Urinary Bladder vs Gastrointestinal Tissue: A Comparative Study of Their Biomechanical Properties for Urinary Tract Reconstruction

Accepted Manuscript Title: Urinary Bladder Versus Gastrointestinal Tissue: a Comparative Study of Their Biomechanical Properties for Urinary Tract Reconstruction Author: Davis N.F., Mulvihill J.J.E., Mulay S., Cunnane E.M., Bolton D.M., Walsh M.T. PII: DOI: Reference:

S0090-4295(17)31229-3 https://doi.org/10.1016/j.urology.2017.11.028 URL 20780

To appear in:

Urology

Received date: Accepted date:

13-8-2017 18-11-2017

Please cite this article as: Davis N.F., Mulvihill J.J.E., Mulay S., Cunnane E.M., Bolton D.M., Walsh M.T., Urinary Bladder Versus Gastrointestinal Tissue: a Comparative Study of Their Biomechanical Properties for Urinary Tract Reconstruction, Urology (2017), https://doi.org/10.1016/j.urology.2017.11.028. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Urinary bladder versus gastrointestinal tissue: A comparative study of their biomechanical properties for urinary tract reconstruction Davis NF1*, Mulvihill JJE2*, Mulay, S2, Cunnane EM2, Bolton DM1, Walsh MT2 1. Department of Urology, The Austin Hospital, Melbourne, Australia 2. School of Engineering, Bernal Institute and the Health Research Institute, University of Limerick, Limerick, Ireland

*Joint first authorship Correspondence: Dr Niall Davis Department of Urology The Austin Hospital Melbourne 3068 Victoria Australia Phone: 03 9496 5000 E-mail: [email protected]

1 Page 1 of 23

Keywords: Reconstructive urology Urinary bladder Mechanical properties Augmentation cystoplasty Neobladder

Abstract word count: 233 Manuscript word count: 2490 Number of references: 15 Number of figures: 3 Number of tables: 1

Acknowledgements: None

Abstract Objective To evaluate the mechanical properties of gastrointestinal (GI) tissue segments and to compare them with the urinary bladder for urinary tract reconstruction.

Methods Urinary bladders and GI tissue segments were sourced from porcine models (n=6, 7 months old [5 male; 1 female]). Uniaxial planar tension tests were performed on bladder tissue and Cauchy stress-stretch ratio responses were compared with stomach, jejunum, ileum and colonic GI tissue.

2 Page 2 of 23

Results The biomechanical properties of the bladder differed significantly to jejunum, ileum and colonic GI tissue. Young’s modulus (kPa - measure of stiffness) of the GI tissue segments was on average 3.07-fold (± 0.21 standard error) higher than bladder tissue (p<0.01), and the strain at Cauchy stress of 50 kPa for bladder tissues was on average 2.27-fold (± 0.20) higher than GI tissues. There were no significant differences between the averaged stretch ratio and Young’s modulus of the horizontal and vertical directions of bladder tissue (315.05 ± 49.64 kPa and 283.62 ± 57.04, respectively, p=0.42). However, stomach tissues were 1.09 (± 0.17) and 0.85 (± 0.03) fold greater than bladder tissues for Young’s Modulus and strain at 50 kPa, respectively.

Conclusion An ideal urinary bladder replacement biomaterial should demonstrate mechanical equivalence to native tissue. Our findings demonstrate that GI tissue does not meet these mechanical requirements. Knowledge on the biomechanical properties of bladder and GI tissue may improve development opportunities for more suitable urological reconstructive biomaterials.

1.0

Introduction

Surgical repair for end stage bladder disease has utilized vascularized, autologous, mucussecreting gastrointestinal (GI) tissue to either replace the diseased organ or to augment autologous bladder tissue1. A second approach has focused on the development of tissue engineered tissues and organs. Methods to evaluate these tissues remains a challenge because the functional organ must maintain specific physiologic, immunologic and biochemical

3 Page 3 of 23

properties to protect the functional urinary tract system, most specifically renal function and urinary continence.

Preservation of renal function and maintenance of low-intravesical pressure in response to increases in intravesical volume are the primary goals when reconfiguring the bowel into a spherical storage reservoir. Although GI tissue is frequently applied in reconstructive urology; there are no studies that accurately characterise and compare its biomechanical properties. This study aims to characterise the baseline mechanical properties of bladder, gastric, jejunal, ileal and colonic tissue segments in order to identify the segment that most suitably conforms to the baseline biomechanical properties of the urinary bladder.

