High-resolution Diffusion-weighted Imaging of the Prostate Using Multiplexed Sensitivity-encoding: Comparison with the Conventional and Reduced Field-of-view Techniques

Abstract

Purpose

To compare objective and subjective image quality, lesion conspicuity, and apparent diffusion coefficient (ADC) of high-resolution multiplexed sensitivity-encoding diffusion-weighted imaging (MUSE-DWI) with conventional DWI (c-DWI) and reduced FOV DWI (rFOV-DWI) in prostate MRI.

Methods

Forty-seven patients who underwent prostate MRI, including c-DWI, rFOV-DWI, and MUSE-DWI, were retrospectively evaluated. SNR and ADC of normal prostate tissue and contrast-to-noise ratio (CNR) and ADC of prostate cancer (PCa) were measured and compared between the three sequences. Image quality and lesion conspicuity were independently graded by two radiologists using a 5-point scale and compared between the three sequences.

Results

The SNR of normal prostate tissue was significantly higher with rFOV-DWI than with the other two DWI techniques (P ≤ 0.01). The CNR of the PCa was significantly higher with rFOV-DWI than with MUSE-DWI (P < 0.05). The ADC of normal prostate tissue measured by rFOV-DWI was lower than that measured by MUSE-DWI and c-DWI (P < 0.01), while there was no difference in the ADC of cancers. In the qualitative analysis, MUSE-DWI showed significantly higher scores than rFOV-DWI and c-DWI for visibility of anatomy and overall image quality in both readers, and significantly higher scores for distortion in one of the two readers (P < 0.001). There was no difference in lesion conspicuity between the three sequences.

Conclusion

High-resolution MUSE-DWI showed higher image quality and reduced distortion compared to c-DWI, while maintaining a wide FOV and similar ADC quantification, although no difference in lesion conspicuity was observed.

Introduction

In recent years, multiparametric MRI has been widely used for the screening and evaluation of prostate cancer (PCa). Diffusion-weighted imaging (DWI) is one of the essential sequences in prostate MRI, and in Prostate Imaging Reporting and Data System (PI-RADS) version 2.1, DWI is used as the dominant sequence for the assessment of peripheral zone (PZ) cancers and is also used to categorize transitional zone (TZ) cancers.¹ Improving the quality of DWI would be helpful in detecting small cancers and more accurately assessing lesion extension.

Currently, single-shot echo-planar imaging (EPI) is routinely used to acquire DWI. However, single-shot EPI is prone to susceptibility artifacts, and image distortion caused by rectal gas is often a problem.² In addition, it is difficult to improve the spatial resolution of DWI with single-shot EPI because increasing the number of encodings prolongs the TE and further increases the distortion. A reduced FOV EPI technique, FOCUS (FOV optimized and constrained undistorted single shot; GE Healthcare, Waukesha, WI, USA), is a relatively new technique that uses a 2D spatially selective echo-planar RF excitation pulse and a 180° refocusing pulse to reduce the FOV in the phase-encoding direction.³ The application of this technique to prostate MRI has been reported to improve image quality, increase spatial resolution, and reduce image distortion.^4–6 However, this technique has a limited FOV and is therefore not suitable for assessing lymph nodes or bone metastases in the pelvis.

