Three-Dimensional Reconstruction of Renovascular-Tumor Anatomy to Facilitate Zero-Ischemia Partial Nephrectomy

Transcription

1 EUROPEAN UROLOGY 61 (2012) available at journal homepage: Case Series of the Month Three-Dimensional Reconstruction of Renovascular-Tumor Anatomy to Facilitate Zero-Ischemia Partial Nephrectomy Osamu Ukimura *, Masahiko Nakamoto, Inderbir S. Gill Center for Image-Guided Surgery and Center for Advanced Robotic and Laparoscopic Surgery, USC Institute of Urology, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA Article info Article history: Accepted July 22, 2011 Published online ahead of print on September 15, 2011 Keywords: Kidney cancer Partial nephrectomy Laparoscopic surgery Robotic surgery Surgical navigation Three-dimensional image Abstract Zero-ischemia robotic and laparoscopic partial nephrectomy, a novel concept, eliminates ischemia to the tumor-free normal kidney. Anatomic microdissection of tertiary/ higher-order tumor-specific arteries is performed to selectively devascularize only the tumor, maintaining normal perfusion of the remaining kidney. A thorough understanding of renovascular tumor anatomy is essential. Based on 0.5-mm-slice thickness computed tomography scans, we developed a novel three-dimensional (3D) reconstruction technique that fuses three key anatomic aspects: surface-rendered tumor, semitransparent kidney, and extra- and intrarenal arterial anatomy. Four central completely intrarenal hilar masses underwent 3D reconstruction for surgical navigation during zero-ischemia partial nephrectomy. Negative surgical margins were obtained in all four cases, with no intraoperative complications or transfusions. For these challenging laparoscopically invisible masses, 3D image navigation precisely identified tumorspecific arterial branches, thus facilitating zero-ischemia partial nephrectomy without hilar cross clamping. # 2011 European Association of Urology. Published by Elsevier B.V. All rights reserved. * Corresponding author. USC Institute of Urology, University of Southern California, 1441 Eastlake Ave, Suite 7416, Los Angeles, CA 90089, USA. Tel ; Fax: address: ukimura@usc.edu (O. Ukimura). 1. Case report Four patients with a central intrarenal hilar small renal mass underwent three-dimensional (3D) reconstruction. These completely intraparenchymal renal masses were invisible to intraoperative laparoscopic view. Median patient age was 64 yr (range: yr), tumor size was 2.6 cm (range: cm), RENAL score was 9.5 (range: 8 10), C-index score was 1.5 (range: ), and PADUA score was 11 (range: 11 12). Preoperative computed tomography (CT) images were acquired on a 64 multidetector-row CT scan (GE Medical Systems, Milwaukee, WI, USA) with 0.5-mm step interval, including arterial phase and venous phase. Using our team s developed prototype computer software, we reconstructed 3D images to fuse three key anatomic aspects: 3D surfacerendered renal tumor, semitransparent kidney, and 3D course of extra- and intrarenal arteries (Fig. 1). Segmentation of kidney surface was performed using arterial phase images with automated thresholding and closing operation using Hounsfield units. The segmentation of renal tumor was performed manually using mainly arterial phase images with referring venous phases, jointly interpreted by an experienced urologist and a radiology technician. The 3D arteriogram was created using an arterial phase by region growing method with seed point on the main renal artery. Image-fusion techniques were used to create a combined display of /$ see back matter # 2011 European Association of Urology. Published by Elsevier B.V. All rights reserved. doi: /j.eururo

2 212 [(Fig._1)TD$FIG] EUROPEAN UROLOGY 61 (2012) Fig. 1 Schematic drawing of the process of creating three-dimensional (3D) reconstruction of renovascular-tumor anatomy. CT = computed tomography. [(Fig._2)TD$FIG] Fig. 2 (a) Case 1: A 57-yr-old woman with a right 3-cm intrarenal posteromedial central hilar mass. Three-dimensional (3D) renovascular and tumor reconstruction image fuses three key surgical anatomies, including 3D surface-rendered renal tumor (green), semitransparent kidney (pink), and 3D course of the extra- and intrarenal arteries (red). This allows the surgeon to see the interrelationships of the tumor vis-à-vis the intrarenal segmental arterial branches. (b) Correlation of intraoperative surgical field with 3D reconstructed image. Intraoperative robotic identification of arterial branches corresponded accurately with the reconstructed 3D renovascular and tumor model. The red vessel loop encircles a secondary renal artery branch (the main posterior branch); the blue vessel loop encircles the tertiary and quaternary branches. Atraumatic retraction of these vessels allows visualization of the tumor surface posterior to the tertiary branches. (c) Intraoperative color Doppler ultrasound. Left, Intraoperative color Doppler ultrasound confirms the abundant peritumoral vascularity that was also identified on 3D reconstructed image. These real-time ultrasound images were obtained intraoperatively, in preparation for superselective clamping. Right, After microbulldog clamping of the two tertiary vessels, the peritumoral vascularity is virtually eliminated while the remainder of the kidney continues to be perfused normally. Identification of these targeted vessels was based on the 3D reconstructed images.

