Association of bi‐parametric MRI measures with continence after robot‐assisted radical prostatectomy

Abstract

Objective

To investigate the association between pre‐ and postoperative magnetic resonance imaging (MRI) measurements of the membranous urethra and the prostate volume and continence following robot‐assisted radical prostatectomy (RARP).

Patients and Methods

From December 2018 to June 2022, 100 continent patients undergoing unilateral nerve‐sparing or non‐nerve‐sparing RARP were included in this cohort study. Bi‐parametric MRI scans were performed before and 12 months after RARP and measurements included the membranous urethral length (MUL) measured in cm (mMUL) and in the number of image slices (sMUL; 3 mm/slice), the membranous urethral diameter (MUD), and the prostate volume. Urinary function was evaluated by the International Consultation on Incontinence Questionnaire‐Urinary Incontinence Short Form (ICIQ‐UI SF) and continence, defined as the use of zero pads and the answer ‘never’ to the ICIQ‐UI SF question regarding incontinence frequency or <8 g urine‐loss per 24 h. Regression with robust variance estimates was used to analyse the association between measurements and outcomes.

Results

At 12 months, continence and MRI data were available for 82 patients. The continence rate was 63% and the median (interquartile range) ICIQ‐UI SF score was 4 (0–9). Both preoperative MUL measurements were associated with continence at 12 months. Every extra 5 mm of MUL increased the likelihood of being continent by 13 percentage points (P = 0.03) and every additional slice of sMUL increased it by 6 percentage points (P = 0.05). Both postoperative MUL measurements were associated with better continence and lower ICIQ‐UI SF scores (P < 0.01). A larger prostate volume was associated with urinary incontinence at 12 months, with a small effect size. The MUD was not associated with continence.

Conclusion

Preoperative mMUL and sMUL are associated with continence at 12 months after RARP. The sMUL may be a useful measurement when only the axial plane is available, and the slice gap is known. Postoperative MUL measurements are strongly associated with continence, while MUD and prostate volume hold minimal prognostic value.

Abbreviations

AUC

area under the curve

bpMRI

bi‐parametric MRI

ICC

interclass correlation coefficient

ICIQ‐UI SF

International Consultation on Incontinence Questionnaire‐Urinary Incontinence Short Form

MD

mean difference

(m)MUL

(measured in cm) membranous urethral length

MUD

membranous urethral diameter

(N)(U)NS

(non‐) (unilateral) nerve‐sparing

RCT

randomised controlled trial

RD

risk difference

ROC

receiver operating characteristic

(RA)RP

(robot‐assisted) radical prostatectomy

sMUL

MUL counted in the number of slices (3 mm/slice)

T2w

T2‐weighted

UI

urinary incontinence

Introduction

Radical prostatectomy (RP) for prostate cancer is associated with urinary incontinence (UI), which can negatively impact health‐related quality of life and may lead to treatment decision regret [1]. In the era of the robot‐assisted radical prostatectomy (RARP), UI affects 4–31% of patients a 1 year after surgery [2]. Postoperative UI is associated with risk factors including age, preoperative LUTS, and body mass index, as well as surgical factors including surgical technique, nerve sparing, and surgeon experience [2, 3, 4, 5].

Preoperative MRI measurements have been reported to be predictive of UI after RARP. These include membranous urethral length (MUL) [6, 7, 8, 9, 10], membranous urethral diameter (MUD) [11], and prostate volume [2]. Increasing MUL has consistently been reported to increase the odds of a return to continence [12]. Continence has also been associated with postoperative MRI measurements including the MUL and the change in length from pre‐ to postoperative scans [13, 14], but this has been less investigated. Some of the challenges likely lie in the accuracy of measuring MUL, in this context the inter‐rater variability differs between studies limiting its clinical value [15]. There are also conflicting reports on the associations between both MUD and prostate volume and continence [16, 17].

