ProCAncer-I results | Prof. Daniele Regge on clinical validation
In this video, we present the results of the clinical validation of AI models that were developed during the ProCAncer-i Project. Clinical validation is a critical step to ensure that AI tools not only perform well in controlled settings, but also truly assist radiologists in daily practice. Two AI models were tested. Radiologists read 160 prostate exams first without AI, then re-read them six weeks later with AI assistance. Key Takeaways: AI improved sensitivity and specificity compared to unassisted readings, though not always with statistical significance. AI reduced reading time by 15 to 21 seconds per exam.
Bottom Line: AI can meaningfully improve radiologist performance, especially for those with lower baseline accuracy, and it can help save valuable reading time.
Watch it here:
