Prospective longitudinal assessment of health-related quality of life in patients with brain metastases undergoing radiation therapy

Abstract

Objectives

We conducted a prospective clinical trial of patients receiving radiation (RT) for brain metastases to identify clinical predictors of pre- and post-RT health-related QoL (hrQoL).

Methods

Patients with brain metastases completed overall (EORTC QLQ C-15-PAL) and brain tumor-specific (QLQ-BN20) hrQoL assessments pre-RT (n=127) and one (n=56) and three (n=45) months post-RT. Linear and proportional-odds models analyzed patient, disease, and treatment predictors of baseline, one-, and three-month hrQoL scores. Generalized estimating equations and repeated measures proportional-odds models assessed predictors of longitudinal hrQoL scores.

Results

Most patients underwent stereotactic radiosurgery (SRS) (69.3%) and had non-small-cell lung (36.0%) metastases. Compared with SRS, receipt of whole brain radiotherapy (WBRT) was associated with a higher odds of appetite loss (baseline p=0.04, one-month p=0.02) and greater motor dysfunction (baseline p=0.01, one-month p=0.003, three-month p=0.02). Receipt of systemic therapy was associated with better emotional functioning after RT (one-month p=0.03, three-month p=0.01). Compared to patients with breast cancer, patients with melanoma had higher odds of better global hrQoL (p=0.01) and less pain (p=0.048), while patients with lung cancer reported lower physical function (p=0.048) three months post-RT. Non-married patients had greater odds of higher global hrQoL (one-month p=0.01), while male patients had lower odds of reporting more hair loss (baseline p=0.03, three-month p=0.045). Patients ≥60 had lower odds of more drowsiness (p=0.04) and pain (p=0.049) over time.

Conclusions

Patients receiving SRS versus WBRT and systemic therapy reported better post-treatment hrQoL. Additionally, melanoma metastases, non-married, male, and older patients reported better hrQoL in various domains after intracranial RT.

Introduction

The incidence of brain metastases is rising, affecting 20 to 40% of all cancer patients¹. Recent advances in tumor detection and systemic therapy have led to considerable improvements in survival among patients with brain metastases². Current management of brain metastases involves a combination of radiotherapy (RT), including both whole brain radiation (WBRT) and stereotactic radiosurgery (SRS), systemic therapy, and surgery³. Clinical trials examining these treatment modalities in the brain metastases population have focused primarily on endpoints such as overall and progression-free survival. However, prospective data on patient-relevant outcomes including health-related quality of life (hrQoL) and functional status after brain metastases treatment are more limited. These have emerged as crucial endpoints in this patient population⁴, particularly given that patients may now survive months to years beyond detection of their intracranial metastases.

HrQoL is a patient-reported outcome that encompasses physical, psychological, and social aspects of well-being. HrQoL outcomes are increasingly being collected and analyzed in clinical trials of cancer patients, as they may guide clinicians, patients, and family members in disease management decisions. Additionally, hrQoL has been linked to survival in studies of patients with colorectal⁵, non-small-cell lung⁶, and pancreatic cancer⁷. Treatment for brain metastases may impact the assorted domains within hrQoL in different ways. For example, though intracranial RT may relieve headaches and seizures⁸, it has also been linked to worsening nausea⁹ and fatigue¹⁰.

