue Inventory; FACT‐P: Functional Assessment of Cancer Therapy‐Prostate; HR: heart rate; METS: metabolic equivalent of task; n = number of participants; PFS: Piper Fatigue Scale; QoL: quality of life; RCT: randomised controlled trial; ROM: range of motion; SD: standard deviation; VAS: visual analogue scale; VO₂ max: maximal oxygen consumption; WHOQOL‐BREF: World Health Organization Quality of Life Questionnaire.

Study Characteristics

Hwang 2008 was conducted in South Korea and enrolled 40 women who had undergone breast cancer surgery (modified radical mastectomy (MRM), breast‐conserving surgery (BCS) with axillary lymph node dissection (ALND), or BCS with sentinel lymph node biopsy (SLNB)) and were on the outpatient waiting list for RT. They received a total dose of 50 Gy at 2 Gy per fraction over five weeks. Three participants randomised to the control group did not wish to have their follow‐up measurements taken; after these dropouts, the control group comprised 20 participants with a mean age of 46.3 (SD 9.5) years. The exercise group comprised 17 participants with a mean age of 46.3 (SD 7.5) years. The control group had slightly better baseline scores for fatigue, physical performance and QoL.

Kulkarni 2013 was conducted in India and enrolled 60 women aged 30 to 60 years who had stage I or stage II breast cancer, had undergone unilateral MRM, and were receiving RT. Four women in the control group and two in the exercise group dropped out of the study before completion. The control group comprised 26 participants with a mean age of 46.42 (SD 8.3) years, and the exercise group comprised 28 participants with a mean age of 44.82 (SD 3.4) years. Both groups had comparable baseline characteristics. The control group had slightly better baseline scores for fatigue and physical performance.

Monga 2007 was conducted in the USA and enrolled 30 individuals with localised prostate cancer who had been referred for primary RT. All participants received doses of 68 Gy to 70 Gy in 34 to 38 fractions at 1.80 Gy to 2.0 Gy per fraction over seven to eight weeks. The control group comprised 10 participants with a mean age of 70.6 (SD 5.3) years and the exercise group comprised 11 participants with a mean age of 68 (SD 4.2) years. Nine participants dropped out before the intervention started. All baseline characteristics, such as ethnicity or medical conditions (hypertension, diabetes, coronary artery disease, chronic obstructive pulmonary disease), were evenly distributed between the groups. The control group had slightly better baseline scores for fatigue, physical performance, and QoL.

Interventions

In Hwang 2008, participants in the exercise group took part in a supervised exercise programme three times per week for five weeks. Each session consisted of a 10‐minute warm‐up, 30 minutes of various exercises (shoulder stretching, walking on a treadmill, cycling, and strengthening exercises), and a 10‐minutes relaxation period. The investigators monitored heart rate (HR) to ensure participant were training at 50% to 70% of their age‐adjusted maximum HR. The participants in the control group received standard care, which included patient education, instruction on shoulder range of motion exercises, and encouragement to continue with normal activities.

Kulkarni 2013 combined conventional physiotherapy with aerobic exercise. The participants in the control group received standard care and conventional physiotherapy. Aerobic training took place under the supervision of a physiotherapist five times per week for six weeks, and each session lasted 30 to 50 minutes. Each training session began with 10 minutes of warm‐up exercises, including light stretching. Participants then walked on a treadmill without incline at a self‐regulated speed. Walking time was 10 minutes at the beginning of the programme, and was increased to 30 minutes over the six weeks of training. Lastly, during a 10‐minute cool‐down period, participants slowed their walking pace on the treadmill then performed some stretching exercises. Investigators calculated the target HR in beats per minute (bpm) for these sessions as follows.

•  Target HR = 0.4 to 0.6 x (maximum HR − resting HR) + resting HR

The maximum HR (bpm) used in this formula was 220 − age of participant in years.

In Monga 2007, the exercise programme was led by a kinesiotherapist under medical supervision. The control group received standard care, which included patient education. Exercise interventions took place in the morning before daily RT three times per week for eight weeks. A session consisted of a 10‐minute warm‐up, a 30‐minute aerobic section (walking on a treadmill), and a 5‐ to 10‐minute cool‐down. Throughout the aerobic section, participants were instructed to maintain their target HR, calculated using the following formula.

• Target HR = 0.65 × (maximum HR − resting HR) + resting HR

The investigators determined the maximum HR before the start of the exercise during a measurement of maximum oxygen consumption. Resting HR was measured weekly and target HR recalculated accordingly. 

Outcomes

All three studies assessed baseline characteristics before the start of RT and the endpoints after completion of RT. 

All three studies examined fatigue, QoL, and physical performance. Hwang 2008 and Monga 2007 reported psychosocial effects; the results for this outcome were unclear in Kulkarni 2013.

Hwang 2008 and Kulkarni 2013 assessed fatigue using the BFI (Mendoza 1999). This is a one‐page questionnaire with nine items and scores from 0 (no fatigue) to 10 (most fatigue imaginable). The simplicity of this instrument makes it very informative. It has a reported reliability of up to 0.97 (Mendoza 1999). Monga 2007 examined fatigue using the revised PFS (PFS‐Revised; Annunziata 2010), which was developed specifically for people with cancer and contains questions related to behaviour/severity, affective meaning, sensory meaning, and cognitive meaning/mood. Participants rate each sub‐aspect on a scale from 0 to 10, with higher scores representing greater fatigue (Reeve 2012).

To assess QoL, Hwang 2008 and Kulkarni 2013 used the WHO QoL questionnaire WHOQOL‐BREF (Whoqol Group 1998). This is the short form of the WHOQOL‐100, which WHO originally designed as a disease‐independent health questionnaire. It includes the four major domains of physiological health, mental health, social relationships, and environment, in addition to general health and general QoL. The domains are divided into a total of 24 sub‐aspects, which are rated from 1 to 5, with higher scores indicating better QoL. Monga 2007 examined QoL using the Functional Assessment of Cancer Therapy‐Prostate (FACT‐P; Esper 1997). This is an extension of the general version (FACT‐G), which examines physical well‐being; relationship with friends, acquaintances, and family; relationship with physicians; psychological well‐being; and functioning. FACT‐P contains a further 12 items specifically related to prostate cancer, including sexuality and bowel or bladder weaknesses (Esper 1997).

To assess physical performance, Hwang 2008 evaluated shoulder mobility and pain on a VAS, and Kulkarni 2013 calculated maximum oxygen consumption (VO₂ max) using the HR and exercise capacity measured during a six‐minute walk test (ATS Committee 2002). Monga 2007 evaluated lower extremity strength (stand‐and‐sit test), flexibility (sit‐and‐reach test), and cardiac performance (Bruce Treadmill Test). 

Monga 2007 examined psychosocial effects using the Beck Depression Inventory (BDI; Beck 1961), while Hwang 2008 examined social well‐being on the WHOQOL‐BREF social subscale.

Studies awaiting classification

We found one study protocol from the UK in the ISRCTN Registry (isrctn.com) that would have met our inclusion criteria, but no results have been published to date (ISRCTN26140710). The study protocol was submitted in September 2004 and the last revision was noted in October 2014. The overall trial start date was 1 November 2003 and the overall trial end date is noted as 1 November 2008. The investigators aimed to randomise 10 women with breast cancer to an exercise intervention or control, and analyse their results to answer the following questions.

• Do fitness levels and perceptions of QoL change during and following RT treatment for breast cancer?

• Does an individualised exercise programme during RT treatment for cancer affect fitness levels?

• Does perception of QoL change during and following RT as a result of exercise intervention?

Owing to the very small number of participants, we did not expect this study to have an important impact on our overall results.

