Cancer-Associated Cachexia: Bridging Clinical Findings with Mechanistic Insights in Human Studies

Abstract

Cancer-associated cachexia (CAC) is a chronic wasting disease typically associated with advanced cancer, resulting in progressive and debilitating loss of function and poor tolerance to anti-cancer therapy. Preclinical animal models have identified various potential mechanisms and mediators, which have had limited translational success in clinical trials. This review focuses on human studies and discusses the clinical phenotyping of CAC using imaging-derived body composition, quality-of-life and functional measures, existing evidence for mediators, current therapeutic options, and future directions to advance the field. Identifying mechanisms driving CAC and targeting them is expected to improve the quality of life, treatment efficacy, and survival.

Introduction

First described in modern medical literature in the 1850s as “cancerous cachexia” (1), cancer-associated cachexia (CAC) is now recognised as a complex and debilitating wasting syndrome associated with underlying malignant disease that induces systemic inflammation and metabolic alterations, leading to involuntary progressive muscle and/or fat tissue loss, anorexia and declining performance status that cannot be completely reversed by conventional nutritional support (2,3) (Fig. 1). It is associated with poor survival rates, reduced anti-cancer therapy efficacy and tolerance, fatigue, lower quality of life (QoL), and prolonged hospitalisations, contributing to increased healthcare costs (3–6). Despite its clinical significance, the term CAC is not widely recognised among patients and caregivers. While clinicians may be familiar with the term, there can be variability in its definition and, therefore, diagnosis (7).

Figure 1.

Interdependent systemic and organ-level consequences of CAC. Cancer induces host response involving inflammation and metabolic dysregulation, resulting in reduced appetite, enhanced metabolic alterations, altered gut microbiome and organ wasting. Created in BioRender. Koh, K. (2025) https://BioRender.com/c10m987

Epidemiologic studies have traditionally defined CAC by the hallmark characteristic of unintentional weight loss, which has shaped our understanding of its prevalence, impact, and association with various diseases, particularly in cancer and metabolic disorders (8–15). However, unintentional weight loss is caused by many physiologic processes that lead to a systemic metabolic imbalance, including treatment-induced anorexia, mechanical obstruction in the gastrointestinal tract, malabsorption, and increased metabolic demands (12,16–21). These processes may differ based on tumour tissue type and treatment modalities, highlighting the complexity of the syndrome and suggesting the potential need for different treatment interventions. Furthermore, CAC may not always present in the classical form of a severely wasted individual but may develop in obese patients with ongoing loss of muscle (22–25). Thus, despite its longstanding recognition in medicine, this constellation of symptoms can make CAC challenging to formally define as a whole, and as a result, effective strategies for intervention are lacking.

While definitions have remained anchored in clinical phenotyping (2,3), much of the work identifying CAC mediators has been conducted in animal models, with limited translational success. Although preclinical models replicate certain aspects of CAC in humans, they frequently involve tumour cells injected subcutaneously in mice with tumours disproportionately large relative to body size compared to human disease, induce levels of inflammation far in excess of those observed in humans, and fail to capture the intricate interplay between metastases, tumour microenvironment, and host responses (26–28). For example, the most commonly used mouse models of CAC, the LLC (Lewis lung carcinoma ) and C26 (Colon-26 carcinoma) allografts, result in exceedingly large tumours (>5% of body weight) and high levels of interleukin(IL)-6 (IL-6) (>600 pg/mL), respectively, which do not match the observed clinical phenotypes (26,28,29). Genetically engineered mouse models (GEMMs) of cancer where tumours arise spontaneously with a physiological microenvironment and tumour burden relative to body size may represent a superior opportunity to address CAC; however, these models are also limited (27,30). For example, GEMMs develop multiple primary tumours consisting of different mutations, heterogeneity of the CAC phenotype, and organ dysfunction due to tumour burden.

There is a critical juncture in CAC research; confronting a condition with a broad and heterogeneous clinical presentation, a complex and likely multifactorial biological origin, and a host (the patient) whose genotype, cancer response, comorbidities, and ethnicity may all shape disease progression and outcomes. To synthesise current knowledge, this review provides an overview of human studies that define the clinical features of CAC, explore its underlying mechanisms, and evaluate therapeutic strategies aimed at its management. While cachexia is also associated with other chronic conditions such as infectious disease, heart and kidney failure (13), this review focuses specifically on the clinical findings and mechanistic insights from human studies in CAC.

Epidemiology

CAC has been associated with poor survival across various cancer types, including colorectal, pancreatic, breast, liver, haematological, and lung cancer (9,19,20,31–34). The prevalence and severity of CAC varies according to cancer type and stage, with lower rates observed in blood cancers (e.g. favourable non-Hodgkin’s lymphoma and acute myeloid leukaemia), breast cancer and prostate cancer, ranging from 15%-30%. Higher incidence is observed in lung cancer, gastric cancer and pancreatic cancer, ranging from 40% in lung to as high as 80% in pancreatic cancer (9,14). Upper gastrointestinal cancers have been shown to exhibit the highest degree of weight loss compared to other cancers, such as breast cancer (9,35). While CAC can manifest in early stage disease, such as gastrointestinal cancers (12,20), it is typically associated with late stage disease across various cancer types (12,14,34,35). In upper gastrointestinal and lung cancers, higher disease stage has been associated with increased weight loss and poorer survival (12,34,36). There may also be sexual dimorphism underlying CAC since the loss of muscle mass in males but not females has been observed in various advanced cancer types (37). The inconsistent definitions used across studies and the challenge of distinguishing CAC from other conditions, such as sarcopenia, which when present at diagnosis alongside muscle loss, has been associated with worsened survival compared to muscle loss alone (38), likely impedes the accurate determination of the prevalence and severity of CAC, and its impact on survival.

Diagnostic and staging criteria of CAC

Over the past two decades, various expert groups have proposed definitions incorporating clinical, biological, and radiological parameters (Table 1). While there is broad agreement on the descriptive definition of CAC, the quantitative diagnostic and staging criteria vary in their emphasis on weight loss thresholds, body composition, metabolic and inflammatory markers, and functional impairment, reflecting the diverse ways in which CAC can present in different patient populations. The Fearon consensus criterion remains among the most widely used diagnostic criteria, likely due to its simplicity and ease of application since it is largely based on weight loss thresholds (3). Evans’s criterion is based on weight loss and some features of CAC, such as fatigue and abnormal biochemistry (2). However, the criterion is not cancer-specific and includes other chronic illnesses, though the authors proposed a shorter timeframe (3–6 months) for weight loss in patients with cancer. The staging of CAC is conceptual and is intended to guide clinical treatment decisions and trial stratification. Zhou et al. (39) and Silva et al. (40) proposed staging criteria for advanced stage disease, whereas Argilés et al. (41) and Bozzetti et al. (42) proposed criteria that are applicable across all stages. Argilés et al.’s staging tool incorporates loss of lean body mass (LBM) and broader metabolic and inflammatory changes, while Silva et al.’s only relies on serum albumin and CRP.

Table 1.

CAC definitions and classifications, ordered in chronological order. BMI, body mass index; SMI, skeletal muscle index; CRP, C-reactive protein; IL-6, Interleukin-6; Hb, haemoglobin; LBM, lean body mass; DEXA, dual X-ray absorptiometry; BIA, bioelectrical impedance analysis; ECOG PS, Eastern Cooperative Oncology Group performance status

Publication	Criteria
Cachexia: a new definition (2)	Weight loss ≥5% within the previous 12 months or BMI <20kg/m2 and at least three of the following criteria: decreased muscle strength, fatigue, anorexia, low fat free mass index, and abnormal biochemistry (increased inflammatory markers CRP>5mg/L, or IL-6>4pg/ml or anaemia Hb <12g/dL or low serum albumin <3.2g/dL)
Defining and Classifying Cancer Cachexia: A Proposal by the SCRINIO Working Group (42)	4 classes based on weight loss (<10% pre-cachexia, ≥10% cachexia) and the presence of at least 1 symptom of anorexia, fatigue, or early satiation
Definition and classification of cancer cachexia: an international consensus (3)	>5% weight loss in the previous 6 months, or >2% weight loss in patients already showing depletion (BMI <20kg/m2, or SMI <7.26kg/m2 in males, and SMI < 5.45 kg/m2 in females
The cachexia score (CASCO) (41)	A CAC staging score from 0–100 that uses weight loss, LBM (DEXA or BIA), inflammatory/metabolic biomarkers, physical function tests, and patient-reported outcomes
A clinically applicable score to classify cachexia stages in advanced cancer patients (39)	A cachexia staging score between 0–12 with five components: weight loss, a questionnaire of sarcopenia, functional assessment by the ECOG PS score, appetite loss, and abnormal biochemistry (white blood cell count, albumin, haemoglobin)
Modified Glasgow Prognostic Score (40)	A score from 0–2 with four stages of CAC based on serum albumin and CRP

There is lack of a unifying diagnostic criterion for CAC. With the numerous criteria available, the exact diagnostic definition used to define CAC is inconsistent between studies (14). Furthermore, despite emerging evidence on the prognostic relevance of adipose tissue loss in lung and pancreatic cancers (31,32), none incorporate adipose tissue loss thresholds, and a universally accepted biomarker is lacking. A unifying criterion may incorporate weight, muscle and adipose tissue loss thresholds, along with inflammatory or metabolic features and may evolve as mechanisms of CAC are elucidated. It is also important to acknowledge that defining CAC thus far has been an iterative process, progressing from the clinical phenotype to an understanding of its underlying biological mechanisms. In CAC, where the clinical presentation is highly heterogeneous and likely represents a spectrum of conditions, a single well-defined clinical phenotype may not be particularly useful in refining disease classification and guiding therapeutic strategies. Perhaps the best approach is to perform deep phenotyping of CAC manifestation in patients to define subtypes representing the spectrum of conditions rather than trying to define it as a single condition (NCT06073431).

Clinical phenotyping of CAC

Given that CAC is a complex syndrome potentially related to multiple clinical factors including anorexia and fatigue, a holistic approach is required to comprehensively phenotype the condition. However, much of the research to date on CAC has overlooked detailed body composition change, including adipose and muscle tissue loss; and relied on weight loss, which remains the predominant metric used to identify CAC in patient studies and is often used as an endpoint in clinical trials (43,44). Various methods of phenotyping have been described in the literature. The following section focuses on the most common approaches used to identify the clinical phenotype of CAC using body composition (including weight), functional assessments, and QoL measures.