4 Page 4 of 23

2.0

Methods

2.1

Overview of experimental design

Urinary bladders and GI tissue segments were sourced from 6 porcine models, all 7 months old (5 male; 1 female) from a local abattoir (Fitzgerald’s, Ballylanders, Limerick) immediately after euthanasia. All other materials were obtained from CABER (Centre for Applied Biomedical Engineering Research, Limerick, Ireland) unless indicated. Ethical approval for ex vivo tissue characterization was approved by the University of Limerick’s ethics approval process. Porcine tissue was stored in phosphate buffer solution (PBS) at 4°C and sample extraction was carried out within 24 hours of euthanasia. Tissue samples were isolated from the urinary bladder, stomach, jejunum, ileum, cecum, ascending colon, transverse colon, and descending colon. The primary endpoint of the study was to evaluate the mechanical properties of the urinary bladder and to compare them with each GI tissue segment. The evaluated mechanical property was the level of ‘strain’, or ‘distension’, that tissues undergo at a certain stress (50 kPa). The stress values equate to a pressure that the bladder tissue undergoes when filling to capacity. Strain is defined as stretch ratio subtracted by one, this helps understand the amount of distension in the tissue (for example in figure 3 where a value of 1.6 stretch ratio equals 0.6 strain, [i.e. the tissues expands by 60% of its original length]).

2.2

Mechanical testing

2.2.1

Measurement of tissue samples

Two perpendicular samples of tissue were sectioned from each anatomical location to evaluate anisotropy, a measure that describes the directional dependence of mechanical behavior. All tissue strips were matched for width and length into approximately 20 x 30 mm segments (width x length). Thickness was measured with Vernier calipers at three locations

5 Page 5 of 23

throughout each section and validated using noncontact photography (mean thickness: bladder: 2.53 ± 0.11 mm, colonic: 1.40 ± 0.05 mm, stomach: 2.75 ± 0.19 mm standard error)2. Tubular tissue samples (i.e. small and large bowel) were investigated along their axial and circumferential axes. Spherical tissue samples (i.e. urinary bladder) were investigated along their horizontal (cranial to caudal, C-C) and vertical (for bladder: dorsal to ventral, DV, for stomach: medial to lateral; M-L) axes. Tissue samples were clamped at each end in a tensile-tester which resulted in testing dimensions of 20 x 10 mm (width x length).

2.2.2

Mechanical testing

Samples were mechanically characterized sequentially under, separate, circumferential (vertical) and axial (horizontal) extension in a PBS bath maintained at 37°C. A dedicated device designed for the extension of biological soft tissues (CellScale, Canada) facilitated measurement of regional mechanical properties for the urinary bladder and GI tissue segments2. Gauge length did not exceed 10 mm in any sample (bladder: 6.53 ± 0.15 mm, colonic: 6.06 ± 0.09 mm, stomach: 5.23 ± 0.15 mm standard error; range: 4.73 - 9.63 mm). Width of the tissue specimens was 18.56 ± 0.47 mm, 17.93 ± 0.35 mm, and 18.8 ± 0.56 mm for bladder, colonic, and stomach tissues, respectively. Samples were therefore in planar tension as the width to length ratio was greater than 2:13. Planar tension testing is advantageous compared to pure tension testing as it provides an enlarged contact area between clamps and the tissue which reduces slippage4.

2.2.2 Measurement of mechanical properties Samples were oriented in their circumferential or axial direction and the edges were placed in clamps that were lined with foam-backed sandpaper. A torque of 7.5 cN.m was applied to the clamps through a single centrally positioned bolt5. The clamped samples were equilibrated to

6 Page 6 of 23

37°C in a PBS bath for 3 minutes prior to testing. The elastic mechanical properties of the samples were characterized by stretching samples to a preload of 0.1 N and then subjecting them to 10 cycles of extension to a stretch ratio equal to 3 times the sample gauge length at a displacement rate of 45 mm/min which equates to a rate of 15 % of the sample gauge length per second. This rate was chosen as it was shown to minimize viscous phenomena (i.e. reduce hysteresis) and to prevent the dynamic effects that arise at higher rates (unloading stress exceeding corresponding loading stress) during preliminary testing6. The initial 9 cycles were performed to minimize the strain softening effect displayed by soft biological tissues7. The final cycle was employed to characterize the elastic response of the tissue to extension. The extension stretch limit of 3 was chosen as this was the limit of stretch prior to sample slippage during preliminary testing. The absence of sample slippage was confirmed by the absence of slack in the samples following testing. Young’s modulus, a measure of stiffness of solid linear elastic materials, was used to compare the various tissue types and the anisotropy of the tissue. As most biological tissue is hyper elastic rather than linear elastic, the Young’s modulus was calculated over the stiff linear region in the high stretch domain of each tissue sample (see insert in figure 1)8.

