Is It Possible to Model the Risk of Malignancy of Focal Abnormalities Found at Prostate Multiparametric MRI?

Eur Radiol. 2012 May;22(5):1149-57. doi: 10.1007/s00330-011-2343-8. Epub 2012 Jan 7.

Abstract

Objective: To evaluate whether focal abnormalities (FAs) depicted by prostate MRI could be characterised using simple semiological features.

Methods: 134 patients who underwent T2-weighted, diffusion-weighted and dynamic contrast-enhanced MRI at 1.5 T before prostate biopsy were prospectively included. FAs visible at MRI were characterised by their shape, the degree of signal abnormality (0 = normal to 3 = markedly abnormal) on individual MR sequences, and a subjective score (SS(1) = probably benign to SS(3) = probably malignant). FAs were then biopsied under US guidance.

Results: 56/233 FAs were positive at biopsy. The subjective score significantly predicted biopsy results (P < 0.01). As compared to SS(1) FAs, the odds ratios (OR) of malignancy of SS(2) and SS(3) FAs were 9.9 (1.8-55.9) and 163.8 (11.5-2331). Unlike FAs' shape, a simple combination of MR signal abnormalities (into "low-risk", "intermediate" and "high-risk" groups) significantly predicted biopsy results (P < 0.008). As compared to "low risk" FAs, the OR of malignancy of "intermediate" and "high-risk" FAs were 4.5 (1.1-18.4) and 52.7 (6.8-407) in the overall population and 5.4 (1.1-27.2) and 118.2 (6.1-2301) in PZ.

Conclusions: A simple combination of signal abnormalities of individual MR sequences can significantly stratify the risk of malignancy of FAs, holding promise of a more standardised interpretation of MRI by readers with varying experience.

Key points: • Using multiparameter(mp)-MRI, experienced uroradiologists can stratify the malignancy risk of prostatic lesions • The shape of prostatic focal abnormalities in the peripheral zone does not help predicting malignancy. • A simple combination of findings at mp-MRI can help less-experienced radiologists.

MeSH terms

  • Computer Simulation
  • France / epidemiology
  • Humans
  • Image Interpretation, Computer-Assisted / methods*
  • Magnetic Resonance Imaging / methods*
  • Magnetic Resonance Imaging / statistics & numerical data*
  • Male
  • Middle Aged
  • Models, Statistical*
  • Prevalence
  • Proportional Hazards Models*
  • Prostatic Neoplasms / diagnosis*
  • Prostatic Neoplasms / epidemiology*
  • Reproducibility of Results
  • Risk Assessment
  • Risk Factors
  • Sensitivity and Specificity