https://orcid.org/0000-0002-3253-3106
Abstract
Cachexia is characterized by weight loss and decline in muscle mass and function and is a poor prognostic factor among patients with cancer. Patients with metastatic EGFR-mutant non-small cell lung cancer (NSCLC) derive remarkable survival benefits with osimertinib, a third-generation EGFR tyrosine kinase inhibitor. It is not known whether patients treated with osimertinib experience any weight loss or whether weight loss impacts patient outcomes. Therefore, we sought to describe the frequency and consequences of weight loss in this patient population.
We conducted a single-center retrospective pilot study of 56 patients treated with first-line osimertinib for metastatic EGFR-mutant NSCLC. We defined on-treatment weight loss as a loss of ≥5% body weight at 6 or 12 months of treatment. We described the characteristics of patients with and without on-treatment weight loss and differences in progression-free survival (PFS), time on treatment with osimertinib, and overall survival (OS).
Forty-six percent (n = 26) of patients met the criteria for on-treatment weight loss. There were no significant differences in patient or disease characteristics between patients with and without weight loss. Compared to patients without weight loss, patients with weight loss had similar PFS and time on treatment with osimertinib. Yet, patients with weight loss had significantly worse overall survival (HR 4.91, 95% CI, 1.56-15.5, P = .007).
Weight loss was observed in nearly half of patients with metastatic EGFR-mutant NSCLC treated with osimertinib, and patients with weight loss had significantly worse overall survival.
Implications for Practice
This study highlights the prevalence of on-treatment weight loss in patients treated with osimertinib for metastatic EGFR-mutant non-small cell lung cancer and demonstrates worse overall survival among patients who experience weight loss. Our results suggest that weight loss is common and is a poor prognostic factor in this patient population. Additionally, our findings underscore the need for effective anti-cachexia therapies for cancer patients.
Introduction
Patients with lung cancer are at high risk of developing cancer cachexia, a syndrome that is characterized by progressive muscle wasting and weight loss and negatively impacts quality of life.1-5 Cachexia is not simply a nutritional disorder, but a process with a complex underlying pathophysiology. It is characterized by systemic metabolic dysfunction and chronic inflammation.6 Moreover, chemotherapy agents such as cisplatin and 5-fluorouracil/oxaliplatin, for instance, have been shown to promote cachexia independent of their effects on tumor growth.7-10 Cachexia is a poor prognostic factor among patients with advanced cancer and is prevalent in pancreatic, gastrointestinal, and lung cancers.11 Although there is wide interest in developing anti-cachexia therapies, there are currently no FDA approved treatments for cancer cachexia.1,12,13 Anamorelin, a ghrelin receptor agonist, was approved for cancer cachexia in Japan in 2020 based on the results of 2 prospective Japanese studies.14 However, it has not gained approval in the US due to insufficient evidence to support its efficacy, as 2 international randomized phase III trials failed to meet their functional co-primary endpoint of improved handgrip strength.14,15 More recently, ponsegromab, a humanized monoclonal antibody against GDF-15, was shown to improve weight, appetite, and physical activity in patients with non-small cell lung cancer, pancreatic cancer, and colorectal cancer in a phase 2 study.16
Despite the importance of weight loss and cachexia in prognosis and quality of life, there is no uniform system for the diagnosis, grading, and assessment of cachexia that is followed in oncology clinics. Weight loss, however, is consistently included among the various criteria that have been proposed since it is usually the first clinical sign of cachexia.3,17 Although the trajectory of weight loss differs for each patient, a threshold of 5% loss is widely accepted to be a marker of high risk for poor outcomes.6 The 2011 Delphi International Consensus, which set forth multiple criteria for cancer cachexia, included weight loss greater than 5% within 6 months.3 The etiology of weight loss in cancer cachexia is multifactorial and varies between patients and cancer types. Anorexia, reduced ability to chew food, or inability to absorb and utilize nutrients are some of the common complications of cachexia in patients.11,18 Even though reduced food intake is a major driver, anorexia is not the sole cause of weight loss, as cachexia cannot be completely reversed with nutritional supplementation alone.18 In patients with cancer, tumors and tumor-associated circulating factors can cause weight loss by inducing catabolic processes like protein degradation and reduced protein synthesis, which subsequently promotes the loss of muscle and adipose tissue.1,19,20