Imaging modalities to assess body composition

Body composition refers to the relative proportions of different tissue compartments, including muscle, fat, and bone. Muscle and fat are key determinants of metabolic health, physical function, and disease risk (45–51). Furthermore, body weight and body composition are strong predictors of mortality in patients with cancer, particularly in advanced disease (19,20,31–34,52). Although low baseline body mass index (BMI) and weight loss are consistently associated with worse survival and may capture loss of body compartments, they do not differentiate the contributions of individual muscle and adipose tissue depots to weight loss (43), both of which independently predict poor outcomes (19,20,31,33,53). Moreover, weight stability may obscure the loss of body compartments, especially in cases of divergent changes such as sarcopenic obesity, where increases in adipose tissue mask muscle loss, potentially leading to the misidentification of CAC (22–25).

Since the mid-2000s, advancements in imaging modalities have shifted the focus from merely determining whether weight loss has occurred to understanding and distinguishing the underlying changes in body composition that contribute to it (23,43,54). Body composition is not routinely measured in clinical practice; however, several modalities exist for this purpose. Imaging modalities such as magnetic resonance imaging (MRI) and computed tomography (CT) are commonly used reference standards for body composition assessment (the gold standard being cadaver measurements from dissections) as they allow for the discrimination and precise quantification of subcutaneous adipose tissue (SAT), visceral adipose tissue (VAT), skeletal muscle (SKM), intermuscular adipose tissue and organs (43,54) (Fig. 2A). Furthermore, CT can provide proxy measures of muscle function by measuring muscle attenuation, where lower muscle attenuation is associated with higher lipid infiltration, reduced muscle strength and poor survival (19,55,56). CT- and MRI-derived measurements have been shown to strongly correlate with each other and with cadaver estimates (57,58).

Figure 2.

Applications of CT-derived body composition in CAC research. (A) CT-derived measurements of SAT, VAT and SKM at the L3 vertebral level, illustrating change of body compartments over time. (B) Potential clinical subtypes of CAC with predominant or isolated loss of SAT, VAT, or SKM. (C) Impact of treatment on body composition, where treatment may contribute to loss of body compartments. (D) Association of tumour burden and body composition, where body compartments may deteriorate as tumour volume increases. Created in BioRender. Koh, K. (2025) https://BioRender.com/c10m987

Body composition can also be estimated by measuring compartments such as total body lean mass (also known as fat-free mass), which excludes fat mass, but includes skeletal muscle, bone, water, and other organs such as heart, liver, skin and connective tissues (59). LBM can be estimated using air displacement (60), hydrostatic weighing (61), dual-energy X-ray absorptiometry (DEXA) (62), or by measuring electrical impedance by bioimpedance analysis (63). However, unlike CT and MRI, these techniques are unable to differentiate between SAT, VAT, and SKM. Furthermore, while lean mass from DEXA strongly correlates with MRI-derived SKM volume at a single time point, there is only a modest correlation when assessing the percentage change over time (64). Patients with cancer routinely undergo CT imaging as part of the standard of care in oncology. The opportunistic use of CT imaging to study CAC has previously been proposed as it poses no additional cost or burden to patients (43). Consequently, although MRI avoids radiation exposure and has higher soft tissue resolution, leveraging routine CT scans to quantify body compartments has emerged as the predominant imaging modality used to study CAC (23,43,54). Ultimately, the choice between DEXA, MRI and CT depends on various factors, including burden to patients, cost and study objectives.

Since whole-body CT is not routinely performed for all cancer types, body composition analysis in the context of CAC is often implemented on cross-sectional, single-slice images (17,23,32,43,65). The L3 vertebra landmark is commonly used and has been shown to correlate strongly with whole-body muscle and adipose tissue measurements (66). Furthermore, with advances in artificial intelligence technologies, the identification and quantification of body composition, termed segmentation, can now be automated and has been validated to have strong concordance with measurements derived from manual segmentation, enabling large-scale and multi-vertebral analysis of body composition (67,68).

Challenges in the clinical implementation of body composition loss thresholds

Despite advances in body composition analysis, there are challenges in implementing these metrics clinically. Firstly, there is an absence of CT-based reference values from multiethnic and multinational populations to identify patients outside the normal range of body composition values and at risk of health consequences. Existing abdominal and thoracic CT-based age- and sex-specific reference values are primarily derived from White non-Hispanic outpatients (69) and study participants of the Framingham Heart Study (70), limiting applications to broader demographics.

Secondly, there is no consensus on adipose tissue and SKM loss thresholds that determine clinically significant deviations in body composition for diagnosing CAC. For example, the thresholds for SKM loss ranged from any loss to 14% loss (71). Cross-sectional reference values do not capture intra-person changes in body composition and represent a limitation when used to assess CAC. Additionally, there is a lack of standardisation in the assessment and reporting of adipose tissue and SKM changes across studies, highlighting variations in factors such as the reporting of the exact time interval between scans, baseline body composition by sex, and the use of multiple body composition change metrics (cm²/m², cm² and %) (71). Thus, there is a need to establish a framework that outlines specific requirements to standardise findings from CT-derived body composition analysis.

Thirdly, precision testing is needed to ensure the repeatability of measurements and distinguish true biological changes from measurement errors using the least significant change (72). In a similar vein, a unified approach is needed to consider factors that can introduce errors in body composition measurements and lead to the misidentification of CAC, such as the quality of automated body composition segmentation, as well as imaging protocols and acquisition parameters (e.g. slice thickness, IV contrast administration and tube current) that often vary across clinical centres (67,73–76). For example, at the L3 vertebral level, thicker image slices increase SKM area by 1% and decrease muscle attenuation by 12% (73), whereas thinner slices increase VAT area by 3% and intermuscular adipose tissue area by 17% (75). Furthermore, scans from patients who are severely wasted or with metal implants deteriorate segmentation quality and should either be excluded or manually corrected (67,74). Thus, future studies should consider these technical variables when analysing body composition, particularly when examining intra-patient changes in CAC. At present, while the loss of adipose tissue and SKM has been associated with poor patient outcomes, the challenges discussed have hindered the integration of body composition loss thresholds in the clinical decision process and as a trial stratification or outcome instrument.

Tracking temporal changes in body composition

Longitudinal body composition analysis is required to ascertain temporal patterns of change during the cancer disease course when CAC may develop, which may be important for early detection of and risk stratification for CAC to help inform clinical decision-making. Focusing on change over time, rather than solely on the absolute amount of tissue mass, is critical, as individuals may begin with varying levels of muscle and fat, influencing how their bodies respond to disease progression and treatment. In patients with pancreatic cancer, CAC has been shown to develop early in the disease course, where the median time to CAC (defined as >5% SKM and/or adipose tissue loss (in the absence of >5% gain in either tissue type) from initial CT scan to final CT scan) onset was approximately 2 months after diagnosis (21). In addition, the loss of SKM, VAT and SAT has been shown to precede clinical diagnosis of pancreatic cancer, suggesting the existence of a pre-CAC phase that can be detected with body composition change (50). In several cancer types, during disease progression in the months prior to death, the rate of body compartment loss has been shown to rapidly increase (17,77,78), with a less likelihood for gain in SKM and adipose tissue, suggesting the need to intervene early in the disease course to enable timely intervention (78). In addition to early diagnosis and management of CAC, understanding the temporal patterns of body composition may also improve palliative care for patients with CAC. Patients with solid gastrointestinal cancer who were referred for palliative care have been reported to experience greater body fat loss than lean tissue loss (62). Moreover, higher whole-body fat, but not lean tissue, was a significant predictor of improved survival. Similarly, higher SAT at diagnosis was associated with improved survival after chemotherapy in female patients only (33), emphasising the importance of investigating the sexual dimorphism, aetiology and consequences of fat loss and gain in CAC.

Despite these advances in body composition analyses, studies have often been performed in retrospective cohorts and lacked granularity in either clinical phenotype, cancer treatments or outcomes (20,31,33,34,48,51,79,80). The prospective UK-wide TRACERx (TRAcking Cancer Evolution through therapy (Rx)) study of operable non-small cell lung cancer (NSCLC) (81–83) highlighted the use of large-scale body composition analysis in the context of CAC (32). Body composition analyses between diagnosis and first relapse after primary surgery identified thresholds for defining CAC as at least one of the following criteria: >10% SKM loss, >20% SAT loss, >20% VAT loss, or grade 4 BMI-adjusted weight loss (32). Jin and colleagues are the first to utilise latent longitudinal trajectories of body composition changes to identify CAC (34). In a cohort of patients with liver cancer, two distinct trajectories for SKM and total adipose tissue area were defined: stable and sharp-falling, using 2138 body composition measurements across all time points from diagnosis through follow-up. Patients with the sharp-falling trajectory for either SKM or total adipose tissue were defined as having CAC and worse overall survival, progression-free survival, and response to treatment.

Clinical CAC subtypes defined by distinct body composition

Given the different patterns of body composition change demonstrated in studies to date, clinical CAC subtypes where patients experience a loss in an isolated or predominant body compartment may exist (Fig. 2B). Such clinical subtypes have previously been proposed in patients with lung and pancreatic cancer. In the TRACERx study, a majority of patients with NSCLC demonstrated loss of SAT, VAT and SKM between diagnosis and first relapse (32). However, some patients experienced a loss in one body compartment only. In another study of patients with advanced pancreatic cancer on combined chemotherapy, two CAC phenotypes were identified – ‘adipose tissue loss only’ and ‘adipose and SKM tissue loss’, alongside a group of patients with no tissue loss, independent of tumour response, sarcopenia or obesity (21). While the initial amount of adipose tissue and SKM can influence the extent of loss, there was no difference in the sarcopenia and obesity status between identified subtypes. However, sarcopenic obesity was associated with the adipose tissue loss only CAC phenotype, and both CAC phenotypes were associated with worse survival. More research is needed to support the existence of distinct clinical CAC subtypes, which may be driven by specific underlying biological mechanisms and require bespoke treatment approaches.

