We designed and implemented a retrospective study on the images of 68 patients who underwent mpMRI in our Institution for PCa suspicion.
The same scans were to be interpreted twice,
first according to the biparametric protocol,
and subsequently according to the multiparametric protocol,
to be considered as reference.
To allow a blinded review of MRI scans,
reducing interpretation biases,
our institution’s database was searched backwards to add to the 45 PCa cases a control sample of 23 men with no diagnosis of PCa (ratio of one control every two case patients) and at least a 2.5-year follow-up negative for PCa,
including at least another MRI scan and/or another negative prostate biopsy plus a stable PSA.
All controls had a negative 12-core transrectal ultrasound (TRUS) guided prostate biopsy (Table 1).
The MRI scans were interpreted by three groups of readers: Group A (two senior radiologists with a record of about 1000 cases analyzed),
Group B (two junior radiologists with about 300 cases) and Group C (two residents,
with about 100 cases).
The inclusion of two readers in each group was done for avoiding biases due to singularities in the individual performing capabilities.
Each of the six observers performed two readings of the same image,
first with the biparametric protocol including diffusion-weighted imaging (DWI), apparent diffusion coefficient (ADC) maps and T2-weighted (T2W) imaging in three planes and,
after at least one month,
with the multiparametric one adding dynamic contrast enhancement (DCE).
The order of the sessions was chosen to prevent possible biases in the biparametric reading due to possible (and not controllable) remembrances of the mpMRI images.
A positive DCE MRI lesion was considered a lesion where the enhancement is focal,
earlier or contemporaneous with enhancement of adjacent normal prostatic tissues,
and corresponds to a finding on T2W and/or DWI.
Scores 4 or 5 were considered as positive findings whereas PI-RADSv2 scores 1,
2 and 3 were considered negative [4,5]
Positive MRI corresponding to histologically confirmed tumors were considered true positives (TP),
whereas those without corresponding histopathologic correlates were considered false positives (FP).
Negative MRI without histological confirmation were considered true negatives (TN),
whereas those corresponding to histologically confirmed tumors were considered missed (false negative,
FN).