We found one study conducted in Poland and published in 2013 (Milecki 2013). According to the abstract, it enrolled 46 people, of whom 25 participated in aerobic training. According to the baseline characteristics table, the study enrolled 66 people, of whom 35 participated in aerobic training. The report provided an adequate description of participants and the exercise intervention. All participants in this study underwent breast surgery then, four to five weeks later, received external beam radiation treatments seven days per week for five weeks. The affected breast and regional lymph nodes received a total dose of 50 Gy in a daily fraction of 2 Gy. Aerobic training consisted of a supervised exercise session five times per week for six weeks. Each session consisted of a two‐minute warm‐up, 40 minutes of cycling, and a 3‐minute relaxation period.

We did not include this study because of the discrepancy in participant numbers and because it is unclear if participants were truly randomised.

Ongoing studies

We found two study protocols in ClinicalTrials.gov (clinicaltrials.gov) that appear to meet our inclusion criteria.

NCT04507789 is a Turkish study investigating the effects of exercise interventions on upper extremity function in 40 women receiving RT to the axillary region after breast cancer surgery. The start date was October 10, 2020, and the anticipated study end date was April 10, 2021. The study authors will compare outcomes related to upper limb function between two groups and also within the groups after completion of RT.

NCT04506476 is a single‐centre, three‐arm RCT being conducted at the University Hospital of Tuebingen, Germany. The study start date was 1 August 2020 and the expected study duration is five years. The study authors aim to enrol 201 participants and investigate the benefit of activity tracker‐based exercise training in relation to CRF during adjuvant RT in women with breast cancer. They will measure QoL and intensity of fatigue using the fatigue subscale of the Functional Assessment of Chronic Illness Therapy (FACIT) questionnaire three months after completion of RT.

Excluded studies

We provided a justification for exclusion of 118 studies excluded during full‐text assessment in the Characteristics of excluded studies table. Forty‐eight studies did not have RT as an inclusion criterion, 43 did not exclude people receiving chemotherapy, 12 did not exclude people receiving hormone therapy, 12 were not RCTs, and three were expert opinions (Figure 1).

Risk of bias in included studies

Due to the insufficient reporting all three studies we judged the overall risk of bias of the included trials as unclear.

For detailed information see the Characteristics of included studies table.

Allocation

We judged the risk of bias in random sequence generation to be low in Monga 2007 and unclear in Hwang 2008 and Kulkarni 2013. All three studies described randomisation and achieved a balanced distribution of baseline participant characteristics, but only Monga 2007 described a stratified randomisation with approximately equal numbers of prior exercisers in both groups. No studies described the exact allocation process and whether there was any concealment of the allocation sequence (unclear bias).

Blinding

We judged Monga 2007 to be at high risk of performance bias as it blinded neither participants nor staff. As the investigators used many subjective questionnaires to measure outcomes, knowledge about participation and group assignment alone could have influenced the response pattern. In contrast, Hwang 2008 did not report whether participants gave their informed consent to participate in an exercise study, and did not indicate who carried out the analyses, so we cannot judge risk of performance and detection bias at this time. Only Kulkarni 2013 reported blinding of the participants: the study included two active groups, but only one with additional aerobic training. The physiotherapists in Kulkarni 2013 were not blinded because they had to schedule the intervention sessions. However, as lack of blinding did not affect the endpoints (the blinded participants completed the questionnaires), we judged the risk of performance and detection bias to be low for this study.

Incomplete outcome data

Monga 2007 reported all measured endpoints. There was no evidence of incompleteness and no report of study dropouts. Therefore, we considered the study at low risk of attrition bias. In contrast, Hwang 2008 and Kulkarni 2013 reported study dropouts without explaining how incomplete outcome data were handled. Furthermore, it was unclear whether all questionnaires were completely filled out. We judged both studies at unclear risk of attrition bias.

Selective reporting

As no study protocol is available for any of the included studies, and we received no response to our requests for additional information, we considered the risk for reporting bias to be unclear for all three trials.

Other potential sources of bias

Monga 2007 administered a general activity level questionnaire before the allocation process so that the groups were balanced in terms of exercise behaviour at baseline, but neither Hwang 2008 nor Kulkarni 2013 evaluated previous exercise habits. We identified no other potential sources of bias. Therefore, we judged risk of other bias to be low in Monga 2007 and unclear in Hwang 2008 and Kulkarni 2013. 

Effects of interventions

See: Table 1

Three two‐arm RCTs with 130 participants were included in this review (Hwang 2008, Kulkarni 2013, Monga 2007). Participants had either a breast or prostate cancer diagnosis. As we did not consider the data sufficiently homogeneous to be pooled, owing to the different cancer diagnoses and assessment tools, we displayed the analyses for each study separately in a collective forest plot. Our analyses and forest plots are based on post‐RT assessments. We were unable to analyse the changes scores between pre‐ and post‐RT assessments because Hwang 2008 and Kulkarni 2013 provided insufficient data, and some data in Monga 2007 appeared to be inaccurate. For some endpoints (fatigue, physical performance, QoL), the control groups showed slightly better scores before RT. However, the exercise group scores had improved after RT, and the control group scores had not. We were unable to account for this difference in our analysis, but it is a factor to consider when interpreting the results. 

Fatigue 

All three studies reported fatigue. Monga 2007 measured this outcome with the PFS‐Revised, while Hwang 2008 and Kulkarni 2013 used the BFI. All three studies provided low‐certainty evidence that exercise improves fatigue (Table 1; Analysis 1.1). The results of the individual analyses are presented below (positive SMD values signify less fatigue).

1.1. Analysis.

Comparison 1: Exercise intervention plus standard of care versus standard of care alone, Outcome 1: Fatigue

• SMD 0.96, 95% CI 0.27 to 1.64; 37 participant (Hwang 2008)

• SMD 2.42, 95% CI 1.71 to 3.13; 54 participants (Kulkarni 2013)

• SMD 1.44, 95% CI 0.46 to 2.42; 21 participants (Monga 2007)

All studies reported reduced fatigue scores in the exercise group at the post‐RT assessment compared to the pre‐RT assessment (P value not reported in Hwang 2008; P < 0.001 in Kulkarni 2013; P = 0.02 in Monga 2007), or a significant difference between the groups in change scores from pre‐ to post‐RT assessment, favouring the exercise group (P < 0.05 in Hwang 2008; P value not reported in Kulkarni 2013; P < 0.001 in Monga 2007). Fatigue increased in the control groups in Hwang 2008 and Monga 2007 and remained unchanged in Kulkarni 2013, where the control group received physiotherapy (without aerobic exercise) plus standard care. Unfortunately, no studies provided accurate exact numbers for further analysis.

Further research might affect the estimate and our confidence in the estimate.

Quality of life

All three studies reported QoL. Hwang 2008 and Kulkarni 2013 used the WHOQOL‐BREF, while Monga 2007 used the FACT‐P. Kulkarni 2013 did not provide sufficient data for analysis. The other two studies provided low‐certainty evidence that exercise may have little or no effect on QoL (Table 1; Analysis 1.2). The results of the individual analyses are presented below (positive SMD values signify better QoL).

1.2. Analysis.

Comparison 1: Exercise intervention plus standard of care versus standard of care alone, Outcome 2: Overall Quality of Life (QoL)

• SMD 0.40, 95% CI −0.26 to 1.05; 37 participants (Hwang 2008)

• SMD 0.47, 95% CI −0.40 to 1.34; 21 participants (Monga 2007)

All three studies reported differences between the two groups in change scores from the pre‐RT assessment to the post‐RT assessment, favouring the exercise group (P < 0.05 in Hwang 2008; P < 0.001 in Kulkarni 2013; P = 0.006 in Monga 2007). Monga 2007 also reported an improvement in QoL in the exercise group at the post‐RT assessment compared to the pre‐RT assessment (P = 0.04). Hwang 2008 and Monga 2007 reported that QoL decreased in the control groups after RT. In Kulkarni 2013, physiotherapy alone was inferior to the exercise group, which received additional aerobic training. Unfortunately, no studies provided accurate exact numbers for further analysis.

