Session: MP44: Prostate Cancer: Detection & Screening II
MP44-06: Tumor characteristics of multiparametric MRI-detected and -undetected lesions in patients with suspected radiorecurrent prostate cancer: an analysis from the FOcal RECurrent Assessment and Salvage Treatment (FORECAST) trial
Introduction: The interpretation of multiparametric MRI (mpMRI) post-radiotherapy can be challenging due to factors like glandular atrophy and diffuse T2 low signal. Using data from the FORECAST paired validation study, we compared the characteristics of mpMRI-detected and -undetected radiorecurrent disease. Methods: The prospective FORECAST study (NCT01883128) recruited men from 6 UK centers with suspected radiorecurrent cancer who underwent prostate mpMRI followed by systematic transperineal template mapping and MRI-targeted biopsies. MRI visibility was defined as a Likert score 3-5. 3 cancer definitions were studied: any cancer; Grade Group =3 and/or maximum cancer core length (MCCL) =6mm (PROMIS definition 1); and Grade Group =2 and/or MCCL =4mm (PROMIS definition 2). Characteristics of MRI-detected vs -undetected tumors were compared. Analyses were performed at the prostate quadrant level using cluster bootstrapping with 10,000 resamples. Results: 578 quadrants (147 men) were included. 88/261 quadrants (33.7%) with any cancer were MRI-undetected, compared to 62/198 quadrants (31.3%) for definition 1, and 80/238 quadrants (33.6%) for definition 2. Sensitivity of MRI for any cancer was 0.66 (95%CI 0.60-0.73), comparable to definition 1 (0.69, 95%CI 0.61-0.76) and definition 2 (0.66, 95%CI 0.60-0.73). Specificity was poorer (any cancer 0.54, 95%CI 0.46-0.62; definition 1: 0.54, 95%CI 0.46-0.62; definition 2: 0.54, 95%CI 0.46-0.62;). MRI-detected cancer comprised more Grade Group 5 tumors (difference 13.5%, p=0.01) (Table 1). MRI-detected tumors also had longer MCCL (median difference 3mm, p<0.001), and yielded more positive cores (median difference 2 cores, p=0.003). Conclusions: Of mpMRI-undetected tumors, there were fewer Grade Group 5 tumors, MCCL was shorter, and number of positive cores were fewer. At the prostate quadrant level, mpMRI had modest sensitivity and specificity, and may miss over 30% of radiorecurrent tumors across cancer definitions. mpMRI alone is therefore insufficient in localizing radiorecurrent prostate cancer. SOURCE OF Funding: Funded by the Pelican Cancer Foundation, the US National Institutes of Health, and the UK Medical Research Council.