The treatment of non-small cell lung cancer (NSCLC) has been transformed by novel targeted agents that improve survival for patients with tumors harboring specific molecular aberrations. Osimertinib, a third-generation EGFR tyrosine kinase inhibitor (TKI), has replaced chemotherapy and early-generation TKI’s as the standard of care for patients with metastatic EGFR-mutant NSCLC. It has a more favorable side effect profile compared to chemotherapy agents and led to a median overall survival of over 30 months for patients with metastatic disease in the landmark FLAURA trial.21
Weight loss was previously shown to be an independent predictor of shorter overall survival (OS) in patients with lung cancer who are undergoing chemotherapy.22 However, the frequency and prognostic significance of cachexia in the context of new targeted therapies like osimertinib have not been studied. Although 20% of patients in FLAURA reported decreased appetite while on osimertinib, the impact of anorexia on patient outcomes was not studied, and other clinical markers of cachexia such as weight loss and changes in muscle mass were not described.21
To the best of our knowledge, the frequency of weight loss in patients on osimertinib for EGFR-mutant NSCLC is not known, and the implications of weight loss on survival in this patient population have not been adequately investigated. To address this knowledge gap, our objective was to explore whether the prevalence and negative prognostic value of cachexia holds true for patients treated with osimertinib for metastatic EGFR-mutant NSCLC. We conducted a pilot retrospective study to describe the frequency of weight loss in patients treated with first-line osimertinib for metastatic EGFR-mutant NSCLC at our cancer center at Columbia University, New York, and to study the impact of weight changes on patient survival. We chose weight loss as a proxy for cachexia since it is the basis of most definitions of cancer cachexia, is the primary method of diagnosing cachexia in the clinical setting, and is an objective finding that is reliably documented during all in-person visits for patients seen in our clinic.23
Even though osimertinib has greatly improved outcomes for patients with EGFR-mutant NSCLC, weight loss has been observed in some patients on osimertinib in our clinical practice. Based on our clinical observations, we hypothesized that weight loss would be observed in a subset of patients on osimertinib and could serve as a determinant of poor outcomes in this patient population.
Materials and methods
Patients
We conducted a single-institution retrospective study in adult patients with metastatic EGFR-mutant NSCLC who were treated with osimertinib in the first-line setting. We obtained records from our institution’s specialty pharmacy and initially identified 85 patients who were prescribed osimertinib as first-line treatment for metastatic EGFR-mutant NSCLC from 2017 to 2022. Out of 85 patients, we included 56 patients for analysis who met the following inclusion criteria: (1) treated with osimertinib for at least 6 months, (2) had a documented weight value in the electronic medical record at the time of starting osimertinib (baseline weight), and (3) had a documented weight value in the electronic medical record from at least one of the following time points to determine the presence of on-treatment weight loss: 6 months and/or 12 months. Twenty-nine patients were excluded due to a lack of weight documentation in the chart.
Patient data
We collected the following information for each patient from the electronic medical record: age at diagnosis, sex, race/ethnicity, ECOG performance score at the time of diagnosis (0-4), date of stage IV diagnosis, sites of metastatic disease at diagnosis (systemic only vs central nervous system involvement), smoking history (never vs ever), comorbidities that may be associated with cachexia (such as chronic obstructive pulmonary disease, heart failure, and chronic kidney disease), concurrent medications that may cause weight changes (such as chronic steroids and psychotropic medications), EGFR mutation type (exon 19 deletion vs exon 21 L858R), histology, osimertinib start date, presence of self-reported anorexia symptoms (such as loss of appetite and decreased food intake) during treatment with osimertinib, date of disease progression (if applicable), date of last clinic visit, date of stopping osimertinib (if applicable), and date of death (if applicable).