Further research might affect the estimate and our confidence in the estimate.

Physical performance

All three studies reported physical performance. Hwang 2008 measured pain on a VAS, Kulkarni 2013 used the six‐minute walk test, and Monga 2007 used the stand‐and‐sit test. Hwang 2008 and Kulkarni 2013 provided very low‐certainty evidence that exercise may improve physical performance, while Monga 2007 provided very‐low certainty evidence of no difference between the groups (Table 1; Analysis 1.3). The analyses of the individual studies are presented below (positive SMD values signify better physical performance).

1.3. Analysis.

Comparison 1: Exercise intervention plus standard of care versus standard of care alone, Outcome 3: Physical performance

• SMD 1.25, 95% CI 0.54 to 1.97; 37 participants (Hwang 2008)

• SMD⁠⁠⁠⁠⁠⁠ 3.13 (95% CI 2.32 to 3.95; 54 participants (Kulkarni 2013)

• SMD 0.00, 95% CI −0.86 to 0.86; 21 participants (Monga 2007)

All three studies reported a difference between the groups in change scores from pre‐RT assessment to post‐RT assessment, favouring the exercise group (P < 0.05 in Hwang 2008; P < 0.001 in Kulkarni 2013; P < 0.001 in Monga 2007). Monga 2007 also reported an improvement in physical performance in the exercise group from pre‐RT assessment to post‐RT assessment (P < 0.001). Unfortunately, no studies provided accurate exact numbers for further analysis.

We are very uncertain about the estimate.

Psychosocial effects

Two studies reported psychosocial effects. Hwang 2008 measured this outcome using the WHOQOL‐BREF social subscale, and Monga 2007 used the BDI. There was very‐low certainty evidence that exercise may have little or no effect on psychosocial effects (Table 1; Analysis 1.4) The analyses of the individual studies are presented below (positive SMD values signify better psychosocial well‐being).

1.4. Analysis.

Comparison 1: Exercise intervention plus standard of care versus standard of care alone, Outcome 4: Psychosocial effects

• SMD 0.48, 95% CI −0.18 to 1.13; 37 participants (Hwang 2008)

• SMD 0.29, 95% CI −0.57 to 1.15; 21 participants (Monga 2007)

Both studies reported differences between groups in change scores from pre‐RT assessment to post‐RT assessment, favouring the exercise group (P < 0.001 in Hwang 2008; P = 0.002 in Monga 2007). Monga 2007 also reported an improvement in social well‐being in the exercise group from pre‐RT assessment to post‐RT assessment (P = 0.02). Both studies reported a decrease in the corresponding scores in the control groups after RT.

We are very uncertain about the estimate.

Overall survival

No studies reported overall survival.

Return to work

Non studies reported return to work.

Adverse events

Hwang 2008 and Kulkarni 2013 recorded exercise‐related adverse events and reported no events. They did not record adverse events that were not related to exercise.

We estimated the certainty of the evidence as very low (Table 1). The evidence is very uncertain about the effect of exercise on adverse events in this group.

Discussion

Summary of main results

In this review, we identified three RCTs evaluating exercise interventions for adults with cancer receiving RT alone (Hwang 2008, Kulkarni 2013, Monga 2007). Two studies are awaiting classification (ISRCTN26140710, Milecki 2013). 

All three included studies had two study arms. They enrolled a total of 130 people who had either a breast cancer or prostate cancer diagnosis. The exercise groups attended an exercise programme several times per week for several weeks. All three studies assessed fatigue, QoL, and physical performance as outcomes. Two studies also reported psychosocial effects (Hwang 2008; Monga 2007), and two recorded exercise‐related adverse events (Hwang 2008; Kulkarni 2013). The investigators of all three studies measured baseline characteristics and scores before the start of RT and measured the outcomes after completion of RT. Our analyses are based on the end scores. In some outcomes (fatigue, physical performance, QoL), the control groups had slightly better baseline scores before RT. However, in most cases, the control groups had worsened or not improved after RT, while the exercise group scores had improved. Kulkarni 2013 reported a smaller improvement in QoL in the control group (physiotherapy) versus the exercise group (physiotherapy plus aerobic exercise). We were unable to analyse differences in change scores between the pre‐ and post‐RT assessments because of limited data. This could explain the differences between our results and those reported by the study authors.

The following list summarises the most relevant findings of this review.

• Exercise interventions are possible in people with breast and prostate cancer undergoing RT without additional systemic cancer treatment.

• Our analyses of follow‐up data from each study showed low‐certainty evidence of reduced fatigue in the exercise group versus the control group. All three studies reported significant decreases in fatigue from baseline in the group that received the exercise intervention in addition to standard care, versus no change or increased fatigue in the control groups. Exercise may reduce fatigue. 

• Our analyses of the follow‐up QoL data from two studies showed low‐certainty evidence of little or no difference between the groups (Hwang 2008; Monga 2007). Kulkarni 2013 provided insufficient data for analysis. However, all three studies reported better change scores in the exercise group compared to the control group. Exercise may improve QoL. 

• Our analyses of the follow‐up data for physical performance from two studies showed very low‐certainty evidence of a difference between the groups in favour of exercise (Hwang 2008; Kulkarni 2013); while our analysis of the data from the third study showed very‐low certainty evidence of little or no difference (Monga 2007). All three studies reported better physical performance change scores in the exercise group compared to the control group. The evidence is very uncertain about the effect of exercise on physical performance.

• Our analyses of the follow‐up data for psychosocial effects from two studies showed very low‐certainty evidence of little or no difference between the groups (Hwang 2008; Monga 2007). However, both studies reported better change scores in the exercise groups compared to the control groups. The evidence is very uncertain about the effect of exercise on psychosocial effects in this cohort.

• Two studies reported that there were no exercise‐related adverse events; they did not identify and record treatment‐related adverse events (Hwang 2008; Kulkarni 2013). Monga 2007 did not report adverse events of any type. The evidence is very uncertain about the effect of exercise on adverse events.

• No studies reported outcome assessments for overall survival, anthropometric measurements, or return to work in any of the studies.

• Whenever there was evidence of a difference between the groups, the results of our analyses generally favoured exercise. However, our analyses showed less evidence of a difference between the groups than reported by the study authors. No study reported accurate exact numbers for every assessed outcome.

• The authors of all studies concluded that a supervised exercise training programme during RT alone for breast or prostate cancer leads to positive physical and psychological benefits and reduces fatigue. The exercise groups had better results than the control groups in all participant‐rated endpoints.

• Kulkarni 2013 reported that conventional physiotherapy alone was not helpful for improving participants' overall physical condition, and Hwang 2008 found no evidence of negative effects associated with exercise in cancer survivors.

• To date, there are no data on overall survival or short‐ and long‐term adverse events in this cohort.

• All three studies were very small and had limitations due to risk of bias and poor documentation.

• The certainty of the evidence for all outcomes was either low or very low.

• Studies investigating the effects of exercise interventions for people with cancer receiving RT alone are still scarce and of limited quality, although exercise has become a very important topic in the area of supportive care.

Overall completeness and applicability of evidence

We conducted a broad‐term literature search, which resulted in a very large number of initial hits.

From this search, we identified three studies that focused on RT in addition to an exercise intervention. We also identified one study from the UK with no published results (ISRCTN26140710). However, the study protocol describes a cohort of only 10 women, so we did not expect this study to have an impact on our overall results. We identified another study with unclear participant numbers and randomisation process (Milecki 2013).