Despite growing recognition of hrQoL’s importance in cancer treatment, there remains a lack of data on comprehensive, validated hrQoL outcomes among patients with intracranial metastases after treatment. Most studies evaluating hrQoL after brain RT have focused on only one RT modality^11–14 or have reported hrQoL outcomes at only one post-treatment timepoint¹⁵. Few studies have investigated the impact of systemic therapy on hrQoL among brain metastases patients, and these have included patients with only one primary tumor type^16,17. Importantly, patient demographic characteristics also influence well-being^18–20, yet their impact on post-treatment hrQoL in brain metastases patients have not been explored in depth. In this prospective clinical trial, we sought to understand the demographic and clinical factors influencing hrQoL in various domains at multiple timepoints among 127 brain metastases patients undergoing treatment using comprehensive, well-validated hrQoL questionnaires. Specifically, we assessed hrQoL prior to and at one and three months after intracranial RT completion. We then investigated demographic, tumor, and treatment predictors of hrQoL outcomes at each timepoint and over time. We hypothesized that patients treated with SRS versus WBRT²¹, receiving systemic therapy²², and with breast cancer versus lung²³ or melanoma²⁴ brain metastases would report higher hrQoL outcomes after treatment. We also hypothesized that older, non-married, and female patients²⁵, as well as patients within a lower socioeconomic status²⁶, would report worse hrQoL outcomes after treatment. Our results highlight patient groups that may benefit from targeted management to improve well-being after brain metastases treatment and may guide the development of hrQoL-directed interventions.

Materials and Methods

Procedure

This was a prospective longitudinal clinical trial of hrQoL in patients with metastatic brain tumors receiving intracranial RT at a single institution between August 2013 and July 2017. Participation was offered at first visit with the radiation oncologist. This study was approved by our institutional review board. All enrolled patients provided written informed consent.

Eligible patients completed baseline overall hrQoL measures (QLQ C-15-PAL) and a brain tumor-specific module (QLQ-BN20) one to two weeks prior to the planned RT start date. Patients repeated these questionnaires one and three months post-treatment, mailed to their home through hard copy. Physicians were blinded from patient responses and patient privacy was maintained by assigning random study identification numbers to each participant.

Patients

Inclusion criteria included age >18, life expectancy >3 months, Eastern Cooperative Oncology Group performance status ≤3 or Karnofsky performance status (KPS) >50, fluency in English or Spanish, and a plan to receive RT for brain metastases. Baseline data collected included age, gender, race, KPS, primary tumor histology, RT type (SRS versus WBRT), receipt of systemic therapy, regional poverty level, and regional median income. Regional poverty level and median income were determined using patient residential zip codes.

Measures

The European Organisation for Research and Treatment of Cancer (EORTC) Quality of Life Group QLQ-C15-PAL is a shortened version of the previously developed EORTC QLQ-C30, a well-validated assessment tool designed for patients undergoing cancer therapy that has been widely used for three decades^27,28. The QLQ-C15 PAL was developed specifically for patients undergoing palliative treatment and correlates well with the QLQ-C30^27,28. The QLQ-C15-PAL consists of 15 questions: two multi-item functional scales (physical and emotional functioning), two multi-item symptom scales (fatigue and pain), five single-item symptom scales (nausea/vomiting, dyspnea, insomnia, appetite loss, and constipation), and one global hrQoL scale, producing 10 subscales. Patients rated each question/item on a four-point scale (from 1=not at all to 4=very much), with the exception of global hrQoL, which was rated on a seven-point scale (from 1=very poor to 7=excellent). A greater score on the physical functioning, global hrQoL, and emotional functioning subscales corresponds to higher hrQoL; a greater score on the remaining subscales indicates lower hrQoL. Raw scores were linearly transformed to a 0 to 100 scale using the QLQ-C30 scoring manual²⁹ and the QLQ-C15-PAL scoring addendum from the EORTC Quality of Life Unit. Higher scores for each symptom scale reflect greater symptom burden; however, higher scores on the functional and global hrQoL scales reflect better functioning and higher hrQoL.

EORTC QLQ-Brain Neoplasm (BN)-20 is a module developed specifically for patients treated for intracranial malignancy and often supplements the QLQ-C15-PAL or QLQ-C30^30,31. The QLQ-BN20 consists of 20 questions: seven single-item symptom scales (headaches, seizures, drowsiness, hair loss, itchy skin, leg weakness, and bladder control) and four multi-item scales (future uncertainty, visual disorder, motor dysfunction, and communication deficit). Higher scores (ranging from 0–100), which were linearly transformed from raw scores, reflect greater symptom burden. Differences ≥10 points are considered clinically meaningful³².