Another technique to increase spatial resolution and reduce distortion in DWI is multi-shot EPI, which divides the k-space into multiple segments along the phase-encoding direction. Because multi-shot EPI collects signal data in multiple segments, it minimizes the TE extension associated with increasing the number of encodings, thereby preventing distortion. Multiplexed sensitivity encoding (MUSE, GE Healthcare) is a newly developed technique for reconstructing multi-shot EPI DWI using self-navigation to correct for motion-induced shot-to-shot phase errors.^7–9 Multi-shot EPI DWI techniques developed by other vendors include readout segmentation of long variable echo-trains (RESOLVE) and image reconstruction using the image-space sampling function (IRIS).^10,11 RESOLVE is a readout-segmented technique, while MUSE and IRIS are phase-segmented techniques, and IRIS uses a 2D navigator echo to reduce motion effects between shots.¹¹ The usefulness of MUSE-DWI has been reported in various organs, including the breast, liver, and rectum,^12–14 and there have also been a few reports of its usefulness in prostate MRI.^15,16 However, to our knowledge, there are no reports evaluating the image quality and lesion conspicuity of prostate DWI using reduced FOV DWI (rFOV-DWI) and MUSE-DWI with higher spatial resolution compared to single-shot EPI DWI. Improving image quality, including DWI, is essential in prostate MRI, and the recently proposed Prostate Imaging Quality (PI-QUAL) score has been reported to correlate with the frequency of PI-RADS-3, the positive predictive value in detecting clinically significant cancer, and the percentage of upgrades in postoperative pathology.^17–20 Therefore, if MUSE-DWI improves the image quality of DWI, it is expected to improve the diagnostic performance of prostate MRI. The purpose of this study was to compare objective and subjective image quality, lesion conspicuity, and apparent diffusion coefficient (ADC) of high-resolution MUSE-DWI with conventional DWI (c-DWI) and rFOV-DWI in prostate MRI.

Materials and Methods

Patients

This retrospective study was approved by our institutional review board, and the requirement for informed consent was waived. This study included 47 consecutive patients with suspected of having a PCa who underwent prostate MRI, including c-DWI, rFOV-DWI, and MUSE-DWI, between December 2019 and December 2020 (age range, 43–85 yrs; mean age, 68.0 yrs). The mean serum prostate-specific antigen level was 9.63 ng/mL (range, 0.76–42.01 ng/mL). Systematic prostate biopsy was performed after MRI in 24 of the 47 cases, whereas biopsy was not performed in the remaining 23 cases. Thirteen PCa were histologically diagnosed in 12 patients, including 6 with Gleason score 3 + 3, 5 with 3 + 4, and 2 with 4 + 3. Two patients had a high suspicion of cancer based on MRI and clinical findings, but a biopsy was not performed due to their poor general condition. These 15 lesions were included in the subsequent analysis. Nine cancers were located in the PZ, five in the TZ, and one cancer was located across the TZ and PZ.

MR examination

MRI was performed using a 3T scanner (SIGNA Architect, GE Healthcare) with the anterior (ventral) flexible surface phased-array coil (adaptive image receive coil, GE Healthcare) combined with the dorsal embedded (built-in) coil. Antispasmodics were not used in all cases. The scan parameters of the three DWI techniques are summarized in Table 1. The in-plane spatial resolution of rFOV-DWI and MUSE-DWI was almost equivalent and higher than that of c-DWI. The acquisition times for rFOV-DWI and MUSE-DWI were almost the same (approximately 4 min). In our institution, all DWIs are acquired with b values of 0 and 1000 s/mm², and images with higher b values are generated using the computed DWI technique, which provides a higher SNR.²¹ ADC maps and b = 2000 computed DWI were calculated from the combination of b = 0 and 1000 s/mm² images for each of c-DWI, MUSE-DWI, and rFOV-DWI using a commercially available workstation (SYNAPSE VINCENT version 5.3.001; FUJIFILM, Tokyo, Japan).

Table 1.

Scan parameters of the three DWI acquisition techniques

	c-DWI	rFOV-DWI	MUSE-DWI
b value (s/mm2)	0, 1000	0, 1000	0, 1000
TR/TE (ms)	7500/68.5	5000/65	5000/79.5
Phase-encoding direction	A-P	A-P	A-P
Number of excitations	5	12	5
Slice thickness/gap (mm)	4/1	4/0	4/0.4
Slice number	26	14	21
Band width (kHz)	250	250	250
Matrix	128 × 128	96 × 76	192 × 192
FOV (mm)	300 × 300	140 × 112	300 × 300
In-plane spatial resolution (mm)	2.34 × 2.34	1.46 × 1.47	1.56 × 1.56
Acquisition time	2 min 8s	4 min 10s	4 min 15s

c-DWI, conventional diffusion-weighted image; MUSE-DWI, multiplexed sensitivity-encoding diffusion-weighted image; rFOV-DWI, reduced FOV diffusion-weighted image.

