Search for the optimized and key nephrometry elements combination in retroperitoneal laparoscopic partial nephrectomy: A retrospective study

BACKGROUND: The nephrometry scoring system plays a key role in the preoperative evaluation of partial nephrectomy, and scoring systems based on anatomical characteristics have high similarity in scoring elements. Currently, there is little research on scoring systems related to retroperitoneal laparoscopic partial nephrectomy, and there is a lack of research on the combination of scoring elements, which requires further investigation. METHODS: We retrospectively analyzed the clinical records of 107 patients who underwent retroperitoneal laparoscopic partial nephrectomy conducted by a single operator at a single center. The score and scoring elements were generated based on imaging. The scoring elements of each scoring system and all combinations of two to five elements were extracted. The predictive ability of different score combinations was evaluated by AUC value, and the key parameters of the score were found by taking the intersection. A nomogram was constructed and evaluated. RESULTS: We observed that with an increase in scoring elements, the strongest combination of elements did not significantly increase the predictive ability of warm ischemia time (P>0.05), postoperative complications (P>0.05), and trifecta achievement (P>0.05). The combination of the maximum tumor diameter and the distance between tumor and collecting system or renal sinus had a good comprehensive predictive ability, and there is no significant difference with the traditional score (P>0.05). The nomogram generated according to this combination has an excellent prediction ability for predicting whether obtain trifecta of partial nephrectomy. CONCLUSIONS: Within the range of two to five elements, the critical degree of elements is more important than the number of elements. The maximum tumor diameter and the distance between the tumor and the collecting system or renal sinus was the key element of the prediction ability.