Data Analysis

Descriptive statistics were calculated for the cohort. Demographic, tumor, and treatment characteristics assessed included sex, age (<60 versus ≥60), metastatic histology (breast, lung, melanoma, other), RT modality (SRS versus WBRT), KPS (<70 versus ≥70), marital status, race (white versus non-white), income (above versus below median income of the cohort), and regional poverty rate (above versus below median poverty rate of the cohort).

Baseline, one-month, and three-month hrQoL outcomes

HrQoL outcomes graded on a scale of ten or more points were analyzed as continuous variables; the remainder were analyzed as ordinal variables. We used multivariable linear and proportional-odds models for continuous and ordinal outcomes, respectively, to assess clinical predictors of hrQoL scores at baseline. Our sample size at one- and three-months post-RT did not meet criteria for multivariable analysis³³; univariable linear and proportional-odds models assessed predictors of hrQoL at these timepoints. We did also perform a limited multivariable analysis of global hrQoL one and three months post-RT with the following covariates: age²⁵, sex²⁵, radiation modality²¹, and marital status²⁵ to explore potential associations. Univariable models of post-RT hrQoL outcomes additionally included the baseline score as a predictor.

HrQoL outcomes over time

We used univariable generalized estimating equations with an exchangeable working correlation to analyze longitudinal hrQoL data. Models included one- and three-month hrQoL scores as the outcome and visit month, baseline hrQoL score, and an interaction term between visit month and covariates of interest as predictors. We included time, each variable of interest, and their interaction as predictors to analyze the impact of each variable on longitudinal trends in hrQoL scores. Gee and repolr R packages were used for continuous and ordinal outcomes, respectively^34,35.

Models that did not converge are not reported. All statistical tests were two-sided, and we considered p-values less than 0.05 statistically significant. All analyses were performed using R version 3.6.1.

Results

Patient characteristics at each time point are shown in Table 1. Among the 127 patients in our cohort, most were <60 years old (59.8%) and female (57.5%). The most common primary tumor histologies included non-small-cell lung (36.0%), breast (17.6%), and melanoma (16.0%). Most patients received SRS (69.3%) and systemic therapy (75.6%). At one-month post-RT, 56 (44%) patients completed surveys; 45 (35%) completed surveys at three months post-RT.

Table 1.

Clinical and treatment characteristics of patients.

	Baseline n=127	Month 1 n=56	Month 3 n=45
Age
< 60 years	76 (59.8%)	36 (64.3%)	28 (62.2%)
≥60 years	51 (40.2%)	20 (35.7%)	17 (37.8%)
Gender
Female	73 (57.5%)	31 (55.4%)	25 (55.6%)
Male	54 (42.5%)	25 (44.6%)	20 (44.4%)
Race
White	94 (74.0%)	41 (73.2%)	35 (77.8%)
Non-white	32 (25.2%)	15 (26.8%)	10 (22.2%)
Unknown	1 (0.78%)	0 (0.0%)	0 (0.0%)
Metastatic Tumor Origin
NSCLC	45 (36.0%)	16 (29.1%)	16 (35.6%)
Breast	22 (17.6%)	10 (18.2%)	8 (17.8%)
Melanoma	20 (16.0%)	12 (21.8%)	7 (15.6%)
Colorectal	9 (7.2%)	4 (7.3%)	2 (4.4%)
Renal	8 (6.4%)	4 (7.3%)	3 (6.7%)
Ovarian	5 (4.0%)	3 (5.5%)	4 (8.9%)
SCLC	3 (2.4%)	0 (0.0%)	0 (0.0%)
Other*	15 (11.8%)	7 (12.5%)	5 (11.1%)
KPS
< 70	8 (6.3%)	1 (1.8%)	1 (2.2%)
≥70	117 (92.1%)	55 (98.2%)	44 (97.8%)
Unknown	2 (1.6%)	0 (0.0%)	0 (0.0%)
Systemic Therapy
No	31 (24.4%)	14 (25.0%)	10 (22.2%)
Yes	96 (75.6%)	42 (75.0%)	35 (77.8%)
Marital Status
Married	87 (68.5%)	40 (71.4%)	32 (71.1%)
Unmarried	37 (29.1%)	15 (26.8%)	13 (28.9%)
Unknown	3 (2.4%)	1 (1.8%)	0 (0.0%)
Radiation Modality
SRS, fSRS†	88 (69.3%)	42 (75.0%)	37 (82.2%)
WBRT	39 (30.7%)	14 (25.0%)	8 (17.8%)
Regional poverty level §
Below cohort median	58 (45.7%)	27 (48.2%)	22 (48.9%)
Above cohort median	68 (53.5%)	29 (51.8%)	23 (51.1%)
Unknown	1 (0.8%)	0 (0.0%)	0 (0.0%)
Regional median income §
Below cohort median	58 (45.7%)	24 (42.9%)	21 (46.7%)
Above cohort median	68 (53.5%)	32 (57.1%)	24 (53.3%)
Unknown	1 (0.8%)	0 (0.0%)	0 (0.0%)