Quantitative analysis

Two radiologists (18 and 7 yrs of experience in abdominopelvic radiology) placed oval ROIs of approximately 50 mm² on PZ and TZ of the prostate and the homogeneous area of the internal obturator muscle on c-DWI, rFOV-DWI, and MUSE-DWI. For patients with PCa, ROIs were placed on the corresponding lesions. If the lesion was too small to place a 50 mm² ROI, the largest possible ROI was placed. ROIs were also placed on PZ, TZ, and PCa on ADC maps generated from c-DWI, MUSE-DWI, and rFOV-DWI. For PZ, TZ, and muscle, three ROIs were placed in different areas, and for PCa, the ROI placement was repeated three times, and overlap between ROIs was allowed if the PCa was too small to place ROIs in three different areas. The mean value within each ROI was then averaged. The standard deviation (SD) of the signal intensity (SI) of the muscle was used as an index of image noise. The SNR of the PZ and TZ and contrast-to-noise ratio (CNR) of the PCa were calculated using the following equations:

$$\mathrm{S}\mathrm{N}\mathrm{R}=\mathrm{S}\mathrm{I}/\mathrm{S}{\mathrm{D}}_{\mathrm{m}\mathrm{u}\mathrm{s}\mathrm{c}\mathrm{l}\mathrm{e}},$$

$$\mathrm{C}\mathrm{N}\mathrm{R}=\left(\mathrm{S}{\mathrm{I}}_{\mathrm{l}\mathrm{e}\mathrm{s}\mathrm{i}\mathrm{o}\mathrm{n}}-\mathrm{S}{\mathrm{I}}_{\mathrm{p}\mathrm{r}\mathrm{o}\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{t}\mathrm{e}}\right)/\mathrm{S}{\mathrm{D}}_{\mathrm{m}\mathrm{u}\mathrm{s}\mathrm{c}\mathrm{l}\mathrm{e}},$$

where SI_prostate and SI_lesion are the SI of the prostate and PCa, respectively. SD_muscle is the SD of the muscle SI. Measurements were performed using a workstation (SYNAPSE VINCENT version 5.3.001; FUJIFILM).

Qualitative analysis

Three DWIs (c-DWI, MUSE-DWI, and rFOV-DWI) from each patient were randomly divided into three datasets for subjective image quality assessment, and the order of patients was also randomized. Two radiologists (18 and 10 yrs of experience in abdominopelvic radiology) independently reviewed the DWI (b = 1000) and assessed image quality in terms of visibility of prostate anatomy, degree of noise, image distortion, and overall image quality. Visibility of prostate anatomy, such as the border between the prostate and surrounding tissue and the border between the PZ and TZ, was graded on a 5-point scale: 1, unclear; 2, slightly unclear; 3, moderate; 4, slightly clearly delineated; and 5, very clearly delineated. The degree of noise was rated on a 5-point scale: 1, unacceptable image noise; 2, slightly excessive noise; 3, moderate noise; 4, slightly less noise; and 5, minimal noise. Image distortion due to susceptibility effects was rated on a 5-point scale: 1, unacceptable image distortion; 2, slightly excessive distortion; 3, moderate distortion; 4, slightly less distortion; and 5, minimal distortion. Overall image quality was also rated on a 5-point scale: 1, poor; 2, fair; 3, moderate; 4, good; and 5, excellent. For cases with at least one cancer, lesion conspicuity on DWI (b = 1000) was graded with reference to computed DWI (b = 2000) and ADC map on a 5-point scale: 1, no abnormal signal; 2, abnormal signal but does not appear to be a lesion; 3, abnormal signal that may be a lesion; 4, there was probably a lesion; and 5, there was definitely a lesion. In addition, the DWI was evaluated to determine if it met the criteria for diagnostic quality according to PI-QUAL.¹⁷ Both readers were blinded to the DWI acquisition technique, although they were informed that the given images were c-DWI, rFOV-DWI, or MUSE-DWI. To minimize the recall bias, the evaluation of each dataset was separated by a time interval of at least 2 weeks.