In this study, we report the predictive value of the preoperative MUL, MUD, and prostate volume from bi‐parametric MRI (bpMRI) scans on postoperative continence in patients undergoing unilateral nerve‐sparing (UNS) or non‐NS (NNS) RARP. In addition, we examine the association between the absolute values of postoperative bpMRI measurements and the change between pre‐ and postoperative scans and continence outcomes. We evaluate the MUL by conventional measurement (mMUL) and as a novel approach, in the number of slices or images of the body (sMUL). If the latter is predictive of continence outcomes, it could be a potential measurement on ultra‐fast bpMRI scans consisting of axial T2‐weighted (T2w) sequences only, which vastly limits the time used on diagnostic scans while maintaining prostate cancer detection accuracy [18]. The data for the present study were obtained in conjunction with a randomised controlled trial (RCT) investigating the effect of intraoperative nerve monitoring on functional outcomes after RARP [19]. The same cohort of patients was used for both studies. However, it was pre‐planned to analyse and publish the bpMRI data separately, focusing specifically on preoperative continence prediction and postoperative continence outcomes.

Patients and Methods

From December 2018 to June 2022, participants were enrolled in a RCT [19] conducted at the Department of Urology, Copenhagen University Hospital, Herlev and Gentofte Hospital. It was approved by the regional ethics committee and the Danish Data Protection Agency (Journal ‐nr.: H‐6‐2013‐003). It was registered at http://www.clinicaltrials.org (study identifier: NCT03721029). All participants underwent bpMRI scans preoperatively and again 12 months after their surgery.

In total, 100 continent men scheduled to undergo NNS or UNS RARP were included, and randomised (1:1) to undergo RARP with the use of nerve monitoring or standard RARP. For inclusion, participants needed to be continent, determined as a score of zero on the International Consultation on Incontinence Questionnaire‐Urinary Incontinence Short Form (ICIQ‐UI SF) [20]. They were excluded if they had previously received radiotherapy to the pelvic region, trauma to the pelvic region, pelvic surgery, or TURP or if they had a neurological disease or diabetes. All procedures were performed by three consultant urologists highly experienced in RARP (>300 procedures/surgeon). During RARP the peritoneum was incised, the bladder dropped, and the endopelvic fascia was incised bilaterally. The bladder neck was dissected, and Denonvilliers’ fascia opened. The vasa deferentia were divided and clipped. The lateral pedicles were clipped with Hem‐o‐lok® clips (Weck Closure Systems, Research Triangle Park, NC, USA). For UNS, the neurovascular bundle was dissected anterograde and retrograde. The dorsal vein complex was sutured, and the urethra was exposed and cut, preserving the maximum urethral stump length. Haemostasis was achieved using 5‐0 poliglecaprone 25 (Monocryl®, Ethicon Inc., Somerville, NJ, USA) and/or PerClot® (Baxter International Inc., Deerfield, IL, USA). The bladder–urethral anastomosis was made with a 3‐0 V‐Loc™ (Medtronic, plc, Dublin, Ireland) or 3‐0 poliglecaprone 25 (Monocryl, Ethicon) suture, followed by catheter placement and a 100 mL NaCl injection to test the anastomosis. The catheter was left in situ for at least 7 days.

Sphincter lesions during surgery, postoperative anastomosis leakage, or biochemical recurrence resulting in further treatment led to exclusion from the final analysis. There were 82 patients available for per‐protocol analyses [19].

Patient demographics, pre‐ and postoperative tumour characteristics, intraoperative events, and complications were recorded. Continence outcomes were collected via office visits at baseline, and at, 3, 6, and 12 months postoperatively, and included the validated ICIQ‐UI SF questionnaire, the number of pads used/day, and a 24‐h pad test. Continence after surgery was defined as the use of zero pads and the answer ‘never’ to the ICIQ‐UI SF question: ‘How often do you experience urinary incontinence?’ and/or a urine loss <8 g on the 24‐h pad test. Reporting followed the STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) guidelines (see Appendix S1). The study data were managed using the Research Electronic Data Capture (REDCap) electronic data capture tools hosted by the Capital Region of Denmark [21].

