Synthesized high b-value diffusion-weighted imaging (sDWI) has been shown to be less prone to artifacts, avoid longer required echo times for high b-value acquisitions, and provide improved contrast to noise ratios for detection of prostate cancer compared to equivalent directly acquired high b-value diffusion-weighted imaging (DWI).^1,2,3,4 We examined whether sDWI for a b-value of 1200 s/mm² provides similar image quality, detection of peripheral zone lesions in the prostate, and PI-RADS version 2 categorization, compared to acquired DWI for the same b-value.

With IRB approval, fifty consecutive patients who underwent prostate MRI followed by biopsy were retrospectively identified. A multiparametric MRI protocol was performed for all patients on a 3.0 Tesla scanner with 1 mg intramuscular glucagon and without an endorectal coil.

DWI was acquired using a reduced field of view single shot EPI sequence, 160 x 80 matrix, 4.2-mm slices, TR 2000 ms, and 24 x 12-cm FOV. The acquired DWI b-values were 0, 25, 50, 800, and 1200 s/mm² with the number of signal averages 2, 12, 12, 24, and 64, respectively. The sDWI images for a b-value of 1200 s/mm² and corresponding ADC maps excluding the acquired b-value of 1200 s/mm² were created using the ADC Map Calculation plugin for OsiriX. Small field-of-view (sFOV) T2 was acquired using FSE, axial orientation, 416 x 224 matrix, 3.6-mm slices, 20-cm FOV, TE 130 ms, TR 4100 ms. Dynamic contrast-enhanced (DCE) was acquired using an axial multiphase post-contrast 3D Dual Echo SPGR.

Two blinded radiologists with 9 and 3 years of experience interpreting prostate MRI studies evaluated either the sDWI or acquired DWI with the corresponding ADC map for each patient in an alternating manner in conjunction with sFOV T2 and DCE images. sDWI or acquired DWI image quality was rated on a five-point Likert scale (1=non-diagnostic, 2=limited, 3=adequate, 4=good, 5=excellent). Readers noted either the absence or presence of peripheral zone lesions; when lesions were present, the size and location of the single largest detected lesion was recorded and assigned a PI-RADS version 2 category. A confidence score was also assigned to the overall study interpretation (1=not confident, 2=poor confidence, 3=somewhat confident, 4=confident, 5=very confident). The null hypothesis of no significant difference in scores was assessed using a Wilcoxon signed rank test and a McNemar test for binomial parameters with a target alpha of 0.05.

Subjective image quality was slightly but significantly lower for the sDWI images compared to acquired DWI images for a b-value of 1200 s/mm² (3.7 ± 0.6 vs. 3.94 ± 0.5, p=0.023). No significant difference in detection of the presence of a peripheral zone lesion (p=0.085), lesion size (13.5 ± 8.4 mm vs. 13.4 ± 5.2 mm, p=0.906), or PI-RADS version 2 categories (3.9 ± 0.7 vs. 4.1 ± 0.7, p=0.268) was present between sDWI and acquired DWI images. Confidence scores for interpretation of images were also not significantly different between sDWI and acquired DWI studies (3.9 ± 0.7 vs. 4.0 ± 0.6, p=0.667).Although subjective image quality was rated slightly but significantly worse for sDWI compared to acquired DWI images, no significant differences in detection of peripheral zone lesions, PI-RADS version 2 lesion categorization, or confidence scores for interpretation were observed. Of note, the study design was biased in favor of images acquired with 1200 s/mm² weighting, since the images with 800 s/mm² diffusion weighting (the main components of the synthesized b-1200 images) were acquired with fewer signal averages (24 averages) than the 1200 s/mm² weighted images (64 signal averages); this may have resulted in the inferior subjective image quality of the synthesized images. Additionally, in this study, the echo time for the b-800 images was higher than what could have otherwise been achieved if no b-1200 images were obtained.sDWI high b-value (1200 s/mm²) images were not significantly different for detection of peripheral zone lesions and PI-RADS version 2 categorization compared to directly acquired DWI images and may allow shorter MR acquisition time.