Role for intravesical prostatic protrusion in lower urinary tract symptom: a fluid structural interaction analysis study

Background Numerous studies indicated that Intravesical prostatic protrusion is relevant to prognosis of LUTS, however, the confounding effect that is brought about by prostate volume, urethra anterior curvature angle and other factors makes it hard to evaluate the role of intravesical prostatic protrusion in clinical observation. Methods We proposed a fluid structural interaction analysis approach. 3D models were constructed based on MRI images, and prostatic urethra diameters were calibrated with urodynamic data. Comparisons of urine flow dynamics were made between models with various degree of intravesical prostatic protrusion, while the intravesical pressure, anterior urethra curvature angle and diameter of prostatic urethra were same among all models to rule out their confounding effects. Results Simulation result showed that the decrement of diameter and increment of variation in cross-sectional area for prostatic urethra were related to the degree of intravesical prostatic protrusion. Such deformation would lead to deterioration of flow efficiency and could compromise the effect of bladder outlet obstruction alleviation treatment. Conclusions These results provided further evidence for intravesical prostatic protrusion being an independent risk factor for bladder outlet obstruction severity and demonstrated that intravesical prostatic protrusion would be a promising marker in clinical decision making. Electronic supplementary material The online version of this article (doi:10.1186/s12894-015-0081-y) contains supplementary material, which is available to authorized users.


Background
Intravesical prostatic protrusion (IPP) is the extent to which the prostate protrudes into the bladder, defined as distance from protruded prostate to the base of bladder, and can be measured in midline sagittal plane of the prostate [1]. Population based data indicated that 10 % of male between 40 to 79 years old had an IPP of 10 mm or greater [2]. IPP is considered as a prognostic factor for LUTS [3,4]. And the fact that IPP can be evaluated with non-invasive trans-abdominal ultra-sound made it a promising candidate for initial assessment of LUTS patient [5]. But the mechanism underlying the relationship between IPP and bladder outlet obstruction is still unclear. One key issue the confounding effect caused by prostate volume variation and urethra curvature angle.
Because they are both risk factors for LUTS severity and are closely related with IPP, it is difficult to control these confounding factors with observational study. Computational modeling on the other hand, is a promising alternate, and would shed a light on understanding the role for IPP in bladder outlet obstruction.
Hydraulic energy is the driven force in voiding process. It is lost due to resistance of urethra. Accurate reconstruction of anatomical feature for lower urinary tract is crucial for calculation of hydraulic energy loss. Computational fluid dynamic (CFD) study was proved to be advantageous in such aspect [6][7][8]. However, rigid wall boundary assumption in previous studies ruled out the interaction between urine flow and urethra wall movement, especially prostatic urethra wall. To overcome this limitation and investigate the role for IPP in bladder outlet obstruction, we carried out a fluid structural interaction analysis in models reconstructed from MRI data with various degree of IPP, then compared the difference in flow efficiency among these models.