2.3

Statistical analysis

Data are represented as a mean ± standard error of mean (SEM). The normality of all variables was examined using Shapiro-Wilk tests. Significant differences were identified between groups of continuous variables using Student’s t-test for normally distributed data. A one-way anova test with Bonferroni correction was used to compare the averaged mechanical properties of the different groups. A p-value (alpha value) of less than 0.05 was considered statistically significant.

7 Page 7 of 23

3.0

Results

3.1

Bladder tissue

The two plots in figure 1 represent the Cauchy stress-stretch ratio curves of all bladder tissue samples tested (3 technical replicates from 6 pigs, black dots) along the horizontal (a) and vertical (b) axes of the spherical bladder. The red dash line in each plot represents the Cauchy stress–stretch ratio curves constructed from the averaged data of all 18 bladder samples, for both directions. There was no significant difference between the averaged Young’s modulus (kPa – measure of stiffness, slope of the line) of the horizontal and vertical directions over the stiff linear region of the curve (315.05 ± 49.64 kPa and 283.62 ± 57.04, respectively, p=0.42). There was also no significant difference between the averaged stretch ratio values of both directions at a Cauchy stress value of 50 kPa, P>0.05. A Cauchy stress of 50 kPa was chosen as the comparison point as it is the highest stress value attained by most tissue segments tested in this study. Similarly, no significant differences were found in the geometric measurements, tables 2.

3.2

Gastrointestinal tissue segments

Figure 2 illustrates the Cauchy stress-stretch ratio curves of gastric, small and large bowel tissue segments (a-h) in perpendicular directions (axial/C-C – grey, circumferential/M-L – black). Gastric tissue layers display relatively low stiffness (a-b). Colonic tissues display a considerably stiffer response in comparison to bladder tissue (c-h; low stretch ratio and high Cauchy stress). Furthermore, the colonic and stomach tissues demonstrate apparent differences in mechanical response between the axial and circumferential directions, except for the jejunum, tables 1 and 2. This anisotropy is further characterised by comparing the ratio of axial/C-C and circumferential/M-L Young’s modulus over the stiff linear region of the curve, table 1.

8 Page 8 of 23

3.3

Bladder tissue versus gastrointestinal tissue segments

Figure 3 demonstrates that the strain of the GI tissues at a Cauchy stress of 50 kPa is on average 2.27-fold (± 0.20 standard error) higher than bladder tissue in both directions. However, there is a similar strain between stomach and bladder tissues (0.85 fold ± 0.03). The tissues that display a strain significantly different to native bladder tissue are denoted with an asterisk. As can be seen, all tissues characterised in this study, except for the inner and outer layer of the stomach, display a significantly different strain under a similar stress value when compared to the native bladder tissue. Furthermore, although no significant anisotropy was identified between the directional responses of the GI tissue (P>0.05), examining the ratio of axial/C-C and circumferential/M-L Young’s modulus demonstrates that the GI tissue exhibits a higher degree of anisotropy relative to the bladder tissue (with the exception of the descending colon), table 1.

9 Page 9 of 23

4.0

Discussion

An ideal bladder substitute graft should accommodate large changes in intravesical volume while maintaining almost constant pressure values. From a biomechanical perspective, a replacement biomaterial should restore normal function to the bladder to preserve the upper urinary

tracts.

In

the

1950s,

non-biodegradable

synthetic

materials

such

as

polytetrafluoroethylene (PTFE), silicone, rubber, polyvinyl, and polypropylene were investigated as bladder replacement grafts but were found to be mechanically unsuitable. Additionally, these materials were susceptible to encrustation, bacterial colonization and foreign body reactions

9, 10

. Consequently, autologous GI tissue has remained the gold

standard replacement biomaterial for the urinary bladder since initially described 100 years ago. Advantages associated with GI tissue are an intact vascular network and widespread intracorporal availability. However, our main finding is that the mechanical properties of GI tissue segments are severely limited relative to the compliant nature of the urinary bladder.

