Keywords:
Neoplasia, Education and training, Cancer, Staging, Contrast agent-intravenous, Biopsy, MR-Diffusion/Perfusion, MR, Pelvis, Oncology, Genital / Reproductive system male
Authors:
F. Gentile1, M. Gatti2, R. Faletti2, J. Giglio2, F. Misischi2, C. M. Berzovini1, G. Calleris2, P. Gontero2, P. Fonio2; 1Torino/IT, 2Turin/IT
DOI:
10.26044/ecr2019/C-2705
Conclusion
Our results on the readers of Group A,
with a record of about 1000 cases,
showed no differences in their performance between biparametric and multiparametric protocol: SNS 0.96 vs 0.91 (p=0.25) and AUC 0.93 vs 0.86 (p=0.10).
Quite different was the outcome for Group B,
with a record of about 300 cases.
Despite the fact that when working with mpMRI they achieved levels of performance comparable to those of the more expert readers of group A (SNS=0.91 and AUC=0.86),
when they could not rely on the DCE addition,
their performance dropped significantly (SNS= 0.58,
AUC=0.73; p<0.0001).
The results on Group C,
formed by young residents with about 100 cases,
showed that they still had to increase their experience to reach an acceptable accuracy even with the multiparametric protocol: their diagnostic parameters are definitely inferior to those of the readers with 300 cases: SNS =0.71 vs 0.91,
p<0.0001 and AUC = 0.77 vs 0.86,
p=0.07.
The study revealed the impact of the readers’ experience when using bpMRI.
The bpMRI without contrast media was a valid alternative for expert readers,
whereas less experienced ones needed DCE to significantly boost SNS and AUC.
Overall,
our results indicate that for reaching the needed accuracy with the bpMRI protocol,
300 cases are not sufficient: based on the almost-linear increase of SNS and AUC with the number of cases (every 100 cases +0.05 points for SNS and +0.02 for AUC),
a minimum of 700-800 cases seems to be necessary.