^*Includes esophageal, anal, endometrial, vaginal, bladder, testicular, leukemia, head/neck, and thyroid

^†SRS, single fraction; fSRS, >1–5 fractions

^§Determined using patient residential zip codes.

Abbreviations: NSCLC, non-small-cell lung cancer; SCLC, small-cell lung cancer; SRS, stereotactic radiosurgery; fSRS, fractionated stereotactic radiosurgery; WBRT, whole brain radiotherapy

Multivariable predictors of hrQoL prior to RT

Mean raw hrQoL subscale scores at baseline are shown in Supplemental Digital Content 1. Multivariable predictors of continuous and ordinal outcomes at baseline are shown in Supplemental Digital Content 2–4. There were multiple factors associated with hrQoL prior to RT. Briefly, patients about to receive WBRT reported worse physical function (β=−24.7, p<0.001) and more motor dysfunction (β=16.4, p=0.01), and had higher odds of reporting more appetite loss (OR=2.71, p=0.04) and leg weakness (OR=3.26, p=0.03) compared to patients about to undergo SRS. Compared to patients with breast cancer, patients with melanoma had lower odds of better global hrQoL (OR=0.12, p=0.01). Male patients had lower odds of reporting more hair loss (OR=0.31, p=0.03).

Univariable predictors of hrQoL at one month post-RT

Univariable predictors of continuous and ordinal hrQoL outcomes at one month post-RT are shown in Figures 1a and 2, respectively (numerical results in Supplemental Digital Content 5, 7, and 8). Compared to SRS, WBRT was associated with worse physical function (β=−21.0, p=0.02), more motor dysfunction (β=21.5, p=0.003), and higher odds of more appetite loss (OR=4.27, p=0.02) and leg weakness (OR=3.49, p=0.04). Patients receiving systemic therapy reported better emotional functioning (OR=4.09, p=0.03). Non-married patients had greater odds of higher global hrQoL (OR=4.91, p=0.01). Men had reduced odds of greater leg weakness (OR=0.27, p=0.04).

Fig. 1.

Univariable patient, tumor, and treatment predictors of continuous hrQoL outcomes at one and three months after radiotherapy One-month

(a) and three-month (b) univariable associations between patient, tumor, and treatment characteristics and continuous hrQoL outcomes. Only significant (p<0.05) associations are shown in color. Blue squares indicate significant positive association, meaning that the variable of interest is associated with a higher hrQoL subscale score; yellow squares indicate significant negative association, meaning that the variable of interest is associated with a lower hrQoL subscale score. Higher future uncertainty, visual disorder, motor dysfunction, and communication deficit subscale scores indicate worse hrQoL. Grey squares indicate no significant association.

Abbreviations: WBRT, whole brain radiotherapy; KPS, Karnofsky performance status

Fig. 2.

Univariable patient, tumor, and treatment predictors of ordinal hrQoL outcomes at one month after radiotherapy

Only significant (p<0.05) associations are shown in color. Dark (OR>5) and light (1<OR≤5) blue squares indicate significantly higher odds of having a higher score; dark (OR≤0.5) and light (0.5<OR≤1) yellow squares indicate significantly reduced odds of having a higher score. Grey squares indicate no significant association.