Weight values in kilograms (kg) were collected at the following time points from the electronic medical record: time of starting osimertinib, 6 months on treatment, 12 months on treatment, and time of disease progression (if applicable). Given prior definitions of cancer cachexia in the literature, we defined on-treatment weight loss in this study as a decrease in relative weight by greater than or equal to 5% from baseline weight at either 6 months or 12 months on treatment with osimertinib.3
Data analysis
Summary statistics were performed to describe the patient cohort. Pearson’s Chi-squared test, Wilcoxon rank sum test, and Fisher’s exact test were used to describe differences between patients with on-treatment weight loss (≥5% from baseline) and those without weight loss.
With on-treatment weight loss as the outcome of interest, a logistic regression model was also used to model the relationship with covariates including age, sex, EGFR mutation type, ECOG score at time of diagnosis, comorbidities, concurrent medications, sites of metastatic disease, and presence of anorexia symptoms while on treatment. We next investigated the impact of weight loss on progression-free survival (PFS), defined as the time from starting osimertinib to either disease progression or death, whichever occurred first, and overall survival (OS), defined as the time from starting osimertinib to death. We also investigated the impact of weight loss on time on treatment, defined as the time from starting osimertinib to either stopping osimertinib or death, whichever occurred first. We felt this was an important clinical outcome to capture for 2 reasons: (1) osimertinib is sometimes discontinued in the setting of drug-related adverse events, and (2) treatment beyond progression is often considered in patients who are believed to derive clinical benefit from osimertinib despite radiographic progression.
A censor date of October 12, 2023 was used for the PFS, OS, and time on treatment analyses. Kaplan-Meier curves and log-rank test statistics were computed to compare PFS, OS, and time on treatment between patients with on-treatment weight loss and those without weight loss. Cox proportional hazards models were fitted for PFS, time on treatment, and OS to investigate the associations with on-treatment weight loss and adjusted for the following covariates: age, sex, ECOG performance status at the time of diagnosis, and sites of metastatic disease at diagnosis.
All analyses were conducted using R version 4.3.1 with the tidyverse (v2.0.0), gtsummary (v1.7.2), ggplot2 (v3.4.3), ggsurvfit (v0.3.0), and survival (v3.5-7) packages. All hypothesis tests were conducted with a significance level at .05.
Results
We identified a cohort of 56 patients who were treated with first-line osimertinib for metastatic EGFR-mutant NSCLC. On-treatment weight loss was defined as weight loss of 5% or greater from baseline at either month 6 or month 12 while on treatment with osimertinib.
Patient characteristics of the 56 patients are noted in Table 1. Among the entire cohort, 64% of patients are female, and 66% are never-smokers. Most (70%) patients had systemic-only metastatic disease at the time of diagnosis, and 64% of patients had disease progression on first-line osimertinib by the time of data cutoff. All patients harbored one of the 2 classical EGFR mutations, deletion exon 19 (52%, n = 29) or exon 21 L858R (48%, n = 27). Three patients harbored L858R with a concurrent atypical EGFR mutation-T790M (n = 2) or E709A (n = 1) and were included in the L858R group. Thirty-four percent (n = 19) of patients reported anorexia symptoms such as loss of appetite during treatment. Forty-six percent (n = 26) of patients met criteria for on-treatment weight loss. When comparing patients with on-treatment weight loss to those without weight loss, there were no significant differences in terms of age, sex, race, smoking history, EGFR mutation type, sites of metastatic disease, ECOG performance score, concurrent weight-impacting comorbidities, and concurrent weight-impacting medications. Although the frequency of self-reported anorexia was double in patients with objective on-treatment weight loss (46%) compared to patients without weight loss (23%), this difference was not significant (P = .072).
Patient characteristics in the overall study population (n = 56) and by weight loss status.
