Search

BackgroundThe cumulative, health system-wide survival benefit of immune checkpoint inhibitors (ICIs) is unclear, particularly among real-world patients with limited life expectancies and among subgroups poorly represented on clinical trials. We sought to determine the health system-wide survival impact of ICIs.MethodsWe identified all patients receiving PD-1/PD-L1 or CTLA-4 inhibitors from 2010 to 2023 in the national Veterans Health Administration (VHA) system (ICI cohort) and all patients who received non-ICI systemic therapy in the years before ICI approval (historical control). ICI and historical control cohorts were matched on multiple cancer-related prognostic factors, comorbidities, and demographics. The effect of ICI on overall survival was quantified with Cox regression incorporating matching weights. Cumulative life-years gained system-wide were calculated from the difference in adjusted 5-year restricted mean survival times.ResultsThere were 27,322 patients in the ICI cohort and 69,801 patients in the historical control cohort. Among ICI patients, the most common cancer types were NSCLC (46%) and melanoma (10%). ICI demonstrated a large OS benefit in most cancer types with heterogeneity across cancer types (NSCLC: adjusted HR [aHR] 0.56, 95% confidence interval [CI] 0.54-0.58, p

PurposeReal world evidence is crucial to understanding the diffusion of new oncologic therapies, monitoring cancer outcomes, and detecting unexpected toxicities. In practice, real world evidence is challenging to collect rapidly and comprehensively, often requiring expensive and time-consuming manual case-finding and annotation of clinical text. In this Review, we summarise recent developments in the use of artificial intelligence to collect and analyze real world evidence in oncology.MethodsWe performed a narrative review of the major current trends and recent literature in artificial intelligence applications in oncology.ResultsArtificial intelligence (AI) approaches are increasingly used to efficiently phenotype patients and tumors at large scale. These tools also may provide novel biological insights and improve risk prediction through multimodal integration of radiographic, pathological, and genomic datasets. Custom language processing pipelines and large language models hold great promise for clinical prediction and phenotyping.ConclusionsDespite rapid advances, continued progress in computation, generalizability, interpretability, and reliability as well as prospective validation are needed to integrate AI approaches into routine clinical care and real-time monitoring of novel therapies.