The MRI Protocol and Measurements

Imaging was performed as bpMRI scans using a 3‐Tesla Philips scanner (Ingenia version 5.3.1, Best, the Netherlands). The measurements were obtained on the axial and sagittal T2w images, and diffusion‐weighted sequences (b values: 0, 100, 800, and 2000) [22], with a slice thickness of 3 mm. The MRI protocol is shown in Table S1.

The mMUL was measured in the sagittal plane as the distance from the lowest point of the prostate apex to the urethral entry of the penile bulbous (Fig. 1). It was measured in centimetres using two decimals. The sMUL was recorded in the number of slices in the axial plane. Slices refer to the individual images of the body captured by the MRI. To obtain this measurement, the number of slices was counted from the first slice that did not show any part of the urethra inside the prostate apex, up to the last slice that did not contain any part of the urethra inside the penile bulbous. The MUD was measured in the axial plane as the outer edges of the hypodense contour that borders the pelvic musculature (Fig. 1). Prostate volume was calculated using the formula for a three‐dimensional ellipse (length × width × height × π/6). On the postoperative scans, mMUL was measured from the lowest point of the bladder at the point of the anastomosis, to the penile bulbous (Fig. 2). The sMUL was counted from the first slice that did not show any of the bladder. The MUD was measured as already described (Fig. 2).

Fig. 1.

Measurements of (A) the preoperative MUL, (B) the preoperative MUD, from study 2.

Fig. 2.

Measurements of (A) the postoperative MUL, (B) the postoperative MUD.

Image readings were performed by a primary expert reader and mMUL, sMUL, and MUD were also measured by a secondary reader to calculate the reliability of the measurements. Only the primary reader was ‘blinded’ to the study population, the original randomisation, and the endpoint data. Measurements from this reader were used for analyses of associations between measurements and functional outcomes.

Statistical Analyses

Analyses were performed using Stata Release 17 (StataCorp. 2021; StataCorp LLC, College Station, TX, USA). Baseline categorical data were described by counts and percentages, and continuous data were described by means and SDs. We conducted multivariable regression analyses with robust variance estimations to assess the predictive power of the primary reader's bpMRI measurements for continence and ICIQ‐UI SF scores at 3, 6, and 12 months. The results were reported as mean differences (MDs) or risk differences (RDs). The RD determines the difference in risk or rates of outcomes between two groups. Additionally, we performed univariate regression analyses with robust variance estimations to investigate the association between baseline characteristics and continence and the ICIQ‐UI SF score at 12 months to identify potential confounders. For each preoperative bpMRI measurement, we performed multivariable regression analyses with robust variance estimations to control for other preoperative measurements and baseline characteristics that were associated with continence or the ICIQ‐UI SF score on univariate analyses. We estimated the optimal cut‐point for preoperative mMUL and sMUL for predicting continence based on the receiver operating characteristic (ROC) curves. The area under the curve (AUC) values were also reported. We used univariate regression with robust variance estimations to analyse the association between both postoperative bpMRI measurements and the change in mMUL and sMUL from before to after RARP and continence and ICIQ‐UI SF scores at 12 months. In addition, we performed univariate regression with robust variance estimations to analyse the association between continence and the change in mMUL, and sMUL from pre‐ to postoperative RARP images. To control for prognostic baseline characteristics, we also conducted multivariable regression analyses with robust variance estimations to evaluate the association between both the postoperative bpMRI measurements, and the changes from pre‐ to postoperative measurements on one side and continence and ICIQ‐UI SF score on the other. We report the inter‐rater reliability by calculating the interclass correlation coefficient (ICC). A P < 0.05 was considered statistically significant.