In the TRACERx study, patients who experienced predominant loss in an isolated body compartment (e.g. >10% SKM loss or <10% SKM loss) demonstrated distinct differential gene expression profiles in primary tumours (32). Sun and colleagues demonstrated distinct metabolic alterations in SAT, VAT, SKM and liver of patients with gastrointestinal cancers and >5% weight loss compared to patients with <5% weight loss (84). In patients with CAC, a greater number of metabolites were elevated in SAT, VAT and liver, while more metabolites were decreased in SKM. Metabolic pathway alterations specific to each organ were also observed, with SAT, VAT and liver exhibiting a higher number of altered metabolic pathways compared to SKM in patients with CAC. These findings suggest that metabolic activity is altered in the liver and adipose tissue but less so in SKM and further support the existence of distinct mediators underlying SKM and adipose tissue wasting.

Quality of life and physical function measures

Patient-reported outcomes (PROs), clinician-reported outcomes (CROs) and objective functional assessments can be used to assess physical function and overall QoL in clinical trials (85–89). The most common PROs used in the assessment of CAC include the Functional Assessment of Anorexia/Cachexia Therapy (FAACT) and the European Organisation for Research and Treatment of Cancer (EORTC QLQ-C30) (90). These questionnaires evaluate features of general QoL, including physical, social, emotional, and functional well-being, fatigue and pain (90–92). FAACT has a cachexia-specific component, the Anorexia/Cachexia Subscale (A/CS), used to assess symptoms of anorexia, such as loss of appetite and stress related to food intake (92). Patient-Generated Subjective Global Assessment (PG-SGA) is recommended as a malnutritional screening tool for CAC in clinical practice, and a worse PG-SGA score has been shown to correlate with the presence of CAC (93,94). PG-SGA contains a patient-reported component that assesses food intake and nutrition, and a physical examination component evaluated by professionals (95). PROs are associated with different body compartments, where higher SKM area is associated with improved physical function scores, and higher intermuscular adipose tissue is associated with increased fatigue (51). Additionally, PROs are independent predictors of overall survival across various cancer types (96). Although the association between PROs and objective physical function measures is unclear, PROs offer patient-centred QoL measures that can supplement CROs and objective physical function measures (90,97).

In the context of CAC, physical function can deteriorate without the loss of lean mass, highlighting the importance of incorporating functional measures when phenotyping CAC (49). CROs such as Karnofsky Performance Score (KPS) and the Eastern Cooperative Oncology Group Performance Status Scale (ECOG-PS) are commonly used to assess physical function in patients with cancer and can predict survival and treatment outcomes (91,98). In addition, objective measures of physical function such as handgrip strength (HGS), stairs climb, and 6-minute walk tests (6MWT) are widely used and overcome limitations of PROs and CROs such as clinician or patient bias and subjectivity (97–99). In a cohort of patients with colorectal or lung cancer, patients with CAC exhibited poorer ECOG scores and weaker HGS (56).

The wide range of physical function measures and their inconsistent use make it challenging to compare and interpret findings across clinical trials. In a recent systematic review, HGS was found to be the most commonly used physical function endpoint and showed statistically significant differences between intervention and control groups in 12 (36%) of clinical trials (98,100). Reduced HGS has been associated with poor survival in patients with CAC in European-based and Chinese populations (101,102) and slow walking speed in elderly Indian population (103). In trials where HGS was not statistically significant, 6MWT, ECOG-PS and KPS show more frequent statistical differences (98), which may indicate higher sensitivity to physical changes compared to HGS. However, statistical significance does not necessarily translate to clinical importance and the relative sensitivities to changes in physical function of these measures have not been assessed. Digital health technologies (DHTs) such as wearable trackers are an emerging approach to measuring real-world physical behaviour as patients conduct their daily activities (98,104). Unlike CROs, PROs, and clinical tests, DHTs are continuous and objective and can measure physical changes such as walking distance and sleep, which are meaningful to patients (105). Taken together, there is a need for consensus on the most appropriate endpoint for different mechanisms of action being targeted, as well as clinically meaningful physical measures to use in CAC studies.

Interplay between anti-cancer therapies and CAC

Treatment-associated cachexia

The pattern of weight loss and body composition change may vary with different cancer types and treatments. Treatments such as chemotherapy, radiotherapy, targeted therapy, and immunotherapy have been implicated in inducing or exacerbating the CAC phenotype (18,52,53,65,80,106–112) (Fig. 2C). This can manifest in weight and SKM loss, which have been shown in patients with nasopharyngeal carcinoma (52), head and neck cancer (18), renal cancer (106) or colorectal cancer (111). Awad and colleagues also reported a higher proportion of patients with sarcopenia after receiving neoadjuvant chemotherapy (110). Loss of SKM after chemotherapy can occur without weight loss in patients with pancreatic and ovarian cancer (53,80). SKM and/or adipose tissue loss after anti-cancer therapy have been shown to independently associate with poor survival across various cancer types (53,65,80,107,109,111). Previous studies have also shown different patterns of loss depending on cancer type after chemotherapy or immunotherapy, where the loss of body composition was confined to SKM in patients with lung cancer (108), whereas patients with metastatic melanoma, oesophagogastric or ovarian cancer experienced significant reductions in all body compartments (65,107,110). Klassen and colleagues demonstrated varying degrees of SKM and/or adipose tissue loss caused by different types of systemic cancer therapy in patients with advanced pancreatic cancer, where FOLFIRINOX (folinic acid, 5-fluorouracil, irinotecan, and oxaliplatin) was associated with greater SKM loss in males while GEM/NAB (gemcitabine and nab-paclitaxel) was associated with greater adipose tissue loss in both sexes, highlighting the differential contribution of treatment type and sex to the manifestation of CAC phenotype (109).

Anti-cancer therapies may elicit weight and body compartment loss through treatment-related side effects such as mucositis, poorer digestive function, constipation, appetite loss, impaired gut barrier function and disrupted gut microbiome, which can lead to intestinal malabsorption and lower nutrient intake (18,112–115). However, weight loss post-treatment may not be entirely due to treatment side effects but rather reflects treatment response in patients with advanced lung cancer (116). Patients who lost SKM after chemotherapy were more likely to be non-responsive to treatment, whereas most patients who gained or had stable SKM mass had stable or responding disease, with improved survival outcomes. Although greater loss of SKM and adipose tissue has been associated with tumour progression in patients with advanced pancreatic cancer, patients in the greatest loss tertile had worse survival, independent of tumour response (109). Thus, more studies are needed to investigate whether CAC is induced by the side effects of treatment, altered gut microbiome or tumour progression owing to non-responsive disease (Fig. 2D).

Pre-treatment body composition and treatment outcomes

Previous studies have shown that weight loss prior to chemotherapy is associated with poor tumour response and severe dose-limiting toxicity in patients with breast and gastrointestinal cancer (9,117,118). As a result of treatment toxicity, patients with weight loss often do not receive the full therapeutic benefits of cancer treatment, as studies have demonstrated that they received lower doses of chemotherapy or had their treatment halted, contributing to poor treatment outcomes (117,118). These findings support the notion that early intervention for the management of CAC is important to maximise the therapeutic benefits of anti-cancer therapies and improve treatment outcomes.

Body surface area (BSA) is typically used to calculate the optimal dose of cytotoxic chemotherapy that can be administered while minimising the risk of toxicity (23,119,120). It has been suggested that some chemotherapy agents are distributed in the LBM, as evidenced by high drug exposure in patients with low SKM (121,122). Consequently, patients with low pretreatment BMI, SKM (or sarcopenia) and muscle attenuation exhibited higher treatment-related toxicity, poor functional status and poor survival outcomes across several types of cancer (52,65,119,120,122–124). Given that BSA correlates poorly with LBM, some studies have proposed the need to incorporate LBM to calculate treatment doses (23,119,123).

Altered energy balance and catabolism

Energy homeostasis, where energy intake and total energy expenditure (TEE) are balanced, is necessary to maintain a stable weight. TEE consists of the thermic effect of food, physical activity energy expenditure and resting energy expenditure (REE), the latter of which is the biggest component of TEE. When CAC is associated with a negative energy balance, energy intake is lower than the TEE, resulting in weight loss. There are three primary explanations for a negative energy balance. Firstly, reduced food intake despite reduced energy expenditure contributes to negative energy balance, with severe cases meeting only ~50% of measured REE (35,125). The prevalence of reduced food intake in patients with advanced cancer has been reported to be up to 80% (11) and is a significant contributor to weight loss and poor overall survival (12,35,125). Several factors contribute to reduced dietary intake, including mechanical food obstruction, particularly in gastrointestinal cancers, anorexia, nausea, and vomiting (12,15,126). Secondly, nutrient malabsorption in the gastrointestinal tract at the organ level, leading to lower energy intake (18,113). Thirdly, hypermetabolism, where an increase in energy expenditure is not compensated by normal or increased food intake (10). Although some patients with CAC may have elevations in REE that contribute to weight loss (10,16,127), human evidence for this is limited, and additional data is urgently needed to quantify the changes in energy balance in humans with cancer.

This negative energy balance, coupled with systemic metabolic disruptions, ultimately results in tissue catabolism. The following sections will explore the factors driving altered host metabolism, and the mechanisms underlying SKM and adipose tissue wasting.

Anorexia and the central melanocortin system

Anorexia, defined as the loss of appetite, is often associated with CAC and can lead to reduced food intake and eventually CAC (15,128,129). Anorexia in the context of CAC is distinct from anorexia nervosa, caloric restriction and simple starvation. For example, weight-losing patients with anorexia nervosa had relatively preserved body cell mass (muscle, organ, blood cells) compared to weight-losing patients with upper gastrointestinal cancer (130). The prevalence of anorexia varies according to cancer site, ranging from 26% in advanced breast cancer to 40% in advanced stomach and colorectal cancer (131). In patients with metastatic oesophagogastric cancer, anorexia (FAACT A/CS) has been shown to be a stronger predictor for survival compared to weight loss in the six months pre-diagnosis (132). Several factors can contribute to anorexia, such as treatment-related side effects (e.g. difficulty swallowing, nausea, vomiting, abnormal taste and smell) (12,15,18,131), as well as disrupted appetite regulation in the brain (32,133–145). While anorexia and reduced food intake are key features of CAC, they do not always correlate with weight loss (8,10,146). Other factors, such as inflammation and disease stage, have been shown to be independently associated with weight loss (12,35). This underscores the multifactorial nature of CAC and may be a reason why nutritional support alone has not been sufficient to reverse CAC (147).

