Tumor-specific supportive care utilization after first-line PD-1 inhibitor therapy: Real-world transfusion and mortality patterns in lung cancer and melanoma.

e20671 Background: Immune checkpoint inhibitors (ICIs) are standard first-line therapy for non–small cell lung cancer (NSCLC) and melanoma, but tumor-specific supportive-care needs remain poorly described. Clinicians often observe different clinical trajectories between these diseases. We conducted a descriptive, real-world comparison of transfusion and mortality after first-line PD-1 inhibitor monotherapy in NSCLC versus melanoma, using transfusion as a tumor-specific supportive-care endpoint. Methods: Using the TriNetX US Collaborative Network (69 healthcare organizations), we identified adults receiving first-line nivolumab or pembrolizumab monotherapy between January 2015 and May 2024. Cohort 1 included NSCLC (ICD-10 C34); Cohort 2 included melanoma (ICD-10 C43–C44). Patients with pre-treatment anemia (D64), neutropenia (D70), transfusion (CPT 36430), hospice/home health (HCPCS Q5001), or palliative care (Z51.5) before index were excluded. Outcomes (transfusion and all-cause mortality) were assessed from 1 day after PD-1 initiation through end of follow-up. Propensity score matching and time-to-event modeling were not used; all estimates are unadjusted cumulative risks and are interpreted descriptively, not causally. Results: The NSCLC and melanoma cohorts included 6,445 and 6,083 patients, respectively. Transfusion occurred in 141 NSCLC patients (2.2%) and 68 melanoma patients (1.1%), risk difference 1.1% (95% CI, 0.6–1.5; p < 0.001), risk ratio 1.96 (95% CI, 1.47–2.61). All-cause mortality was 38.5% in NSCLC (2,481/6,445) and 20.5% in melanoma (1,249/6,083), risk difference 18.0% (95% CI, 16.4–19.5; p < 0.001), risk ratio 1.88 (95% CI, 1.77–1.99). Mortality represents cumulative incidence over variable follow-up and not survival estimates from time-to-event analyses. In patients treated after 2020 (n = 4,847), transfusion patterns were similar, suggesting these descriptive findings are stable across more contemporary PD-1 practice. Conclusions: In this unmatched, federated EHR analysis, NSCLC patients had higher cumulative transfusion and mortality risks after first-line PD-1 therapy than melanoma patients, despite identical supportive-care exclusions. These differences are most plausibly explained by underlying tumor biology, comorbidities, and care patterns, and they should not be interpreted as causal treatment effects. Transfusion, although infrequent, may function as a marker of higher supportive-care intensity in select ICI-treated populations such as NSCLC, supporting future matched, time-to-event studies to design tumor-specific supportive-care strategies and earlier monitoring during immunotherapy.