3D CEST imaging for differentiating malignant and benign prostatic areas: Diagnostic performance relative to DWI

In contrast to diffusion-weighted imaging (DWI) and dynamic contrast-enhanced perfusion MRI, different information from chemical exchange saturation transfer (CEST) imaging has been suggested to have potential for diagnosis of prostate cancer. The purpose of this study was to compare the descriptor capability for differentiating malignant from benign prostatic areas among three-dimensional (3D) amide proton transfer weighted (APTw) imaging using a newly developed 3D CEST sequence, DWI, and the combined descriptor method.

Fifty-two prostate cancer patients underwent DWI and 3D APTw imaging, surgical treatments, and pathological examinations. Then, 154 areas were identified as malignant, and 154 out of 624 areas were computationally selected as benign. From 3D APTw imaging and DWI, magnetization transfer ratio asymmetry (MTR_asym) in the 3.5 ppm map and apparent diffusion coefficient (ADC) within malignant and benign prostatic areas were determined by ROI measurements. Then, a logistic regression analysis was conducted to evaluate the descriptors. An ROC analysis and McNemar's test were conducted to compare the diagnostic performance of MTR_asym, ADC, and the combined discriminator method.

The logistic regression analysis identified MTR_asym (unit odds ratio: 3.6, p < 0.0001) and ADC (unit odds ratio: 0.03, p < 0.0001) as significant descriptors. The area under the curve (AUC) of the combined descriptor method (AUC = 0.86) was significantly larger than that of MTR_asym (AUC = 0.81, p = 0.001) and ADC (AUC = 0.76, p < 0.0001).

3D APTw imaging is considered at least as valuable as DWI to distinguish malignant from benign prostatic areas. Combined descriptors could improve diagnostic performance, relative to each sequence alone.