3-dimensional ultrasound in the evaluation of female pelvic organ prolapse of the posterior compartment

Studies in the USA suggest that approximately 1 in 9 women will require an operation for pelvic organ prolapse (POP) by the age of 80, and of significant concern is that up to 30% of these women will require re-operation. This implies that either our assessment or our treatment of POP is unsatisfactory. Comparison of therapeutic interventions in the research setting has been difficult due to a lack of standardisation of assessment. Most clinical methods to quantify POP have not been rigorously tested. In an attempt to standardise the assessment of pelvic organ descent a pelvic organ prolapse quantification system (POPQ) has been developed, which has been validated for reliability and reproducibility. However, only 40% of clinicians who have a special interest in urogynaecology actually utilise this system in routine clinical practice. In addition it has been difficult to quantify defects of the lateral vaginal support mechanisms. The place of two-dimensional ultrasound for the investigation and management of women with gynaecological problems is now well established. Both transabdominal and transvaginal scanning permit the identification of the uterus and adnexae with excellent resolution to more accurately direct appropriate management. However transabdominal ultrasonography provides poor imaging of the lower pelvic structures, due to the depth of the tissues from the transducer. Transvaginal ultrasound permits excellent visualisation of the bladder, urethra and the posterior compartment but distorts the pelvis thereby precluding accurate assessment of POP. Transperineal ultrasonography avoids these problems. Ultrasonography may therefore improve assessment for both clinical and research purposes by adding a further objective measuring system to vaginal examination. The addition of threedimensional ultrasonography provides the opportunity for visualisation of the axial plane of the pelvis (the third dimension), the ability to store volumes of data (data as opposed to images are stored) and therefore allow offline, independent audit by remote observers. The objective of this thesis was to study the use of 3–dimensional ultrasonography in the investigation of disorders of the posterior compartment of the vagina. The initial experiment was designed to obtain normal values for the incidence and measurement of rectocele prolapse and other posterior compartment defects of the rectovaginal septum, descent of the perineum of the pelvis, and measurements of the levator hiatus. After initial training in the technique and confirmation of the protocol, a test-retest series together with intraclass correlation test was undertaken in 12 women who were undergoing objective measurements for suburethral sling position, as part of a large, randomised sub-urethral sling outcome study and in whom it was clinically noted that they had at least a grade 2 (Baden-Walker grading system) rectocele prolapse. This first experiment confirmed the reliability of measurements with intraclass correlation (ICC) of between 0.68 and 0.79 between observation measurements and test-retest series of between 0.72 and 0.84. Mean depth of defects was 1.45cm, width 0.95cm with a mean perineal excursion of 0.46cm. Mean levator area dimensions increased from 15.6 cm² to 28.4 cm². The technique was therefore felt to be reliable enough for the following experiments. Ninety-seven nulliparous women then underwent the same standardised technique and methodology of ultrasound evaluation. Eighty-nine datasets were available for analysis. Results demonstrated a wide variation of dimensions for levator hiatal diameters and area at rest and on straining. Rectocele protrusions were seen in 11 women (12%), 10 women had protrusions less than 1cm in depth. Median depth was 0.69cm (SD 0.3), width 0.45cm (SD 0.3). There was no correlation with bowel symptoms (p=0.12 ANOVA). Levator area ranged from a mean 17.76 cm² at rest to 20.9 cm² on strain. Again the technique seemed reliable and confirmed in a nulliparous population the presence of congenital or acquired rectocele defects. In view of the reproducibility and reliability of the techniques a further experiment was undertaken to compare US against Magnetic Resonance Imaging (MRI). Eleven women with rectocele prolapse underwent 3D US evaluation and rapid sequence MRI scans using T2 image acquisition at rest and on straining. Intra-class correlation using dot-plots were analysed for all the measurements. Women underwent examination using the POPQ scoring system. Pre-op A(p) was 0.2cm (SD 1.2), B(p) was -0.3cm (SD 1.4), and overall clinical grading was 2. US was significantly better at diagnosing enterocele prolapse versus clinical (p=0.3) but less so MRI (p=0.11). There was no correlation between symptoms and rectocele parameters on either MRI nor US. Correlation between US and MRI for all levator dimensions and rectocele measurements was poor precluding formal analysis. ICC was also poor at 0.485 for levator area at rest and 0.16 for levator area on strain. This study found poor correlation between the two techniques, however small numbers were involved in a group of women with multiple compartmental prolapse. To test 3D US prospectively, 33 women with a diagnosis of rectocele awaiting surgery were recruited and prospective data gathered on ultrasound parameters, POPQ and symptoms. Twenty-nine datasets were available for analysis. Mean age was 59 years and average ordinal staging was 2 with point B(p) was 0. All women had fascial repair and composite mesh graft. Enterocele prolapse was again diagnosed more commonly on US than clinically (3 vs. 1). All symptoms improved with surgery. However again there was poor correlation between clinical measurements, ultrasound quantification and symptomatology in relation to successful treatment. Although there was a trend towards more accurate diagnosis of enterocele with ultrasound as defined by surgical staging this did not reach statistical significance. Why is there such a poor correlation between symptoms and both clinical and ultrasound findings. This may relate to the complex interaction between rectal function, neurological damage and prolapse of the vagina. It is proposed that larger studies be conducted to further identify the place of ultrasonography and assessment of POP in women and in particular the pre-operative factors that may influence outcome