Statistical analysis

The SNR of PZ and TZ, CNR of PCa, and ADC of PZ, TZ, and PCa were compared between c-DWI, rFOV-DWI, and MUSE-DWI using Friedman’s test, followed by post hoc Wilcoxon signed rank test with Bonferroni correction (n = 3). The intraclass correlation coefficient (ICC) was calculated to assess the interobserver agreement. The visual scores of the qualitative analyses were also compared using Friedman’s test, followed by post hoc Wilcoxon signed rank test with Bonferroni correction (n = 3). The frequencies of cases meeting the PI-QUAL diagnostic quality criteria were compared using the chi-square test with Bonferroni correction (n = 3). The weighted kappa statistic was used to assess inter-reader agreement. All statistical analyses were performed using R (version 4.2.2, R Foundation for Statistical Computing, Vienna, Austria). For all tests, a P value < 0.05 was considered significant.

Results

Quantitative analysis

The SNR of PZ and TZ, CNR of PCa, and ADC of PZ, TZ, and PCa are summarized in Table 2. For both PZ and TZ, SNR was significantly higher on rFOV-DWI than on c-DWI and MUSE-DWI (P ≤ 0.01), and higher on c-DWI than on MUSE-DWI for both radiologists (P < 0.001). CNR was also higher on rFOV-DWI than on c-DWI and MUSE-DWI, and the difference was significant between rFOV-DWI and the other two techniques for Radiologist 1 and between rFOV-DWI and MUSE-DWI for Radiologist 2 (P < 0.001). For both PZ and TZ, ADC was significantly lower on rFOV-DWI than on c-DWI and MUSE-DWI for both radiologists (P < 0.01), whereas ADC of PCa was not significantly different between the three DWIs for both radiologists. The ICCs of the two readers for SNR of PZ and TZ were 0.478 and 0.319, 0.424 and 0.411, and 0.483 and 0.314 for c-DWI, rFOV-DWI, and MUSE-DWI, respectively (fair to moderate reproducibility). The ICCs of the two readers for CNR of PCa were 0.804, 0.844, and 0.775 for c-DWI, rFOV-DWI, and MUSE-DWI, respectively (substantial to almost perfect reproducibility). The ICCs of the two readers for ADC of PZ, TZ, and PCa were 0.882, 0.676, and 0.904, 0.848, 0.776, and 0.864, and 0.887, 0.724, and 0.870 for c-DWI, rFOV-DWI, and MUSE-DWI, respectively (substantial to almost perfect reproducibility).

Table 2.

SNR, CNR, and ADC of normal prostate tissue and PCa

	c-DWI	rFOV-DWI	MUSE-DWI	P value
				c-DWI vs rFOV-DWI	c-DWI vs MUSE-DWI	rFOV-DWI vs MUSE-DWI
Radiologist 1
SNR
PZ	42.7 ± 7.5	47.2 ± 6.5	24.0 ± 3.6	< 0.001	< 0.001	< 0.001
TZ	42.1 ± 6.8	49.3 ± 7.6	24.9 ± 3.4	< 0.001	< 0.001	< 0.001
CNR	17.8 ± 8.5	25.8 ± 11.5	13.7 ± 7.5	< 0.001	< 0.001	< 0.001
ADC
PZ	1.88 ± 0.24	1.76 ± 0.21	1.88 ± 0.25	< 0.001	> 0.99	< 0.001
TZ	1.46 ± 0.15	1.38 ± 0.14	1.46 ± 0.14	< 0.001	> 0.99	< 0.001
Lesion	0.88 ± 0.18	0.87 ± 0.19	0.85 ± 0.17	> 0.99	> 0.99	> 0.99
Radiologist 2
SNR
PZ	41.5 ± 9.7	45.9 ± 6.9	23.7 ± 4.6	0.01	< 0.001	< 0.001
TZ	41.6 ± 7.7	48.4 ± 6.2	24.4 ± 4.2	< 0.001	< 0.001	< 0.001
CNR	18.1 ± 11.1	22.2 ± 11.6	11.6 ± 7.3	0.25	0.04	< 0.001
ADC
PZ	1.89 ± 0.23	1.81 ± 0.20	1.93 ± 0.22	< 0.001	0.17	< 0.001
TZ	1.45 ± 0.16	1.41 ± 0.14	1.48 ± 0.15	0.009	0.08	< 0.001
Lesion	0.93 ± 0.19	0.94 ± 0.18	0.92 ± 0.18	> 0.99	> 0.99	> 0.99