| Characteristic | Overall (n = 56)1 | Patients without weight loss (n = 30)1 | Patients with weight loss (n = 26)1 | P-value2 |
|---|---|---|---|---|
| Sex | .9 | |||
| Female | 36 (64%) | 19 (63%) | 17 (65%) | |
| Male | 20 (36%) | 11 (37%) | 9 (35%) | |
| Age at diagnosis in years (range) | 68 (62, 76) | 66 (62, 73) | 72 (66, 78) | .13 |
| Ethnicity | .062 | |||
| Asian | 5 (8.9%) | 4 (13%) | 1 (3.8%) | |
| Black | 6 (11%) | 6 (20%) | 0 (0%) | |
| Declined | 4 (7.1%) | 2 (6.7%) | 2 (7.7%) | |
| Other | 11 (20%) | 4 (13%) | 7 (27%) | |
| White | 30 (54%) | 14 (47%) | 16 (62%) | |
| Smoking | .5 | |||
| Former | 18 (32%) | 11 (37%) | 7 (27%) | |
| Never | 37 (66%) | 18 (60%) | 19 (73%) | |
| Unknown | 1 (1.8%) | 1 (3.3%) | 0 (0%) | |
| EGFR mutation type | .2 | |||
| Exon 19 deletion | 29 (52%) | 18 (60%) | 11 (42%) | |
| Exon 21 L858R | 27 (48%) | 12 (40%) | 15 (58%) | |
| Sites of metastatic disease at diagnosis | .3 | |||
| Systemic only | 39 (70%) | 19 (63%) | 20 (77%) | |
| CNS involvement | 17 (30%) | 11 (37%) | 6 (23%) | |
| ECOG performance score at diagnosis | .11 | |||
| 0-1 | 49 (88%) | 24 (80%) | 25 (96%) | |
| ≥2 | 7 (13%) | 6 (20%) | 1 (3.8%) | |
| Concurrent weight-impacting comorbidity3 | 9 (16%) | 4 (13%) | 5 (19%) | .7 |
| Concurrent weight-impacting medication(s)4 | 7 (13%) | 3 (10%) | 4 (15%) | .7 |
| Patient-reported anorexia symptoms | 19 (34%) | 7 (23%) | 12 (46%) | .072 |
| Dose reduction required | 9 (16%) | 3 (10%) | 6 (23%) | .3 |
| Disease progression | 36 (64%) | 20 (67%) | 16 (62%) | .7 |
| Characteristic | Overall (n = 56)1 | Patients without weight loss (n = 30)1 | Patients with weight loss (n = 26)1 | P-value2 |
|---|---|---|---|---|
| Sex | .9 | |||
| Female | 36 (64%) | 19 (63%) | 17 (65%) | |
| Male | 20 (36%) | 11 (37%) | 9 (35%) | |
| Age at diagnosis in years (range) | 68 (62, 76) | 66 (62, 73) | 72 (66, 78) | .13 |
| Ethnicity | .062 | |||
| Asian | 5 (8.9%) | 4 (13%) | 1 (3.8%) | |
| Black | 6 (11%) | 6 (20%) | 0 (0%) | |
| Declined | 4 (7.1%) | 2 (6.7%) | 2 (7.7%) | |
| Other | 11 (20%) | 4 (13%) | 7 (27%) | |
| White | 30 (54%) | 14 (47%) | 16 (62%) | |
| Smoking | .5 | |||
| Former | 18 (32%) | 11 (37%) | 7 (27%) | |
| Never | 37 (66%) | 18 (60%) | 19 (73%) | |
| Unknown | 1 (1.8%) | 1 (3.3%) | 0 (0%) | |
| EGFR mutation type | .2 | |||
| Exon 19 deletion | 29 (52%) | 18 (60%) | 11 (42%) | |
| Exon 21 L858R | 27 (48%) | 12 (40%) | 15 (58%) | |
| Sites of metastatic disease at diagnosis | .3 | |||
| Systemic only | 39 (70%) | 19 (63%) | 20 (77%) | |
| CNS involvement | 17 (30%) | 11 (37%) | 6 (23%) | |
| ECOG performance score at diagnosis | .11 | |||
| 0-1 | 49 (88%) | 24 (80%) | 25 (96%) | |
| ≥2 | 7 (13%) | 6 (20%) | 1 (3.8%) | |
| Concurrent weight-impacting comorbidity3 | 9 (16%) | 4 (13%) | 5 (19%) | .7 |
| Concurrent weight-impacting medication(s)4 | 7 (13%) | 3 (10%) | 4 (15%) | .7 |
| Patient-reported anorexia symptoms | 19 (34%) | 7 (23%) | 12 (46%) | .072 |
| Dose reduction required | 9 (16%) | 3 (10%) | 6 (23%) | .3 |
| Disease progression | 36 (64%) | 20 (67%) | 16 (62%) | .7 |
1n (%); Median (IQR).