We defined 'RT alone' as RT without adjuvant systemic therapy (chemotherapy, immunotherapy, or hormone therapy; surgery was not an exclusion criterion). This likely explains why only three studies met our eligibility criteria. Most cancer types are treated with a multimodal treatment approach. There are still indications for RT alone for which exercise might be beneficial, some of which are reflected in our review. In two studies where the participants had surgery and adjuvant RT, the surgical procedure may have influenced study outcomes (Hwang 2008, Kulkarni 2013). 

The studies enrolled 130 participants, of whom 112 completed follow‐up. One notable reason for dropouts was participants randomised to the control group deciding they would rather participate in the exercise group.

Hwang 2008 and Kulkarni 2013 included pre‐surgical women with breast cancer with adjuvant RT, while Monga 2007 included men with prostate cancer and primary RT. Thus, the participants have different baselines, which weakens the comparability of the results.

Duration of follow‐up was very limited in all three studies, so we are unable to draw any long‐term conclusions from the results.

Only Hwang 2008 and Kulkarni 2013 recorded adverse events (specifically, those related to exercise during the intervention period) and reported that none occurred. There were no reports on short‐ or long‐term adverse events, RT‐related adverse events, or any comparisons between the control and exercise groups.

All three studies excluded severe pre‐existing medical conditions. The results are therefore not directly applicable to participants with severe pre‐existing medical conditions. Most studies on exercise enrol people who are capable of completing the exercise sessions.

Nether Hwang 2008 nor Kulkarni 2013 collected baseline activity level.

In all three studies, the randomisation process was unclear and the information on the study design was superficial. The mere knowledge of being allocated to the exercise group could increase the participant's motivation and thus falsely improve their results. Only Kulkarni 2013 blinded participants by treating both the control and the exercise group with conventional physiotherapy. The addition of aerobic training was not recognisable as an intervention for the exercise group participants.

The studies collected data at only two time points, once before and once after the RT. An interim assessment or longer‐term follow‐up would have made the results more reliable. It remains unclear at what point in time the benefits of exercise interventions are detectable.

Both Hwang 2008 and Kulkarni 2013 presented some results in graphs without absolute numbers, which made the reading and the analyses imprecise. We did not analyse the endpoints for which we were unable to obtain scores including SD. The tables in Monga 2007 initially appeared very detailed and explicit, but we found several inconsistencies between the different tables (prostate cancer symptoms, FACT‐P values, varying SDs) and the reported average weight of the control group was clearly erroneous (80.1 lb (36.3 kg)). Consequently, we were unable to analyse change scores, and our analyses only reflect the follow‐up scores without taking into account the baseline differences between the groups. This may partly explain why the results of our analyses differ from the results reported in the individual studies. Because they only provided P values (and inaccurate numbers in the case of Monga 2007), further analyses were not possible.

There is scarce evidence on the effects of exercise in people with cancer undergoing RT alone. Furthermore, no studies reported data on adverse events not related to exercise or on survival rates, which were among our secondary outcomes.

Quality of the evidence

RCTs provide high certainty in effect estimates. Serious limitations such as risk of bias, imprecision, inconsistency between studies, indirectness, or publication bias lower our certainty that the effect estimates reflect the true effects in our target population. There were no study protocols available for any of the included studies, so we were unable to assess risk of bias in some categories. Overall, we found poor documentation and resulting uncertainties in the assessment of the systematic error, and we graded the certainty of the evidence as low or very low.

For fatigue and QoL, we downgraded for lack of blinding of study participants and personnel, possible outcome reporting bias due to missing study protocols, and serious imprecision due to small sample sizes in a small number of studies. The certainty of the evidence is low.

For physical performance the studies reported different endpoints, which limited between group comparability. We downgraded for lack of blinding of study participants and personnel, possible outcome reporting bias due to missing study protocols, and very serious imprecision due to very small sample size in a small number of studies. The certainty of the evidence is very low. 

For psychosocial effects, we were unable to analyse the data from Kulkarni 2013 due to poor presentation of results. We downgraded for lack of blinding of study participants and personnel, possible outcome reporting bias due to missing study protocols, serious imprecision due to small sample sizes in a small number of studies, and indirectness of outcomes. The certainty of the evidence is very low.

No studies reported overall survival, anthropometric measurements, or return to work.

Reported adverse events only affected the exercise groups because studies only reported exercise‐related adverse events. Monga 2007 did not record adverse events of any type. We downgraded for lack of blinding of study participants and personnel, possible outcome reporting bias due to missing study protocols, serious imprecision due to small sample sizes in a small number of studies, and indirectness of outcomes. The certainty of the evidence is very low.

For further details please see the Table 1, Figure 2, and Figure 3.

Potential biases in the review process

Two or three review authors double‐checked all processes in this review to detect any possible errors at an early stage.

The main limitation of this systematic review is the very small number of appropriate RCTs that met our inclusion criteria. We conducted an extensive literature search with no language restriction, but it is still possible that we overlooked RCTs (e.g. in foreign languages, unpublished, in progress, or published only as dissertations). Therefore, we cannot rule out publication bias or language bias.

Agreements and disagreements with other studies or reviews

Strong evidence exists on the benefits of exercise for people with cancer. One 2017 systematic review of exercise studies in the cancer literature noted that exercise is beneficial before, during, and after cancer treatment, for all cancer types and for a variety of cancer‐related adverse events (Stout 2017). The authors concluded that exercise interventions are generally feasible and safe, but that people with cancer should be screened and that clinical practice guidelines need further development. 

Consistent with our review, the most commonly studied cancer types in the literature on exercise interventions are breast, prostate, and colorectal cancer (Stout 2017). Our three included studies addressed prostate cancer (N = 1) and breast cancer (N = 2). Many systematic reviews have focused on a single cancer type (Baumann 2012; Capozzi 2016; Cheema 2008; Cramer 2014; Cramer 2017; Granger 2011; Keogh 2011; McNeely 2006; Van Dijck 2016; van Vulpen 2016). We decided to include all cancer types, as most reviews have shown positive results from exercise regardless of the primary cancer diagnosis.

To our knowledge, this is the first systematic review to focus on exercise interventions for people with cancer undergoing RT alone. Several systematic reviews have evaluated the effectiveness of exercise for people with cancer undergoing systemic and multimodal treatment approaches (Zeng 2019, Cave 2018, Loughney 2018). We considered that adjuvant systemic treatment would have a major impact on the measured outcomes.

Cancer treatment modalities and combinations are usually based on the type of cancer and the severity of the disease, and are associated with specific adverse events. Some reviews focus on the effect of exercise on a single physical impairment (Meneses‐Echavez 2015, Kwan 2011, Mustian 2017), while most provide sub‐analyses for specific adverse events. We included all types of physical impairments.

Cancer‐related fatigue (CRF) is the most commonly studied physical impairment in systematic reviews investigating the role of exercise in people with cancer. Emerging evidence indicates a significant benefit of exercise for reducing CRF (Capozzi 2016; Cramp 2012; Keogh 2011; Larkin 2014; Loughney 2018; Meneses‐Echavez 2015; Mustian 2017; van Vulpen 2016). In line with these conclusions, our results also suggest that exercise interventions during RT alone can improve fatigue.

The literature provides evidence that exercise usually has a positive effect on QoL in people with cancer (Capozzi 2016; Keogh 2011; Knips 2019; McNeely 2006; Smits 2015; Van Dijck 2016). We also identified studies that found no evidence of a significant impact on QoL (Cramer 2014, Granger 2011). In one of our included studies, the authors considered that because the exercise intervention improved physical performance, the improvement in QoL could be due to an improved physical and psychological state during RT (Kulkarni 2013).

