Iyama et al reported a model for distinguishing a transition zone cancer from benign prostatic hyperplasia (BPH). This relies on multi-parametric prostate magnetic resonance imaging. The authors are from Kumamoto Chuo Hospital, Kumamoto University and National Hospital Organization Kumamoto Medical Center in Japan.
Patient selection: multi-parametric prostate magnetic resonance imaging including T2-weighted images
Parameters:
(1) mean apparent diffusion coefficient (ADC), from 1.8 to 0.5
(2) normalized T2 weighted signal from 1.5 to 9.0
points for mean ADC =
= 138.461 - (76.9231 * (mean ADC))
points for T2 weighted image signal =
= 48 - (5.3333 * (T2WI signal))
total score =
= SUM(points for the 2 parameters)
Interpretation:
• minimum score: 0
• maximum score: 140
• The higher the score the greater the risk of cancer.
Total Score
Probability of Cancer
< 56.3
< 1%
56.3 to 103.9
X = (0.19326 * (score)) - 15.472
probability of cancer = 1/(1+EXP((-1)*X))
> 103.9
> 99%
Performance:
• The area under the ROC curve is 0.97.
To read more or access our algorithms and calculators, please log in or register.