2Pearson’s Chi-squared test; Wilcoxon rank sum test; Fisher’s exact test.
3Comorbidities include those that may be associated with cachexia: chronic obstructive pulmonary disease, chronic kidney disease, and heart failure.
4Concurrent medications include those that may be associated with weight gain: steroids, mirtazapine, megestrol acetate, paroxetine, and sertraline.
Patient characteristics in the overall study population (n = 56) and by weight loss status.
| Characteristic | Overall (n = 56)1 | Patients without weight loss (n = 30)1 | Patients with weight loss (n = 26)1 | P-value2 |
|---|---|---|---|---|
| Sex | .9 | |||
| Female | 36 (64%) | 19 (63%) | 17 (65%) | |
| Male | 20 (36%) | 11 (37%) | 9 (35%) | |
| Age at diagnosis in years (range) | 68 (62, 76) | 66 (62, 73) | 72 (66, 78) | .13 |
| Ethnicity | .062 | |||
| Asian | 5 (8.9%) | 4 (13%) | 1 (3.8%) | |
| Black | 6 (11%) | 6 (20%) | 0 (0%) | |
| Declined | 4 (7.1%) | 2 (6.7%) | 2 (7.7%) | |
| Other | 11 (20%) | 4 (13%) | 7 (27%) | |
| White | 30 (54%) | 14 (47%) | 16 (62%) | |
| Smoking | .5 | |||
| Former | 18 (32%) | 11 (37%) | 7 (27%) | |
| Never | 37 (66%) | 18 (60%) | 19 (73%) | |
| Unknown | 1 (1.8%) | 1 (3.3%) | 0 (0%) | |
| EGFR mutation type | .2 | |||
| Exon 19 deletion | 29 (52%) | 18 (60%) | 11 (42%) | |
| Exon 21 L858R | 27 (48%) | 12 (40%) | 15 (58%) | |
| Sites of metastatic disease at diagnosis | .3 | |||
| Systemic only | 39 (70%) | 19 (63%) | 20 (77%) | |
| CNS involvement | 17 (30%) | 11 (37%) | 6 (23%) | |
| ECOG performance score at diagnosis | .11 | |||
| 0-1 | 49 (88%) | 24 (80%) | 25 (96%) | |
| ≥2 | 7 (13%) | 6 (20%) | 1 (3.8%) | |
| Concurrent weight-impacting comorbidity3 | 9 (16%) | 4 (13%) | 5 (19%) | .7 |
| Concurrent weight-impacting medication(s)4 | 7 (13%) | 3 (10%) | 4 (15%) | .7 |
| Patient-reported anorexia symptoms | 19 (34%) | 7 (23%) | 12 (46%) | .072 |
| Dose reduction required | 9 (16%) | 3 (10%) | 6 (23%) | .3 |
| Disease progression | 36 (64%) | 20 (67%) | 16 (62%) | .7 |
| Characteristic | Overall (n = 56)1 | Patients without weight loss (n = 30)1 | Patients with weight loss (n = 26)1 | P-value2 |
|---|---|---|---|---|
| Sex | .9 | |||
| Female | 36 (64%) | 19 (63%) | 17 (65%) | |
| Male | 20 (36%) | 11 (37%) | 9 (35%) | |
| Age at diagnosis in years (range) | 68 (62, 76) | 66 (62, 73) | 72 (66, 78) | .13 |
| Ethnicity | .062 | |||
| Asian | 5 (8.9%) | 4 (13%) | 1 (3.8%) | |
| Black | 6 (11%) | 6 (20%) | 0 (0%) | |
| Declined | 4 (7.1%) | 2 (6.7%) | 2 (7.7%) | |
| Other | 11 (20%) | 4 (13%) | 7 (27%) | |
| White | 30 (54%) | 14 (47%) | 16 (62%) | |
| Smoking | .5 | |||
| Former | 18 (32%) | 11 (37%) | 7 (27%) | |
| Never | 37 (66%) | 18 (60%) | 19 (73%) | |
| Unknown | 1 (1.8%) | 1 (3.3%) | 0 (0%) | |
| EGFR mutation type | .2 | |||
| Exon 19 deletion | 29 (52%) | 18 (60%) | 11 (42%) | |
| Exon 21 L858R | 27 (48%) | 12 (40%) | 15 (58%) | |
| Sites of metastatic disease at diagnosis | .3 | |||
| Systemic only | 39 (70%) | 19 (63%) | 20 (77%) | |
| CNS involvement | 17 (30%) | 11 (37%) | 6 (23%) | |
| ECOG performance score at diagnosis | .11 | |||
| 0-1 | 49 (88%) | 24 (80%) | 25 (96%) | |
| ≥2 | 7 (13%) | 6 (20%) | 1 (3.8%) | |
| Concurrent weight-impacting comorbidity3 | 9 (16%) | 4 (13%) | 5 (19%) | .7 |
| Concurrent weight-impacting medication(s)4 | 7 (13%) | 3 (10%) | 4 (15%) | .7 |
