RESUMO
OBJECTIVE: To evaluate predictive radiological elements for adherent perinephric fat (APF) and the Mayo adhesive probability (MAP) score in the setting of open partial nephrectomy, and to assess their reproducibility. PATIENTS AND METHODS: We performed a retrospective case-control study involving 86 patients who had open partial nephrectomies performed by a single surgeon between January 1, 2009 and April 1, 2015. Radiological elements were assessed by 4 readers blinded to patient APF status. Univariate and multivariate analyses were performed for all radiological and clinical factors. Reproducibility was analyzed using agreement coefficients. RESULTS: On univariate analysis for radiological findings, lateral and posterior fat thickness (odds ratio [OR]: 1.084 [1.033, 1.138], P < .001), stranding (OR: 2.179 [1.431, 3.318], P < .001), -80 HU fat area, and the MAP score (OR: 1.797 [1.332, 2.424], P < .001) were predictive of APF. On multivariate analysis, only age and the MAP score remained statistically significant (OR: 1.060 [1.005, 1.118], P = .03; and OR: 1.560 [1.137, 2.139], P = .0058, respectively). The reproducibility of the MAP score was fair (AC1 = 0.367 and kappa F = 0.353), as was that of stranding (AC1 = 0.499, kappa F = 0.376). The agreement was important if we defined a "low" (0 to 3) or "high" (4 or 5) score (AC1 = 0.706 and kappa F = 0.681). CONCLUSION: The MAP score was the element that best predicted APF in our study, although its reproducibility among our readers was only fair. The agreement becomes important if we defined "low or high" score.