Abstract 776P
Background
Early cross-over of Kaplan-Meier survival curves seen in clinical trials of immunotherapy (IO) suggests that some patients with cancer treated with IO are at risk for Early Mortality (EM). To prevent potential harm and guide best practice, we sought to estimate the proportion of patients who die soon after starting IO in the real world setting and to examine potential demographic and disease-related factors associated with EM.
Methods
We conducted a retrospective cohort study using linked health administrative data from Ontario, Canada. EM was defined as death from any cause within 60 days of IO initiation. Patients with solid tumours (melanoma, lung, bladder, head and neck or kidney cancer) treated with IO between 2011 and 2020 were included. Multivariable logistic regression was used to identify factors associated with EM. Sensitivity analyses assessed EM at 30 and 90 days.
Results
A total of 7,127 patients treated with IO were evaluated. Median age was 67 (IQR 60-74), 42% were of female sex, and 58% had lung cancer. EM at 60 days was 15% (1,076/7,127) overall with the highest mortality observed in bladder and head and neck tumors (»21% each). In multivariable analysis, older patients, patients with history of recent hospital admission or emergency department visit, receipt of chemotherapy or radiation therapy prior to IO, stage 4 disease at diagnosis, lower hemoglobin (<10 g/dl), higher white blood cell count (>11,000/mm3) and higher baseline Edmonton Symptoms Assessment System (ESAS) scores (>=2) were at a higher risk of dying within 60 days post IO start. Conversely, patients with lung and kidney cancer (compared to melanoma), lower neutrophil-to-lymphocytes ratio (NLR<5), who received steroids after IO start, and with higher body-mass index (BMI>=25) were less likely to die within 60 days post IO initiation. In a sensitivity analysis, the 30 and 90-day mortality were 7% (519/7,127) and 22% (1,583/7,127), respectively with comparable adjusted analyses results.
Conclusions
EM is common among patients treated with IO in the real-world setting and is associated with several patient and tumor characteristics. Development of a validated tool to predict EM in this setting may facilitate better patient selection for treatment with IO in routine clinical practice.
Clinical trial identification
N/A
Editorial acknowledgement
N/A
Legal entity responsible for the study
The authors.
Funding
Medical Oncology Research Fund, London Regional Cancer Program.
Disclosure
J. Raphael: Financial Interests, Personal, Advisory Board: Lilly, AstraZeneca, Merck, Novartis; Financial Interests, Personal, Invited Speaker: Roche. All other authors have declared no conflicts of interest.