Quantitative comparison of luminal water imaging with DWI for characterization of prostate cancer aggressiveness: early experience
Stefanie Hectors1,2, Daniela Said1,2, Jeffrey Gnerre1,2, Ashutosh Tewari3, and Bachir Taouli1,2

1Translational and Molecular Imaging Institute, Icahn School of Medicine at Mount Sinai, New York, NY, United States, 2Department of Radiology, Icahn School of Medicine at Mount Sinai, New York, NY, United States, 3Department of Urology, Icahn School of Medicine at Mount Sinai, New York, NY, United States


Luminal water imaging (LWI) is an emerging technique for noninvasive characterization of prostate cancer (PCa) aggressiveness. The goal of our ongoing study is to compare the diagnostic performance of LWI to DWI for assessment of PCa aggressiveness. We observed that the luminal water fraction (LWF) from LWI showed high diagnostic performance for differentiation between Grade Group (GG) 2 (i.e. Gleason 3+4) and GG 3 and higher (i.e. Gleason 4+3 and higher) cancers (AUC=0.86), while ADC showed an AUC of 0.62. These initial results suggest additional value of LWI for PCa characterization, which will be verified in a larger cohort.


Luminal water imaging (LWI), a recently introduced MRI method that employs multicomponent modeling of T2 mapping data, has shown promise for the detection and grading of prostate cancer (PCa) 1. While initial results are promising, it has not yet been established how this technique compares to DWI for PCa characterization. ADC quantification is valuable for the characterization of PCa, although there is substantial overlap of ADC values between histopathological grades 2. The goal of our study is to compare the diagnostic performance of LWI to DWI for characterization of PCa.


Thirteen PCa patients scheduled to undergo prostatectomy were recruited prospectively in this IRB-approved ongoing study (mean age 60 y, range 46-70 y). MRI was performed at 3T (Skyra, Siemens) with an 18 element phased-array coil and included a LWI T2 mapping protocol that was acquired using a non-fat-suppressed TSE sequence (TR 3680 ms, 30 TE values between 30 and 900 ms, matrix 128x128, FOV 250x250 mm2, 8 slices, slice thickness/gap 4/0.8 mm, GRAPPA x 4, acquisition time 6:05 min). DWI was performed using a fat-suppressed SS EPI sequence (TR 8700 ms, TE 72 ms, b-values 50, 1000, 1600 and 2000 s/mm2, matrix 114x114, FOV 250x250 mm2, 39 slices, slice thickness/gap 3/0 mm, GRAPPA x 2, acquisition time 9:53 min). Index lesions were identified and annotated by a body MRI radiologist, with reference to the prostatectomy report. LWI parameters (geometric mean of the short, long components and entire distribution T2 short,T2 long, gmT2 and luminal water fraction LWF) were quantified in index tumors and PZ tissue using a regularized non-negative least squares algorithm. ADC values were also quantified in index tumors and PZ. Gleason score (GS) and Grade Groups (GG) were retrieved from prostatectomy reports. Spearman correlation analysis was employed to assess the correlation of the MRI parameters with GS/GG and the correlation of LWI parameters with ADC and lesion size. Wilcoxon signed rank tests, Mann Whitney U tests and ROC analysis were used to assess differences between MRI parameters in tumors and PZ and between GS/GG scores.


Thirteen index lesions were analyzed in the 13 included patients, located in PZ (n=9) and TZ (n=4). Mean tumor size was 1.8±0.7cm (range 0.7 – 3.7 cm). A representative example of LWF and ADC maps is shown in Figure 1. Most patients [12/13 (92%)] had GS 7 cancer, separated into GS 3+4 (i.e. GG2, n=7) and GS 4+3 (i.e. GG3, n=5) (Table 1). All LWI parameters and ADC were significantly different between PZ and PCa (P<0.001 for all parameters; Table 2). LWI parameters were not significantly correlated with GG/GS (P>0.096), probably related to the small sample size in this pilot study. LWI parameters also did not correlate with ADC (P>0.325) or lesion size (P>0.303). LWF was significantly different between GG 2 and GG ≥ 3 lesions (P=0.035, Table 3), while ADC was not (P=0.534). The strongest diagnostic performance for differentiation between GG 2 and GG ≥ 3 lesions was observed for LWF (AUC=0.86), while ADC showed an AUC of 0.62.


Our study confirms the results from a previous study 1, showing significant differences in LWF between benign PZ and PCa and association of LWF with Gleason score. While the previous study did not observe significant differences between benign and malignant tissue using any of the other LWI parameters, we observed that all parameters were significantly different between these tissue types. The discrepancies between the studies may be due to different LWI acquisition protocols. The effect of acquisition protocol on LWI parameter estimation needs to be further investigated. The lack of significant difference in ADC values between GG 2 and GG 3 cancers has been demonstrated previously 3. The significant difference in LWF between GG 2 and GG ≥ 3 lesions observed in our study warrant further investigation of this technique for PCa characterization.


Our preliminary results suggest value of LWI for characterization of PCa aggressiveness. These findings need to be validated in a larger patient cohort that will be recruited in our ongoing study.


This research was supported by a Prostate Cancer Foundation Young Investigator Award.


1. Sabouri S, Chang SD, Savdie R, Zhang J, Jones EC, Goldenberg SL, Black PC, Kozlowski P. Luminal Water Imaging: A New MR Imaging T2 Mapping Technique for Prostate Cancer Diagnosis. Radiology 2017;284(2):451-459.

2. Wibmer AG, Sala E, Hricak H, Vargas HA. The expanding landscape of diffusion-weighted MRI in prostate cancer. Abdominal radiology 2016;41(5):854-861.

3. Jyoti R, Jain TP, Haxhimolla H, Liddell H, Barrett SE. Correlation of apparent diffusion coefficient ratio on 3.0T MRI with prostate cancer Gleason score. European journal of radiology open 2018;5:58-63.


Figure 1. T2WI, LWF and ADC maps in a 58 year old PCa patient with a GS7/GG2 lesion in PZ, outlined in blue. The lesion exhibited decreased LWF (mean LWF 0.07) and ADC (mean ADC 0.8*10-3 mm2/s) compared to non tumoral PZ tissue (mean LWF 0.48 and mean ADC 1.5*10-3 mm2/s).

Table 1. Distribution of prostatectomy Gleason scores (GS) and Grade Groups (GG) in 13 patients.

Table 2. Average parameter values in PZ and index PCa tumors in 13 patients. The P-values originate from Wilcoxon signed rank tests.

Table 3. Average parameter values in GG 2 lesions (n=7) and GG ≥ 3 (n=6). The P-values originate from Mann Whitney U tests.

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)