Predictors of Early Durvalumab Discontinuation After Chemoradiotherapy for Non-Small Cell Lung Cancer

Abstract

Based on the PACIFIC trial, adjuvant durvalumab is now recommended after concurrent chemoradiotherapy (CRT) for locally advanced non-small cell lung cancer (LA-NSCLC), with an intended course duration of one year. In PACIFIC, approximately 50% of subjects did not complete the planned durvalumab course. Here, we explore causes and predictors of early durvalumab discontinuation to test the hypothesis that extent of normal tissue irradiation may be associated with durvalumab treatment duration.We reviewed patients treated for LA-NSCLC with definitive CRT at our institution who began adjuvant durvalumab between 2017 and 2020. Duration of durvalumab receipt and causes for early discontinuation (defined as a course duration shorter than 11.5 months) were tabulated. Cox proportional hazards models were utilized to test clinical factors (age, performance status, tumor histology, presence of known driver mutations, PD-L1 expression), measures of disease burden (stage, metabolic tumor volume [MTV] on pre-treatment FDG-PET), and treatment characteristics (mean radiotherapy doses received by the esophagus, heart, and lungs) as predictors of early durvalumab discontinuation. These variables were also tested as predictors of progression-free survival (PFS) duration after durvalumab initiation. The Kaplan-Meier method was utilized to estimate rates of durvalumab completion and PFS for the entire cohort and for patient subgroups.Fifty-three patients met inclusion criteria. Nineteen patients completed a one-year course of durvalumab, twenty-eight patients discontinued durvalumab early, and six are still receiving durvalumab. The actuarial rate of completing a one-year durvalumab course was 36%. Reasons for early durvalumab discontinuation were disease progression (n = 10), lung toxicity (n = 8), other toxicity (n = 6), and death unrelated to cancer progression or treatment toxicity (n = 4). Univariate and multivariate Cox models identified mean heart dose as the only statistically significant predictor of early immunotherapy discontinuation (HR 1.11 per Gy, 95% CI 1.01 to 1.21, P = 0.025). The rate of durvalumab completion for patients with mean heart dose below the cohort median (9.7 Gy) was 55%, compared to only 20% for patients with higher mean heart doses (log rank P = 0.020). Mean heart dose was also the only statistically significant predictor of PFS after immunotherapy initiation in a multivariable Cox model (HR 1.19 per Gy, 95% CI 1.07 to 1.32, P = 0.001). Median PFS duration for patients with mean heart dose below the cohort median was 32 months, compared to only 8 months for patients with higher mean heart doses (log rank P < 0.001).For LA-NSCLC patients treated with chemoradiotherapy followed by immunotherapy, extent of cardiac irradiation may be a risk factor for early immunotherapy discontinuation and disease recurrence or death. A multi-institutional validation study is underway.