S. A. Hilton2, M. M. Hodges2, L. C. Dewberry2, D. M. Mirsky1,3, M. Handler1,4, C. C. Wilkinson1,4, H. L. Galan1,5, M. V. Zaretsky1,5, N. Behrendt1,5, A. I. Marwan1,2, L. Pyle6, J. L. Wagner7, J. G. Jacot7, C. Zgheib2, T. M. Crombleholme1,2, K. W. Liechty1,2 1Children’s Hospital Colorado,Colorado Institute For Fetal & Maternal Health,Aurora, CO, USA 2University Of Colorado Denver,Laboratory For Fetal And Regenerative Biology,Aurora, CO, USA 3University Of Colorado Denver,Department Of Radiology,Aurora, CO, USA 4University Of Colorado Denver,Department Of Neurosurgery,Aurora, CO, USA 5University Of Colorado Denver,Division Of Maternal Fetal Medicine,Aurora, CO, USA 6Children’s Hospital Colorado,Department Of Pediatrics,Aurora, CO, USA 7University Of Colorado Denver,Department Of Bioengineering,Aurora, CO, USA
Introduction: In utero repair of myelomeningcele (MMC) has become the gold standard to decrease rates of ventriculoperitoneal (VP) shunting in selected cases. It is thought that repair of the MMC in utero restores normal cerebrospinal fluid circulation by decreasing leakage from the MMC defect. In large MMC or myeloschisis defects, primary skin closure is not possible and an acellular human dermal matrix is used. The use of a patch requires additional time and is only determined at the time of the procedure. In addition, there has been no assessment of whether patch closure is as effective in reversing the hindbrain herniation (HBH) or decreasing the rate of VP shunts. We hypothesized that 3-Dimensional (3D) printing could be used to standardize the use of a patch as well as be predictive of the need for a patch prior to surgery. In addition, we hypothesized that the use of a patch closure would be associated with decreased rates of improvement in HBH and increased rates of shunt placement at 1 year compared to primary closure.
Methods: Retrospective review of patients who underwent fetal closure of MMC at a quaternary referral center for fetal surgery between 2013 and 2016. 3D models of the spinal defect were created and the size of the template correlated with the need for a patch. The degree of HBH was assessed on MRI preop and two weeks post-repair. The need for ventricular shunt placement at 1 year was assessed by phone survey. Univariate analyses were performed using Fisher’s exact test and multivariate analyses were performed using logistic regression.
Results: Of 41 patients, 24 (59%) were diagnosed preoperatively with MMC and 17 (41%) with myeloschisis. 58% of MMC and 53% of patients with myeloschisis required patch closure. 3D printed templates with total defect area >6cm2 accurately predicted over 90% of the cases that required a patch (ROC curve with area under the curve 0.8). In univariate analysis, primary closure was associated with higher rates of HBH improvement than patch closure (100% vs 48%, p<0.001), however, there was no significant difference in rate of shunt placement by closure type (38% vs 42%, p=NS). Consistent with the literature, only maximum ventricle size was associated with need for shunt placement on multivariate analysis (p=0.04).
Conclusion: This is the first direct comparison of outcomes following primary versus patch closure of MMC. Our findings indicate that primary closure of MMC and myeloschisis is associated with significantly higher rates of HBH reversal. The decreased rates of HBH improvement following patch closure suggests the need for either an improved surgical technique or an improved, impermeable biomaterial patch to better protect the exposed neural elements. We found that custom, 3D printed models of fetal MMC defects aid in the preoperative planning of fetal intervention and have the potential to reduce operative times, preserve tissue and function, and improve patient outcomes.