Data are mean score ± SD. The significant results are shown in bold fonts. ADC, apparent diffusion coefficient; c-DWI, conventional diffusion-weighted image; CNR, contrast-to-noise ratio; MUSE-DWI, multiplexed sensitivity-encoding diffusion-weighted image; PZ, peripheral zone; rFOV-DWI, reduced FOV diffusion-weighted image; SD, standard deviation; TZ, transitional zone.

Qualitative analysis

Table 3 shows the mean visual scores for each DWI technique. The visibility of the prostate anatomy on MUSE-DWI was significantly better than that on c-DWI or rFOV-DWI for both readers (P < 0.001). In terms of subjective image noise, rFOV-DWI had the highest score (i.e., lower image noise) for both readers, and the difference between c-DWI and rFOV-DWI was statistically significant for Reader 1. MUSE-DWI scored significantly higher than c-DWI and rFOV-DWI in terms of distortion (i.e., less distorted image), and the differences were statistically significant for Reader 2 (P < 0.001) (Fig. 1). Lesion conspicuity was not significantly different between the three DWI techniques for either reader (Fig. 2). Overall image quality on MUSE-DWI was significantly better than that on c-DWI or rFOV-DWI for both readers (P < 0.001). The percentages of each score for each rating are shown in Fig. 3. The percentages of cases classified as diagnostic quality according to PI-QUAL criteria for c-DWI, rFOV-DWI, and MUSE-DWI were 46.8%, 55.3%, and 44.7%, respectively, for Reader 1, and 48.9%, 68.1%, and 74.5%, respectively, for Reader 2. A significant difference between c-DWI and MUSE-DWI was observed in Reader 2 (P = 0.01).

Table 3.

Mean visual scores of the qualitative analysis

	c-DWI	rFOV-DWI	MUSE-DWI	P value
				c-DWI vs rFOV-DWI	c-DWI vs MUSE-DWI	rFOV-DWI vs MUSE-DWI
Reader 1
Visibility of prostate anatomy	3.0 ± 0.7	2.9 ± 0.7	4.3 ± 0.8	> 0.99	< 0.001	< 0.001
Subjective image noise	3.2 ± 0.7	3.5 ± 0.5	3.2 ± 0.6	0.04	> 0.99	0.05
Distortion	3.3 ± 1.0	3.3 ± 1.3	3.5 ± 1.2	> 0.99	0.22	0.08
Lesion conspicuity	4.6 ± 0.6	4.7 ± 0.6	4.5 ± 0.9	> 0.99	> 0.99	> 0.99
Overall image quality	3.1 ± 0.7	2.9 ± 0.8	3.7 ± 0.8	0.76	< 0.001	< 0.001
Reader 2
Visibility of prostate anatomy	3.0 ± 0.6	3.3 ± 0.7	4.7 ± 0.5	0.008	< 0.001	< 0.001
Subjective image noise	2.9 ± 0.7	3.2 ± 0.5	3.0 ± 0.6	0.06	> 0.99	0.19
Distortion	2.8 ± 0.7	3.1 ± 0.8	4.1 ± 0.8	0.06	< 0.001	< 0.001
Lesion conspicuity	3.7 ± 1.0	4.1 ± 1.1	4.0 ± 1.3	0.22	0.83	> 0.99
Overall image quality	2.9 ± 0.5	3.2 ± 0.7	4.5 ± 0.5	0.02	< 0.001	< 0.001

Data are mean score ± SD. The significant results are shown in bold fonts. c-DWI, conventional diffusion-weighted image; MUSE-DWI, multiplexed sensitivity-encoding diffusion-weighted image; rFOV-DWI, reduced FOV diffusion-weighted image; SD, standard deviation.