| Patient-reported anorexia symptoms | 19 (34%) | 7 (23%) | 12 (46%) | .072 |
| Dose reduction required | 9 (16%) | 3 (10%) | 6 (23%) | .3 |
| Disease progression | 36 (64%) | 20 (67%) | 16 (62%) | .7 |
1n (%); Median (IQR).
2Pearson’s Chi-squared test; Wilcoxon rank sum test; Fisher’s exact test.
3Comorbidities include those that may be associated with cachexia: chronic obstructive pulmonary disease, chronic kidney disease, and heart failure.
4Concurrent medications include those that may be associated with weight gain: steroids, mirtazapine, megestrol acetate, paroxetine, and sertraline.
A logistic regression model was used to identify predictors of on-treatment weight loss. No covariates, including age, sex, EGFR mutation type, presence of anorexia symptoms, ECOG performance status, comorbidities, concurrent medications, sites of metastatic disease, or disease progression, were significantly associated with on-treatment weight loss status.
Cox proportional hazard models were used to compare PFS, time on treatment, and OS between patients with and without on-treatment weight loss (Figures 1-3). After adjusting for age, sex, ECOG performance status, and sites of metastatic disease, there was no significant difference in PFS (Figure 1; HR 1.18, 95% CI, 0.56-2.46, P = .7). Although there was a trend toward shorter time on treatment in patients with on-treatment weight loss, this was also not significant (Figure 2; HR 2.08, 95% CI, 0.90-4.80, P = .088). However, patients with on-treatment weight loss had worse overall survival compared to those without on-treatment weight loss, and this was statistically significant (Figure 3; HR 4.91, 95% CI, 1.56-15.5, P = .007).
Kaplan-Meier curve for progression-free survival by weight loss status. Using a Cox proportional hazards model and adjusting for age, sex, ECOG performance score, and sites of metastasis at diagnosis, patients with on-treatment weight loss (“Yes”) did not have significantly different PFS compared to patients without weight loss (“No”); HR 1.18, 95% CI, 0.56, 2.46, P = .7.
Kaplan-Meier curve for time on osimertinib by weight loss status. Using a Cox proportional hazards model and adjusting for age, sex, ECOG performance score, and sites of metastasis at diagnosis, patients with on-treatment weight loss (“Yes”) did not have significantly different time on osimertinib compared to patients without weight loss (“No”); HR 2.08, 95% CI, 0.90, 4.80, P = .088.
Kaplan-Meier curve for overall survival by weight loss status. Using a Cox proportional hazards model and adjusting for age, sex, ECOG performance score, and sites of metastasis at diagnosis, patients with on-treatment weight loss (“Yes”) had significantly worse overall survival compared to patients without weight loss (“No”); HR 4.91, 95% CI, 1.56-15.5, P = .007.
Discussion
Although the treatment landscape for lung cancer has evolved tremendously in the context of novel targeted therapies, cancer cachexia remains a largely unaddressed problem with negative implications on patient survival and quality of life. It is important to characterize cancer cachexia in lung cancer patients as novel therapies are introduced and treatment paradigms change, since a key objective of ongoing research efforts is to develop effective anti-cachexia therapies to complement cancer-directed treatment and improve patient outcomes.
