Introduction

LWI is a new application of quantitative T₂ mapping that has shown promising results for detection, and grading of prostatic carcinoma (PCa)¹. This technique measures the fractional amount of luminal water which has been shown to be proportional to the percentage area of luminal space in prostatic tissue ². Due to the fact that the amount of luminal space decreases as tissue transforms from normal to malignant, and the decrease in luminal space increases with Gleason score (GS), LWI can potentially provide valuable information for both detection and grading of PCa. This study was designed with the purpose of evaluating the diagnostic accuracy of LWI directly compared to DCE and DW-MRI, which are the two most commonly used techniques in mp-MRI protocols, and evaluating the significance of adding LWI in to an mp-MRI protocol in detection and grading of PCa.

Methods

15 patients with biopsy proven PCa were examined with LWI, DWI, and DCE MRI at a 3T scanner [Philips Medical Systems, Best, The Netherlands] prior to undergoing radical prostatectomy. MR signals were acquired with a combined endorectal/pelvic phased-array coils. Sequence parameters are presented in Table 1. From LWI data, signal decay curve for each voxel was extracted and fitted to multi-exponential function to generate maps of 7 parameters called: T_2-short, T_2-long, gmT₂, LWF, A_short, A_long, and N as described previously by Sabouri et al.¹. From the set of four DWI images, with b values: 0, 100, 500, and 1000 s/mm², apparent diffusion coefficient (ADC) was calculated by using the Stejskal-Tanner equation. For processing DCE MRI data, concentration of contrast agent in the tissue was calculated from PD and T1W images as described by Parker et al.³. By fitting the concentration of contrast agent in tissue to the extended Kety model⁴ values of pharmacokinetic parameters: K_trans, ν_e, and ν_p were calculated. Maps of all MR parameters were correlated to whole-mount histology through digital registration. Receiver operating characteristics (ROC) analysis and logistic generalized linear mixed effect models (GLMMs) were used to evaluate the diagnostic accuracy of individual and combined MR parameters in detection of PCa. In multi-parametric ROC analysis only the parameters that contributed significantly to the GLMM model were used. Correlations between MR parameters and GS were evaluated with application of Spearman’s rank correlation test for each patient and averaging over the entire correlation coefficients. Significance of the correlation between MR parameters and GS was evaluated with application of the one sample t-test. P<0.05 was considered statistically significant.

Results

Representative MR parametric maps are shown in Figure 1. The values of sensitivity, specificity and AUC calculated from single and multi-parameter ROC analyses are summarized in Tables 2 and 3. Average Spearman's rank correlation coefficients between MR parameters and GS, as well as the results of t-test are presented in in Table 4. Among the single MRI parameters, the highest area under ROC curve (AUC) was obtained for: ADC and A_short in the entire prostate, gmT₂, LWF, and A_long in the peripheral zone (PZ), and gmT₂ and LWF in transition zone (TZ). In our multi-parametric analysis, the highest AUC value in the entire prostate was obtained for combination of LWI, DCE, and DWI, while in PZ the highest value of AUC was obtained for combination of LWI with either of the other two techniques. In TZ, there was no significant difference in AUC value obtained from different combinations. In PZ, the strongest correlation with GS was achieved by LWF (ρ:-0.81±0.09, P<0.001).

Discussion

Comparing diagnostic accuracies obtained from individual parameters of each technique showed that values of AUC obtained from LWI are equally good or better than those obtained from DCE and DWI in all regions of investigation. Comparing the diagnostic accuracies of combined techniques showed that for both DCE and DWI the accuracy is improved when combined with LWI. In correlations with GS, the highest correlation coefficients were obtained from LWI, suggesting that in clinical settings LWI can provide higher contrast than DCE and DWI between tumors of different GS.

Conclusion

Results of this pilot study show that LWI alone performs equally well or better than DCE and DWI in detection of PCa, and performs better than the other two techniques in correlation with GS. This technique can potentially be combined with either DCE or DWI in clinical MP-MRI protocols to increase the accuracy of diagnosis of PCa. To engage LWI in clinical settings, a study with higher number of patients and broader range of GS is required.

Acknowledgements

This study has been supported by the Canadian Institutes of Health Research.