The model and boundary conditions
A retrospective revision of the clinical data for all patients, presenting with lower urinary tract symptoms secondary to benign prostate hyperplasia (LUTS/BPH), who also completed MRI scan of pelvic region and pressure flow study before surgery, in the time period from January 2000 to December 2014 was carried out. Diagnosis of LUTS/BPH was established if criteria of the 5th International Consensus Committee on BPH [9] was met. The data from MRI scanning (Discovery MR750, GE Healthcare) are needed for model reconstruction, and the pressure flow study are needed for calibration of arbitrary determined parameters of the model. Patients with a history of neurogenic bladder, previous pelvic surgery or urinary cancer were excluded, detrusor insufficiency was also ruled out. Ten male patients were included for the study after providing informed consent. Approval for the study was granted by the ethics committee of Second Affiliated Hospital of Guangzhou Medical University.
Organ contouring for prostate, bladder and surrounding connective tissue was done in Mimics (Materialise, Leuven Belgium) by studying axial T2 MRI images of each patient. This was conducted by one senior urologist and confirmed by another radiologist. Degree of IPP was measured in mid-sagittal plane. First a line was drawn from the anterior to posterior intersections of the bladder base and prostate, then the distance between the protruded prostate to this line is defined as IPP (Fig. 1a, f), and categorized as grade I(<5 mm), grade II(5~10 mm) and grade III(>10 mm) [10]. Then three-dimensional models were constructed from contouring region of each slice (Fig. 1e), and optimized ( Fig. 1f) with SolidWorks (DS Solidworks, Massachusetts, USA).
Since prostatic urethra can't be clearly identified in MRI images [11], model was reconstructed with arbitrary parameters. The urethra model was divided into three parts. The proximal part (Fig. 2a) was a translational zone from bladder to urethra, 10 mm in length with a diameter decreasing from the width of normal bladder neck(8 mm) to prostatic urethra width [6]. The distal part (Fig. 2a) was another 10 mm long translational zone between distal prostatic and anterior urethra. The urethra in-between started at bladder neck, curved anteriorly at veru montanum and ended near at the prostatic apex [11]. These key points were marked specifically during organ contouring to define a fitting spline that would be closest to the prostatic urethra course ( Fig. 1a-d). Then this part of urethra was modeled as a cylindrical structure running along the spline.
The diameter for anterior urethra(d urethra ) was set to 5 and 4 mm for meatus, corresponding to the average cross-sectional area of around 20 mm 2 for anterior urethra [12,13]. Diameter for prostatic urethra during voiding process was a parameter paramount for accurate simulation. The value should be between 1 mm and the diameter of anterior urethra [14]. Five candidate diameters (1, 1.5, 2, 3 and 4 mm) were proposed for initial diameter of prostatic urethra (d prostate ). Result of Abrams-Griffiths nomogram in our pilot study showed that these 5 candidate diameter covers the pressure-flow relationship from obstructed to unobstructed scenario (Additional file 1: Figure S1). This coincided with previous result [15]. In this way, we could preserve the anatomic feature of anterior urethra curvature, and find the optimum diameter which would represent the obstruction level in prostatic urethra at the same time.
Fluid structural interaction analysis in ANSYS (ANSYS, Inc. Canonsburg, USA) was employed to study the deformation of prostate and its influence on urine flow. Boundary conditions were configured as follows ( Fig. 1f ): (1) Superior wall of bladder was set as inlet with preset total pressure; (2) Meatus of the urethra was set as outlet with 1 atm (101,325 Pa) static pressure; (3) Fluid-structural interface region included the bladder wall which lay over the protruded prostate, the surrounding connective tissue, the bladder neck and prostatic urethra; (4) Fixed support was added to the lateral wall of prostate and the surrounding connective tissue, representing the supportive structure around the prostate, along with fascia and pelvic supportive structures. The properties of the fluid were set as water. Pilot studies indicated that for all models in our simulation, the reynolds number, defined as Re = ρvd/η (ρ is the density of fluid, η is the dynamic viscosity of fluid, v and d are the velocity and hydraulic diameter of flow, respectively), range from 7000 to 13,000 (Additional file 2: Table S1), which were greater than 4000. So κ-ε turbulence model was used for CFD analysis. For the structural analysis, prostate was assumed to be linear elastic [16] (Poisson ratio:0.4, Young's modulus: 21kPa). The same assumption was applied to connective tissue (Poisson ratio:0.4, Young's modulus:15kPa).

Simulation planning
By adopting five candidate diameter of prostatic urethra to all ten patients, 50 candidate models were reconstructed. Intravesical pressure (P ves ) measured at maximal urine flow rate (V m ) was used as total pressure for inlet, then flow rate (V c ) was calculated with fluid structural analysis model. Bladder volume was set to be more than 200 ml in all models. Calculated flow rate (V c ) and measured flow rate (V m ) were compared to select the optimal d prostate .
One model from a patient with grade 2 IPP was selected by random, and marked as model 2. Another two models with different grades of IPP were created by adjusting the shape of the protruded part. These two models were marked as model 1 (grade 1 IPP model) and model 3 (grade 3 IPP model). Shape of the bladder neck, prostatic urethra diameter and anterior urethra curvature angle were same among three models (Additional file 3: Figure S2). Same boundary conditions were applied to all models. For each model, Fluid structural interaction(FSI) analysis were conducted under three different inlet pressure(7840.8, 17081.1, 13721.4 Pa), results of prostatic deformation and flow efficiency were compared among three models. Then, to simulate the obstruction alleviated scenario, initial diameter of prostatic urethra was increased to 2 and 3 mm, respectively. Then urine flow