To the best of our knowledge, this pilot study is the first to assess the frequency of weight loss in patients treated with osimertinib for metastatic EGFR-mutant NSCLC and examine the relationship between on-treatment weight loss and patient survival. We found that nearly half of the patients in our cohort met the definition for on-treatment weight loss. Importantly, patients who lost weight during treatment had significantly worse overall survival.
Weight loss was observed in patients in this study and has been observed in our clinical practice during treatment with osimertinib. However, the specific cause(s) of weight loss in patients treated with osimertinib remain unclear and will be explored in future studies. While a drug-related toxicity may contribute to weight loss, weight loss was not noted to be a common adverse event in clinical trials of osimertinib or earlier-generation EGFR-TKI’s.24–26 Decreased appetite (mostly grades 1-2) was observed in 20% of patients treated with osimertinib in the FLAURA trial and was common with erlotinib and gefitinib, suggesting a class effect on anorexia.24–26 However, the frequency of weight loss with these agents has not been reported in clinical trials. Interestingly, weight increase is common with the ALK-TKI’s alectinib and lorlatinib, although this may be due in part to increased edema.27,28 Both weight increase and decreased appetite were observed in clinical trials with lorlatinib, highlighting some discrepancies that may arise when using weight alone as a marker for cachexia.28
In this pilot study, the presence of weight loss in osimertinib-treated patients did not correlate with age, sex, race, smoking history, EGFR mutation type, comorbidities, concurrent medications, sites of metastatic disease, or ECOG performance score. As such, our data do not point towards a specific cause of weight loss and cannot distinguish between weight loss that results from the cancer cachexia syndrome, from anorexia alone, from a possible drug-related side effect, or from a combination of factors. Since we did not observe any differences in disease progression or sites of metastasis between patients with and without weight loss, we are unable to say whether tumor burden or response to therapy correlates with weight loss in this patient population. Self-reported anorexia symptoms, which were observed in 34% of patients, were also not predictive of weight loss. This may suggest that the weight loss observed in these patients was not explained by loss of appetite or low food intake alone. The impact of a drug-related effect on weight might be elucidated by revisiting data from prior clinical trials to identify associations between drug pharmacokinetics and weight change. Additional retrospective and prospective studies that examine the relationship between weight loss and other clinical measures of cachexia like anorexia are needed.
The negative relationship between weight loss and OS in our patient cohort is not unexpected given the well-known association between cachexia and poor prognosis in patients with cancer. Nonetheless, our findings reinforce the prevalence of weight loss in patients on osimertinib and its striking impact on patient survival. Although patients with on-treatment weight loss had worse OS, there were no significant differences in PFS or time on treatment with osimertinib. This suggests that while weight loss may be a negative prognostic factor in this patient population, it does not appear to be predictive of tumor response to osimertinib in this pilot study. Since time on treatment was not significantly impacted by the presence of weight loss, weight loss may not affect the tolerability of osimertinib. Prior work has demonstrated that cachexia and weight loss are associated with lower response rates to chemotherapy and inability to tolerate chemotherapy.6,29,30 However, it is not clear whether cachexia directly mediates tumor response to chemotherapy or whether cachexia is simply a reflection of aggressive disease and poor patient functional status. Our findings may suggest the latter for patients with NSCLC treated with osimertinib.