Although technology in urology has developed rapidly over the last two-decades, the development of an ideal urological biomaterial that can reliably augment or replace the urinary bladder remains challenging due to the highly elastic and unique biomechanical properties of the organ. To our knowledge, this is the first study to comparatively investigate the urinary bladder and GI tissue segments from a biomechanical perspective. Our mechanical findings were consistently reproducible and may have important clinical implications as they emphasize the persistent clinical challenge faced when constructing mechanically equivalent biomaterials. Knowledge on the biomechanical properties of normal bladder tissue and ‘gold-standard’ GI segments may improve development opportunities for more suitable urological reconstructive biomaterials in the future.

10 Page 10 of 23

Recently, there has been extensive interest in developing tissue-engineered urological scaffolds derived from biodegradable synthetic meshes and decellularized extracellular matrices (ECMs). ECMs are decellularized, biocompatible, biodegradable biomaterials usually derived from porcine organs11, 12. They are prepared by mechanical, chemical and enzymatic treatments to yield tissue that is minimally immunogenic and retains its basic structural elements13.

A recent phase 2 clinical trial utilized autologous cell seeded polyglycolide/polylactide (PGA/PLA) biodegradable scaffolds for augmentation cystoplasty in patients with spina bifida (n=10)14. Notably, there was no improvement in bladder capacity on urodynamics after 1-year or 3years with 5 patients requiring re-operation in the form of a conventional ileocystoplasty14. Limitations with mechanical durability and poor compliance in vivo need to be addressed before tissue-engineered biomaterials can reliably replace autologous GI tissue in reconstructive urology.

Although the long-term metabolic adverse effects of GI tissue segments are well described and should be considered; our results provide new information on their mechanical limitations in a urological environment. Ileal tissue is the most frequently utilised bowel segment for reconstructing the urinary bladder; however, it is considerably stiffer compared to stomach, transverse colon and descending colon tissue, figure 3. Table 1 lists the Cauchy stress value for each tissue at a stretch ratio of 1.6 (i.e. 60% of stretch, or 0.6 strain) and the strain at Cauchy stress of 50 kPa offers further insight into the mechanical response of the various tissues. GI tissues are found to have lower values of strain at 50 kPa, and the converse trend for Cauchy stress values at 1.6 stretch ratio (or 0.6 strain). This trend

11 Page 11 of 23

highlights the innate distensibility of the urinary bladder as it fills and the pronounced difference in its mechanical response relative to GI tissues. Although gastric tissue is associated with significant metabolic side-effects, our mechanical findings indicate that it conveys ideal mechanical properties for bladder replacement purposes.

Our secondary findings also highlight that most GI tissue segments demonstrate increased anisotropy over the stiff linear region relative to the urinary bladder, albeit not statistically significant. The apparent anisotropic behaviour of GI tissue noted herein has not previously been characterised until the present study where we demonstrate that GI tissue segments generally exhibit a higher Young’s modulus when extended in the axial direction relative to the circumferential direction. This finding may have important clinical implications as it suggests that GI tissue should be preferentially oriented along its circumferential axis when applied to reconstruct the urinary bladder as this orientation is likely to accommodate larger volume increases at lower pressure values relative to axially orientated GI tissues. Other important factors such as blood supply, three-dimensional spherical characteristics and neuromechanical properties should also be considered in this clinical setting.

. However, porcine tissue samples were initially selected because their biomechanical properties closely resemble those of human tissue11, 12. It is also arguable that mechanical properties may have been altered due to the ex vivo nature of the experimentation. However, we have previously demonstrated that ex vivo organ tissue characterisation techniques can compromise cellular properties while mechanical properties remain largely unaffected7. A further limitation of this study is that testing is conducted on 2-dimensional samples rather than the more anatomically accurate 3-dimension spherical bladder

12 Page 12 of 23

configuration. However, in a previous study by our group we evaluated the 2-dimensional mechanical properties of tissue-engineered ECM scaffolds and assessed their impact on 3dimensional spherical bladder capacity and compliance after partial cystectomy in ovine models. Our results demonstrated that bladder capacity decreases by approximately 40% and compliance values (ΔV/ΔP) also deteriorate by approximately 40% with mechanically limited ECM scaffolds in place of distensible bladder tissue. Therefore, a deterioration in capacity and compliance values is likely to occur when the 2-dimensional mechanical properties of GI tissue are translated into the 3-dimensional bladder structure (15).

