RSNA 2015

Abstract Archives of the RSNA, 2015


GU219-SD-MOB4

Clinical Impact of Prostate Cancer Detection with Extrapolated High b-value DWI

Monday, Nov. 30 12:45PM - 1:15PM Location: GU/UR Community, Learning Center Station #4



Sadhna Verma, MD, Cincinnati, OH (Presenter) Nothing to Disclose
Jason W. Young, MD, Cincinnati, OH (Abstract Co-Author) Nothing to Disclose
Sarad Sarkar, Grass Valley, CA (Abstract Co-Author) Employee, Eigen
Rajesh Venkataraman, PhD, Grass Valley, CA (Abstract Co-Author) Employee, Eigen
Xu Yang, Grass Valley, CA (Abstract Co-Author) Nothing to Disclose
Krishnanath Gaitonde, MD, CIncinnati, OH (Abstract Co-Author) Nothing to Disclose
PURPOSE

To assess the clinical impact of prostate cancer detection using acquired versus extrapolated high b-value diffusion weighted imaging (DWI) computed using 4 diffusion models.

METHOD AND MATERIALS

50 sequential patients from 2013-2015 with pathologically proven prostate cancer (CaP) were chosen for analysis. 3T Multiparametric prostate MRI exams of the patients included one of 2 low b-value DWI protocols (b=100, 600, 1200 or b=15, 250, 800, 1200) and a high b-2000 DWI. Additionally, high b-2000 DWI was extrapolated from the lower b-value images using 4 diffusion models - Monoexponential, IVIM, Stretched exponential and Kurtosis. All images were scored on subjective quality and readability independently by 2 radiologists and 1 resident. Lesions were identified by consensus on all images by the 3 readers and subjectively graded for lesion conspicuity. Lesion-to-background contrast ratios were computed for each lesion on all images. Pathological ground truth was established using MRI-Ultrasound fusion prostate biopsy of the identified lesions. Logistic regression analysis was conducted to compare the CaP predictive capabilities of acquired b-2000 DWI versus computed b-2000 DWI from the 4 models.

RESULTS

All extrapolated b-2000 series demonstrated unanimously higher ratings for subjective quality and readability then acquired b-2000 except the Kurtosis model (Wilcoxon Rank Test, p<0.0001). All extrapolated DWI (except Kurtosis) also demonstrated better lesion conspicuity in a direct comparison with acquired b-2000 DWI (T-test, p < 0.0001). Mathematical computation demonstrated higher lesion to background contrast ratio (LBCR) for all extrapolated DWI compared to acquired b-2000 DWI (ANOVA, p<0.0001). Logistic regression analysis determined that the LBCR of extrapolated b-2000 DWI was a better predictor of CaP than the LBCR of acquired b-2000 DWI (p-value ~ 0.05). Receiver Operator Curve (ROC) analysis demonstrated higher area under the curve for exponential b-2000 DWI (72%) as compared to acquired b-2000 DWI (65%) or PSA (57%) alone

CONCLUSION

The increased lesion conspicuity of extrapolated DWI vs acquired high b-value DWI may be a major advantage in CaP detection.

CLINICAL RELEVANCE/APPLICATION

The increased lesion conspicuity of extrapolated DWI vs acquired high b-value DWI may be a major advantage in CaP detection

Honored Educators

Presenters or authors on this event have been recognized as RSNA Honored Educators for participating in multiple qualifying educational activities. Honored Educators are invested in furthering the profession of radiology by delivering high-quality educational content in their field of study. Learn how you can become an honored educator by visiting the website at: https://www.rsna.org/Honored-Educator-Award/

Sadhna Verma, MD - 2013 Honored Educator