4.5 Article

Prostate Cancer: Utility of Whole-Lesion Apparent Diffusion Coefficient Metrics for Prediction of Biochemical Recurrence After Radical Prostatectomy

Journal

AMERICAN JOURNAL OF ROENTGENOLOGY
Volume 205, Issue 6, Pages 1208-1214

Publisher

AMER ROENTGEN RAY SOC
DOI: 10.2214/AJR.15.14482

Keywords

apparent diffusion coefficient; DWI; MRI; prostate cancer; radical prostatectomy

Funding

  1. Hitachi-Aloka
  2. Joseph and Diane Steinberg Charitable Trust
  3. Healthtronics

Ask authors/readers for more resources

OBJECTIVE. The purpose of this study was to investigate the additional value of whole-lesion histogram apparent diffusion coefficient (ADC) metrics, when combined with standard pathologic features, in prediction of biochemical recurrence (BCR) after radical prostatectomy for prostate cancer. MATERIALS AND METHODS. The study included 193 patients (mean age, 61 +/- 7 years) who underwent 3-T MRI with DWI (b values, 50 and 1000 s/mm(2)) before prostatectomy. Histogram metrics were derived from 3D volumes of interest encompassing the entire lesion on ADC maps. Pathologic features from radical prostatectomy and subsequent BCR were recorded for each patient. The Fisher exact test and Mann-Whitney test were used to compare ADC-based metrics and pathologic features between patients with and patients without BCR. Stepwise logistic regression analysis was used to construct multivariable models for prediction of BCR, which were assessed by ROC analysis. RESULTS. BCR occurred in 16.6% (32/193) of patients. Variables significantly associated with BCR included primary Gleason grade, Gleason score, extraprostatic extension, seminal vesicle invasion, positive surgical margin, preoperative prostate-specific antigen level, MRI tumor volume, mean whole-lesion ADC, entropy ADC, and mean ADC of the bottom 10th, 10-25th, and 25-50th percentiles (p <= 0.019). Significant independent predictors of BCR at multivariable analysis were primary Gleason grade, extraprostatic extension, mean of the bottom 10th percentile ADC, and entropy ADC (p = 0.002-0.037). The AUC of this multivariable model was 0.94 for prediction of BCR; the AUC of pathologic features alone was 0.89 (p = 0.001). CONCLUSION. A model integrating whole-lesion ADC metrics had significantly higher performance for prediction of BCR than did standard pathologic features alone and may help guide postoperative prognostic assessments and decisions regarding adjuvant therapy.

Authors

I am an author on this paper
Click your name to claim this paper and add it to your profile.

Reviews

Primary Rating

4.5
Not enough ratings

Secondary Ratings

Novelty
-
Significance
-
Scientific rigor
-
Rate this paper

Recommended

No Data Available
No Data Available