Exercise interventions have shown a strong positive impact on several physical fitness measures, particularly VO₂ max (McNeely 2006; Van Dijck 2016), strength (Capozzi 2016; Cheema 2008; Keogh 2011), flexibility (Cheema 2008), and several measures of cardiorespiratory fitness (Cheema 2008; Cramer 2014). The studies included in this review also reported positive effects on physical performance.

Data on the effects of exercise interventions on psychological functioning are diverse. There are reviews demonstrating positive effects (Cheema 2008, Cramer 2017), but there are also inconclusive reports on this topic (Knips 2019). Some studies also report only moderate to nonexistent improvements of cognitive functions with exercise interventions (van Vulpen 2016, Zimmer 2016). A wide range of instruments are available for measuring psychosocial effects, and the concept of psychosocial effects covers a range of cognitive impairments. The results of this review suggest that exercise interventions in people with cancer receiving RT alone may have a positive effect on specific impairments as measured by the corresponding tools, but the evidence is very uncertain.

Authors' conclusions

Implications for practice.

For fatigue, which was the primary outcome of our review, there was low‐certainty evidence of a difference between the exercise and control groups favouring exercise in all three included studies. For physical performance, two studies provided very low‐certainty evidence of a difference between the groups favouring exercise, while one study provided very‐low certainty evidence of little or no difference. For the other assessed outcomes, our analyses showed low or very‐low certainty evidence of little or no difference between the groups in post‐RT scores. Two studies reported that no exercise‐related adverse events occurred during the exercise intervention period. There were no further reports about adverse events.

All three studies reported improvements in all assessed outcomes in the exercise groups. Due to insufficient available data and the lack of accurate exact numbers, we were unable to consistently support these results in our analyses. 

In summary, exercise interventions may be beneficial for people with cancer undergoing RT alone, but we only have low‐ to very low‐certainty evidence to support this statement. Although there are indications that exercise may have physical and psychological benefits for people with breast or prostate cancer receiving radiation therapy (RT) alone, we cannot reach solid conclusions or provide recommendations for clinical practice based on the scientific data currently available. 

Implications for research.

Exercise has become a very important topic in the area of additional supportive care. To date, there are no data on survival rates or short‐ and long‐term adverse events. Overall, the evidence is limited and there is a need for large, well‐conducted randomised controlled trials and high‐quality reviews on this topic. Future studies should clearly and fully report the most important outcomes, including overall survival, quality of life, physical and social well‐being, improvements in symptoms, and return to work. They should also record all adverse events, examining whether they occur as a result of particular exercise programme, in which types of cancer they occur, whether exercise could exacerbate radiation‐induced adverse events, and, conversely, whether exercise could prevent or mitigate adverse events. Since many people with breast cancer receive RT alone, studies should report RT‐induced adverse events affecting lung function, as these could impede patients' ability to exercise. It is also crucial to investigate the effects of exercise interventions in more types of cancer requiring RT alone. To date, it is difficult to assess when and how exercise interventions are appropriate and beneficial for people with cancer undergoing RT alone. We need additional high‐quality studies on this topic to guide healthcare professionals when making decisions about optimal supportive therapy. 

History

Protocol first published: Issue 10, 2019

Appendices

Appendix 1. CENTRAL search strategy

#1. MeSH descriptor: [Neoplasms] explode all trees
#2. neoplas* or carcinoma* or adenocarcinoma* or malignan* or cancer* or tumor* or tumour* or oncolog*
#3. #1 or #2 
#4. MeSH descriptor: [Radiotherapy] explode all trees
#5. Any MeSH descriptor in all MeSH products and with qualifier(s): [radiotherapy ‐ RT] 
#6. radiotherap* or irradiat* or radiat* or stereotactic* or radiosurgery* or cyberknife* or brachytherapy* or "rapid arc" or EBRT or "external beam radiation therapy" or VMAT or "volumetric modulated arc therapy" or IMRT or "intensity modulated radiotherapy"
#7. #4 or #5 or #6 
#8. #3 and #7
#9. MeSH descriptor: [Exercise] explode all trees 
#10. MeSH descriptor: [Exercise Therapy] explode all trees 
#11. MeSH descriptor: [Exercise Movement Techniques] explode all trees 
#12. MeSH descriptor: [Physical Fitness] this term only 
#13. MeSH descriptor: [Physical Endurance] explode all trees
#14. MeSH descriptor: [Muscle Strength] explode all trees
#15. exercis* or resistance* or movement* or stretch* or aerobic* or anaerobic* or flexibility*
#16. ((physical* or resistance*) near/3 (activ* or therap* or exercise* or endurance* or education* or fitness* or train*))
#17. ((balance* or coordination* or strength*) near/3 (train* or exercise*))
#18. #9 or #10 or #11 or #12 or #13 or #14 or #15 or #16 or #17
#19. #8 and 18#

Appendix 2. MEDLINE search strategy

1. exp Neoplasms/
2. (neoplas* or carcinoma* or adenocarcinoma* or malignan* or cancer* or tumor* or tumour* or oncolog*).ti,ab.
3. 1 or 2
4. exp Radiotherapy/
5. radiotherapy.fs.
6. (radiotherap* or irradiat* or radiat* or stereotactic* or radiosurgery* or cyberknife* or brachytherapy* or "rapid arc" or EBRT or "external beam radiation therapy" or VMAT or "volumetric modulated arc therapy" or IMRT or "intensity modulated radiotherapy").ti,ab.
7. 4 or 5 or 6
8. 3 and 7
9. exp Exercise/
10. exp Exercise Therapy/
11. exp Exercise Movement Techniques/
12. Physical Fitness/
13. exp Physical Endurance/
14. exp Muscle Strength/
15. (exercis* or resistance* or movement* or stretch* or aerobic* or anaerobic* or flexibility*).ti,ab.
16. ((physical* or resistance*) adj3 (activ* or therap* or exercise* or endurance* or education* or fitness* or train*)).ti,ab.
17. ((balance* or coordination* or strength*) adj3 (train* or exercise*)).ti,ab.
18. 9 or 10 or 11 or 12 or 13 or 14 or 15 or 16 or 17
19. 8 and 18
20. randomized controlled trial.pt.
21. controlled clinical trial.pt.
22. randomized.ab.
23. placebo.ab.
24. drug therapy.fs.
25. randomly.ab.
26. trial.ti.
27. groups.ab.
28. 20 or 21 or 22 or 23 or 24 or 25 or 26 or 27
29. (animals not (humans and animals)).sh.
30. 28 not 29
31. 19 and 30

Appendix 3. Embase search strategy

1. exp neoplasm/             
2. (neoplas* or carcinoma* or adenocarcinoma* or malignan* or cancer* or tumor* or tumour* or oncolog*).ti,ab.         
3. 1 or 2                
4. exp radiotherapy/      
5. radiotherapy.fs.           
6. (radiotherap* or irradiat* or radiat* or stereotactic* or radiosurgery* or cyberknife* or brachytherapy* or "rapid arc" or EBRT or "external beam radiation therapy" or VMAT or "volumetric modulated arc therapy" or IMRT or "intensity modulated radiotherapy").ti,ab.               
7. 4 or 5 or 6        
8. 3 and 7             
9. exp exercise/                
10. exp kinesiotherapy/                
11. fitness/         
12. exp endurance/        
13. exp muscle strength/              
14. (exercis* or resistance* or movement* or stretch* or aerobic* or anaerobic* or flexibility*).ti,ab.    
15. ((physical* or resistance*) adj3 (activ* or therap* or exercise* or endurance* or education* or fitness* or train*)).ti,ab.     
16. ((balance* or coordination* or strength*) adj3 (train* or exercise*)).ti,ab.   
17. 9 or 10 or 11 or 12 or 13 or 14 or 15 or 16          
18. 8 and 17        
19. crossover procedure/             
20. double‐blind procedure/       
21. randomized controlled trial/                
22. single‐blind procedure/         
23. random*.mp.             
24. factorial*.mp.             
25. (crossover* or cross over* or cross‐over*).mp.          
26. placebo*.mp.             
27. (double* adj blind*).mp.      
28. (singl* adj blind*).mp.            
29. assign*.mp. 
30. allocat*.mp.                
31. volunteer*.mp.         
32. 19 or 20 or 21 or 22 or 23 or 24 or 25 or 26 or 27 or 28 or 29 or 30 or 31                
33. 18 and 32

key:

mp=title, abstract, original title, name of substance word, subject heading word, keyword heading word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier
pt=publication type
ab=abstract
fs=floating subheading

