Effect of concomitant medications with immune-modulatory properties on the outcomes of patients with advanced cancer treated with immune checkpoint inhibitors: development and validation of a novel prognostic index.

ABSTRACT

BACKGROUND: Concomitant medications are known to impact on clinical outcomes of patients treated with immune checkpoint inhibitors (ICIs). We aimed weighing the role of different concomitant baseline medications to create a drug-based prognostic score. METHODS: We evaluated concomitant baseline medications at immunotherapy initiation for their impact on objective response rate (ORR), progression-free survival (PFS) and overall survival (OS) in a single-institution cohort of patients with advanced cancer treated with ICIs (training cohort, N = 217), and a drug-based prognostic score with the drugs resulting significantly impacting the OS was computed. Secondly, we externally validated the score in a large multicenter external cohort (n = 1012). RESULTS: In the training cohort (n = 217), the median age was 69 years (range 32-89), and the primary tumours were non-small-cell lung cancer (70%), melanoma (14.7%), renal cell carcinoma (9.2%) and others (6%). Among baseline medications, corticosteroids (hazard ratio [HR] = 2.3; 95% confidence interval [CI] 1.60-3.30), systemic antibiotics (HR = 2.07; 95% CI 1.31-3.25) and proton-pump inhibitors (PPIs) (HR = 1.57; 95% CI 1.13-2.18) were significantly associated with OS. The prognostic score was calculated using these three drug classes, defining good, intermediate and poor prognosis patients. Within the training cohort, OS (p < 0.0001), PFS (p < 0.0001) and ORR (p = 0.0297) were significantly distinguished by the score stratification. The prognostic value of the score was also demonstrated in terms of OS (p < 0.0001), PFS (p < 0.0001) and ORR (p = 0.0006) within the external cohort. CONCLUSION: Cumulative exposure to corticosteroids, antibiotics and PPIs (three likely microbiota-modulating drugs) leads to progressively worse outcomes after ICI therapy. We propose a simple score that can help stratifying patients in routine practice and clinical trials of ICIs.