Association between pretreatment emotional distress and immune checkpoint inhibitor response in non-small-cell lung cancer

Abstract

Emotional distress (ED), commonly characterized by symptoms of depression and/or anxiety, is prevalent in patients with cancer. Preclinical studies suggest that ED can impair antitumor immune responses, but few clinical studies have explored its relationship with response to immune checkpoint inhibitors (ICIs). Here we report results from cohort 1 of the prospective observational STRESS-LUNG study, which investigated the association between ED and clinical efficacy of first-line treatment of ICIs in patients with advanced non-small-cell lung cancer. ED was assessed by Patient Health Questionnaire-9 and Generalized Anxiety Disorder 7-item scale. The study included 227 patients with 111 (48.9%) exhibiting ED who presented depression (Patient Health Questionnaire-9 score ≥5) and/or anxiety (Generalized Anxiety Disorder 7-item score ≥5) symptoms at baseline. On the primary endpoint analysis, patients with baseline ED exhibited a significantly shorter median progression-free survival compared with those without ED (7.9 months versus 15.5 months, hazard ratio 1.73, 95% confidence interval 1.23 to 2.43, P = 0.002). On the secondary endpoint analysis, ED was associated with lower objective response rate (46.8% versus 62.1%, odds ratio 0.54, P = 0.022), reduced 2-year overall survival rate of 46.5% versus 64.9% (hazard ratio for death 1.82, 95% confidence interval 1.12 to 2.97, P = 0.016) and detriments in quality of life. The exploratory analysis indicated that the ED group showed elevated blood cortisol levels, which was associated with adverse survival outcomes. This study suggests that there is an association between ED and worse clinical outcomes in patients with advanced non-small-cell lung cancer treated with ICIs, highlighting the potential significance of addressing ED in cancer management. ClinicalTrials.gov registration: NCT05477979 .

Fig. 1. Flowchart diagram of STRESS-LUNG-1.

The data cutoff was 30 November 2023.

Fig. 2. Kaplan–Meier curve and subgroup analysis of investigator-assessed PFS by baseline ED.

a, Investigator-assessed PFS in the ED and no ED group. P values were calculated using a two-sided log-rank test. The HR and the corresponding 95% CI were calculated using a Cox proportional-hazards regression. b, The forest plots show HR and 95% CI of the subgroup analysis of PFS. Data are presented as the HR with error bars showing 95% CI. mPFS, median PFS; NR, not reached; SCC, squamous carcinoma.

Fig. 3. Kaplan–Meier curve of investigator-assessed PFS by baseline ED after propensity score analyses.

a, Population based on PSM analysis. The exact P value was 0.0002. b, Population based on IPTW. In a and b, the final covariates were sex and pathology for PSM and IPTW, which were independently associated with ED on the basis of the multivariable logistic regression model. P values were calculated using a two-sided log-rank test. No adjustment was made for multiple comparisons. mPFS, median PFS.

Fig. 4. Survival analyses by Time 2 ED.

a, The PFS analysis in the ED and no ED groups by Time 2 assessments. The exact P value was 0.0007. b, The OS analysis in the ED and no ED groups by Time 2 assessments. In a and b, P values were calculated using a two-sided log-rank test. mPFS, median PFS; mOS, median OS; NR, not reached.

Extended Data Fig. 1. Study design of STRESS-LUNG-1.

Stage IIIB-IV NSCLC patients who are about to receiving ICIs or combination therapy with chemotherapy as first-line treatment are screened for enrollment in this study. Emotional distress at baseline is assessed using the PHQ-9 and GAD-7 scales, while QoL is evaluated based on the EORTC QLQ-C30 questionnaire; A second assessment (Time 2) of emotional distress is conducted during follow-up. The efficacy of ICIs is evaluated every two cycles. Abbreviations: NSCLC, non-small-cell lung cancer; ICIs, immune checkpoint inhibitors; PHQ-9, Patient Health Questionnaire-9; GAD-7, Generalized anxiety disorder 7-item; EORTC QLQ-C30, European organization for research and treatment of cancer core quality of life questionnaire; ACTH, adrenocorticotropic hormone; RECIST, Response Evaluation Criteria in Solid Tumors.

