A Validation Study of the Deep-Learning-Based Prostate Imaging Reporting and Data System Scoring Algorithm

Purpose: The Prostate Imaging Reporting and Data System (PI-RADS) was introduced to standardize prostate cancer diagnosis by MRI. However, the inter-reader agreement PI-RADS scoring is not always high. purpose of this study validate a deep-learning-based diagnostic algorithm PI-RADS. Methods: We applied Siemens Healthineers Artificial Intelligence (AI) prototype (work in progress) for fully automated lesion detection, classification reporting. More than 2000 bi-parametric MRI studies along with reports were included as training, validation, test data. This prospective validation includes 101 consecutive patients suspected cancer, 100 analysis. All subjects underwent noncontrast-enhanced including T2-weighted diffusion-weighted imaging. Two board-certified radiologists independently scored PI-RADS, if there disagreements; another radiologist confirmed diagnosis. compared AI results interpretation radiologists. Results: sensitivity our model ≥ 4 0.76, specificity 0.76. For cases 3, 0.69, In lesion-based analysis, detection rates 4, 5 lesions peripheral zone 43%, 63%, 100%, respectively. transition zone, 30%, 54%, Conclusion: Our has been validated shown help score