5.0

Conclusion

This study characterizes and compares the mechanical properties of the urinary bladder with GI tissue for bladder substitution purposes. An ideal urinary bladder replacement biomaterial should demonstrate mechanical equivalence to the native tissue. Our findings demonstrate that autologous GI tissue does not meet these requirements. The continued development of tissue-engineered urological substitutes may ultimately replace GI tissue segments in the future if mechanical equivalence to the urinary bladder can be reliably demonstrated.

6.0

Conflicts of interest: None

7.0

References

1.

Flood HD, Malhotra SJ, O’Connell HE, Ritchey MJ, Bloom DA, McGuire EJ. Longterm results and complications using augmentation cystoplasty in reconstructive urology. Neurourol Urodyn. 1995;14:297–309.

2.

O’Leary SA, Doyle BJ, McGloughlin TM. Comparison of methods used to measure the thickness of soft tissues and their influence on the evaluation of tensile stress. J

13 Page 13 of 23

Biomech. 2013;46:1955–60. 3.

Mulvihill JJ, Walsh MT. On the mechanical behaviour of carotid artery plaques: The influence of curve-fitting experimental data on numerical model results. Biomech Model Mechanobiol. 2013;12:975–85.

4.

Hollenstein M, Ehret AE, Itskov M, Mazza E. A novel experimental procedure based on pure shear testing of dermatome-cut samples applied to porcine skin. Biomech Model Mechanobiol. 2011;10:651–61.

5.

Mulvihill JJ, Cunnane EM, McHugh SM, Kavanagh EG, Walsh SR, Walsh MT. Mechanical, biological and structural characterization of in vitro ruptured human carotid plaque tissue. Acta Biomater. 2013;9:9027–35.

6.

Natali AN, Carniel EL, Frigo A, Pavan PG, Todros S, Pachera P, et al. Experimental investigation of the biomechanics of urethral tissues and structures. Exp Physiol. 2016;101:641–56.

7.

Walsh MT, Cunnane EM, Mulvihill JJ, Akyildiz AC, Gijsen FJH, Holzapfel GA. Uniaxial tensile testing approaches for characterisation of atherosclerotic plaques. J Biomech. 2014;47:793–804.

8.

Gasser TC, Ogden RW, Holzapfel GA. Hyperelastic modelling of arterial layers with distributed collagen fibre orientations. J R Soc Interface. 2006;3:15–35.

9.

Kudish HG. The use of polyvinyl sponge for experimental cystoplasty. J Urol. 1957 Sep;78:232-5.

10.

Agishi T, Nakazono M, Kiraly RJ, Picha G, Nose Y. Biodegradable material for bladder reconstruction. J Biomed Mater Res. 1975 Jul;9:119-31.

11.

Matoka DJ, Cheng EY. Tissue engineering in urology. Can Urol Assoc J. 2009 Oct;3:403-8.

12.

Crapo PM, Gilbert TW, Badylak SF. An overview of tissue and whole organ

14 Page 14 of 23

decellularization processes. Biomaterials. 2011; 12: 3233–43. 13.

Gilbert TW, Sellaro TL, Badylak SF. Decellularization of tissues and organs. Biomaterials. 2006; 27: 3675–83.

14.

Joseph DB, Borer JG, De Filippo RE, Hodges SJ, McLorie GA. Autologous cell seeded biodegradable scaffold for augmentation cystoplasty: Phase II study in children and adolescents with spina bifida. J Urol. 2014;191:1389–94.

15.

Davis NF, Callanan A, McGuire BB, Mooney R, Flood HD, McGloughlin TM. Porcine extracellular matrix scaffolds in reconstructive urology: An ex vivo comparative study of their biomechanical properties. J Mech Behav Biomed Mater. 2011 Apr;4(3):375-82.

15 Page 15 of 23

Figure legends Figure 1: Cauchy stress-stretch ratio response curves of porcine bladder tissue (black dash line, n=6, 3 technical replicates) in two perpendicular directions (a: cranial to caudal; C-C, b: dorsal to ventral; D-V). Legend: Red dash line represents the averaged Cauchy stress/stretch ratio response of the bladder tissue. Insert illustrates the region (line denoted as ‘Stiff’) over which Young’s modulus was calculated from the Cauchy stress/stretch ratio response plot.

