Prostate magnetic resonance imaging (MRI) shows excellent detection of clinically significant prostate cancer (csPCa) and should be performed before biopsy in men at risk of having prostate cancer (PCa) [1], [2], [3], [4]. The introduction of prebiopsy MRI has revolutionised PCa diagnosis by enabling targeted MRI-ultrasound fusion biopsy, increasing the number of csPCa cases detected and reducing overdiagnosis of indolent disease [3], [5], [6]. However, diagnostic accuracy is limited by a false-positive rate of approximately 45%, while excellent sensitivity has been demonstrated (91%) [7], [8]. Hence, a substantial number of men undergo unproductive targeted biopsies (TBs) along with potential complications, which need to be minimised. Moreover, radiological image analysis requires high expertise, leading to substantial interobserver variability [9], [10], [11], [12]. The high number of false positives and the qualitative variations in reporting are substantial limitations for urologists in their biopsy decision-making [13], [14]. Hence, there is a continued need for additional assistance in biopsy planning to reduce unnecessary biopsies while maintaining high detection sensitivity.

Artificial intelligence (AI) computer-aided diagnosis (CAD) tools aim to address the need for improved diagnostic accuracy including assisting clinicians in their biopsy management tasks [15]. Good diagnostic accuracy has been demonstrated for detecting csPCa using AI models in retrospectively collected MRI data, paving the way for the approval of several CAD tools [16], [17], [18], [19], [20], [21], [22]. To date, the actual diagnostic value of approved CAD tools for MRI-based PCa detection has not been evaluated in routine clinical practice. Thus, this study aimed to document the diagnostic utility of biopsy planning strategies (BPSs) for csPCa detection at transperineal biopsy using radiological Prostate Imaging Reporting and Data System (PI-RADS) alone or in combination with an image-based CAD tool.