Fig. 1.

DWI (b = 1000) acquired with c-DWI (a), rFOV-DWI (b), and MUSE-DWI (c) in a male in his 70s. Distortion due to rectal gas (arrows) is prominent on c-DWI, while it is reduced on rFOV-DWI, and further reduced on MUSE-DWI. c-DWI, conventional diffusion-weighted image; MUSE-DWI, multiplexed sensitivity-encoding diffusion-weighted image; rFOV-DWI, reduced FOV diffusion-weighted image.

Fig. 2.

DWI (b = 1000) acquired with c-DWI (a), rFOV-DWI (b), and MUSE-DWI (c) in a male in his 70s. Although MUSE-DWI has more noise, the prostate anatomy is most clearly delineated. An area of high-signal intensity in the left PZ (arrows), suspicious for PCa, is clearly visible on all images. c-DWI, conventional diffusion-weighted image; MUSE-DWI, multiplexed sensitivity-encoding diffusion-weighted image; PZ, peripheral zone; rFOV-DWI, reduced FOV diffusion-weighted image.

Fig. 3.

The graphs show the percentage of each score for visibility of prostate anatomy (a), subjective image noise (b), distortion (c), lesion conspicuity (d), and overall image quality (e).

The weighted kappa coefficients for c-DWI, rFOV-DWI, and MUSE-DWI were 0.39, 0.44, and 0.30 for anatomy visibility (fair to moderate agreement), 0.26, 0.25, and 0.27 for image noise (fair agreement), 0.53, 0.56, and 0.51 for distortion (moderate agreement), 0.46, 0.30, and 0.47 for lesion conspicuity (fair to moderate agreement), and 0.41, 0.54, and 0.26 for overall image quality (fair to moderate agreement), respectively.

Discussion

DWI is an essential sequence in prostate MRI, and the acquisition of high-quality DWI is crucial for the accurate diagnosis of PCa. In recent years, the usefulness of bi-parametric MRI, which obviates contrast-enhanced dynamic MRI, has been reported.²² It would be more important to obtain high-quality DWI in a bi-parametric MRI. We, therefore, evaluated the utility of two relatively new DWI techniques, rFOV- and MUSE-DWI, in prostate MRI compared with c-DWI. Our results of quantitative analysis showed that rFOV-DWI had the highest SNR, followed by c-DWI. The CNR of PCa was also higher with rFOV-DWI than with the other two DWI techniques. The higher SNR and CNR of rFOV-DWI are likely due to the higher number of excitations (i.e., 12 for rFOV-DWI and 5 for other DWIs). MUSE-DWI uses a smaller number of excitations to keep the acquisition time in line with rFOV-DWI, which probably results in a higher noise level. MUSE-DWI uses the same number of excitations as c-DWI, although the noise level is thought to be higher, reflecting the higher spatial resolution. In the qualitative analysis, the rFOV-DWI had the highest noise score (i.e., less noise) for both two readers, although a significant difference between rFOV-DWI and c-DWI was found only for Reader 1. On the other hand, a previous report showed that the qualitative SNR scores were higher for MUSE-DWI compared with rFOV-DWI.¹⁵ However, in that report, the acquisition time for MUSE-DWI was twice that of rFOV-DWI. In our study, we used comparable acquisition times, which may have made it difficult to maintain a high SNR with MUSE-DWI.