The central melanocortin system integrates signals from peripheral organs and the central nervous system (CNS) to regulate feeding behaviour and energy expenditure through the anorexigenic (appetite-suppressing) and orexigenic (appetite-stimulating) centres in the brain (148,149) (Fig. 3A). The system acts through multiple neuronal populations and is a therapeutic target for obesity and CAC, where appetite regulation and energy homeostasis are disrupted (148–151). In mammalian brain, the anorexigenic proopiomelanocortin (POMC)-expressing neurones are located in the ARC (arcuate nucleus) of the hypothalamus and the NTS (nucleus of the tractus solitarius) of the brainstem, while the orexigenic neuropeptide Y (NPY)/agouti-related peptide (AgRP)-expressing neurones are found in the ARC (149,152,153). POMC-expressing neurones express melanocortin peptides such as α-melanocyte stimulating hormone (α-MSH), which exert their effects through downstream targets that express a family of five G protein-coupled melanocortin receptors (MC1R-MC5R) (149). In the CNS, melanocortin peptides activate MC3R and MC4R, whereas AgRP inhibits these receptors. These neuronal populations can be modulated by inflammatory cytokines and hormones.

Figure 3.

Illustrative depiction of the pathophysiology of CAC. (A) Tumour/host-derived inflammatory and hormonal factors act centrally in the central nervous system to reduce appetite. (B) Tumour-derived factors promote the catabolism of adipose and muscle tissue, possibly to derive host nutrients for its growth. It also alters host metabolism by promoting the Cori cycle, a futile cycle. (C) Lipolysis can be stimulated by tumour/host-derived catecholamines, natriuretic peptides and inflammatory cytokines and promote adipose tissue loss. The evidence for adipose tissue browning in humans is lacking and controversial. (D) Muscle loss can be mediated by protein degradation by MURF1 or Atrogin-1 and can be stimulated by tumour/host-derived glucocorticoids, Activin A and inflammatory cytokines. Mitochondrial dysfunction has also been implicated in muscle wasting. Created in BioRender. Koh, K. (2025) https://BioRender.com/c10m987

The role of the hypothalamus in CAC has recently been demonstrated in humans (141,145). In 13 patients with lung cancer, hypothalamic activity has been shown to differ between patients with and without anorexia (but without weight loss) after oral nutritional supplementation, indicating the role of the CNS in the development of cancer-associated anorexia (141). Simoes and colleagues showed for the first time, structural and functional changes in the CNS in 22 patients with colorectal CAC, which may lead to neuroinflammation and anorexia. This human mapping of CNS responses to CAC provides a foundation to mechanistically investigate these changes as potential therapeutic targets (145). Both groups employed functional MRI (fMRI), an emerging neuroimaging technique that can be used to non-invasively study in vivo spatial and temporal patterns of brain activity (141,145). This is of particular interest in the context of CAC as it assesses changes in neurological activity in response to visual food stimuli and may elucidate neural pathways underlying appetite regulation. Key appetite-regulating mediators, including growth differentiation factor 15 (GDF-15), ghrelin and leptin, are discussed in subsequent sections.

Cancer-induced alterations in host metabolism

Hypermetabolism may in part be driven by cancer’s unique metabolic adaptations (63,127,154–156) (Fig. 3B). The Warburg effect, where cancer exhibits high glucose uptake and favours glycolysis over tricarboxylic acid (TCA) to derive ATP, is commonly assumed to contribute to hypermetabolism (154,157). However, in mouse pancreatic cancer, the increase in glycolysis flux is not sufficient to compensate for the decrease in TCA-derived ATP (158). Moreover, tumours have been shown to produce ATP at a slower rate than normal tissues and adapt to this by downregulating energy-expensive processes such as protein synthesis. Hypermetabolism may instead be due to tumour-driven derangements in host metabolism and substrate utilisation that lead to increased REE and catabolic processes in distant tissues (63,127,154,155). Cancer may elevate metabolic demands by rewiring host metabolism to favour energetically expensive futile cycles such as the Cori cycle, whereby ATP is consumed to recycle glucose from lactate in the liver, which is subsequently utilised by the tumour (154,155). It has also been associated with increased peripheral substrate mobilisation through increased hepatic gluconeogenesis, proteolysis and fat oxidation, which may ultimately contribute to progressive host wasting (63,127,154).

Cancer cells may secrete factors that directly stimulate tissue catabolism, such as LMF (lipid mobilising factor) (159) and parathyroid hormone-related protein (PTHrP) (156) (Fig. 3B). PTHrP has been shown to increase the gene expression of Uncoupling protein 1 (UCP1) in primary brown and white adipocyte cultures and has been implicated in adipose tissue browning and elevating energy expenditure (156). The same study found that patients with NSCLC who had detectable blood PTHrP levels measured by immunoassay exhibited lower LBM and higher REE. Tumour-derived factors such as MMPs (matrix metalloproteinases) have been shown to be amplified in the primary tumours of patients with CAC (32) and implicated in fat body and muscle wasting in CAC drosophila models (160). Consequently, the degree of wasting in CAC has been correlated with tumour burden. For instance, Lieffers and colleagues demonstrated that increased liver weight (including metastases), a highly metabolic organ, is associated with increased REE in patients with advanced colorectal cancer (17) (Fig. 3B). Additionally, in patients with lung and gastrointestinal cancer, weight loss has been associated with an increasing number of metastatic sites or higher cancer disease stage (12,36), potentially suggesting that metastases may contribute to CAC by amplifying the metabolic reprogramming of distant organs.

Emerging technologies to assess systemic metabolism in humans, such as metabolic imaging and heavy isotope labelling, offer exciting potential not only to further the understanding of altered metabolism in cancer and CAC but also as potential predictive tools for its onset. Dynamic metabolic imaging using PET and MRI allows for the non-invasive visualisation and quantification of spatiotemporal nutrient dynamics (i.e. carbohydrate, amino acid, and fatty acid) across multiple organs (161). Mitamura and colleagues demonstrated for the first time in patients with oesophageal cancer that increased tumour glucose uptake on PET correlates with higher energy expenditure and weight loss (162). Subsequent studies in NSCLC have also shown higher tumour glucose uptake correlating with weight loss at diagnosis and poor survival (163). Using whole-body PET/CT imaging, a machine-learning model built on a retrospective cohort of patients with lung cancer was able to identify CAC at diagnosis with 81% accuracy, with uptake values from spleen, pancreas, liver and adipose tissue emerging as the most predictive features (164). This demonstrates the clinical utility of PET imaging in the early detection of potentially catabolic tumour phenotypes resulting in CAC, supporting early intervention.

The assessment of tumour metabolism at the cellular level in humans and how that translates to changes in metabolic flux and fuel usage alterations in tumours and peripheral tissues is now possible with techniques such as in vivo heavy isotope tracing (165,166). For example, lactate is commonly assumed to be a waste product of tumour metabolism, which can be recycled through the Cori cycle. However, in recent years, lactate has been shown to be utilised by tumours to fuel the TCA cycle in patients with NSCLC (166). Additionally, in patients with NSCLC, tumours exhibited higher glucose oxidation than adjacent normal tissue (165). These findings highlight the importance of studying metabolic flux in humans to better understand nutrient use and potential nutrient competition between host-cancer, identify potential pathways that fuel nutrients to the tumour as potential therapeutic targets, and characterise potential distinct metabolic alterations that differentiate patients with cancer who do and do not develop CAC.

Tissue wasting mechanisms

Adipose tissue wasting

Lipolysis has been proposed as the primary mechanism underlying adipose tissue loss rather than apoptosis or inhibition of lipogenesis in human and preclinical studies (167,168) (Fig. 3C). In patients with gastrointestinal CAC, the size, as opposed to number, of adipocytes from abdominal SAT has been shown to be smaller than that of weight-stable patients (167). Furthermore, plasma glycerol and fatty acids have been shown to be elevated in CAC. In contrast to preclinical studies (169), there is limited and conflicting evidence supporting impaired lipid anabolism in humans. Insulin-stimulated lipogenesis has been demonstrated to be similar between patients with gastrointestinal CAC and those who were weight-stable (167). Conversely, a reduction in the enrichment of adipogenesis genes has been observed in the VAT of patients with gastrointestinal CAC (169). In addition, decreased lipid anabolism has only been observed in the adipose tissue adjacent to the tumour but not in distant adipose tissue in patients with colorectal CAC (170). Thus, impaired lipogenesis may only be present in adipose tissue in the vicinity of the tumour. These conflicting results may be attributed to differences in adipose tissue depots studied, cancer type, markers of lipolysis, comparisons between cancer weight loss versus weight-stable, or cancer versus non-cancer patients, as well as variations in CAC definitions. These findings may be alluding to, and more research is needed to interrogate the potential differences in mechanisms underlying wasting in SAT (primarily enhanced lipolysis) and VAT (primarily reduced lipogenesis).

Adipose tissue lipolysis can be regulated by hormonal, neural, and intracellular signalling mechanisms that control the breakdown of stored triglycerides into free fatty acids and glycerol (171–174). Insulin inhibits lipolysis by activating phosphodiesterase 3B (PDE3B), which degrades cAMP, and by promoting the inactivation of hormone-sensitive lipase (HSL) (173,175). Other regulators include natriuretic peptides (176), glucocorticoids (177), growth hormones (178), inflammatory cytokines like tumour necrosis factor alpha (TNF-α) (173,179), and the metabolite lactate (180). Catecholamines (epinephrine and norepinephrine) stimulate lipolysis via β-adrenergic receptor activation, leading to cAMP production and protein kinase A (PKA) activation, which phosphorylates HSL and perilipin, enhancing triglyceride breakdown (181,182). The lipolytic effects of catecholamines and natriuretic peptides on HSL have been shown to be stronger in patients with CAC. In preclinical studies, cytokine-mediated lipolysis (i.e. IL-6) has been shown to act through adipose triglyceride lipase (ATGL) and is independent of PKA and HSL phosphorylation (183). While studies in mice CAC models have suggested ATGL as the primary mediator of CAC-associated lipolysis (168,183), this has yet to be strongly supported by human studies. Taken together, more studies are needed to ascertain the relative contribution of lipolysis through hormonal regulation and cytokine regulation.