Results

Preoperative bpMRI scans and continence rates at 12 months were available for all 82 patients. Secondary exclusions were due to treatment following recurrence (seven patients), anastomosis leakage (five), sphincter lesion (one), consent withdrawal due to COVID‐19 lockdown (two), non‐muscle‐invasive bladder cancer diagnosis (one), logistical problems in the operating room (one), and one patient receiving radiotherapy instead of RP after inclusion. Postoperative bpMRI scans were available for 79 patients. The three additional patients lost to follow‐up were all due to the COVID‐19 lockdown. Table 1 summarises the patients’ characteristics. The continence rates were 27% (n = 22) at 3 months, 54% (n = 44) at 6 months, and 63% (n = 52) at 12 months. The median (interquartile range) ICIQ‐UI SF scores at 3, 6, and 12 months were 11 (4–10), 7 (3–11), and 4 (0–9), respectively.

Table 1.

Baseline characteristics.

Characteristic	Value
N (%)	82 (100.0)
Demographic, median (IQR)
Age, years	68 (62–72)
Body mass index, kg/m2	26 (24–28)
Comorbidity, n (%)
None	43 (52)
Hypertension, coronary artery disease	25 (31)
Other	14 (17)
ISUP Grade Group, n (%)
2	37 (45)
3	34 (42)
4	10 (12)
5	1 (1)
Clinical tumour stage (cT stage), n (%)
cT1	34 (42)
cT2	39 (48)
cT3	8 (10)
cTx	1 (1)
PSA level, ng/mL, median (IQR)	9.1 (5.8–15)

IQR, interquartile range; ISUP, International Society of Urological Pathology.

The mean preoperative mMUL was 1.58 cm and the mean preoperative sMUL was 4.8 slices. The mean MUD was 1.09 cm, and the mean (SD) prostate volume was 44 (16) mL. On the postoperative bpMRI scans the mean mMUL was 1.33 cm, the mean sMUL was 4.1 slices and the mean MUD was 1.02 cm. We found a significant change in mean mMUL from pre‐ to postoperative measurements of 2.6 mm (P < 0.01). There was also a significant change in the mean sMUL of 0.75 slices (P < 0.01).

On univariate analyses, statistically significant associations were found between mMUL and continence at 12 months (P = 0.03), prostate volume and continence at 12 months (P < 0.01), and sMUL and ICIQ‐UI SF score at 12 months (P = 0.05). The mMUL was not statistically significantly associated with the ICIQ‐UI SF scores at 12 months (P = 0.09), and neither was sMUL with continence at 12 months (P = 0.06). The univariate analyses of the association between the preoperative bpMRI measurements and continence as well as ICIQ‐UI SF scores at 3, 6, and 12 months are shown in Table S2. Of the baseline characteristics, age and PSA levels showed statistically significant associations with continence outcomes at 12 months on univariate analyses (data not shown).

Table 2 shows the results of the multivariable analyses of the association between preoperative bpMRI measurements and continence and ICIQ‐UI SF scores at 12 months. Greater preoperative mMUL and sMUL were predictors of continence at 12 months. For the mMUL, there was a RD of 0.13 per 5 mm of extra length (95% CI 0.01–0.24, P = 0.03), meaning an increase of 13 percentage points of being continent at 12 months per 5 mm. For sMUL, there was an RD of 0.06 per slice (95% CI 0.00–0.13, P = 0.05). Conversely, a larger prostate volume was associated with a decrease in continence rate, with an RD of −0.004 per 5 mL (95% CI −0.01 to −0.001, P = 0.02).

Table 2.

Multivariate regressions with robust variance estimations of identified prognostic preoperative measurements (mMUL, sMUL, MUD, and prostate volume) and characteristics (age and PSA level) and the association with continence and ICIQ‐UI SF score at 12 months.