Abbreviations: WBRT, whole brain radiotherapy; KPS, Karnofsky performance status

Univariable predictors of hrQoL at three months post-RT

Univariable predictors of continuous and ordinal hrQoL outcomes at 3 months are shown in Figures 1b and 3, respectively (numerical results in Supplemental Digital Content 6, 9, and 10). Compared to SRS, WBRT was associated with worse physical function (β=−31.2, p=0.01), greater motor dysfunction (β=15.4, p=0.02), and higher odds of itchy skin (OR=5.79, p=0.04). Patients with melanoma had higher odds of better global hrQoL (OR=17.5, p=0.01) and lower odds of more pain (OR=0.07, p=0.048) compared to patients with breast cancer. Patients with lung cancer had worse physical functioning (β=−20.6, p=0.048), lower odds of higher emotional functioning (OR=0.04, p=0.02), and greater odds of more drowsiness (OR=7.08, p=0.046) compared to patients with breast cancer. Patients receiving systemic therapy had higher odds of better emotional functioning (OR=8.53, p=0.01). White patients reported less motor dysfunction (β=−11.6, p=0.048). Male patients had lower odds of more hair loss (OR=0.24, p=0.045).

Fig. 3.

Univariable patient, tumor, and treatment predictors of ordinal hrQoL outcomes at three months after radiotherapy

Only significant (p<0.05) associations are shown in color. Dark (OR>5) and light (1<OR≤5) blue squares indicate significantly higher odds of having a higher score; dark (OR≤0.5) and light (0.5<OR≤1) yellow squares indicate significantly reduced odds of having a higher score. Grey squares indicate no significant association.

Abbreviations: WBRT, whole brain radiotherapy; KPS, Karnofsky performance status

Multivariable predictors of global hrQoL one and three months post-RT

Results of the multivariable analysis of global hrQoL one and three months post-RT are demonstrated in Table 2. There were no significant associations between global post-RT global hrQoL and any of the covariates included in the analysis.

Table 2.

Multivariable association between clinical and treatment factors and global hrQoL at one and three months post-RT

Characteristic	1 month post-RT			3 months post-RT
	OR	95% CI	p-value	OR	95% CI	p-value
Age			0.96			0.43
<60 years	Ref	-		Ref	-
≥60 years	0.63	−2.30–4.36		0.62	−2.94–4.19
Sex			0.44			0.24
Female	Ref	-		Ref	-
Male	1.03	−2.92–4.19		0.49	−3.08–4.06
Radiation modality			0.12			0.10
SRS, fSRS	Ref	-		Ref	-
WBRT	0.35	−3.53–4.22		0.29	−3.91–4.48
Marital status			0.10			0.85
Married	Ref	-		Ref	-
Unmarried	2.63	−0.87–6.12		0.89	−2.80–4.58

Univariable predictors of hrQoL outcomes over time

Significant univariable patient, tumor, and treatment predictors of longitudinal hrQoL outcomes are demonstrated in Figure 4 (numerical results in Supplemental Digital Content 11–13). Patients ≥60 years of age had lower odds of worsening drowsiness (OR=0.22, p=0.04) and pain (OR=0.38, p=0.049) over time. Patients with cancers other than breast, lung, or melanoma had higher odds of worsening drowsiness over time compared to patients with breast cancer (OR=10.5, p=0.02). White patients had higher odds of worsening fatigue over time (OR=7.92, p=0.001). WBRT patients had higher odds of more insomnia (OR=10.1, p<0.001) and pain (OR=3.63, p=0.01) over time. Male patients (OR=0.28, p=0.04) had lower odds of more pain over time. Lower socioeconomic status was associated with improved bladder control over time (median income greater than cohort median OR=0.18, p=0.03; regional poverty level greater than cohort median OR=5.35, p=0.04).

Fig. 4.

