The Image Grading Study For Prostate Cancer Diagnosis In Patients With PSA<20μg/L Using T2WI And DWI

ObjectiveTo explore the significance of1.5-Tesla MRI in early diagnosis of prostate cancer patients with PSA level less than20μg/L, and analyze the image features of prostate cancer on MRI. To assess the significance of the image gradings of T2WI and DWI for the diagnosis of prostate cancer, and evaluate the predictive capability of MRI for Gleason score and number of positive biopsy cores at early stage of prostate cancer.Patients and MethodsSection1:We retrospectively analyzed the clinical data of patients who received prostate biopsies in our hospital between Feb.2006and Jan.2011. The patients with PSA<20μg/L who had complete MRI image data were included in this study. We evaluated the correlation between T2WI image features (including the signal of peripheral zone and seminal vesicles, prostate capsule and the relationship with surrounding tissues) and prostate cancer with reference to histopathologic evaluation of prostate biopsies.Section2:133patients with PSA<20μg/L who received prostate biopsies in our hospital were prospectively recruited between Dec.2011and Jan.2013. T2WI and DWI image features on peripheral zone of prostate were evaluated according to the pre-established image grading standards. We compared the significance of the grading system for the diagnosis of prostate cancer with either scanning methods alone or in combination, and evaluated the predictive capability of MRI for Gleason score and number of positive biopsy cores at early stage of prostate cancer. Histopathologic evaluation of prostate biopsies was the reference standard.ResultsSection1:The retrospective study included359cases, of which137patients (38.16%) were diagnosed prostate cancer.27cases (7.52%) were found with homogeneous high signal on peripheral zone of T2WI and the other332cases (92.48%) were found with low signals. Univariate logistic regression analysis showed that low signals on peripheral zone images of T2WI were more helpful for the diagnosis of prostate cancer (P=0.015). Number, shape, margin of lesions, the status of prostatic capsule, the boundary of peripheral zone and transitional zone were significantly associated with the diagnosis of prostate cancer. Through multivariate logistic regression analysis it was found that lesion shape (flaky vs. nodular, P-0.0027; diffuse vs. nodular, P=0.0096) and the status of prostatic capsule (disappearing vs. completely clear, P=0.0372; indistinct vs. completely clear, P=0.0569) on MR images were risk factors which can affect the diagnosis of prostate cancer. The area under receiver operating characteristic curve was0.833(P<0.0001). The incidence of abnormal prostate rectal angle, abnormal prostate seminal vesicle angle, the neurovascular bundle involvement, invasion to the surrounding fat of prostate and abnormal prostate morphology on T2WI was2.11%,4.52%,3.92%,2.71%,1.81%respectively, and the corresponding positive biopsy rate was100%,80%,84.62%,77.78%,100%respectively.Section2:In prospective study,60patients (45.11%) out of133were diagnosed prostate cancer. The peripheral zone of the prostate of each patient on MR images was divided into six regions and each region corresponds to2out of12biopsy cores. Positive biopsy rate of grade0～5on T2WI image grading was2.78%,13.45%,20.35%,34.64%,52.63%and88.89%respectively, and the rate was11.09%,16.10%,48.48%,70.97%and86.11%respectively on grade1-5of T2WI+DWI image grading. We found that with the increments of pre-established T2WI and T2WI+DWI image gradings, positive rates of prostate biopsies showed a significant growth by Cochran-Armitage trend test (P<0.0001, P<0.0001). ROC analysis found that the ability of T2WI combined with DWI was greater than that of T2WI alone in detection of prostate cancer (AUC=0.743vs. AUC=0.715, P=0.0072). A significant positive correlation was found between the grading of DWI and Gleason score by Spearman rank correlation analysis (p=0.27507, P=0.0014). This correlation was also found between Gleason score and the grading of T2WI combined with DWI (ρ=0.23236, P=0.0071). The gradings of T2WI, DWI and T2WI combined with DWI were all found to have significant positive correlations with the number of positive biopsy cores (ρ=0.39633, P<0.0001; ρ=0.59211,P<0.0001; ρ=0.57148, P<0.0001). ConclusionAmong patients with PSA<20μg/L, lesion shape of prostatic peripheral zone and the status of prostatic capsule on T2WI are risk factors which can affect the diagnosis of prostate cancer. The grading of T2WI can predict the risk of prostate cancer for patients with PSA<20μg/L. On the basis of T2WI, DWI can further enhance the ability of predicting prostate cancer. A significant positive correlation exists between the grading of DWI and Gleason score, this correlation also exists between Gleason score and the grading of T2WI combined with DWI. All the gradings of T2WI, DWI and T2WI combined with DWI have significant positive correlations with the number of positive biopsy cores.