Results
The demographic data of ten patients was listed in Table 1. Pressure flow data (Additional file 4: Table S2) showed that mean maximal urine flow rate(Q max ) was 9.5 ml (7~12 ml), detrusor pressure at Q max (P det,Qmax ) was 56.7 cmH2O (42~70 cmH2O), voided volume was 188.5 ml (153~245 ml). This result indicated that all patients can be categorized as obstructed according to Blaivas' criteria [17]. For each patient, a model was constructed using MRI images, and urine flow rate was calculated. The deviation between calculated and measured flow rate in each candidate model were charted in a 3D mesh. Minimal deviation was acquired when d prostate = 1.5 mm, and maximal deviation was found when d prostate = 3 mm (Fig. 2b). Initial diameter of prostatic urethra for model 1, model 2 and model 3 were set to be 1.5 mm.
During voiding, we found that deformation of the prostate would lead to urethra constriction (Fig. 3a-c), and its magnitude increased with intravesical pressure. The constriction of urethra was most prominent near bladder neck, then loosened gradually and returned to its initial diameter at prostatic apex. Under each intravesical pressure, urethra diameter of different models was compared. Such comparison leads to an interesting discovery. In region near bladder neck, the widest urethra was found in model 1 (grade 1 IPP model), followed by model 2 and model 3(grade 3 IPP model), but this order was reversed in distal prostatic urethra. Such pattern of constriction indicated that the variation of cross sectional area for urethra was most prominent in model 3, and lowest in model 1 (Fig. 3d-f ).
Total pressure, defined as P ¼ ρ 2 v 2 þ p, was a combination of static pressure and dynamic pressure, often used to evaluate flow efficiency. As urine flow runs from bladder to urethra meatus, total pressure decreased due to urethra resistance. In our simulation, such pressure loss was most prominent in model 3 (Fig. 4a-c), most of which occurred in the constricted urethra near bladder neck. In the other two models, less total pressure was loss around the bladder neck. For these two models, the majority of total pressure loss took place in distal prostatic urethra (Fig. 4d-f ).
Vorticity, defined as curl of velocity [18]: was used to study the pattern of flow energy dissipation. Consistent with pattern in total pressure loss, highest magnitude of vorticity for model 3 was found in bladder neck region while the highest vorticity magnitude for model 1 located in veru montanum and prostatic apex (Fig. 4g). Flow velocity reached its peak as urine run through bladder neck, and then it decreased gradually (Fig. 5a-c). Histograms for flow velocity distribution in sagittal plane were compared among models (Fig. 5d-f ). For each model, two peaks were found, corresponding to flow velocity in prostatic urethra and anterior urethra, respectively. Comparison among models indicated that flow velocity in majority part of model 1 was higher than the other two models.
Urine flow rate at the urethra meatus in model 1 was greater than the other two models, and increasing  intravesical pressure would further widen this gap (Fig. 5g). Then, to simulate obstruction alleviated scenario, prostatic urethra diameter was increased to 2 and 3 mm for the three models. Urine flow rate was calculated in the diameter increased models under the pressure of 80 cmH2O (7840.8 Pa). We found that urine flow rate in all models increased as initial urethra diameter widen, but at different rates. The gap in flow rate between model 1 and model 3 went up to 4 ml/s when the diameter was 2 mm. Then it went down to 1 ml/s when diameter was 3 mm (Fig. 5h).

Discussion
In patient with LUTS symptom, the confounding effect between prostate volumes, anterior urethra curvature angle, detrusor muscle contractility and intravesical prostatic protrusion (IPP) make it hard to understand IPP's role with clinical observation. Invasive methods such as pressure flow study are generally not applicable to routine clinical practice. Computational fluid dynamic(CFD) was already proved to be an effective method in investigating flow dynamic in studies regarding airway flow [19], circulation [20] and urine transport [21]. It is an attractive alternative to elucidate the role for IPP in LUTS manifestation.
This is the first study to investigate voiding behavior of lower urinary tract with fluid structural interaction analysis (FSI) and provide a scope for better understanding of prostate deformation. Our results showed that intravesical pressure above 7840.8 Pa was enough to cause prominent deformation of the prostate in model 3, which would lead to constriction of prostatic urethra. Clinical studies showed that maximum intravesical pressure in LUTS patient is usually between 8820.9 to 14701.5 Pa [22,23]. So it was obvious that for LUTS patient with severe intravesical prostatic protrusion, intravesical pressure during voiding would cause the prostatic deformation which would lead to severe constriction of prostatic urethra.
Structural analysis showed that the most severe constriction in bladder neck and greatest variation of cross-sectional area for urethra were found in model 3. The anatomy feature of fascia surrounding the prostate might shed a light on this. Prostate was attached anteriorly by pubo-prostatic ligaments, laterally by endopelvicfascia, and posteriorly by Denonvilliers' fascia [24]. As these supportive structures go superiorly, they fuse with other fascia, leaving the protruded portion of the prostate susceptible to radial component of intravesical pressure. And this may be the underlying cause for the difference in deformation between three models. The variation for urethral cross-sectional area in model 3 was the main reason for its flow energy dissipation [25]. Such variation was most prominent near the bladder neck, which coincided with the distribution of pressure loss and vorticity magnitude. As degree of intravesical protrusion decreased, location for major pressure loss shifted towards distal part of prostatic urethra, and the amount of total pressure drop also decreased. Since factors other than intravesical prostatic protrusion were same among three models, our results indicated that intravesical prostatic protrusion could affect the flow efficiency independently.