Retrospective study,
approved by the ethics committee of our institution.
All multiparametric prostate MRI (mpMRI) reports were identified in our PACS between January 2016 and December 2017 (48 months).
Inclusion criteria for the analysis were:
1) Radiological report with a category PI-RADS 3.
2) Histological result of transrectal biopsy available
Data such as age,
prostate-specific antigen and prostatic volume (calculated at the time of the radiological report) were extracted.
Prostate-specific antigen density was calculated (PSA / Prostatic Volume = PSA Density).
All biopsies results were extracted from the electronic medical record.
2416 mpMRI were identified,
424 corresponded to PI-RADS 3 cases (17.5%).
267 patients (63%) corresponded to institutional (controls and biopsies available) and 157 (37%) to external patients to our institution (no follow-up/biopsies available).
Among these 267 patients,
134 biopsies were performed, equivalent to 50.1% of the institutional PI-RADS 3 mpMRI.
Fig. 4
Fig. 4: Methods and Materials
Data acquisition: All MR acquisitions were performed in 3.0 T magnet.
(Siemens Magnetom Skyra©),
with a maximum gradient of 45 mT/m and slew rate of 200 T/m/sec.
Patients were placed in prone position,
and two coils were placed (dorsal and ventral,
of 32 and 30 channels,
respectively).
The MRI protocol was performed using PI-RADS v2.0,
consisted of T1-weighted axial,
T2-weighted in three planes,
DW-images with b-values 0,
50,
500,
and 1000 s/mm2,
3 mm slices,
FOV between 16 -18 cm.
ADC map and DWI b1600 interpolated by software.
Dynamic Contrast Enhanced (DCE) and subsequent subtraction images were obtained.
Statistical Analysis: Continuous variables are presented as means with standard deviations (SD) or median with interquartile range (IQR),
categorical variables as frequencies and percentages.
Cut-off values were defined according to the Youden index,
using histological results as reference standard.
They were determined with ROC analyses and the calculation included area under the curve (AUC),
specifity,
positive predictive value,
and negative predictive value.
Results were regarded as significant when the p-value was below 0.05.