Simplified Luminal Water Imaging for the Detection of Prostate Cancer From Multiecho T2 MR Images. Issue 3 (19th December 2018)

• Record Type: Journal Article
• Authors: Devine, William; Giganti, Francesco; Johnston, Edward W.; Sidhu, Harbir S.; Panagiotaki, Eleftheria; Punwani, Shonit; Alexander, Daniel C.; Atkinson, David
• Is Part Of: Journal of magnetic resonance imaging. Volume 50:Issue 3(2019)
• Page Start: 910
• Page End: 917
• Publication Date: 2018-12-19
• Subjects: microstructure -- modeling -- multiecho T2 -- prostate -- cancer; Magnetic resonance imaging -- Periodicals; 616
• DOI: 10.1002/jmri.26608
• Languages: English
• ISSNs: 1053-1807

Abstract

Abstract : Background: Luminal water imaging (LWI) suffers less from imaging artifacts than the diffusion‐weighted imaging used in multiparametric MRI of the prostate. LWI obtains multicompartment tissue information from a multiecho T2 dataset. Purpose: To compare a simplified LWI technique with apparent diffusion coefficient (ADC) in classifying lesions based on groupings of PI‐RADS v2 scores. Secondary aims were to investigate whether LWI differentiates between histologically confirmed tumor and normal tissue as effectively as ADC, and whether LWI is correlated with the multicompartment parameters of the vascular, extracellular, and restricted diffusion for cytometry in tumors (VERDICT) diffusion model. Study Type: A subset of a larger prospective study. Population: In all, 65 male patients aged 49–79 were scanned. Field Strength/Sequence: A 32‐echo T2 and a six b‐value diffusion sequence (0, 90, 500, 1500, 2000, 3000 s/mm 2 ) at 3T. Assessment: Regions of interest were placed by a board‐certified radiologist in areas of lesion and benign tissue and given PI‐RADS v2 scores. Statistical Tests: Receiver operating characteristic and logistic regression analyses were performed. Results: LWI classifies tissue as PI‐RADS 1, 2 or PI‐RADS 3, 4, 5 with an area under curve (AUC) value of 0.779, compared with 0.764 for ADC. LWI differentiated histologically confirmed malignant from nonmalignant tissue with AUC, sensitivity, and specificity values of 0.81, 75%, and 87%, compared with 0.75, 83%, and 67% for ADC. The microstructural basis of the LWI technique is further suggested by the correspondence with the VERDICT diffusion‐based microstructural imaging technique, with α, A 1, A 2, and LWF showing significant correlations. Data Conclusion: LWI alone can predict PI‐RADS v2 score groupings and detect histologically confirmed tumors with an ability similar to ADC alone without the limitations of diffusion‐weighted MRI. This is important, given that ADC has an advantage in these tests as it already informs PI‐RADS v2 scoring. LWI also provides multicompartment information that has an explicit biophysical interpretation, unlike ADC. Level of Evidence : 3 Technical Efficacy : Stage 2 J. Magn. Reson. Imaging 2019;50:910–917.