Measurement and characteristics	RD	Lower 95% CI	Upper 95% CI	P
Continence
mMUL (5 mm)	0.13	0.01	0.24	0.03
sMUL (3 mm/slice)	0.06	<−0.001	0.13	0.05
MUD (5 mm)	0.05	−0.21	0.32	0.7
Prostate volume (5 mL)	−0.004	−0.01	−0.001	0.02
Age (years)	−0.02	−0.03	0.00	0.1
PSA	−0.01	−0.03	0.01	0.33
ICIQ‐UI SF	MD
mMUL (5 mm)	−1.04	−2.30	0.21	0.1
sMUL (3 mm/slice)	−0.84	−1.53	−0.16	0.02
MUD (5 mm)	−1.08	−3.88	1.72	0.44
Prostate volume (5 mL)	0.01	−0.02	0.04	0.49
Age (years)	0.12	−0.08	0.31	0.23
PSA	0.22	−0.001	0.44	0.05

The association of mMUL with continence is shown per 5 mm and the association of prostate volume with continence is shown per 5 mL. mMUL was not controlled for sMUL, as they measured the same structure. slice).

The only preoperative bpMRI measurement that was predictive of the ICIQ‐UI SF score at 12 months on multivariable analyses was the sMUL, with a MD of −0.84 points per slice (95% CI −1.53 to −0.16, P = 0.017).

The AUC on ROC‐curve analysis for preoperative mMUL predicting continence was 0.65. The optimal cut‐off using Youden's index was 1.39 cm balancing sensitivity (0.83) and specificity (0.47).

For preoperative sMUL predicting continence, the AUC on ROC‐curve analysis was 0.63. The optimal cut‐off was five slices using Youden's index balancing sensitivity (0.73) and specificity (0.53).

Table 3 summarises the uni‐ and multivariable analyses of the association between postoperative bpMRI measurements and continence and ICIQ‐UI SF score.

Table 3.

Uni‐ and multivariable regressions of postoperative bpMRI measurements and the association with continence and ICIQ‐UI SF score at 12 months.

Measurement and characteristics	Univariate				Multivariable
	Crude	Lower 95% CI	Upper 95% CI	P	Age, PSA	Lower 95% CI	Upper 95% CI	P
Continence	RD				RD
mMUL (5 mm)	0.24	0.12	0.36	<0.001	0.22	0.11	0.34	<0.001
sMUL (3 mm/slice)	0.13	0.05	0.21	0.001	0.11	0.03	0.19	0.005
Change in MUL (5 mm)	0.10	−0.05	0.25	0.18	0.08	−0.07	0.22	0.29
Change in sMUL (3 mm/slice)	0.07	−0.02	0.15	0.11	0.05	−0.03	0.14	0.24
ICIQ‐UI SF	MD				MD
mMUL (5 mm)	−2.46	−3.70	−1.23	<0.001	−2.09	−3.30	−0.88	0.001
sMUL (3 mm/slice)	−1.41	−2.12	−0.70	<0.001	−1.19	−1.91	−0.47	0.002
Change in mMUL (5 mm)	−1.32	−2.61	−0.03	0.05	−0.85	−2.14	0.45	0.20
Change in sMUL (3 mm/slice)	−0.65	−1.60	0.30	0.18	−0.25	−1.14	0.64	0.6

The association between mMUL and continence is shown per 5 mm. The changes in the length of mMUL and sMUL from pre‐ to postoperative scans are included. Change in mMUL and sMUL, the change from preoperative to postoperative measurements.

On multivariable analyses, greater postoperative mMUL and sMUL were associated with better continence and lower ICIQ‐UI SF score at 12 months. For mMUL, there was an RD of 0.22 (95% CI 0.11–0.34, P = 0.01) per 5 mm and the MD in the ICIQ‐UI SF score was −2.09 (95% CI −3.30 to −0.88, P = 0.01) per 5 mm. For the association between sMUL and continence, the RD was 0.11 (95% CI 0.03–0.19, P = 0.01) per slice and the MD in ICIQ‐UI SF score was −1.19 (95% CI −1.91 to −0.47, P = 0.01) per slice. The changes in measurements from pre‐ to postoperative scans were not associated with continence or the ICIQ‐UI SF score.

