Interpreting nephrometry scores with three-dimensional virtual modelling for better planning of robotic partial nephrectomy and predicting complications.

OBJECTIVE

3D models are increasingly used as additional preoperative tools for renal surgery. We aim to evaluate the impact of 3D renal models in the assessment of PADUA, RENAL, Contact Surface Area (CSA) and Arterial Based Complexity (ABC) for the prediction of complications after Robot assisted Partial Nephrectomy (RAPN).

METHODS AND MATERIALS

Overall, 57 patients with T1 and 1 patient with T2 renal mass referred to RAPN, were prospectively enrolled. 3D virtual modelling was obtained from 2D computed tomography (CT). Two radiologists recorded PADUA, RENAL, CSA and ABC by evaluation of 2D images; two bioengineers recorded PADUA, RENAL CSA and ABC by evaluation of the 3D model, using MeshMixer software. To evaluate the concordance between 2D and 3D nephrometry scores, Cohen's j coefficient was calculated. Receiver-operating characteristic (ROC) curves were generated to evaluate the accuracy of 3D and 2D nephrometry scores to predict overall complications. Finally, the impact of 3D model on clamping approach during RAPN was compared to 2D imaging.

RESULTS

PADUA, RENAL CSA and ABC scores had a significant different distribution compared to PADUA, RENAL, CSA and ABC (all p≤0.03). 2D nephrometry scores may be unchanged, reduced or increased after assessment by 3D models: CSA, PADUA, RENAL and ABC were reduced in14%, 26%, 29% and 16% and increased in 16%, 36%, 38% and 29% of cases, respectively. At ROC curve analysis, PADUA, RENAL and ABC showed were significantly better accuracy to predict complications compared to PADUA, RENAL and ABC. PADUA (OR: 1.66), RENAL (OR: 1.69) and ABC (OR: 2.44) revealed a significant correlation with postoperative complications (all P ≤0.03).

CONCLUSION

Nephrometry scores calculated via 3D models predict complications after RAPN with higher accuracy than conventional 2D imaging.