Appendix 4. CINAHL search strategy

1. (neoplasms).mh
2. (neoplas* OR carcinoma* OR adenocarcinoma* OR malignan* OR cancer* OR tumor* OR tumour* OR oncolog*).ti,ab                
3. (1 OR 2)          
4. (Radiotherapy).mh     
5. (radiotherap* OR irradiat* OR radiat* OR stereotactic* OR radiosurgery* OR cyberknife* OR brachytherapy* OR "rapid arc" OR EBRT OR "external beam radiation therapy" OR VMAT OR "volumetric modulated arc therapy" OR IMRT OR "intensity modulated radiotherapy").ti,ab            
6. (4 OR 5)           
7. (3 AND 6)        
8. (Exercise).mh               
9. (Physical Fitness).mh 
10. (Physical Endurance).mh       
11. (Muscle Strength).mh            
12. (exercis* OR resistance* OR movement* OR stretch* OR aerobic* OR anaerobic* OR flexibility*).ti,ab           
13. (physical* N3 activ* OR therap* OR exercise* OR endurance* OR education* OR fitness* OR train*).ti,ab     
14. (resistance* N3 activ* OR therap* OR exercise* OR endurance* OR education* OR fitness* OR train*).ti,ab                
15. (balance* N3 train* OR exercise*).ti,ab          
16. (coordination* N3 train* OR exercise*).ti,ab                
17. (strength* N3 train* OR exercise*).ti,ab        
18. (8 OR 9 OR 10 OR 11 OR 12 OR 13 OR 14 OR 15 OR 16 OR 17)    
19. (7 AND 18)   
20. (randomized controlled trials).mh     
21. (double‐blind studies).mh    
22. (single‐blind studies).mh       
23. (random assignment).mh     
24. (pretest‐posttest design).mh              
25. (cluster sample).mh                
26. (randomised OR randomized).ti         
27. (random*).ab             
28. (trial).ti          
29. (sample size).mh      
30. (assigned OR allocated OR control).ab             
31. (29 AND 30) 
32. (placebos).mh            
33. (randomized controlled trial).pt         
34. (control W5 group).ab            
35. (crossover design OR comparative studies).mh           
36. (cluster W3 RCT).ab 
37. (20 OR 21 OR 22 OR 23 OR 24 OR 25 OR 26 OR 27 OR 28 OR 31 OR 32 OR 33 OR 34 OR 35 OR 36)                
38. (19 AND 37)ere]

Data and analyses

Comparison 1. Exercise intervention plus standard of care versus standard of care alone.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Fatigue	3		Std. Mean Difference (IV, Fixed, 95% CI)	Subtotals only
1.2 Overall Quality of Life (QoL)	3		Std. Mean Difference (IV, Fixed, 95% CI)	Subtotals only
1.3 Physical performance	3		Std. Mean Difference (IV, Random, 95% CI)	Subtotals only
1.4 Psychosocial effects	3		Std. Mean Difference (IV, Random, 95% CI)	Subtotals only

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Hwang 2008.

Study characteristics
Methods	Study design: RCT Study grouping: parallel group
Participants	Control group (n = 23) Age (years): mean 46.3 (SD 9.5) Exercise group (n = 17) Age (years): mean 46.3 (SD 7.5) Overall Age (years): mean 46.3 (SD 8.52) Inclusion criteria Women on outpatient waiting list for radiotherapy for breast cancer Exclusion criteria Concurrent major health problems Uncontrolled hypertension Cardiovascular disease Acute or chronic respiratory disease Cognitive dysfunction
Interventions	Exercise group (n = 17): supervised exercise programme 3 times/week with HR monitoring. Programme consisted of 10‐minute warm up; 30 minutes shoulder stretches, aerobic exercise, and strengthening exercise; and 10‐minute cool down Control group (n = 23): standard care that included showing participants how to perform shoulder ROM exercises and encouraging them to continue with normal activities
Outcomes	QoL (overall, psychological, environmental, social, physical; WHOQOL‐BREF) Fatigue (BFI) ROM of shoulder (forward flexion, abduction, and internal rotation and external rotation) Pain (VAS) Adverse events
Identification	Sponsorship source: none Country: South Korea Setting: Department of Physical Medicine and Rehabilitation, Division of Sports Medicine, Departments of Radiation Oncology and Surgery; Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, South Korea Comments: no detailed participant characteristics Corresponding author: Dr Hyun Jung Chang Institution: Department of Physical Medicine & Rehabilitation, Samsung Medical Center, Sungkyunkwan University School of Medicine Email: reh.chj@gmail.com Address: Department of Physical Medicine & Rehabilitation, Samsung Medical Center, Sungkyunkwan University School of Medicine, 50 Ilwon‐dong, Gangnam‐gu, Seoul 135‐710, South Korea
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Missing details on random sequence generation.
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not reported.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Unclear whether participants knew they were participating in a study.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Outcome assessment not reported.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Study did not report number of participants for each outcome; completeness unclear.
Selective reporting (reporting bias)	Unclear risk	No study protocol available.
Other bias	Unclear risk	No activity survey before randomisation.

Kulkarni 2013.

Study characteristics
Methods	Study design: RCT Study grouping: parallel group
Participants	Control group (n = 30) Age (years): mean 46.42 (SD 8.286) Height (cm): mean 157.92 (SD 2.682) Weight (kg): mean 49.73 (SD 4.601) BMI (kg/m2): mean 19.93 (SD 1.526) Exercise group (n = 30) Age (years): mean 44.82 (SD 3.370) Height (cm): mean 158.03 (SD 3.237) Weight (kg): mean 51.14 (SD 5.930) BMI (kg/m2): mean 20.42 (SD 1.872) Overall Age (years): mean 45.59 (SD 7.332) Height (cm): not reported Weight (kg): not reported BMI (kg/m2): not reported Inclusion criteria Age 30–60 years Clinical diagnosis of stage I or II breast cancer Unilateral modified radical mastectomy Radiotherapy Exclusion criteria Any cardiovascular abnormality Impaired cognitive function Musculoskeletal disorders Neurological disorders Emotional instability Very low level of activity
Interventions	Exercise group (n = 30): aerobic training combined with conventional physiotherapy, 5 times/week for 6 weeks. Aerobic training consisted of a 10‐minute warm‐up (mild stretching of large muscle groups), 10–30 minutes treadmill walking at self‐adjusted speeds, and 10‐minute cool‐down. Control group (n = 30): standard care plus conventional physiotherapy 5 times/week for 6 weeks, supervised by a physiotherapist
Outcomes	Fatigue (BFI) Exercise capacity (6‐minute walk test) VO2max QoL (WHOQOL‐BREF) Adverse events
Identification	Sponsorship source: none Country: India Setting: rural tertiary medical centre (Department of Community Health and Rehabilitation, College of Physiotherapy, Pravara Rural Hospital, Loni, India) Comments: no detailed participant characteristics Corresponding author: Nupoor Kulkarni Institution: College of Physiotherapy, Pravara Institute of Medical Sciences, Loni Email: kulkarninupoor@yahoo.in Address: College of Physiotherapy, Pravara Institute of Medical Sciences, Loni‐413736, Tal: Rahata, India
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Missing details on random sequence generation.
Allocation concealment (selection bias)	Unclear risk	No allocation concealment reported.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Blinding of participants (no blinding of personnel, but irrelevant to the outcomes).
Blinding of outcome assessment (detection bias) All outcomes	Low risk	The blinded participants were the outcome assessors, as they filled in the questionnaires. The other outcomes were not affected by knowledge of randomisation.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Study did not report number of participants per outcome.
Selective reporting (reporting bias)	Unclear risk	No study protocol available.
Other bias	Unclear risk	No activity survey before the randomisation.