Univariable patient, tumor, and treatment predictors of hrQoL outcomes over time

Only significant (p<0.05) associations are shown in color. Dark (OR>0.5) and light (1<OR≤5) blue squares indicate significantly higher odds of having a higher score over time; dark (OR≤0.5) and light (0.5<OR≤1) yellow squares indicate significantly reduced odds of having a higher score over time. Grey squares indicate no significant association over time. There were no significant predictors of longitudinal continuous hrQoL outcomes.

Abbreviations: WBRT, whole brain radiotherapy; KPS, Karnofsky performance status

*Includes esophageal, anal, endometrial, vaginal, bladder, testicular, leukemia, head/neck, and thyroid

Discussion

Brain metastases comprise the most common intracranial tumor in adults and are increasing in prevalence due to earlier detection and improved systemic therapy¹. These advancements have also heralded marked improvements in survival². As such, many clinicians and researchers have focused increasingly on patient-reported outcomes, including hrQoL, in the management of these patients⁴. In this prospective clinical trial, we analyzed demographic and clinical predictors of hrQoL in patients with brain metastases prior to and up to three months after RT using comprehensive, well-validated hrQoL questionnaires. We found that SRS correlated with better hrQoL in multiple domains compared to WBRT. Moreover, numerous additional factors, including age, sex, and primary tumor histology, among others, influenced pre- and post-treatment hrQoL.

The addition of WBRT to SRS has been linked to greater cognitive decline than SRS alone in brain metastases management¹¹; however, few studies have explored longitudinal non-cognitive hrQoL outcomes in depth following treatment with each of these modalities. We found that patients about to receive SRS reported better hrQoL in multiple domains at baseline, including physical functioning and appetite loss. This likely reflects reduced intracranial disease burden among patients selected for SRS. Importantly, the higher hrQoL among SRS patients in our cohort persisted at one and three months post-RT. It is unclear whether this is secondary to lower disease burden at baseline or the hrQoL-sparing effects of SRS itself. Indeed, in a randomized trial of patients with one to three brain metastases, SRS was associated with better overall, physical, functional, and brain tumor-specific hrQoL three months after RT compared to combined WBRT and SRS¹¹. For this reason, SRS has gained momentum as an alternative to WBRT in patients with up to ten brain metastases³⁶. Several randomized controlled trials comparing hrQoL outcomes after WBRT and SRS in brain metastases patients are currently underway^37,38, which may help clarify whether SRS promotes preservation of post-treatment hrQoL independent of intracranial metastasis burden.

One prior prospective, observational study comparing hrQoL among patients treated with WBRT versus SRS found that WBRT patients had worse fatigue, appetite loss, and future uncertainty than SRS patients; however, the authors excluded patients who received chemotherapy from their analysis¹⁵. Systemic therapy is frequently used in combination with RT in the management of brain metastases and extraneural systemic disease, and may carry significant ramifications for tumor response and patient wellbeing³⁹. Indeed, patients in our cohort receiving systemic therapy reported better emotional functioning after RT. This may be attributed to a greater sense of agency and hope among patients for whom active treatment is still an option⁴⁰. This could also reflect improved tumor response with systemic therapy, which has been associated with greater hrQoL in patients with metastatic breast, gastrointestinal, and lung cancer⁴¹.

We also found that metastatic tumor histology had implications for post-treatment hrQoL, likely related to patient subgroups affected by each cancer, the toxicities of disease-dependent systemic therapies, and the natural history of specific cancer types. Specifically, patients with intracranial lung metastases reported lower physical and emotional functioning and more drowsiness after RT compared with metastatic breast cancer patients. Metastatic lung cancer portends a high symptom burden with severe consequences for physical and emotional hrQoL⁴². Despite significantly reduced hrQoL at baseline, patients with melanoma displayed higher global hrQoL and less pain than patients with breast cancer after RT, contrary to our hypothesis. Immunotherapy is now standard of care for metastatic melanoma, and has been associated with better global hrQoL compared to chemotherapy⁴³. Meanwhile, chemotherapy and hormone therapy remain first-line treatments for metastatic breast cancer, which may explain the relatively lower hrQoL observed among these patients.