The inter‐rater reliability demonstrated strong agreement between readers for the pre‐ and postoperative mMUL measurements, underlined by MDs of 0.7 and 1.1 mm, respectively, and ICC values of 0.92 and 0.87, respectively. The pre‐ and postoperative sMUL measurements showed a more moderate but good agreement with a MD of 0.39 and 0.64 slices, and ICC values of 0.77 and 0.77, respectively. Finally, the pre‐ and postoperative mean MUD differed by −0.01 and 0.04 mm, respectively, and the ICC values were 0.63 and 0.78, respectively. The inter‐rater reliability for mMUL, sMUL, and MUD measurements is given in Table S3.

Discussion

In the present study, both greater preoperative mMUL and sMUL were predictors of continence after 12 months. The findings reaffirm what has been reported in the existing literature regarding long‐term continence. Thus, a systematic review and meta‐analysis from 2021 by van Dijk‐de Haan et al. [12] reported that a larger MUL was associated with the return of continence in 12 out of 15 studies. The pooled results demonstrated an overall positive association between a greater MUL and the return to continence at 12 months, with an odds ratio of 1.19 (95% CI 1.10–1.29). This has been backed in a recent meta‐analysis from 2023 by Mac Curtain et al. [9] who also found that a greater preoperative MUL was associated with a return of continence at 3 months. In contrast to much of the existing literature, we found no association between MUL and continence before 12 months. The lack of an association before 12 months in our study may be caused by the relatively low proportion of patients who were continent. We attribute this to our strict definition of continence and the fact that we studied patients who had undergone either UNS or NNS procedures, both of which are associated with delayed recovery of continence [23]. This assumption is supported by the individual trials included in the meta‐analyses by van Dijk‐de Haan et al. [12] looking at short‐term continence rates. The studies that found a greater MUL to be associated with continence at 3 and/or 6 months, all used less stringent definitions of continence by either using a simple no‐pad definition or even allowing a ‘safety pad’. Accordingly, the continence rates varied from 37% to 92% at 3 months and from 68% to 96% at 6 months. The point is further illustrated by examining the studies in the meta‐analyses that assessed the 12‐month association [12]. Here, one of the only studies that reported no association between MUL and continence used a definition of continence of <2 g loss on three consecutive 24‐h pad tests [24]. However, such a strict definition may underestimate actual continence rates as patients may perspire more than 2 g per day. We believe our definition reflects no‐leak continence of clinical relevance to patients as opposed to simply the use of zero pads or a safety pad alone [25]. In summary, the present study demonstrates that the MUL is associated with continence, even when applying a stricter definition of continence, than what is generally applied.

We also evaluated a novel method of assessing the MUL, the sMUL, as a possible alternative measurement to be used with ultra‐fast bpMRI scans using only the axial plane [18, 26]. The sMUL showed an association with continence in the present setting where the slice thickness was 3 mm and there was no slice gap. This suggests its potential as an alternative measurement. Yet, considering that each slice in the present study jumped by 3 mm, the few millimetres of MUL potentially overlooked while assessing MUL as the number of slices might be important for its predictive value. It tended to underestimate MUL compared to the mMUL (measured in the sagittal plane). In that context, a previous systematic review and meta‐analysis found that every extra millimetre of conventionally measured MUL increased the odds of a return to continence [8]. Based on these findings, further analyses in larger trials are warranted before potentially implementing the sMUL in clinical practice. For further validation of the sMUL measurement, it could be of value to correlate measurements to measurements obtained in both the coronal and sagittal plane.