Monga 2007.

Study characteristics
Methods	Study design: RCT Study grouping: parallel group
Participants	Control group (n = 10) Age (years): mean 70.6 SD 5.3) Weight (lb): mean 180.1 (SD 28.8) PSA (ng/mL): mean 6.4 (SD 5.0) Combined Gleason: mean 5.1 (SD 1.2) Exercise group (n = 11) Age (years): mean 68 (SD 4.2) Weight (lb): mean 177.3 (SD 29.1) PSA (ng/mL): mean 7.4 (SD 5.7) Combined Gleason: mean 5.5 (SD 1.0) Overall (n = 21) Age (years): mean 69.24 (SD 4.82) Weight (lb): 178.7 (SD 28.9) PSA (ng/mL): mean 6.92 (SD 5.27) Combined Gleason: mean 5.31 (SD 1.09) Inclusion criteria Localised prostate cancer Only first‐time cancer diagnosis Ambulatory Able to complete self‐report measures Exclusion criteria Concurrent chemotherapy  Major health problems (uncontrolled hypertension, uncontrolled insulin‐dependent diabetes mellitus, severe arthritis, obvious cognitive dysfunction)  Recent history of sudden onset of shortness of breath on exertion or recent history of dizziness, blurred vision, or fainting spells Recent history of unstable angina, coronary artery disease, myocardial infarction, or cardiac failure Bone, back, or neck pain of recent origin, or inability to exercise
Interventions	Exercise group (n = 11): aerobic (cardiovascular conditioning) exercise programme with 10‐minute warm‐up, 30‐minute aerobic segment (treadmill walking), and 5–10‐minute cool‐down period, 3 times/week for 8 weeks Control group (n = 10): standard care that included patient education and radiotherapy without exercise prescription
Outcomes	Cardiovascular fitness (Bruce treadmill test + METs + target HR) Flexibility (modified sit‐and‐reach test) Strength (stand‐and‐sit test) Fatigue (PFS) QoL (FACT‐P) Prostate cancer symptoms Depression (BDI)
Identification	Sponsorship source: no commercial party involved Country: Texas, USA Setting: Houston Veterans Affairs Medical Center for radiation treatment of localized prostate cancer Corresponding author: Kuno P Zimmermann Institution: Houston Veterans Affairs Medical Center Email: zimmermann.kunop@med.va.gov Address: Department of Veterans Affairs Medical Center, 2002Holcombe Blvd, Houston, TX 77030
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Stratified randomisation: "Subjects who reported engaging in regular physical exercise (2 times a week) and those who reported not engaging in regular exercise were randomized separately to the 2 groups, so that both would have approximately equal numbers of prior exercisers."
Allocation concealment (selection bias)	Unclear risk	No details of allocation concealment.
Blinding of participants and personnel (performance bias) All outcomes	High risk	No blinding of participants.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not reported.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Detailed documentation of results, no indication of incompleteness.
Selective reporting (reporting bias)	Unclear risk	No study protocol available.
Other bias	Low risk	Activity survey before randomisation: physical fitness assessment before the start of the study; balanced fitness levels in both groups.

BDI: beck depression inventor; BFI: brief fatigue inventory; BMI: body mass index; FACT‐P: functional assessment of cancer therapy – prostate; HR: heart rate; MET: metabolic equivalent of task; PFS: Piper fatigue scale; PSA: prostate‐specific antigen; QoL: quality of life; RCT: randomised controlled trial; ROM: range of motion; SD: standard deviation; VAS: visual analogue scale; VO₂max: maximal oxygen consumption; WHOQOL‐BREF: World Health Organization Quality of Life Questionnaire.

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
ACTRN12618000225213 2018	Expert opinion/study protocol
Adair 2018	Radiotherapy not required
Aghili 2007	Wrong study design (not RCT)
Ahlberg 2011	Wrong study design (not RCT)
Almstedt 2016	Wrong study design (not RCT)
Banerjee 2007	Chemotherapy not excluded
Ben‐Josef 2017	Hormone therapy not excluded
Berglund 2007	Radiotherapy not required
Bertram 2011	Radiotherapy not required
Beurskens 2007	Radiotherapy not required
Box 2002	Radiotherapy not required
Brdareski 2012	Radiotherapy not required
Brown 2006	Chemotherapy not excluded
Brown 2015	Radiotherapy not required
Burnham 2002	Radiotherapy not required
Cadmus 2009	Radiotherapy not required
Cantarero Villanueva 2011	Radiotherapy not required
Chandwani 2010	Chemotherapy not excluded
Chen 2013	Chemotherapy not excluded
Chen 2016	Radiotherapy not required
Cheville 2010	Chemotherapy not excluded
Cho 2006	Radiotherapy not required
Cho 2016	Radiotherapy not required
Chock 2013	Chemotherapy not excluded
Cinar 2008	Radiotherapy not required
Clark 2013	Chemotherapy not excluded
Cormie 2013	Radiotherapy not required
Danhauer 2009	Radiotherapy not required
Desbiens 2017	Radiotherapy not required
DiBlasio 2016	Radiotherapy not required
Dieli‐Conwright 2018	Radiotherapy not required
Dieperink 2013	Hormone therapy not excluded
Dieperink 2017	Hormone therapy not excluded (ADT, Tamoxifen, etc.)
Dimeo 2008	Wrong study design (not RCT)
Dong 2019	Radiotherapy not required
Douglass 2012	Wrong study design (not RCT)
Drouin 2002	Chemotherapy not excluded
Drouin 2005	Chemotherapy not excluded
Ehlers 2018	Radiotherapy not required
Emslie 2007	Radiotherapy not required
Galvao 2013	Hormone therapy not excluded
Grote 2018	Chemotherapy not excluded
Guerreiro Godoy 2010	Wrong study design (not RCT)
Haddad 2011	Chemotherapy not excluded
Haines 2010	Radiotherapy not required
Hajdu 2017	Expert opinion/study protocol
Hayes 2017	Radiotherapy not required
Ho 2014	Chemotherapy not excluded
Ho 2016	Radiotherapy not required
Ho 2018	Chemotherapy not excluded
Hojan 2016	Hormone therapy not excluded
Hojan 2017	Hormone therapy not excluded
Ibrahim 2017	Chemotherapy not excluded
Ibrahim 2018	Chemotherapy not excluded
Irdesel 2007	Wrong study design (not RCT)
Janelsins 2016	Radiotherapy not required
Jazi 2017	Chemotherapy not excluded
Kapur 2010	Hormone therapy not excluded
Kim 2006	Radiotherapy not required
Kneis 2018	Wrong study design (not RCT)
Kyungjin 2014	Wrong study design (not RCT)
Lauridsen 2005	Radiotherapy not required
Leal 2016	Chemotherapy not excluded
Lee 2007	Chemotherapy not excluded
Lee 2019	Radiotherapy not required
Liu 2017	Radiotherapy not required
Mariano 2015	Chemotherapy not excluded
McQuade 2017	Hormone therapy not excluded
Miki 2014	Radiotherapy not required
Milbury 2019	Chemotherapy not excluded
Mina 2014	Radiotherapy not required
Mock 1997	Chemotherapy not excluded
Mohua 2017	Chemotherapy not excluded
Mustian 2004	Radiotherapy not required
Mustian 2006	Chemotherapy not excluded
Mustian 2009	Chemotherapy not excluded
Mutrie 2012	Radiotherapy not required
Odynets 2018	Wrong study design (not RCT)
Oldervoll 2011	Radiotherapy not required
Oliveira 2009	Chemotherapy not excluded
Penttinen 2014	Radiotherapy not required
Pernar 2017	Radiotherapy not required
Pruthi 2012	Radiotherapy not required
Raghavendra 2009	Chemotherapy not excluded
Rahnama 2010	Chemotherapy not excluded
Ratcliff 2016	Chemotherapy not excluded
Reis 2013	Chemotherapy not excluded
Rief 2014a	Chemotherapy not excluded
Rief 2014b	Chemotherapy not excluded
Rief 2014c	Chemotherapy not excluded
Rief 2016	Chemotherapy not excluded
Rogers 2013	Chemotherapy not excluded
Rosenberger 2020	Chemotherapy not excluded
Rummans 2006	Chemotherapy not excluded
Sagen 2009	Radiotherapy not required
Sandel 2005	Radiotherapy not required
Schuler 2017	Radiotherapy not required
Segal 2001	Radiotherapy not required
Segal 2009	Hormone therapy not excluded
Siedentopf 2013	Radiotherapy not required
So 2006	Chemotherapy not excluded
Speck 2010	Radiotherapy not required
Sprave 2019	Chemotherapy not excluded
Sprod 2010	Chemotherapy not excluded
Steindorf 2014	Chemotherapy not excluded
Sweeney 2019	Radiotherapy not required
Taaffe 2017	Radiotherapy not required
Taaffe 2018	Hormone therapy not excluded
Tomasello 2017	Wrong study design (not RCT)
Tome Boisan 2010	Radiotherapy not required
Vadiraja 2009	Chemotherapy not excluded
van der Horst 1985	Radiotherapy not required
VanderWalde 2018	Chemotherapy not excluded
Windsor 2004	Hormone therapy not excluded
Wong 2017	Expert opinion
Yang 2015	Wrong study design (not RCT)
Yennu 2017	Hormone therapy not excluded
Zissiadis 2015	Chemotherapy not excluded

