Prostate cancer: apparent diffusion coefficient map with T2-weighted images for detection--a multireader study

Abstract

Purpose: To retrospectively assess the incremental value of an apparent diffusion coefficient (ADC) map combined with T2-weighted magnetic resonance (MR) images compared with T2-weighted images alone for prostate cancer detection by using a pathologic map as the reference standard.

Materials and methods: This retrospective study was approved by the institutional review board; informed consent was waived. The study included 52 patients (mean age, 65 years +/- 5 [standard deviation]; range, 48-76 years) who underwent endorectal MR imaging and step-section histologic examination. Three readers with varying experience levels reviewed T2-weighted images alone, the ADC map alone, and T2-weighted images and ADC maps. The prostate was divided into 12 segments. The probability of prostate cancer in each segment on MR images was recorded with a five-point scale. Areas under the receiver operating characteristic curve (AUCs) were compared by using the Z test; sensitivity and specificity were determined with the Z test after adjusting for data clustering.

Results: AUC of T2-weighted and ADC data (reader 1, 0.90; reader 2, 0.88; reader 3, 0.76) was greater than that of T2-weighted images (reader 1, 0.79; reader 2, 0.75; reader 3, 0.66) for all readers (P < .0001 in all comparisons). AUC of T2-weighted and ADC data was greater for readers 1 and 2 than for reader 3 (P < .001). Sensitivity of T2-weighted and ADC data (reader 1, 88%; reader 2, 81%; and reader 3, 78%) was greater than that of T2-weighted images (reader 1, 74%; reader 2, 67%; reader 3, 67%) for all readers (P = .01 for reader 1; P = .02 for readers 2 and 3). Specificity of T2-weighted and ADC data was greater than that of T2-weighted images for reader 1 (88% vs 79%, P = .03) and reader 2 (89% vs 77%, P < .001).

Conclusion: The addition of an ADC map to T2-weighted images can improve the diagnostic performance of MR imaging in prostate cancer detection.