We also identified several patient demographic characteristics that correlated with post-RT well-being. While living with a spouse is associated with improved survival among primary and metastatic brain tumor patients⁴⁴, we found that married patients displayed lower global hrQoL than their non-married counterparts after RT. Marriage is consistently used as a proxy for social support in the cancer literature^45,46. Yet, unmarried adults often have more social ties, closer relationships with family, and greater community involvement than do married patients⁴⁷. Moreover, the responsibility of care among married patients often falls predominantly upon the spouse, which may contribute to greater self-perceived burden and relationship strain⁴⁸. We also found that men reported less leg weakness and hair loss following RT. While we do not expect cancer treatment to differentially affect male and female hair loss, females may have been more likely to report hair loss than males. Cancer treatment-related hair loss has been associated with anxiety, depression, and loss of sense of self, particularly among women⁴⁹. Non-white patients reported more motor dysfunction after RT. This suggests that supportive resources such as physical and occupational therapy may not have been accessed equally by all patients in our cohort, and that efforts to optimize physical functioning in non-white patients are critical.

Finally, we found that several demographic, tumor, and treatment characteristics portended a greater rate of hrQoL change after treatment. Consistent with prior studies^50,51, patients ≥60 years of age were less likely to report more pain and drowsiness over time. Older patients may view their health in comparison with that of similar-aged peers, leading to improved self-reported hrQoL. Additionally, the impact of brain metastases and treatment on well-being may be tempered by different demands and expectations in older patients. Male patients were less likely to report more pain over time. The higher prevalence and severity of pain in female patients has been well-documented in the cancer literature^52,53; however, it is unclear if this is because women seek out health care services more frequently⁵⁴, are more sensitive to pain⁵⁵, or are less likely to be adequately treated for pain⁵⁶ than men. Patients who underwent WBRT were more likely to report more pain and insomnia over time than SRS patients. Again, this likely reflects a greater disease burden among patients receiving WBRT. WBRT patients also often receive higher doses of steroids to control RT-mediated neuroinflammation, which may precipitate insomnia⁵⁷.

This study has several limitations. Our sample size declined at post-RT timepoints, which may have introduced participant bias. Attrition is expected in a cohort of metastatic brain tumor patients given their limited life expectancy⁵⁸, especially as post-treatment assessments were mailed to participants’ homes, and not performed in clinic. Certainly, participants with poorer performance status or who passed away would not have been captured longitudinally. This limited our ability to assess predictors of post-treatment hrQoL on multivariable analysis. Nevertheless, we were able to analyze changes in hrQoL outcomes over time in this cohort. While we identified multiple predictors of longitudinal hrQoL in this cohort, we lacked information on the number, volume, and location of brain metastases in the WBRT cohort, specific systemic therapies used, and burden of systemic disease for patients in this cohort, all of which may impact the hrQoL outcomes assessed in this study. Additionally, KPS has been found to be associated with hrQoL in brain metastases patients³⁰. As in prior studies^59,60, we used a cutoff KPS score of 70 to group patients in this study; however, our cohort included only one patient with a KPS score <70 at the one- and three-month post-RT timepoints. We were thus unable to include KPS in our post-RT analysis. Large-scale randomized trials that stratify patients by clinical and treatment characteristics are needed to further characterize hrQoL outcomes following radiotherapy for intracranial metastases.

In summary, we prospectively identified patient, tumor, and treatment characteristics that predicted pre- and post-RT hrQoL among 127 patients with intracranial metastases. These data suggest that SRS may prevent worsening of hrQoL deficits present at baseline. Furthermore, we identified gender, race, metastatic histology, systemic therapy, and marital status as important factors that influence hrQoL outcomes. As therapeutic advances yield longer survival times in brain metastases patients, randomized trials are needed to better elucidate post-treatment hrQoL outcomes and the factors that influence them. Ultimately, this will better inform the management of brain metastases and the development of interventions aimed at improving well-being in this patient population.