Characteristics of studies awaiting classification [ordered by study ID]

ISRCTN26140710.

Methods	Interventional RCT with an individualised exercise programme for the exercise group (n = 5)
Participants	10 women with diagnosed breast cancer
Interventions	Individualised exercise programme
Outcomes	Not reported
Notes

Milecki 2013.

Methods	RCT with supervised exercise programme for the exercise group (n = 25 or 35)
Participants	46 or 66 women with diagnosed breast cancer
Interventions	Aerobic training including cycling
Outcomes	Physical performance (6‐minute walk test)
Notes

Characteristics of ongoing studies [ordered by study ID]

NCT04506476.

Study name	Trial evaluating the benefit of a fitness tracker based workout during adjuvant radiotherapy of breast cancer (OnkoFit I)
Methods	Single‐centre RCT with 3 study arms.
Participants	Breast cancer patients (estimated n = 201)
Interventions	Experimental group 1: participants receive a fitness tracker, a booklet ('Physical training, exercise and cancer') and an in‐person briefing on physical activity during cancer therapy, with suggested daily step count. Participants receive weekly feedback and a new goal with the aim to reaching 6000 daily steps. Experimental group 2: participants receive a fitness tracker, a booklet ('Physical training, exercise and cancer') and an in‐person briefing on physical activity during cancer therapy, with no recommended daily step count. Participants self‐document their daily step count during radiotherapy. Control: participants receive a booklet ('Physical training, exercise and cancer') and an in‐person briefing on physical activity during cancer therapy, but no fitness tracker or recommended daily step count.
Outcomes	Fatigue syndrome, physical well‐being, social/family well‐being, emotional well‐being, functional well‐being, additional factors, all measured with the FACIT‐F questionnaire 3 months after adjuvant radiotherapy.
Starting date	1 August 1 2020
Contact information	Cihan Gani, MD, PD +49 (0) 7071 29‐82165 cihan.gani@med.uni-tuebingen.de Daniel Zips, MD, Prof +49 (0) 7071 29‐82165 ro-info@med.uni-tuebingen.de
Notes

NCT04507789.

Study name	Exercise therapy during radiotherapy
Methods	RCT with 2 study arms
Participants	People with breast cancer (estimated n = 40)
Interventions	Intervention: special exercise sets for upper extremity problems affecting mastectomy patients. Control: routine radiotherapy protocol
Outcomes	Upper extremity functional status, shoulder joint range of motion, hand grip force measurement, fear of movement, avoidance reaction during radiotherapy protocol, physical activity level, pain and sensory impairment, scapular dyskinesia, upper extremity kinaesthesia and shoulder joint position sensation, lymph oedema, QoL, breast cancer‐specific QoL.
Starting date	October 10, 2020
Contact information	Damlagül Aydin Özcan +90 0312 3052525 damlagulozcan@gmail.com Güçlü Sezai Kılıçoğlu +90 0312 3360909 sezaikilicoglu1@gmail.com
Notes

FACIT‐F: Functional Assessment of Chronic Illness Therapy‐Fatigue; QoL: quality of life; RCT: randomised controlled trial.

Differences between protocol and review

We changed the outcome 'reintegration into work life' to 'return to work'.

Owing to considerable clinical heterogeneity between studies (various types of disease), we did not pool results through meta‐analysis with the fixed‐effect model, and we did not use the random‐effects model as a sensitivity analysis for the primary outcome. We performed analyses according to guidance provided in Chapter 9 of the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2011), and we used the statistical software Review Manager 5 for analysis (Review Manager 2020).

We did not analyse quality components (high risk of bias versus low risk of bias). We did not analyse full‐text publications versus abstract publications.

Had we considered the data sufficiently similar to be combined, we would have estimated treatment effect measures of individual studies as risk ratios (RRs) for dichotomous data and mean differences (MDs) for continuous data, each with their 95% confidence intervals (CIs); we would have assessed heterogeneity of treatment effects between trials and explored potential causes of heterogeneity by sensitivity and subgroup analyses where possible; we would have explored potential reporting bias by generating a funnel plot and statistically testing this by conducting a linear regression test; and we would have considered different characteristics for subgroup analyses (age of included participants; type of radiation therapy; type, duration, and intensity of physical exercise; and cancer type).

Contributions of authors

MT conceived the protocol and the review.
MT, NS, and RR designed the protocol and the review.
MT, NS and FTB coordinated the protocol and the review.
MT, RR, and TN designed the search strategies and screened the literature.
MT wrote the protocol and the review.
SM, NS, JM, ST, MvBB, CB, and FTB provided general advice on the protocol and the review, and provided clinical or methodological expertise and advise.
MT, SM, RR, TN, ST, MvBB, CB, and FTB performed previous work that was the foundation of the current study.

Sources of support

Internal sources

• Department of Radiation Oncology and Cyberknife Center, University Hospital of Cologne, Germany

External sources

• None, Other

Declarations of interest

MT: none known
SM: none known
NS: none known
RR: none known
TN: none known
JM: none known
ST: none known
MvBB: none known
CB: none known
FTB: none known

New