Adipose tissue browning

The browning of white adipose tissue (WAT) is a physiologic change that increases tissue heat production via mitochondrial uncoupling by UCP1. In mice, browning of the WAT has been shown to increase REE (184) and has been correlated with lower BMI in patients with various cancer types (185). Petruzzelli and colleagues demonstrated the presence of UCP1 in the WAT of 7 out of 8 patients with CAC but not in patients without CAC (184). Therefore, this process has been implicated in CAC as a potential explanation for weight loss due to hypermetabolism. However, a retrospective study in a cohort of 14,134 patients with various cancer types demonstrated a higher prevalence of BAT in patients without CAC compared to patients with CAC at diagnosis (186). The authors also reported no association between the presence of BAT and survival. This is further supported by the lack of difference in gene expression and protein level of UCP1 in the VAT of patients with advanced pancreatic ductal adenocarcinoma (187). To date, there is no evidence that WAT browning contributes to increases in energy expenditure that result in weight loss in humans with CAC.

Muscle wasting

Muscle mass is regulated by the balance between protein synthesis and degradation, influenced by anabolic signals like insulin, insulin-like growth factor 1 (IGF-1), and resistance exercise, which activate the mTOR pathway to promote protein synthesis (188,189). Catabolic factors such as glucocorticoids (190) and inflammatory cytokines drive muscle protein breakdown (191), while myostatin negatively regulates muscle growth by inhibiting satellite cell activation and protein synthesis in preclinical models (192,193).

Numerous preclinical studies show increased proteolysis underlying muscle wasting primarily through the ubiquitin-proteasome system (UPS) and the autophagy-lysosomal system (ALS) pathways (194,195) (Fig. 3D). However, human studies present conflicting evidence on the mechanisms underlying protein catabolism. In particular, the expression of UPS markers such as muscle RING finger-containing protein 1 (MURF1) and Atrogin-1, ubiquitinated proteins, and proteasome proteolytic activity have been shown to be either elevated (196–199) or unchanged (200,201) in patients with gastrointestinal and lung CAC compared to weight-stable patients or healthy controls. Similar to conflicting evidence observed in adipose tissue wasting, this may stem from differences in cancer type, heterogenous patient groups, patient demographics, disease stage, cohort size, comparison groups and CAC definitions. For example, Stephens and colleagues defined CAC as more than 5% weight loss (200), whereas Khal and colleagues defined 1–11% loss as moderate CAC and more than 11% as severe CAC (196). Similarly, autophagy markers such as Beclin-1 and microtubule-associated protein 1 light chain 3 beta (LC3B) are either elevated (197,199,202) or remain unaltered (198) in patients with gastrointestinal and lung CAC compared to weight-stable patients or healthy controls. Unfortunately, levels of Beclin1 and LC3B are not a measure of autophagy flux and are thereby poor indicators of autophagic activity. In addition to proteolysis, apoptosis has been reported to be activated in CAC, as evidenced by the increased expression of apoptosis-related proteins such as caspases 8 and 9 in patients with gastrointestinal CAC who experienced a median of 15% WL compared to weight-stable patients (47). Conversely, no difference in the number of apoptotic myonuclei was observed in patients with gastric cancer who experienced <10% weight loss compared to patients with non-malignant disease (203). Thus, apoptosis may be a process that manifests at a later stage of CAC.

Decreased muscle protein synthesis has been hypothesised to contribute to chronic muscle wasting in preclinical CAC models (204). However, MacDonald and colleagues reported that while muscle myofibrillar protein synthesis was higher in patients with gastrointestinal CAC compared to weight-stable patients or healthy controls, the rate of muscle protein synthesis was lower than the rate of muscle protein breakdown, resulting in net catabolism (205). Furthermore, muscle protein synthesis is blunted in response to feeding among patients with colorectal and pancreatic CAC, in contrast to healthy controls where muscle protein synthesis is stimulated in the postprandial state (206,207). This suggests an impairment to the normal anabolic stimuli in response to feeding in patients with CAC. Taken together, these findings suggest that muscle wasting can also be attributed to the failure of muscle protein synthesis to compensate for the increased rate of proteolysis rather than being solely due to reduced protein synthesis.

Mitochondrial dysfunction

Mitochondrial dysfunction, especially in the SKM, has also been broadly implicated in contributing to SKM wasting in CAC (Fig. 3D). Across various cancer types, impaired mitochondrial function has been exhibited in patients with CAC, as evidenced by higher reactive oxygen species levels and lower ATP production, which contributes to a negative energy balance (49). Untargeted metabolomics of plasma and muscle tissue revealed lower levels of membrane phospholipid metabolites (i.e. phosphatidylcholine and phosphatidylethanolamine) in CAC, which are important for mitochondrial membrane integrity and function (49). Furthermore, mitochondrial morphology and quality control have been shown to be altered in CAC, as characterised by an increase in markers of mitochondrial fragmentation, apoptosis, impaired mitophagy and autophagy (47,208). There are limited studies on mitochondrial dysfunction in the context of CAC in humans; further research is needed to determine the role of mitochondrial dysfunction in CAC.

The interplay between gut microbiome, systemic inflammation and metabolic alterations

The gut microbiome has been implicated in promoting systemic inflammation in the context of CAC. Patients with CAC among various cancer types have been found to have an enrichment in specific gut microbiota, such as Veillonella, an unknown genus of Enterobacteriaceae and Peptococcus, which is correlated with intestinal inflammation (209). Furthermore, increased expression of gut microbiome lipopolysaccharides (LPS) biosynthesis pathway (210) and elevated serum LPS-binding protein (211) have been correlated with inflammation and impaired intestinal barrier function in patients with lung and colorectal CAC. An impaired gut barrier function allows bacterial translocation into the bloodstream, potentially contributing to systemic inflammatory and metabolic derangements in CAC. This is supported by previous findings that demonstrated increased bacterial translocation and a compromised immune system in patients with colorectal CAC (212). Notably, LPS-binding protein has been demonstrated to be predictive of future CAC onset in lung and colorectal cancer (211).

In addition to contributing to systemic inflammation, altered gut microbiome metabolic pathways have been identified in CAC. Ni and colleagues demonstrated lower gut microbiome anabolism of essential branched-chain amino acids (BCAAs; e.g. isoleucine, leucine) in patients with lung cancer, which are important for SKM synthesis (210). Elevated gut microbiome methane metabolism has been observed in patients with CAC (210), where methane has been shown to reduce appetite by acting through glucagon-like peptide-1 (GLP-1) in mice studies (213). Additionally, chemotherapy-associated alteration in gut microbiota has been demonstrated to correlate with serum metabolites, further supporting the potential contributions of the gut microbiome to systemic metabolic alterations (115). Much of the research suggesting a role for the gut microbiome in the pathogenesis of CAC has been performed in preclinical studies (211,214). Though the specific mechanisms and contributions of the gut microbiome in CAC are unclear, these findings suggest an interplay between alterations in gut microbiota, systemic inflammation and metabolic changes.

Proposed CAC mediators

Although strong evidence for definitive mediators and mechanisms of CAC is currently limited in human studies, specific molecular mediators and biological processes have been associated with features of CAC. The challenge in identifying drivers and their underlying mechanisms of action stems from the multifactorial nature of CAC. Multiple mediators or pathways may converge or act in parallel in a highly integrated fashion to drive the pathogenesis of CAC. Moreover, drivers of CAC may vary by cancer type, stage and host genotype, among other factors (6,12,18,20,21,35,215,216).

There is a growing recognition of the role of central mediators of CAC that act within the CNS (e.g., the melanocortin system and hypothalamic-pituitary-adrenal (HPA) axis), where they influence neuroinflammation, appetite control, and energy balance. These mediators have garnered increasing attention, in part due to promising results from clinical trials. Here, we discuss inflammatory and hormonal mediators that underlie energy imbalance and tissue wasting in CAC, with a particular focus on central mediators.

Inflammatory mediators

Inflammation is a universal defence response to disease that initiates healing processes; however, chronic inflammation can cause damage to healthy cells. In the context of CAC, it is thought to precede weight loss and mediate tissue wasting (3). The inflammatory signalling pathway in primary lung tumour tissue has been shown to be enriched in patients with CAC, supporting the role of tumour-driven systemic inflammation in the development of CAC (32). Several inflammatory factors have been identified to contribute to the development of CAC in preclinical models, such as IL-6, TNF-α, and IL-1. IL-6 promotes SKM wasting through JAK/STAT signalling in preclinical studies (191) and adipose tissue wasting by increasing lipolysis and inducing adipocyte browning (184,217). IL-6 has also been implicated in driving apathy in the C26 CAC mice model (218). Leukaemia inhibitory factor (LIF), part of the IL-6 cytokine family, has been associated with ATGL-mediated lipolysis and muscle atrophy (215,219). LIF has been shown to activate POMC neurons in the hypothalamus and stimulate the release of α-MSH, which may contribute to anorexia (137). TNF-α can contribute to SKM wasting by activating NF-kB (nuclear factor-kappa B), which activates the UPS (220,221) and is implicated in altered adipocyte metabolism (222), increased lipolysis (173,179), adipose tissue apoptosis (223), and is associated with insulin resistance (224).

IL-1 has recently garnered increasing interest as a target for CAC (225,226). The IL-1 family includes IL-1α, IL-1β, and IL-18 (227), and has been shown to trigger hypothalamic inflammation, disrupt appetite regulation and induce muscle catabolism in mice studies (133,135,138,228). In particular, IL-1β has been shown to activate the central melanocortin system by stimulating POMC neurones in the ARC, which express IL-1 receptor (IL-1R), leading to the dose-dependent release of α-MSH (135). Blocking MC4Rs has been shown to prevent anorexia induced by IL-1β in rats (136). Additionally, hypothalamic inflammation through IL-1β has been shown to be sufficient to induce muscle catabolism mediated by the HPA axis, independent of peripheral inflammation, although this is not well characterised in humans (228). IL-1α has been correlated with reduced food intake in tumour-bearing rats (133), while IL-18 knockout mice exhibited increased body weight, reduced energy expenditure, and lower food intake, indicating the role of IL-18 in suppressing appetite and increasing energy expenditure (138).