The results of the qualitative analysis showed that MUSE-DWI scored significantly higher than c-DWI and rFOV-DWI in terms of visibility of prostate anatomy and overall image quality. In addition, MUSE-DWI was rated by one reader as significantly less distorted than c-DWI and rFOV-DWI. These results are consistent with previous reports.^15,16 In addition, the percentage of cases meeting the PI-QUAL diagnostic quality criteria was significantly higher for MUSE-DWI than for c-DWI for one reader. Therefore, MUSE-DWI is a technique that can provide DWI with high image quality and spatial resolution while reducing image distortion. Since image distortion due to rectal gas is often a problem in prostate DWI, MUSE-DWI is expected to be particularly useful in delineating dorsal prostate lesions. Although rFOV-DWI also improves spatial resolution and reduces distortion, it has a limited FOV and is primarily used to evaluate the prostate only. As DWI is also useful in evaluating lymph node and bone metastases of PCa,^23–25 MUSE-DWI, which can improve spatial resolution while maintaining a wide FOV, would be a promising technique that could adequately replace c-DWI in prostate MRI. Inter-reader agreement for qualitative analysis was generally fair to moderate. The relatively low kappa coefficients may be due to the 5-point scale and the difficulty in establishing objective evaluation criteria. Moderate agreement was found for distortion, suggesting a relatively high reproducibility of increased scores with MUSE-DWI, although a significant difference was found in only one reader.

Despite the superior anatomical visibility and overall image quality of MUSE-DWI compared to c-DWI and rFOV-DWI, no difference in lesion conspicuity was observed in our study. The main reason for this may be that PCa usually has very high contrast on DWI, so lesions are sufficiently visible even if the image quality and resolution are not so high. Another reason may be that this study did not include cases of PZ cancer that were poorly delineated on c-DWI due to susceptibility artifacts. A previous study evaluating prostate DWI using the multi-shot EPI technique developed by another vendor reported increased lesion conspicuity as a result of reduced distortion and blurring with multi-shot EPI compared with single-shot EPI, although there was no significant difference in diagnostic performance.¹¹ Thus, we believe that DWI with less distortion will be useful in delineating dorsal PZ cancers, and that the use of MUSE-DWI will improve lesion delineation in cases where the distortion caused by rectal gas is significant. Further studies with more cases would be needed to validate the utility of rFOV-DWI and MUSE-DWI in the detection of PCa.

Our results showed that the ADC measured on rFOV-DWI was lower than MUSE-DWI and c-DWI in both normal PZ and TZ, whereas there was no difference in the ADC of cancers. This result is consistent with a previous report showing lower ADC in non-cancerous regions on rFOV-DWI than on other sequences.¹⁵ Although the noise floor effect has been reported as one of the causes of underestimation of ADC, it does not appear to be the cause in our study, as the SNR of rFOV-DWI was the highest. The exact cause is unknown; however, the difference with other sequences is approximately 0.04–0.12 × 10^–3 mm²/s, which does not seem to be a major clinical problem. Regarding MUSE-DWI, previous reports have shown no difference in ADC compared with c-DWI,^15,16 and our results are consistent with these reports. Therefore, we believe that c-DWI can be replaced by MUSE-DWI in the measurement of ADC, and the results of the quantitative evaluation previously reported on c-DWI can be applied.

This study had several limitations. First, the number of cases was relatively small, especially the number of patients with cancer. In addition, few patients underwent total prostatectomy, thus making it difficult to compare diagnostic performance. To determine the clinical utility of rFOV-DWI and MUSE-DWI, it will be necessary to compare their diagnostic performance using a larger number of cancer cases. Second, due to the small number of cancer cases, it was not possible to study clinically significant and non-significant cancers separately. Third, rFOV-DWI and MUSE-DWI had longer acquisition times than c-DWI, and comparisons were not made for matched acquisition times. Finally, because rFOV-DWI has a different FOV than the other two DWI techniques, a fully blinded evaluation was not possible.

Conclusion

High-resolution MUSE-DWI demonstrated higher image quality and reduced distortion compared to c-DWI while maintaining a wide FOV and similar ADC quantification.