More than 300 million women worldwide suffer from urinary incontinence and the prevalence is increasing with the demographic changes in the aging society [1,2]. There is a lack of noninvasive diagnostic methods as well as regenerative therapy options to identify and treat the underlying condition [3]. Current treatment of stress urinary incontinence (SUI) is either conservative (physiotherapy), medicinal or surgical—including bulking agents and implantation of slings or artificial sphincters [4]. All therapies are associated with limited effectiveness or relevant side effects over time and further do not treat the underlying pathophysiological causes of SUI [5]. The regenerative approach is an emerging field, focusing on the repair and replacement of human tissues and organs using stem cells [6,7]. This approach may, for instance, be focused on muscle precursor cells (MPCs) to regenerate and strengthen urinary sphincter function in female patients with SUI [8]. Still, an objective, noninvasive, and easy-to-use method for the measurement of sphincter function before and after regenerative therapy in patients with SUI is desired.

Among modalities used for imaging of lower urinary tract, including multiplanar 3-dimensinal (3D) ultrasound, magnetic resonance imaging (MRI) provides so far the most accurate and detailed assessment of complex pelvic floor anatomy [9]. Due to high spatial and contrast resolutions, as well its precision in distinguishing tissues, MRI has become an imaging modality of choice for suspected pelvic pathologies as well as in the preoperative planning of female patients [10]. Recently, T2-weighted turbo spin echo (TSE) and balanced steady-state free precision (TrueFISP) sequences were proposed for imaging of pelvic floor weakness [11]. Other techniques like diffusion tensor imaging (DTI), allow the assessment of water proton diffusion anisotropy within structured tissues [12]. Although DTI was firstly designed for neuroscientific applications [12-14], it was soon after acknowledged as a clinically applicable method [15] and described in skeletal muscles [16,17]. Feasibility of DTI was also shown for the visualization of the female pelvic floor [18,19] and recently also applied for assessment of superficial muscles of the human perineum [20] or course and directionality of levator ani fibers in females [21]. However, the presented approaches did not include fast assessment of the muscle function. From a physical point of view, biological structures, such as neuronal or muscle fibers, restrict the random (Brownian) motion of the water molecules less along their longitude than in other directions. The most commonly assessed indices are: mean diffusivity (MD) characterizing the average motion of water molecules disregarding the direction, size and integrity of the cell, and fractional anisotropy (FA) describing the anisotropy and thus the structural integrity of cellular structures within the fibers [14].

Hence, our study aimed to: (1) confirm the feasibility of rapid DTI in evaluation of pelvic floor muscle structures; (2) compare the MD and FA parameters, derived from the urinary sphincter and levator ani muscle regions; based on a rapid acquisition of isometric spatial resolution DTI during rest and active contraction of the lower female pelvis.

Demographic characteristics of the subjects are presented in Table 1. No concomitant disease or previous obstetric history was reported by any of the female volunteers. All scans were completed and participants did not report any significant discomfort during the MRI acquisitions. No significant motion or other artefacts influencing both qualitative and quantitative assessment of the acquired images were found among all the anatomical and diffusion sequences. A representative 3D DTI reconstruction of urethral sphincter fiber tracts during the ‘rest’ and ‘contraction’ phases for a single subject is shown in the Fig. 3.

In the current study, we confirmed the feasibility of DTI in the evaluation of pelvic floor muscles. To date, previous reports described application of DTI for evaluation of female pelvic floor muscles [18-20,22] as well as in the internal and external urinary sphincter muscles in healthy males [23]. However, to our knowledge, we are the first to investigate application of DTI acquired in 30-second frames for isometric resolution of 3×3×3 mm3. Furthermore, we assessed DTI metrics derived from the urinary sphincter and levator ani muscles during 2 contradictory states (rest and contraction) in 10 healthy female volunteers. By evaluation of 2 independent readers, we have found the average MD parameters changing between the rest and contraction in the urinary sphincter and levator ani muscles, when assessed jointly. No change in FA was, however, observed between the rest and contraction phases. It is also worth highlighting, that the major changes in MD were observed in the majority in levator ani muscle regions, by both readers. Overall, the average MD among the healthy female volunteers was significantly higher during both contractions, compared to both resting phases (ΔMD=20.6% increase). At the same time, no change was observed between rest-1 vs. rest-2 and contraction-1 vs. contraction-2 phases for both MD and FA, and a moderate to high correlation was found between the compared phases by 2 independent readers. The obtained results were supported by high ICC between a trained urology resident and a specialist in uroradiology (0.85 for MD and 0.79 for FA).