The role of several of these inflammatory factors, however, remains poorly defined in human studies. Exploratory studies of the plasma proteome in the TRACERx lung study identified no significant difference at the time of diagnosis when comparing patients with CAC to patients without CAC. However, in samples taken at first relapse after primary surgical resection, two proteins that have previously been identified as associated with CAC (IL-15 and hormone GDF-15) were significantly elevated (32). IL-15, however, has been shown to be anabolic and correlates with less weight loss (229) and instead, may likely be correlated with tumour growth and metastasis (230). IL-1β protein expression has been shown to be elevated in the tumour and SAT of patients with CAC (139) and identified as a better predictor of CAC development compared to other cytokines such as IL-6 and TNF-α in patients with pancreatic cancer (140) and has been shown to be elevated in elderly patients with cachexia associated with chronic non-malignant disorders (231). Studies of individual cytokines have provided mixed results in patients with incurable cancer; in a systematic review in 2022, Paval and colleagues identified TNF-α, IL-8, and IL-6 as the only cytokines consistently elevated in patients with incurable cancer compared to non-cancer controls. Notably, only IL-6 was elevated in patients with CAC compared to patients with incurable cancer without CAC (232). Leptin, IL-1β, adiponectin and ghrelin demonstrated no difference between these groups. Circulating IL-6 has also been suggested to correlate with disease progression rather than weight loss in patients with pancreatic cancer (233). However, clinical trials targeting IL-6 and TNF-α have yielded limited success (85,234,235). In a recent study, D’Lugos and colleagues demonstrated inflammation and pathological remodelling of SKM, such as collagen remodelling through the infiltration of plasma proteins and immune cells in the SKM of treatment-naïve PDAC patients with CAC (236).

Pro-inflammatory cytokines, particularly IL-6, have been shown to promote inflammation and metabolic dysregulation (237). This results in a liver-mediated increase in acute phase proteins, including CRP, fibrinogen and lipocalin, while albumin synthesis is decreased (237). Furthermore, the acute phase response is associated with increased REE in patients with pancreatic cancer and CAC (238). Albumin and prealbumin are linked to nutritional status; low levels are independent poor prognostic markers in patients with CAC and are associated with reduced QoL, particularly in advanced disease stages, although the association varies by tumour type (239). However, in pancreatic cancer, albumin synthesis is not decreased despite hypoalbuminemia, indicating alternative mechanisms may contribute to low albumin (240).

Activin A and myostatin, members of the transforming growth factor beta (TGF-β) superfamily, are myokines secreted by SKM and have been shown to induce UPS-mediated degradation and apoptosis in SKM by activating the activin A receptor type 2 (ActRII) in preclinical models of CAC (192,241). Increased serum and tumour levels of Activin A have been found in patients with gastrointestinal and lung cancer who experience weight, SAT, VAT and SKM loss (56,187,242). This increase in Activin A has been demonstrated to correlate with reduced HGS, increased muscle lipid infiltration, tumour progression, and poor overall survival (56,242–244). Furthermore, Xu and colleagues demonstrated increased fibrotic remodelling of VAT in patients with pancreatic cancer who had increased Activin A levels (187). In contrast, myostatin, which inhibits muscle growth under normal conditions, was found to be lower (56) or unchanged (245) in patients with gastrointestinal or hepatic cancer and weight loss, suggesting that Activin A, rather than myostatin, may be the more predominant tissue driver of CAC. However, clinical trials targeting myostatin in pancreatic cancer (NCT01505530) and, ActRII in pancreatic or NSCLC cancer (NCT01433263) have been unsuccessful. Activin A has been associated with a more severe SKM wasting phenotype in male mice and men with early stage PDAC (216). When comparing men and women, only men demonstrated SKM loss following treatment with GEM/NAB, further supporting the need to consider sexual dimorphism in clinical trials to develop effective and tailored therapies.

Hormonal mediators

Neuroendocrine factors

GDF-15, originally termed MIC-1 (macrophage inhibitory cytokine 1), was first identified in 1997 (246) as a distant member of the TGF-β superfamily and has since been increasingly investigated in various diseases, including CAC and obesity (134,142,247). GDF-15 has been shown to induce weight loss in obese nonhuman primates and rodents on a high-fat diet (142,247,248) and protect against insulin resistance (247,248). Furthermore, it is negatively correlated with HGS, LBM and fat mass, and is associated with reduced survival (242,249). In preclinical studies, GDF-15 has been shown to act through glial cell-derived neurotrophic family receptor-alpha-like (GFRAL) receptors located in the hindbrain (142) and modulate the HPA axis (144), leading to suppressed appetite, a decrease in food intake (134), sickness behaviour (143) and enhanced energy expenditure in the muscle (247). It has consistently been found to be elevated in patients with CAC across a variety of tumour types (32,242,249). The TRACERx study observed an increase in plasma protein abundance of GDF-15 in patients with CAC, and a significant correlation between increased normalised abundance of GDF-15 and body compartment and weight loss at relapse (32).

Ghrelin is largely produced by the stomach and acts as a feeding-stimulating hormone by suppressing POMC neurons and stimulating NPY/AgRP neurones (149,250,251). However, in patients with CAC, plasma ghrelin levels are elevated despite weight loss, indicating ghrelin resistance (252–255). Ghrelin resistance may be overcome by increased ghrelin or ghrelin receptor agonism, for example, with the use of ghrelin receptor agonists (256,257). In tumour-bearing rats, the administration of brainstem GLP-1 receptor antagonist into the fourth ventricle has been shown to increase food intake and reduce body weight loss, illustrating the central role of GLP-1 signalling in mediating anorexia (258). While the association between GLP-1 and CAC is not well defined in human studies, GLP-1 receptor agonists are highly effective in treating obesity (259), indicating GLP-1 as a potential therapeutic target in CAC.

Adipokines and their role in CAC: appetite regulation and insulin resistance

Hormones secreted from fat cells, known as adipokines, have been implicated in insulin resistance (e.g. resistin) (260) and regulation of food intake (e.g. leptin and adiponectin) (261,262). While adiponectin has been shown to stimulate food intake (262), leptin has been shown to suppress appetite, reduce weight (263), and regulate food intake through leptin receptors on POMC and NPY neurons in mice (264,265). Smiechowska and colleagues demonstrated a decrease in plasma levels of leptin with no changes in adiponectin and resistin in patients with CAC across various cancer types (4). Although decreased leptin levels typically stimulate appetite, low leptin levels were associated with poor appetite, insulin resistance and increased IL-6 in these patients, indicating either leptin resistance and disruption to the feedback mechanism in the hypothalamus or a compensatory response to weight loss. This may be attributed to the counterbalancing response by leptin rather than leptin resistance, as the administration of leptin to mice treated with LIF (induces weight and fat loss) led to further decreases in fat mass and weight, which indicates a functioning feedback mechanism (215). Several studies have also shown decreased leptin serum levels in patients with CAC in lung cancer (266) and in females (but not males) with breast and colorectal cancer (253). Conversely, other studies either demonstrated higher serum leptin, adiponectin and resistin levels in patients with gastric cancer and CAC (254) or found no association between leptin and weight loss in patients with gastrointestinal cancer (267). Leptin, adiponectin, and resistin are commonly studied in CAC, however their role remains to be fully understood and may vary with tumour type and sex.

Insulin resistance

Insulin is an anabolic hormone that increases the production and reduces the degradation of all major macromolecules in the skeletal muscle and adipose tissue. Insulin is known to increase body weight, and insulin suppression promotes weight loss (268,269). Although human studies on insulin resistance and CAC are limited, some patients with cancer are known to have insulin resistance, which can contribute to CAC (270). Fouladiun and colleagues demonstrated increased serum insulin in patients with CAC across several cancer types (62). Furthermore, patients with lung and gastrointestinal cancers exhibit lower or unchanged glucose uptake, even in states of high insulin, indicating an impaired insulin response (46,271,272). This reduction in glucose uptake has been shown to correlate with lower glucose storage and higher fat oxidation, reflecting a change in substrate utilisation and depletion of energy reserves (271). Additionally, reduced glucose uptake in patients with CAC correlates with increased tumour volume and inflammation, but not weight loss (273). This suggests that insulin resistance may develop earlier in disease progression before any measurable weight loss occurs (45) and reflects a shift from an anabolic to a catabolic state in CAC.

Clinical trials and therapeutic interventions

Pharmacological treatments

In recent years, the landscape of pharmacological treatment for CAC has improved. Several positive trials on anamorelin, ponsegromab and olanzapine, as well as multiple agents targeting different pathways, provide grounds for optimism that CAC now has the potential for effective treatment. These breakthroughs are largely driven by compounds targeting the CNS, underscoring the brain as the most promising target for anti-CAC pharmaceuticals at present. The following section highlights some therapies of interest and draws connections back to the underlying mechanisms and mediators discussed earlier in the review.

Ghrelin Receptor Agonists

The hormone ghrelin, which stimulates appetite, has long been envisaged as a target for treating cancer cachexia. The most studied ghrelin receptor agonist, anamorelin, was examined in two large phase 3 clinical trials and demonstrated improvements in lean mass and CAC symptoms, but not physical function (274). Subsequently, regulatory approval was not given in Europe and remains unapproved. However, it does now have regulatory approval in Japan following further trials (275). Recent work has demonstrated that there may be subgroups where anamorelin is more effective, which requires further examination (276). Therefore, while anamorelin remains the most widely studied ghrelin receptor agonist, further data are awaited on recent phase 3 trials before wider regulatory approval is sought.

Macimorelin, another ghrelin receptor agonist has also been proposed as a potential treatment for cachexia. A recent early phase trial demonstrated that it improved weight and QoL in patients with CAC, and further trials are awaited (277).

GDF-15

Similar to ghrelin, GDF-15 is another pathway of interest that also acts through the CNS (278). A successful translation from pre-clinical work to the trials resulted in a phase 2 clinical trial demonstrating that ponsegromab and GDF-15 antagonists improved weight, SKM, physical activity and patient-reported QoL (i.e. FAACT-ACS) and appetite (89). This work represents one of the first trials in CAC where all meaningful endpoints have been achieved, and like anamorelin, it is of interest to see if there may be subgroups where response could be enhanced (279). Visogromab is another agent targeting GDF-15, which in addition to targeting appetite, may overcome PD-1 (programmed cell death protein 1) resistance, potentially meaning a successful cachexia and cancer treatment combined (280). Several other agents targeting GDF-15/GFRAL include anti-GFRAL monoclonal antibody NGM120 (NCT04068896) and anti-GDF-15 monoclonal antibody AV380 (NCT05865535).