When attempting to provide a useful cut‐off point of preoperative MUL to predict continence, we found a reasonable sensitivity but a low specificity, and as such they cannot be recommended for clinical use. Interestingly, while many previous studies have sought to find cut‐off values for the preoperative MUL in continent vs incontinent groups, they typically fail to disclose the AUC, accuracy, sensitivity, or specificity values. Specifically, no studies in the meta‐analysis reported sensitivity or specificity for their reported cut‐off value at 12 months. Overall, this implies a limited utility as an independent predictor in clinical practice and considering that the average MUL ranged from 7.4 to 17.3 mm across all studies it seems improbable that a universal cut‐off can be readily identified based on this single parameter [12]. Nonetheless, it may serve as part of predictive models. A recent prospective cohort study [27] offered patients a personalised UI risk assessment 6 months after RARP, incorporating MUL, inner levator distance, and NS. Patients provided with this information were more open to exploring alternative treatment options, especially when faced with a high risk of UI. In our study, a larger prostate volume was inversely associated and decreased with continence rate at 12 months, but the effect size was minimal, and MUD was not a significant predictor of urinary outcomes. This is similar to what was reported in prior studies [12, 16, 28, 29, 30].

Lastly, we also investigated the association between postoperative mMUL and sMUL, and continence and the ICIQ‐UI SF score at 12 months. The association between continence and both the mMUL and sMUL seemed to be stronger on postoperative scans. Furthermore, both postoperative mMUL and sMUL were associated with the ICIQ‐UI SF score. These add to the argument that effort should be given to retain membranous urethral length during RARP [31]. Meanwhile, in our study, the change in length was not associated with continence. One possible explanation is that the low rate of continence may have masked the association between the change in length and continence. It is also possible, that the remaining MUL may represent a more accurate functional length of the urethra after surgery. Alternatively, the postoperative measurement may simply be more challenging to obtain due to anatomical changes after surgery. The resulting less accurate or consistent measurements could potentially impact the strength of the association observed. In this context, we observed greater inter‐rater variability in the postoperative MUL measurements than in the preoperative.

Strengths and Limitations

The main strength of this study is that we addressed both pre‐ and postoperative bpMRI within the framework of a RCT. This secured a high degree of standardisation and allowed us to minimise the influence of confounding variables. Furthermore, the definition of continence used in our study might be more reflective of true continence than a no‐pad or safety‐pad definition. The main limitation of our study is the relatively small sample size providing limited ability to detect statistical associations. However, increasing this size would do little to improve the more clinically applicable parameters namely the predictive value of preoperative measurements on postoperative functional outcomes. Further limitations include that it was a single‐centre study, which potentially limits the generalisable of the findings.

Conclusion

Preoperative mMUL and sMUL are associated with continence after UNS or NNS RARP. However, we cannot recommend either as standalone predictors. The sMUL can be valuable when only the axial plane is available and the slice gap is known, but future studies are needed to confirm this. Postoperative measurements of MUL are strongly associated with continence. The MUD and prostate volume have limited clinical value as prognostic factors.

Funding

The study received funding from the government's health strategy for the focus area ‘Increased capacity for more minimally invasive surgical treatment’, and unrestricted Grants from Christian Larsen and Dommer Ellen Larsens Legat, Aage og Johanne Louis‐Hansens Fond, and Dagmar Marshalls Fond.

Disclosure of Interests

No authors have any conflicts of interest to disclose. However, author Lars Boesen is a speaker for Ipsen, and author Mikkel Fode is a speaker for Astellas and Boston Scientific.

Supporting information

Appendix S1. STROBE checklist.

Table S1. Th bpMRI sequence parameters.

Table S2. Regressions with robust variance estimations analyses of preoperative bpMRI measurements and the association with continence and ICIQ‐UI SF score controlled for the time at 3, 6, and 12 months.

Table S3. The inter‐rater variability for mMUL, sMUL, and MUD on preoperative scans, and for mMUL and sMUL on postoperative scans.