Hiatal and Level III Structural Changes After Native Tissue Repair on MRI

Objectives: A persistently enlarged hiatus after prolapse surgery is associated with operative failure. Posterior repair aims to surgically restore the urogenital hiatus, however, little is known regarding the effects of surgery on level III support. This pilot study seeks to develop a strategy to quantify hiatal and level III structural changes on 3D rest and stress MRI after native tissue surgery including posterior repair, and to assess effect sizes for planning a larger study.

Methods: Ten women who underwent prolapse surgery had 3D rest and stress MRI pre-operatively and 3 months postoperatively. The following level III measurements were made with 3D-Slicer software using 3D-curve tool: the urogenital (UGH) and levator (LH) hiatus (area and A-P dimension); curved length of the pubovisceral (PVM), puborectal (PRM) and iliococcygeal (ICM) muscles (Fig. 1). Paired t-test was used to compare surgical changes and effect sizes (Cohen_d) were calculated (considering >=.8 as large, >=0.3 as medium and < 0.3 as small).

Results: Figure 2A shows an example of postoperative UGH and LH changes. There is a noticeable reduction of UGH_area and UGH_A-P (with lesser degree), while LH_area only had minor changes, mainly laterally. UGH_area at rest was smaller after surgery (from 23.5±4.0 to 21.0±4.6 cm2, average reduction of 2.5±3.0 cm2, p=.026), with a large effect size of 0.84 (Figure 2B). We found medium effect sizes for UGH_area at strain, LH_area at rest and UGH_A-P at rest. In particular, we found average reduction of 3.4±7.8 cm2 (d =0.43, p=0.2) in UGH_area at strain from 28.1±4.1 to 24.8±5.3 cm2, although not statistically significant in this small sample due to large variance. The reduction of UGH_A-P showed medium effect size, smaller than UGH_area, as did LH_area at rest (d=0.3). Changes in LH_A-P had small effect sizes, and changes in PRM, PVM and ICM length had minimal effect sizes.

Conclusions: This pilot study demonstrates the feasibility of comparing multiple level III parameters at rest and strain and the effect sizes of pre- to post-operative change caused by native tissue repair. In this small sample there is a significant UGH_area reduction after surgery at rest and a moderate effect size for this at strain. Additional parameters with moderate effect sizes were UGH_A-P and LH_area at rest. The reduction in UGH_area comes without a change in muscle length indicating that other factors are responsible for this change.