Extended Data Fig. 2. The distribution of emotional distress.

a, Distribution of severity of emotional distress; b, Distribution of emotional distress, including depression and anxiety symptoms; c, Number of patients of each PHQ-9 score; d, Number of patients of each GAD-7 score; Abbreviations: ED, emotional distress. PHQ-9, Patient Health Questionnaire-9; GAD-7, Generalized anxiety disorder 7-item.

Extended Data Fig. 3. The associations between severity of emotional distress and progression-free survival.

a, The overall population. b, PSM-based population; The exact P value was 0.0008 in comparison between no emotional distress (ED) group and mild ED group. c, IPTW-based population. a–c, P values were calculated using the two-sided log-rank test. b, c, The final covariates were sex and pathology for PSM and IPTW, which were independently associated with emotional distress based on multivariable logistic regression model. No adjustment was made for multiple comparisons. Abbreviations: ED, emotional distress; PSM, propensity score matching; IPTW, inverse probability of treatment weighting; HR, hazard ratio; CI, confidence interval; Ref, reference.

Extended Data Fig. 4. The associations between emotional distress, objective response rate and overall survival.

a, Waterfall plots for best percentage change in target lesion size for no emotional distress (ED) group; b, Waterfall plots for best percentage change in target lesion size for ED group. c, The overall survival of ED and no ED group in overall population; d, The overall survival of ED and no ED group in PSM-based population; e, The overall survival of ED and no ED group in IPTW-based population; f, The association between severity of emotional distress and overall survival in the overall population. d, e, The final covariates were sex and pathology for PSM and IPTW, which were independently associated with emotional distress based on multivariable logistic regression model. c–f, P values were calculated using the two-sided log-rank test. No adjustment was made for multiple comparisons. Abbreviations: ED, emotional distress; ORR, objective response rate; PSM, propensity score matching; IPTW, inverse probability of treatment weighting; HR, hazard ratio; CI, confidence interval; Ref, reference.

Extended Data Fig. 5. Emotional distress dynamics from baseline to follow-up.

a, Longitudinal changes of emotional distress; b, The consistent analysis of PHQ-9 score between baseline and Time 2 assessments (n = 187); c, The consistent analysis of GAD-7 score between baseline and Time 2 assessments (n = 187); d, The associations between emotional distress dynamics and objective response rate, P values were calculated using the two-sided χ² test. Abbreviations: ED, emotional distress; PHQ-9, Patient Health Questionnaire-9; GAD-7, Generalized anxiety disorder 7-item; SD, standard deviation; ORR, objective response rate.

Extended Data Fig. 6. The associations between emotional distress dynamics and survival outcomes.

a, The progression-free survival analysis in emotional distress (ED) remission and persistence subgroups; b, The progression-free survival analysis in never and onset subgroups; c, The overall survival analysis in remission and persistence subgroups; d, The overall survival analysis in never and onset subgroups. a–d, P values were calculated using the two-sided log-rank test. Abbreviations: ED, emotional distress; PFS, progression-free survival; OS, overall survival; HR, hazard ratio; CI, confidence interval.

Extended Data Fig. 7. The association of peripheral blood stress hormones with emotional distress and clinical outcomes.

a, Serum cortisol level in emotional distress (ED, n = 107) and no emotional distress (no ED, n = 103) groups; b, Serum ACTH level in ED (n = 107) and no ED (n = 103) groups; c, The progression-free survival analysis in different cortisol levels groups; d, The overall survival analysis in different cortisol levels groups. a, b. For each boxplot, the box is presented as median (interquartile range), whiskers extend from the hinges to the largest (or smallest) value no further than 1.5 interquartile range from the hinge. P-values were calculated using a two-sided nonparametric Mann–Whitney U test. c, d. P values were calculated using the two-sided log-rank test. Abbreviations: HR, hazard ratio; CI, confidence interval.