T-stage migration by routine pre-biopsy MRI staging may affect risk assessment with current risk classification systems
Session type: Poster / e-Poster / Silent Theatre session
With routine pre-biopsy prostate MRI, patients diagnosed with prostate cancer are being upstaged from their clinical digital rectal examination (DRE) stage. T-stage migration was assessed in our prospective laparoscopic radical prostatectomy (LRP) database.
All consecutive patients undergoing LRP at a single institution between 2007 – 2017 had pre-operative staging data collected. Clinical stage was recorded by the operating surgeon; MRI and pathological stages were recorded after multidisciplinary cancer team review (MDT). Concordance between clinical and MRI stages with final pathological stage was assessed using cross tabulation, Pearson’s Chi2 and Cramer’s V statistical tests.
DRE stage available in 571/571 men underoing LRP (20% D’Amico low risk, 43% medium risk, and 37% high risk) had significant concordance with the pathological stage (Pearson’s Chi2 of p<0.0001, Cramer V value:0.17 indicating weak concordance).
DRE and MRI staging data were available in 218/571 patients. Cramer V values were 0.31 for T2 disease and 0.37 for T3 disease indicating a moderate level of concordance with the pathological stage (Pearson’s Chi2 of p<0.0001). For T2 disease, MRI had 51% sensitivity, 79% specificity, 64% PPV, and 69% NPV. For T3 disease, the sensitivity was 76%, specificity 60%, PPV 70%, and NPV 67%.
Of 218 patients, 36/91 (40%) with T1 disease on DRE stage were upstaged to T2 on MRI and 44/91 (48%) to T3 on MRI. 28/65 (43%) with T2 disease on DRE stage were upstaged to T3 on MRI. Overall, 108/156 (69%) patients with T1 or T2 disease on DRE were upstaged based on MRI.
Our results show significantly better concordance to final pathology with MRI staging compared to DRE clinical stage. Current risk calculators (D’Amico, NCCN) are based on historical data incorporating DRE clinical stage. MRI staging could lead to significant stage migration and impact on treatment choice, highlighting an urgent need for developing MRI-optimised risk calculators.