TY - JOUR
T1 - The Barcelona Predictive Model of Clinically Significant Prostate Cancer
AU - Morote Robles, Juan
AU - Borque-Fernando, Ángel
AU - Triquell, Marina
AU - Celma, Ana
AU - Regis, Lucas
AU - Escobar, Manel
AU - Mast, Richard
AU - de Torres, Inés
AU - Semidey, María E.
AU - Abascal, Jose Maria
AU - Servian, Pol
AU - Salvador Hidalgo, Daniel
AU - Santamaría, Anna
AU - Planas, Jacques
AU - Esteban, Luis M.
AU - Trilla Herrera, Enrique
PY - 2022
Y1 - 2022
N2 - Magnetic-resonance-imaging-based predictive models (MRI-PMs) improve the MRI prediction of clinically significant prostate cancer (csPCa) in prostate biopsies. Risk calculators (RC) provide easy individual assessment of csPCa likelihood. MRI-PMs have been analysed in overall populations of men suspected to have PCa, but they have never been analysed according to the prostate imaging-report and data system (PI-RADS) categories. Therefore, the true clinical usefulness of MRI-PMs regarding the specific PI-RADS categories is unknown. A new and externally validated MRI-PM for csPCa was developed in the metropolitan area of Barcelona, and a web-RC designed with the new option of selecting the csPCa probability threshold. The development cohort comprised 1486 men scheduled to undergo a 3-tesla multiparametric MRI (mpMRI) and guided and/or systematic biopsies in one academic institution of Barcelona. The external validation cohort comprised 946 men in whom the same diagnostic approach was carried out as in the development cohort, in two other academic institutions of the same metropolitan area. CsPCa was detected in 36.9% of men in the development cohort and 40.8% in the external validation cohort (p = 0.054). The area under the curve of mpMRI increased from 0.842 to 0.897 in the developed MRI-PM (p < 0.001), and from 0.743 to 0.858 in the external validation cohort (p < 0.001). A selected 15% threshold avoided 40.1% of prostate biopsies and missed 5.4% of the 36.9% csPCa detected in the development cohort. In men with PI-RADS <3, 4.3% would be biopsied and 32.3% of all existing 4.2% of csPCa would be detected. In men with PI-RADS 3, 62% of prostate biopsies would be avoided and 28% of all existing 12.4% of csPCa would be undetected. In men with PI-RADS 4, 4% of prostate biopsies would be avoided and 0.6% of all existing 43.1% of csPCa would be undetected. In men with PI-RADS 5, 0.6% of prostate biopsies would be avoided and none of the existing 42.0% of csPCa would be undetected. The Barcelona MRI-PM presented good performance on the overall population; however, its clinical usefulness varied regarding the PI-RADS category. The selection of csPCa probability thresholds in the designed RC may facilitate external validation and outperformance of MRI-PMs in specific PI-RADS categories
AB - Magnetic-resonance-imaging-based predictive models (MRI-PMs) improve the MRI prediction of clinically significant prostate cancer (csPCa) in prostate biopsies. Risk calculators (RC) provide easy individual assessment of csPCa likelihood. MRI-PMs have been analysed in overall populations of men suspected to have PCa, but they have never been analysed according to the prostate imaging-report and data system (PI-RADS) categories. Therefore, the true clinical usefulness of MRI-PMs regarding the specific PI-RADS categories is unknown. A new and externally validated MRI-PM for csPCa was developed in the metropolitan area of Barcelona, and a web-RC designed with the new option of selecting the csPCa probability threshold. The development cohort comprised 1486 men scheduled to undergo a 3-tesla multiparametric MRI (mpMRI) and guided and/or systematic biopsies in one academic institution of Barcelona. The external validation cohort comprised 946 men in whom the same diagnostic approach was carried out as in the development cohort, in two other academic institutions of the same metropolitan area. CsPCa was detected in 36.9% of men in the development cohort and 40.8% in the external validation cohort (p = 0.054). The area under the curve of mpMRI increased from 0.842 to 0.897 in the developed MRI-PM (p < 0.001), and from 0.743 to 0.858 in the external validation cohort (p < 0.001). A selected 15% threshold avoided 40.1% of prostate biopsies and missed 5.4% of the 36.9% csPCa detected in the development cohort. In men with PI-RADS <3, 4.3% would be biopsied and 32.3% of all existing 4.2% of csPCa would be detected. In men with PI-RADS 3, 62% of prostate biopsies would be avoided and 28% of all existing 12.4% of csPCa would be undetected. In men with PI-RADS 4, 4% of prostate biopsies would be avoided and 0.6% of all existing 43.1% of csPCa would be undetected. In men with PI-RADS 5, 0.6% of prostate biopsies would be avoided and none of the existing 42.0% of csPCa would be undetected. The Barcelona MRI-PM presented good performance on the overall population; however, its clinical usefulness varied regarding the PI-RADS category. The selection of csPCa probability thresholds in the designed RC may facilitate external validation and outperformance of MRI-PMs in specific PI-RADS categories
KW - Clinically significant prostate cancer
KW - Magnetic resonance imaging
KW - Predictive model
KW - Risk calculator
U2 - 10.3390/cancers14061589
DO - 10.3390/cancers14061589
M3 - Article
C2 - 35326740
SN - 2072-6694
VL - 14
JO - Cancers
JF - Cancers
ER -