Figure 2: Cauchy stress-stretch ratio response curves of porcine stomach layers (a-b, n=3 replicates) and colonic tissues (c-h, n=6 replicates) in two perpendicular directions (axial/C-C – grey dots, circumferential/M-L – black dots) harvested from porcine gastrointestinal systems.

Figure 3: Averaged strain at 50 kPa of bladder, stomach and colonic tissues of porcine GI system (n=6, 3 technical replicates). Legend: Error bars indicate standard error of the mean, and asterisk (*) denotes statistical significance (one-way anova, P≤0.05)

16 Page 16 of 23

Table 1: Comparative assessment of uniaxial tensile testing results for bladder tissue and gastrointestinal tissue segments Parameter Bladde Bladde s r r (± C-C D-V Standard Error) Average Cauchy Stress 1.04 1.36 (kPa) at a (0.17) (0.38) Stretch Ratio of 1.6 Average 1.68 1.59 Strain at (0.09) (0.08) 50 kPa Average Young’s modulus 315.05 283.62 over the (49.64) (57.09) stiff linear region (kPa) Average 18.70 18.44 Width (0.90) (0.68) (mm) Average 2.59 2.47 Thickness (0.17) (0.15) (mm) Average Gauge 6.51 6.54 Length (0.27) (0.2) (mm)

Ileum Axial

Ileum Circ

Jejunu m Axial

Jejunu m Circ

Cecum Axial

Cecu m Circ

Asc. Colon Axial

Asc. Colon Circ

Trans Trans Colon Colon Axial Circ

Desc Colon Axial

Desc Colon Circ

117.93 (39.78)

71.14 (16.79 )

20.81 (7.41)

47.42 (15.68)

6.77 (3.66)

7.39 (4.39)

11.23 (6.91)

7.02 (3.14)

19.27 (9.06)

21.54 (4.32)

6.49 (4.25)

8.41 (5.44)

0.57 (0.06)

0.58 (0.04)

0.72 (0.08)

0.65 (0.07)

0.72 (0.08)

0.76 (0.04)

0.86 (0.07

0.81 (0.07)

0.70 (0.06)

0.66 (0.01)

0.91 (0.09)

0.86 (0.06)

789.51 (161.4 0)

520.4 1 (74.41 )

607.90 (90.64)

458.94 (96.31)

606.86 (117.1 0)

409.7 8 (72.02 )

476.3 2 (67.92 )

394.42 (115.4 0)

394.4 346.0 2 7 (66.21 (68.36 ) )

323.54 (106.3 0)

321.1 5 (89.45 )

18.91 (1.42)

18.82 (0.70)

19.64 (0.84)

18.49 (1.05)

16.85 (1.27)

18.16 (1.41)

17.22 (1.32)

17.90 (1.56)

18.07 (0.48)

17.47 (0.79)

15.51 (1.58)

15.32 (2.14)

1.39 (0.16)

1.5 (0.28)

1.22 (0.13)

1.28 (0.08)

1.14 (0.08)

1.62 (0.16)

1.33 (0.12)

1.23 (0.09)

1.63 (0.21)

1.55 (0.11)

1.5 (0.16)

1.64 (0.13)

6.15 (0.43)

6.78 (0.50)

5.68 (0.17)

5.72 (0.20)

6.86 (0.86)

6.52 (0.36)

5.75 (0.19)

6.46 (0.62)

5.77 (0.34)

5.67 (0.34)

6.06 (0.07)

6.15 (0.04)

17 Page 17 of 23

Average Width to Length Ratio

2.96 (0.20)

2.89 (0.16)

3.11 (0.25)

2.86 (0.22)

3.47 (0.15)

3.23 (0.13)

2.59 (0.30)

2.87 (0.32)

3.01 (0.26)

2.86 (0.32)

3.18 (0.18)

3.11 (0.17)

2.32 (0.17)

2.15 (0.09)

18 Page 18 of 23

Fig 1_bestsetConverted.png

19 Page 19 of 23

20 Page 20 of 23

Fig 2_bestsetConverted.png

21 Page 21 of 23

22 Page 22 of 23

Fig 3_bestsetConverted.png

23 Page 23 of 23