AI-enhanced detection of prostate cancer
AI-enhanced detection of prostate cancer
Prostate cancer poses a pressing public health challenge with an increasing elderly population in whom prostate cancer-specific mortality is highest. For the diagnosis of prostate cancer, magnetic resonance imaging (MRI) scans are recommended, leading to an increasing demand and a need to use current diagnostic resources more efficiently. This thesis of Stefan Fransen explored an AI-enhanced diagnostic pathway to handle the increasing demand for prostate cancer diagnosis. The focus was on evaluating AI-enhanced lesion detection and AI-based reconstruction of accelerated MRI, assessing the effectiveness compared to conventional methods, and discussing the potential impact on clinical practice.
This thesis demonstrates the potential of AI to reduce MRI scan time and assist radiologists in assessing prostate MRI. The proposed prostate cancer diagnostic pathway could reduce MRI scan time by up to 71% without significantly compromising diagnostic quality. Furthermore, AI could reduce the radiologists’ workload by up to 20%, which might increase with continued improvements in AI performance. The use of AI for prostate cancer detection was also accepted by patients, especially with AI performances approaching expert level.
These findings indicate an AI-enhanced diagnostic pathway that benefits the patient, society, and healthcare professionals. The proposed pathway could potentially handle the increasing demand for prostate cancer diagnosis, but before clinical deployment, prospective studies are needed to validate the generalization of the findings in this thesis.