JAK-STAT and related pathways

As CAC has been shown to have a strong systemic inflammatory genesis, it has been rightly postulated that targeting this through potent anti-inflammatory mediators would be beneficial. Several immunotherapies and other drugs have been examined in CAC. Of these, the only immunotherapy to have completed phase III trials assessing muscle is Bermekimab, which targets IL-1α (225,281). These trials reported beneficial effects on LBM. IL-1β is another potential target which has been shown to have a clear role in the systemic inflammatory response, where exposure to IL-1β results in an acute illness response in animal models (anorexia, weight loss, skeletal muscle atrophy) (226,282,283). The CANTOS (canakinumab) trial has demonstrated a clear benefit in targeting IL-1β in atherosclerotic disease and, in post-hoc analyses, has also demonstrated that incident lung cancer was reduced in the study population (284). However, its role has not yet been confirmed in clinical studies of CAC, and currently, no trials are underway examining this.

Anti-IL-6 therapies have also been examined, including Clazakizumab/ALD518. While a phase 2 trial showed stabilisation of LBM, no phase 3 trials are underway (235). Anti-TNF-α therapies have been examined where Infliximab, through two phase 2 trials showed no beneficial effects on weight but caused increased fatigue and decreased QoL compared to placebo. Similarly, Etanercept was assessed in a phase 2 trial but did not improve weight and resulted in higher rates of neurotoxicity (85,234,285). This highlights the limitations of targeting a single mediator where other mediators may converge on the same pathways. Instead, targeting the pathway of multiple mediators may be more effective, and a phase 1 trial on a JAK/STAT signalling pathway inhibitor, ruxolitinib, is underway (NCT04906746). It has been widely used in oncological and non-oncological contexts, with known but manageable side effects and has been shown to delay CAC-related anorexia and adipose tissue loss and improve survival in preclinical models of CAC (286).

Melanocortin

The melanocortin system, particularly the melanocortin-4 receptor (MC4R), when stimulated, has been shown to contribute to hypothalamic inflammation, anorexia and the development of CAC in animal models and represents a target of interest (135,228,282). In humans, a frameshift mutation in MC4R has been shown to be associated with obesity (287,288). In animal models with CAC and chemotherapy-induced anorexia and weight loss, peripheral treatment of melanocortin antagonists has been shown to significantly attenuate weight, fat and LBM loss, improve food intake and lower hypothalamic inflammatory gene expression (151,289). Phase 1 trials of TCMCB07, an MC3R and MC4R antagonist peptide that importantly can cross the blood-brain barrier, have now been completed and showed promising results in tolerability, weight gain, and improved appetite (NCT05529849), with phase 2 trials currently underway.

Anabolic agents

The selective androgen receptor modulator, Enobosarm, has been evaluated in two phase 2 trials and showed improvements in LBM and physical function as measured by the stair climb test in patients with cancer and weight loss, and in healthy participants (86,290). Results from phase 3 trials in patients with NSCLC are expected (NCT01355484, NCT01355497], and no further trials are currently planned. Espindolol, an anabolic and anti-catabolic agent, was examined in a phase 2a trial of patients with advanced NSCLC or colorectal CAC. Patients on high-dose espindolol demonstrated significant weight gain, improvements in LBM and HGS, preservation of fat mass, and non-significant improvement in stair climbing power and 6MWT compared to placebo (291). Phase 2b/3 trials are currently being planned.

Corticosteroids and Progestins

Off-label pharmacological treatments for CAC recommended by the American Society of Clinical Oncology (ASCO) and European Society of Medical Oncology (ESMO) include corticosteroids (e.g. dexamethasone) for their anti-inflammatory effects and progestins (e.g. megestrol acetate and medroxyprogesterone acetate) for their appetite stimulation effects (94,292). Although dexamethasone and megestrol acetate both have shown improvement in appetite and weight, dexamethasone is associated with higher toxicity and patient discontinuation, while megestrol acetate is associated with higher rates of deep vein thrombosis (293). Medroxyprogesterone acetate has demonstrated improvement in weight and anorexia but not LBM (294).

Olanzapine

Olanzapine, an atypical antipsychotic with appetite-stimulating properties, has been investigated as a treatment for CAC due to its antiemetic and orexigenic effects. In a randomised controlled trial, low-dose, daily olanzapine demonstrated a significant increase in weight and improvement in appetite and QoL in patients with advanced cancer and unintentional weight loss compared to placebo (295). The study demonstrated that olanzapine was well tolerated and offers a promising therapeutic option for managing cancer-associated weight loss and anorexia.

Cannabinoids

Traditionally, cannabinoids have been thought to increase appetite and thus have been proposed as a treatment for CAC (296,297). However, clinical trials to date have been unsuccessful (296). Notably, a current phase 2 trial, the cancer appetite recovery study (CAReS; EudraCT 2020–000-464–27) is examining a synthetic endocannabinoid for its potential to stimulate appetite and improve LBM in patients with CAC, building on previous studies that observed weight gain in healthy subjects (297).

Nutritional, exercise and multimodal interventions

Clinical management of CAC ideally involves a multimodal approach, providing nutritional and psychosocial support while treating the symptoms that can impact appetite, such as nausea and pain. ASCO, the ESMO, and ESPEN (European Society for Clinical Nutrition and Metabolism) have produced guidelines with similar recommendations consisting of screening patients with cancer for malnutrition before and during treatment, dietary counselling, short term courses of progestins or corticosteroids, olanzapine as an appetite stimulant, and structured exercise training (292,298–300). Enteral tube feeding and parenteral nutrition are not advised, except in cases that involve physical obstruction or malabsorption (292,298–300).

Clinical trials on nutritional and exercise-based interventions have reported mixed findings. A systematic review reported no effect on weight or LBM but an improvement in QoL and survival in patients who received omega-3 fatty acid (anti-inflammatory) supplementation (301). Dietary counselling and protein supplementation in patients with colorectal cancer undergoing radiotherapy both increased energy and protein intake (302). However, after radiotherapy only dietary counselling caused sustained improvement in QoL, functional capacity, and appetite (302). Other supplements, such as BCAAs and creatine supplementation, are also being explored; however, the benefit in mitigating CAC remains to be conclusively demonstrated (303,304). Exercise is associated with increased muscle mass; however, a recent review by Baguley and colleagues found limited evidence supporting the effectiveness of exercise-only interventions in mitigating SKM loss, with included clinical trials using various interventions (e.g. aerobic exercise, tai chi, resistance training) (305). The limited success of nutritional- and exercise-only interventions in significantly improving lean body mass and muscle strength reinforces the need for a targeted, multimodal approach for the effective treatment and management of CAC.

Despite the strong theoretical rationale for multimodal interventions, clinical evidence supporting their efficacy remains limited. One notable study, the MENAC (Multimodal Exercise, Nutrition, and Anti-inflammatory medication for Cachexia) trial by Solheim et al., demonstrated an improvement in weight but failed to show significant gains in lean mass or physical function across the overall study population (306). However, in patients who fully adhered to the intervention protocol, positive changes in lean mass, body weight, and physical activity were observed. This suggests that treatment efficacy may be influenced by patient compliance and highlights the need for further research to optimise intervention strategies. Future studies should focus on refining multimodal treatment protocols, identifying patient subgroups most likely to benefit, and integrating personalised approaches that consider genetic, metabolic, and clinical factors. Large-scale trials with standardised outcome measures are essential to establish the efficacy of multimodal therapies and their impact on survival, quality of life, and treatment tolerance in CAC.

New efforts are being undertaken to understand diet and its role in cancer metabolism, immune regulation, and therapy (307). A ketogenic diet has been shown to slow tumour growth but reduce survival as a result of accelerated CAC onset in mice (308). However, administrating dexamethasone to mice fed with a ketogenic diet has been found to delay CAC onset and improve survival. Additionally, calorie restriction has been shown to enhance the effectiveness of radiotherapy in triple-negative breast cancer by reducing immune suppression (309). Thus, it is important to consider both cancer metabolism and CAC when investigating nutritional interventions.

Challenges to CAC diagnosis and clinical management

Clinical management of CAC is impeded by the lack of diagnosis, standardised treatment, and approved therapies for CAC. One of the contributing operational factors is the difficulty of objective measurement of weight over time – patients may not know their usual healthy weight, and measurements may not be routinely or accurately recorded (310). Although ASCO and ESMO recommend regular screening, surveys suggest that there is a high prevalence of underdiagnosis of CAC. According to the Fearon criteria, CAC may be diagnosed with as low as 2% weight loss (3). However, Muscaritoli and colleagues showed that 46% of practitioners would only diagnose CAC when weight loss was over 10%, and 48% if weight loss was over 15% (7). In a retrospective study, Sun and colleagues identified CAC, according to the Fearon Criteria, in 140 out of 390 patients with cancer. However, based on physician assessment, only 33 out of 140 were diagnosed with CAC (311). As there are no specific treatment recommendations for CAC, physicians may not see the value in an official diagnosis, despite recognising symptoms (310). CAC may present with overlapping symptoms with sarcopenia, malnutrition, frailty, and unintentional weight loss, which have distinct causes and may respond to treatment differently, contributing to misdiagnosis (310).

The difficulty in making a definitive CAC diagnosis can also impact clinical trial recruitment and assessment of endpoints. In a retrospective study of 36 patients with colorectal cancer, 17 out of 36 experienced <5% weight loss and would typically be excluded from CAC clinical trials. However, 7 (41%) of these patients were found to experience >5% SKM loss (24). Given the prognostic relevance of adipose tissue loss previously shown in both lung and pancreatic cancer (21,32), consideration of fat loss, in addition to muscle and weight loss, as part of trial eligibility criteria, may help capture a greater proportion of patients with CAC. This suggests that relying on weight loss alone for trial eligibility is likely to not capture all patients with CAC.

Demonstrating improvement in functional capacity is essential for regulatory approval. However, currently, there is no universally agreed-upon optimal endpoint for physical function measures in CAC studies, and there are variations in the endpoints used between cachexia trials, impacting the reproducibility of effective interventions (98). The authors posit that trial endpoints should be selected based on the mechanism of action of the intervention to ensure that the endpoints are relevant and meaningful. For example, in the ROMANA trials, patients treated with Anamorelin demonstrated improvement in lean muscle mass but not HGS (274). Given that Anamorelin is an appetite stimulant, HGS may not be the most appropriate endpoint. Alignment of functional capacity test procedures, along with training of test administrators, can help ensure valid comparisons between trials and facilitate the approval of treatments (98).

Since systemic metabolic, inflammatory, hormonal and behavioural changes can precede overt loss of weight and body compartments (3,6,48), sample collection in trials for circulating mediators associated with CAC could serve as potential biomarkers. These potential biomarkers may offer an opportunity for objective and minimally invasive approaches for detection and risk stratification to facilitate early effective intervention and management of CAC (312). For example, metabolic and inflammatory alterations, such as increased IL-6, TNF-α, CRP, lactate and low albumin have been shown to correlate with QoL and appetite (313), and are incorporated in the CASCO tool (35,41,314). Furthermore, GDF-15 has emerged as a promising diagnostic and prognostic biomarker for CAC (32,242,249,315). Recent advances and applications of metabolomics have proposed novel lipid and amino acid metabolites as potential biomarkers for CAC. Serum metabolites have been shown to identify CAC with similar accuracy to metabolites extracted from SAT, VAT, SKM and liver (84). In addition to cross-sectional studies, longitudinal studies are needed to identify biomarkers predictive of future tissue loss. BCAAs have been demonstrated to be elevated at diagnosis in patients with pancreatic cancer who experience SKM loss but not adipose tissue loss 2–4 months post-diagnosis. Notably, although IL-6 has been associated with CAC in other cross-sectional studies (232), IL-6 is not predictive of future muscle loss (79).

International efforts to address the unmet clinical needs in cachexia

Collaborative and international efforts are currently underway to further advance our understanding of the mechanistic underpinnings of CAC and the means by which clinical phenotyping can help identify patients at risk of CAC. The Routine EValuatiOn of people LivIng with cancer (REVOLUTION) study aims to characterise body composition, physical activity, systemic inflammatory response, symptoms of CAC, and QoL in patients with incurable cancer (316). This will facilitate the understanding of advanced CAC phenotypes and therefore potential subtypes and their prevalence. The CANcer Cachexia Action Network (CANCAN) is a Cancer Grand Challenge initiative funded by Cancer Research UK and the National Cancer Institute, which alongside preclinical studies has established two clinical observational studies: 1) the Longitudinal Observational sTudy to UnderStand cancer cachexia (LOTUS-CC) study aiming to characterise CAC phenotypes using surveys, functional tests, blood and tumour measures, and activity/sleep trackers, with follow up of medical records at 3 months and 1 year (NCT06073431), and 2) the CANcer on neurobehavioral mechanisms and COGnition in cachexia (CANCOG) study aiming to characterise the neurological changes associated with CAC using fMRI, questionnaires and computer-based tasks to assess food reward and motivation (NCT06112964).

There is a desperate need to study CAC from a cancer-intrinsic and -extrinsic angle, but also at the level of the whole-body. Ongoing longitudinal studies, such as TRACERx and TRACERx EVO (NCT05628376), offer the opportunity to integrate genomic and transcriptomic features of the tumour and blood with body composition and symptomatology to holistically profile CAC and shed light on the disease process over time.

Conclusion and future perspectives

The limited mechanistic understanding and poor clinical trial outcomes in CAC are in part driven by the limited number of longitudinal studies in humans to understand the development of CAC, limitations in precedent approaches to phenotyping CAC and the inconsistent use of CAC definitions across studies (43,44). Incorporating body composition loss thresholds, QoL, and functional measures is essential to clinically phenotype CAC in a comprehensive fashion (39,317). This may aid risk stratification, allowing for early intervention, further understanding of distinct molecular mechanisms underlying potential clinical subtypes and improvements to trial design to include endpoints that are reproducible and can lead to future effective and meaningful interventions for patients.

While the mechanistic understanding of CAC is still evolving, strong mechanistic links exist. For example, the mechanism of signalling to the brain leading to food aversion behaviour is known for GDF-15 and is consistent with the clinical presentation of reversible loss of appetite for patients with elevated GDF-15 (32,89,134,142–144,242,247,249). Promising clinical trial results on the central mediators of CAC (Table 2) and the recent inclusion of olanzapine in the ASCO guidelines by rapid recommendation to improve weight and appetite, although not regulatory approved (292), strongly supports the role of central mechanisms in the development of CAC.

Table 2.

Overview of selected phase 2 or 3 CAC clinical trials where findings are published. Results from the latest phase are reported. mAb, Monoclonal antibody; NSCLC, non-small cell lung cancer; WL, weight loss; Hb, haemoglobin; CRP, C-reactive protein; LBM, lean body mass; CRC, colorectal cancer; OS, overall survival; HGS, handgrip strength; FACIT-F, Functional Assessment of Chronic Illness Therapy; VAS, visual analog scale; FAACT-ACS, Functional Assessment of Anorexia/Cachexia Therapy-Anorexia/Cachexia Subscale; EORTC QLQ-C30, European Organisation for Research and Treatment of Cancer

Pathway	Intervention	Trial description	Primary endpoints	Primary results
Ghrelin	Ghrelin receptor agonist anamorelin (274)	Phase 3. Inoperable advanced NSCLC and CAC	Change in LBM (DEXA) and HGS over 12 weeks	Significant increases in LBM in ROMANA 1 and ROMANA 2. No difference in HGS
GDF-15	Anti-GDF-15 mAb ponsegromab (89)	Phase 2. NSCLC, pancreatic cancer, or CRC with elevated serum GDF-15, CAC, ECOG 0–3	Change in weight over 12 weeks	Significantly greater weight gain
JAK-STAT & Related	Anti-IL-6 mAb clazakizumab (235)	Phase 2. Incurable advanced NSCLC with WL>5% over 3 months, ECOG 0–3, Hb >7g/dl, CRP>10mg/L	Change in safety parameters (adverse events, lad safety tests)	Safe and well tolerated. Additional observations: stabilisation of LBM (DEXA)
	Anti-IL-1α mAb MABp1 (87)	Phase 3. Advanced CRC, ECOG 1–2, either CAC or IL-6 ≥ 10pg/ml and either anorexia, fatigue, pain, emotional and social function	Composite of change in LBM (DEXA) ≥ 0kg or health status (EORTC-QLQ-C30) ≥ 2 symptoms stable or improved, over 8 weeks	Higher proportion of patients achieved primary endpoint
	Anti TNF-α infliximab (234)	Phase 2. Advanced incurable NSCLC, ECOG < 2	Weight gain ≥ 10%	No significant improvements, trial closed early
	Anti TNF-α etanercept (85)	Phase 3. Incurable advanced malignancy, ECOG ≥ 2 and WL ≥ 2.27kg over 2 months or <20 calories/kg body weight/day, appetite concern	Weight gain ≥ 10%	No significant improvements
Anabolic Agents	Selective androgen receptor modulator - Enobosarm (290)	Phase 2. Male (>45 years) and female (post-menopausal) with cancer, not obese and >2% WL in 6 months	Change in LBM from baseline at 113 days or end of study	Significant increase in LBM
	Anabolic/anti-catabolic (β-adrenergi modulation) - Espindolol (291)	Phase 2. Advanced CRC or NSCLC and CAC	Rate of weight change over 16 weeks between high dose espindolol and placebo	Significant increase in rate of weight change
Anti-psychotics	Olanzapine (295)	Phase 3. Advanced gastric, hepatopancreaticobiliary, or lung cancer.	>5% weight gain, appetite improvement (VAS, FAACT-ACS)	Significantly higher proportion of patients with weight gain and improvement in appetite

Further research is needed for other mediators, such as IL-6, where its correlation with CAC is strong in human studies (318), and preclinical studies suggest hepatic programming in IL-6-associated CAC (319); however, trials targeting IL-6 have been unsuccessful (Table 2). More work is required to understand cancer-driven metabolic alterations, tumour-host interactions, and tissue wasting mechanisms underlying CAC. WAT browning has been suggested in preclinical studies, however human evidence is controversial (156,184–187). Preclinical and clinical studies have presented inconsistent results, highlighting the need for validation in clinical cohorts to ensure that preclinical findings are relevant to patients (Table 2). With the emerging use of technologies such as metabolic imaging (161–164), automated image segmentation (67,68), in vivo heavy isotope labelling (165,166) and fMRI (141,145), alongside international collaborative efforts like REVOLUTION and CANCAN, the field will undoubtedly advance in the clinical management and mechanistic understanding of CAC.

Despite extensive efforts in clinical trials, there are currently no globally approved, effective treatments for CAC. Pharmacological interventional trials are exploring new avenues to identify patients that are most likely to benefit from treatment, such as in the Phase 2 ponsegromab study, which included only patients with elevated GDF-15 and has yielded positive results (89). Similarly, the Phase 1 study of ruxolitinib (NCT04906746), a JAK 1/2 inhibitor which targets JAK/STAT signalling is exploring the therapeutic potential of targeting a signalling pathway rather than a single mediator. Exercise and nutritional-based interventions are also being explored. However, these interventions have seen limited success.

Current clinical studies on CAC are often disease-site specific, which may not reflect the potential transverse relevance of CAC across different cancer types (17,21,32,52,87,88,167,210). An alternative approach could be to focus on identifying shared mechanisms in CAC, irrespective of cancer type, to potentially develop more unified treatments. Given the multifactorial nature of CAC, a single-target therapeutic approach is unlikely to be effective. Instead, a multimodal strategy addressing the diverse biological and functional impairments associated with CAC has been widely advocated (306,320). The treatment of CAC should be considered together with standard cancer management as ultimately, cancer is the driver of CAC.

Significance.

CAC represents a significant clinical unmet need. Despite its high prevalence and associated mortality and morbidity, there are currently no globally approved effective therapies. This review provides a comprehensive overview of human studies aimed at defining CAC clinically and identifying mediators underlying it that are revealing effective health interventions. Furthermore, we highlight ongoing international efforts to advance our understanding of CAC.