Background: Unanticipated deaths in the modern era of anticancer treatment pose a forensic and clinical challenge as the terminal manifestations of treatment toxicity, tumour progression, infection and comorbidity may be similar.

Methods: A PRISMA-directed systematic review of clinical, histopathological, immunohistochemical, toxicological and molecular-autopsy evidence published between January 2016 and August 2026. 269 records were identified from PubMed/MEDLINE and Google Scholar, of which 258 were removed as duplicates, 72 were assessed for full-text and 69 were included in qualitative synthesis.

Results: Immune-checkpoint inhibitor-associated myocarditis had the most consistent treatment-associated signature, while CAR-T neurotoxicity and pneumonitis were pathologically less specific. Of the 42 patients in the largest whole-body cohort of the ICI autopsy series, 23 (54.8%) had inflammatory immune-related adverse events, with 7 (16.7%) dying directly from these events. Five out of seven ICI-related fatalities were due to myocarditis and 5/8 autopsy-confirmed cases were clinically unsuspected. CAR-T neuropathology was heterogeneous; only 2/6 patients had neurological symptoms and CAR-T DNA was found in only 1/6 brains. Molecular analysis clearly confirmed the diagnosis of DPYD-related fluoropyrimidine toxicity and pulmonary tumour thrombotic microangiopathy became the leading tumour-related mimic. Case-report predominance, variable autopsy depth and the use of different causal terms were limitations to evidence quality.

Conclusion: Postmortem investigation should be a part of the investigation of unexplained deaths and not restricted to exceptional cases, since many of the autopsy-identified events were not identified ante mortem. Standardised organ sampling, microbiology, toxicology, pharmacogenetics and residual-tumour quantification plus multidisciplinary adjudication to enhance oncology pharmacovigilance.