Detection of crossing vessels in pediatric ureteropelvic junction obstruction: Clinical patterns and imaging findings

Background: Pediatric ureteropelvic junction obstruction (UPJO) is caused by congenital intrinsic narrowing and/or a lower pole-crossing renal artery. When a crossing renal vessel (CRV) is missed at the time of pyeloplasty, a redo-pyeloplasty is often required.

Objectives: The aims were to analyze clinical predictors for the presence of a CRV in UPJO and the utility of functional magnetic resonance urography (fMRU) in preoperative identification of a crossing vessel.

Methods: Using an Institutional Review Board approved registry database, we identified 166 patients from July of 2007 until January of 2014 who had undergone open, laparoscopic, or robotic assisted laparoscopic pyeloplasty at our institution. We abstracted data including age at surgery, preoperative symptoms, preoperative imaging findings, and whether or not a CRV was identified intraoperatively. Statistical analysis was performed on SPSS using the Mann-Whitney U test.

Results: Of the 166 patients identified, 78 were found to have a CRV at the time of surgery and 88 did not. The surgical approach was distributed as 104 robotic assisted laparoscopic, 51 open, and 11 pure laparoscopic. On univariate analysis, older age at presentation and pain at presentation predicted the presence of a CRV; antenatal hydronephrosis was a negative predictor, though 20 of 68 (25.6%) infants diagnosed with UPJO antenatally did have a CRV. Subgroup analysis of patients undergoing preoperative MRU showed a sensitivity of 88.2% and specificity of 91.7% for the detection of CRVs.

Discussion: This study confirmed the importance of looking for a crossing vessel in all cases, with the knowledge that increased age and pain at presentation were more likely to be associated with a crossing vessel. In addition, fMRU is a valuable source of information in the preoperative identification of the presence of a crossing vessel. The study has limitations including being retrospective in nature, and that the sensitivity of fMRU to identify CRVs was based on the read of an experienced uroradiologist who specializes in MRU, so may not correlate with the standard clinical read of an fMRU.

Conclusion: This study confirms the need to maintain a high index of suspicion for the presence of a CRV when intervening in a clinically symptomatic older child, although 25% of infants with antenatally detected UPJO did have one too. Our subset analysis demonstrated that MRU is a reliable method of detecting crossing vessels.