Three-dimensional digital reconstruction of renal model to guide preoperative planning of robot-assisted partial nephrectomy.

To maximize the functional advantage of PN, different selective clamping approaches have been proposed.1 However, conventional 2-D cross-sectional images are suboptimal to identify the exact intrarenal vascular anatomy and to predict the real tumor blood supply from segmental branches. Digital 3-D models of renal masses with the complete anatomical reconstruction of the kidney can be elaborated from conventional 2-D preoperative imaging2 and can be used as an additional tool to improve the understanding of the size, location, depth of a renal mass and vascular anatomy before PN.3–6 We aim to report our preliminary experience with preoperative use of 3-D digital reconstruction of renal models to guide the surgical approach during RAPN. We prospectively enrolled 20 consecutive patients with clinical diagnoses of renal mass, submitted to RAPN at a single tertiary institution. Before surgery, all patients underwent chest and abdominal contrast-enhanced CT. Furthermore, 3-D virtual reconstruction of an anatomical model for supporting surgical planning including the lesion, the renal parenchyma, the urinary collecting system and vascular anatomy (including extra- and intra-renal arterial and venous branches) was created by the Laboratory of Bioengineering of the University of Bologna in Bologna, Italy. For the creation of 3-D anatomical renal models, preoperative high-quality contrast-enhanced CT images were collected (slice thickness: 1.25/2.5 mm, step interval: 0.8/2.0 mm). Multiple imaging series with different contrast levels were used for the selective identification of each anatomical structure of interest. Segmentation was achieved using D2 software (DICOM to PRINT; 3D Systems, Rock Hill, SC, USA), to convert DICOM medical images into 3-D digital models. The segmented anatomical structures were then combined into one file using alignment of common regions. The obtained 3-D virtual model is navigable; that is, the surgeon can interact with it changing the transparency of each structure in the 3-D rendering. The preoperative surgical plan to define the level of arterial clamping (namely, no clamping, main artery, selective or super-selective) was recorded by the surgeon based on conventional imaging blinded to the 3-D model. The surgical plan regarding the level of arterial clamping was re-assessed after revision of the 3-D model immediately before intervention. The intraoperative effective clamping technique after revision of the 3-D virtual model was compared with the preoperative plan based on conventional imaging using the v2 -test. All statistical tests were carried out using SPSS 22.0 (IBM, Armonk, NY, USA). RAPN was carried out using the DaVinci Xi Surgical System (Intuitive Surgical, Sunnyvale, CA, USA) in a four-arm configuration with the integrated Firefly fluorescence-imaging mode, as previously described.7 One single experienced surgeon carried out all the RAPN cases. Overall, 16 (80%) and four (20%) of the tumors were clinical T1a and T1b stage, respectively. The median WIT was 10 min (IQR 7–12 min). No intraoperative and postoperative complications were observed. The median postoperative creatinine levels was 0.9 mg/dL (IQR 0.84–1.09 mg/dL) at 6 months’ follow up (Table 1). Based on conventional imaging, the preoperative plan of arterial clamping was: no clamping in four patients (20%), clamping of the main artery in 12 patients (60%) and selective clamping in four patients (20%). During surgery, the 3-D virtual model revision induced the surgeon to modify the clamping plan in eight patients (40%), towards a selective clamping approach, resulting in no cases of unclamped approach, clamping of the main artery in eight patients (40%) and selective clamping in 12 patients (60%; P = 0.01; Table 2). The preoperative plan based on 3-D reconstruction remained unchanged during surgery in 60% of patients, with intraoperative modification of the preoperative clamping plan in eight patients (40%) during RAPN (Video S1). Of note, due to better reconstruction of vascular anatomy, the use of the 3-D model influenced the surgeon to carry out selective clamping in most of the patients in our cohort. However, due to intraoperative findings of complex renal vasculature and adhesive fatty tissues in the hilum region, four patients planned for selective clamping were finally shifted to non-selective preoperative surgeon’s plan of arterial clamping in 40% ofpatients during RAPN toward a more selective clamping approach