NKI ARTIMES AI Outperforms Physicians in Measuring Mesothelioma Volume
Researchers at the Netherlands Cancer Institute have developed ARTIMES, an artificial intelligence model that significantly outperforms human physicians and current international standards in assessing pleural mesothelioma treatment response. Published recently in Lancet Oncology, the breakthrough addresses a longstanding diagnostic challenge: pleural mesothelioma typically grows in thin, irregular sheets along the lung lining, making traditional diameter-based RECIST criteria unreliable for tracking tumor progression or regression. ARTIMES evaluates treatment response by calculating complete volumetric data from sequential CT scans, a task too complex and time-consuming for manual review. The model was trained and validated using over 11,000 CT scans from more than 2,000 patients across 121 hospitals worldwide. Testing demonstrated that volumetric analysis via ARTIMES yields more accurate survival predictions and treatment response metrics than established RECIST diameter measurements. Physicians retain final review authority over all AI-generated assessments to ensure clinical safety. The clinical implications are substantial. By detecting treatment resistance earlier, the model enables physicians to discontinue ineffective therapies promptly and transition patients to alternative regimens. This approach minimizes unnecessary adverse effects, improves patient certainty, and reduces healthcare expenditures. Furthermore, the enhanced measurement precision is expected to streamline clinical trials by providing more reliable endpoints for evaluating novel therapeutic interventions. Currently deployed under a regulatory exemption exclusively at the Netherlands Cancer Institute, ARTIMES is publicly accessible online for external research use. Institution leaders are actively pursuing broader clinical approvals and advocating for streamlined European Union regulatory pathways for comparable medical AI devices. The successful integration of volumetric AI analysis into mesothelioma care establishes a scalable framework for oncology, with the NKI already expanding similar AI architectures to lung cancer and brain metastasis research. This development signals a paradigm shift in radiological assessment, positioning artificial intelligence as a foundational tool in precision oncology and therapeutic development.