Previous work suggested factors including age, gender, race/ethnicity, socioeconomic status, and comorbidities such as COPD to be associated with cancer cachexia.5,31,32 Although we did not find any predictors of on-treatment weight loss in this study, we are limited by a small sample size. Future work in larger cohorts should revisit whether risk factors for weight loss and/or other features of cachexia can be identified in patients with metastatic EGFR-mutant NSCLC. In addition to examining demographic predictors, it would be interesting to see if disease-related factors such as EGFR mutation type or the presence of TP53 co-mutations and other co-mutations associated with poor prognosis are related to weight loss in this cohort.33
Our study has several limitations since it is a single-center retrospective study of a small cohort of patients. Although we initially identified a larger patient cohort, several patients were excluded from analysis due to a lack of documented weights. This was in part due to the replacement of in-person visits with telemedicine visits during the COVID-19 pandemic. There was a particularly small number of patients with brain metastases, which limits the generalizability of our results. Furthermore, although we adjust for performance status and metastatic disease sites in our survival comparisons, there are other factors, such as volume of tumor burden, that are difficult to quantify and control for but may have impacted both weight loss and survival. Nevertheless, our study suggests that in this population of patients with a highly treatable subtype of lung cancer with a highly specific and well-tolerated targeted agent (osimertinib), cachexia-related symptoms such as weight loss still exist in close to half of the patients under treatment and hold independent prognostic value. Given the importance of weight loss in this patient population, our objective is to expand this study to a larger multicenter patient cohort and incorporate other measures of cachexia such as inflammatory markers on laboratory testing and muscle mass loss by CT imaging.34
This study was designed to create awareness of weight loss and cachexia in oncology clinics and highlight the need for anti-cachexia therapeutics to improve patient quality of life and survival. There are currently no FDA approved treatments for cancer cachexia, and nutritional support remains the mainstay of clinical management.23 However, there are a growing number of investigational agents that seek to reverse cancer cachexia by targeting pathways involving appetite regulation and skeletal muscle mass catabolism.14,16,35 For example, our laboratory has identified ZIP14 as a promising therapeutic target to treat cachexia in metastatic lung cancer.1,36 We anticipate that these efforts will eventually lead to the implementation of effective anti-cachexia treatments alongside cancer-directed therapies for patients with lung cancer, including patients with EGFR-mutant NSCLC. Our findings can serve as a foundation for additional research efforts that will enhance patient selection for anti-cachexia therapies in the future.
Acknowledgments
This work was supported by the NCI R01 CA231239, Irma T. Hirschl Monique Weill-Caulier Trust Award, Pershing Square Sohn Award (to S.A.), HICCC and Irving Institute Pilot Awards (to S.A.) funded through HICCC P30CA013696. These studies utilized the resources from the Herbert Irving Comprehensive Cancer Center (HICCC) core facilities, which are funded in part through Center Grant P30CA013696.
Author contributions
Lanyi Nora Chen: Conceptualization, Formal Analysis, Investigation, Methodology, Visualization, Writing—original draft, Writing—review & editing; Xin Ma: Data curation, Formal Analysis, Methodology, Visualization, Writing—original draft, Writing—review & editing; Benjamin Herzberg: Writing—review & editing; Brian S. Henick: Writing—review & editing; Anup K. Biswas: Conceptualization, Writing—review & editing; Swarnali Acharyya: Conceptualization, Investigation, Methodology, Resources, Supervision, Writing—original draft, Writing—review & editing; Catherine A. Shu: Conceptualization, Investigation, Methodology, Supervision, Writing—original draft, Writing—review & editing.
Conflicts of interest
B.H. reports personal fees from Astella, Eli Lilly, Amgen, Eisai, MJH HealthSciences, Ideology Health, Boxer Capital, and Guardant as well as institutional research funding from Amgen, Astellas, Repare Therapeutics, Ideaya Therapeutics, Revolution Medicines, Monte Rosa Therapeutics, AstraZeneca, Seagen, Nested Therapeutics, Prelude Therapeutics. B.S.H. serves on an advisory board for AstraZeneca, Ideaya, Jazz Pharmaceuticals, Sorrento Therapeutics, Genentech-Roche, Regeneron, BMS, OncLive and reports consulting fees from Veeva, Athenium, Boxer Capital, SAI-Med, and DAVA Oncology. He reports institutional research funding from NexImmune, Genentech-Roche, Johnson & Johnson, BMS Foundation/VCU, Stand Up 2 Cancer, V Foundation, National Cancer Institute. C.A.S. reports personal fees from Arcus Biosciences, AstraZeneca, Daiichi-Sankyo, Genentech, Janssen, and Takeda. L.N.C., X.M., A.K.B., and S.A. report no financial COI.
Data availability
No new data were generated or analyzed in support of this research.
