Safety Evaluation of Combination Oncology Devices

Abstract

Combination oncology devices -- products that integrate therapeutic agents with medical device components to treat cancer -- have become one of the most active categories of medical innovation, spanning drug-eluting embolic beads for hepatocellular carcinoma, photodynamic therapy systems combining photosensitiser drugs with activation light sources, drug-loaded brachytherapy seeds, targeted radionuclide therapies delivered through radiopharmaceutical-device combinations, and emerging electroceutical platforms where electrical field therapy combines with cytotoxic or immunotherapeutic agents. The safety evaluation challenge for these products exceeds what pure drug or pure device assessment addresses: the clinical risk profile arises from the combined biological effect of the therapeutic agent and the device-mediated delivery mechanism, with device-drug interactions producing safety signals that neither component alone would generate. This study presents the Combination Oncology Device Safety Assessment Framework (CODSAF), evaluating five safety evaluation approaches -- traditional separated drug and device safety evaluation (TSE-DD), integrated combination product safety assessment per FDA OCP framework (ICPS-FDA), EU MDR Article 117 integrated assessment (EUMDR-117), AI-enhanced pooled safety analysis across combination product classes (AI-PSA), and real-world evidence-augmented post-market safety surveillance (RWE-PMS) -- across four combination oncology device contexts: hepatic transarterial chemoembolisation devices, photodynamic therapy systems, targeted radionuclide combination products, and electroceutical oncology combinations. Performance was scored using the Combination Safety Evaluation Score (CSES), a weighted composite of safety signal detection sensitivity (0.30), regulatory pathway clarity (0.20), device-drug interaction assessment capability (0.20), clinical risk-benefit integration (0.15), and post-market safety monitoring support (0.15). AI-enhanced pooled safety analysis achieved the highest CSES (0.884) through the best safety signal detection sensitivity (0.940) and strong device-drug interaction assessment (0.900). Integrated combination product safety assessment per FDA OCP framework ranked second (0.873) with the best regulatory pathway clarity (0.920). EU MDR Article 117 integrated assessment (0.861) provides systematic combination product evaluation in the EU context. RWE-augmented post-market surveillance (0.853) offers unique post-market capability. Traditional separated evaluation (0.822) remains prevalent but demonstrates systematic limitations for true combination products.