We believe that DTI might become a supportive diagnostic imaging tool for the evaluation of muscle training state, pelvic floor function and maybe even tracking of implanted MPCs. This is supported by a physical perspective, where DTI provides both quantitative and qualitative assessment of microstructure and physiological processes ongoing in the macroscopic muscle architecture [14]. In the muscle fibers, the diffusivity of water is larger in the direction parallel to dominant muscle fibers. Since MD expresses the average and FA the largest directionality of water molecules in random motion, the muscle architecture and function can be assessed indirectly [20,24,25]. As we observed an increase in MD during contraction of pelvic muscles, the effect might have been caused by higher uniformity of the muscle orientation compared to that during rest. Further, during contraction the length of sarcomeres is being reduced and muscle fiber diameter increases with the reduction in intermyocital space. This may lead to a relative increase in measurable MD [25]. Significant change in both MD and FA values was also found between passive shortening and elongation of human calf muscles [26]. However, the investigated volumes of the muscle structures were much larger and of different spatial structure compared to the urinary sphincter or levator ani muscle.

Thus, our findings are affected by a few limitations. First, the limited statistical power presented was connected with acquisition in only 10 young healthy volunteers. Hence, we applied a nonparametric approach along with correction for FDR due to both the log-normal character of the assessed data as well as the manual tissue segmentation applied. In the other studies, DTI of the female pelvic floor and the male urethra or the visualization of periprostatic nerval tracts were also described in 5–11 volunteers with an acquisition time of 4–18 minutes [18,20,22,23]. Additionally, alterations in proton diffusivity of muscle during extension and contraction were also demonstrated [25]. In further research, we aim to perform measurements in larger cohorts as well as to compare them to patients with functional pelvic disorders. Second, to shorten the acquisition during contraction phases, we have performed DTI in only 6 directions with 2 b-values (0 and 400 sec/mm²) measured with no increase in the number of averages. This provided a suboptimal signal-to-noise ratio for DTI measurements [27,28]. Although this prospectively resulted in residual bias during calculation of MD and FA parameters, we were able to acquire a single DTI image in a rapid time of 30 seconds only. It is worth highlighting, that this assured higher repeatability of the contraction performed by the volunteers along with the break periods between each DTI acquisition and stability of physiological response to the muscle action. The acquisition of higher resolution DTI with an increased number of averages, diffusion directions, or reduced acceleration factor would be desired for accuracy of the measurements as shown by other authors [18,21]. However, such protocol would not be useful in applications with focus on functional aspects due to the long scanning time—especially in patients with SUI. Third, the limitations of applied DTI preclude an isolated statement about the urinary sphincter. Since this is a small anatomical structure, the measurements of contractions rely on a few pixels only. Obviously, high-resolution DTI would bring more information, but its application is not feasible due to time constraints during contraction. Therefore, we would potentially be able to increase differentiation by reducing the noise influence with extended acquisition during the rest. The other possibility, was to choose a surrogate muscle structure continence, such as the levator ani, to receive reliability and reproducibility within a larger anatomical region. By applying the latter approach, we found MD and FA parameters to provide sufficient information for a fast clinical DTI assessment. On the contrary, potential findings in patients with SUI may provide lower differentiation between rest and contraction based on the evaluation in urethral sphincter solely. Last, MRI has its limitations in patients with claustrophobia. Therefore, results could be relevantly influenced (i.e., anxiety), altered (i.e., elevated muscle tonus) or the examination not be performed in some individuals. However, many comforting strategies have been implemented, such as the use of ear-plugs or headphones with soothing music, soft blankets, convenient positioning, adjusting of ventilator, shortening of acquisition time, and vocal contact to the examiners with the possibility of ending the investigation via emergency button at any given moment.

In conclusion, our findings demonstrate that rapid DTI acquisition may serve as a supportive noninvasive tool for the assessment of urinary sphincter and levator ani muscle integrity and contractility in healthy female volunteers. Despite the short acquisition in isometric 3×3×3-mm³ resolution, our DTI protocol allows differentiating between the rest and contraction of the pelvic floor muscles. Rapid DTI-based strategy is promising, may improve fast assessment of the urinary sphincter function, and bring additional value for the diagnosis as well as the follow-up of patients suffering from SUI.