3 [(_)TD$FIG][(_)TD$FIG] EUROPEAN UROLOGY 61 (2012) Fig. 2. (Continued ). the 3D arteriogram with 3D surface-rendered tumor and kidney models. During preoperative surgical planning, as well as intraoperatively, snapshots of 3D images from various angle views (eg, anterior, lateral, posterior, and laparoscopic) and 3D video clips were created and presented to the surgeon (Fig. 2 5) Renovascular and tumor three-dimensional navigation Renovascular and tumor 3D image navigation significantly facilitated performance of zero-ischemia robotic and laparoscopic partial nephrectomy [1]. Specifically, precise microdissection to identify the targeted tertiary or higher-order

4 214 [(Fig._3)TD$FIG] EUROPEAN UROLOGY 61 (2012) Fig. 3 Case 2: A 58-yr-old man with right 2.4 cm completely intrarenal central hilar mass. This tumor is supplied by tertiary renal arterial branches, which are not delineated by conventional computed tomography or easily identified by laparoscopic direct vision. Renovascular and tumor three-dimensional (3D) reconstruction provided the necessary 3D road map of the renal vasculature in relation to the tumor, which significantly aided the surgeon. renal arterial branches was aided (Fig. 2a and 2b). If necessary, guided by 3D reconstructed images, a short radial nephrotomy incision, 1 2 cm in length, was created on the medial concave hilar edge of the kidney to identify an intrarenal arterial branch directly supplying the tumor. This minimized unnecessary ischemia or sacrifice of healthy renal parenchyma. Intraoperative color Doppler ultrasonography, comparing pre- and postselective arterial control, confirmed specific devascularization of tumor and immediate peritumoral area only while maintaining uninterrupted arterial perfusion of the healthy renal parenchyma (Fig. 2c) Surgical outcomes Zero-ischemia partial nephrectomy was completed successfully without hilar cross clamping in any case (Table 1). Indication for nephron-sparing surgery was absolute in one patient (renal insufficiency and bilateral renal masses) and relative in the other three. Three cases (cases 1, 3, and 4) required pelvicaliceal system suture repair because the tumor was directly abutting the collecting system. Following arterial microdissection, a neurosurgical aneurysm micro bulldog (Bear Disposable Vascular Clamp, Aros Surgical, Newport Beach, CA, USA) was transiently applied superselectively on each tertiary or higher-order renal arterial branch suspected to be supplying the tumor. The main renal artery or its secondary branches were not occluded in any case. Median operating time was 286 min (range: min); estimated blood loss was 250 ml (range: ml). No intraoperative complications occurred. Surgical pathology reported clear cell renal cell carcinoma with negative margins in all cases. No significant deterioration in renal function was noted over a median follow-up of 3 mo (range: 1 6 mo). No postoperative complications were noted, except for heparin-induced thrombocytopenia in one case (Clavien grade 2). 2. Discussion An increasingly important issue during partial nephrectomy is undesirable ischemic injury to the renal remnant

5 [(Fig._4)TD$FIG] EUROPEAN UROLOGY 61 (2012) Fig. 4 Case 3: A 69-yr-old woman with a right 4-cm central, upper-midpole medial hilar mass abutting collecting system and hilar vessels. Renovascular and tumor three-dimensional reconstruction demonstrated the mass is supplied by not only by the upper renal branch but also by the intrarenal tertiary branch from the middle renal artery branch. Table 1 Demographics and perioperative and renal function data of the four patients Parameters Median (range) or No. (%) Demographics No. of patients 4 Age, yr, median (range) 64 (57 81) Gender: male; female, n 2; 2 Body mass index, kg/m 2, median (range) 25.7 ( ) American Society Anesthesiologists 3 (75); 1 (25) score 2; 3, n (%) PN indication, absolute; relative, n (%) 1 (25); 3 (75) Tumor characteristics CT tumor size, cm, median (range) 2.6 ( ) C-index, median (range) 1.5 ( ) RENAL score, median (range) 9.5 (8 10) PADUA score, median (range) 11 (11 12) Collecting system contact, n (%) 3 (75) Laterality right, n (%) 3 (75) Central, n (%) 3 (75) Completely intrarenal, n (%) 3 (75) Hilar, n (%) 3 (75) Perioperative Approach: Robotic; laparoscopic, n (%) 3 (75); 1 (25) Operative time, min, median (range) 286 ( ) Blood loss, ml, median (range) 250 ( ) Warm ischemia time, min, n (%) 0 (0) Table 1 (Continued ) Parameters Median (range) or No. (%) Pelvicaliceal system repair, n (%) 3 (75) Bolster used, n (%) 3 (75) Percentage kidney excised, median (range) 15 (5 50) Vascular microdissection performed, n (%) 4 (100) Number of micro-bulldogs deployed, 1 (1 3) median (range) Complications *, n (%) 1 (25) Hospital stay, d, median (range) 3 (3 12) Pathology Renal cell carcinoma, n (%) 4 (100) Margins negative, n (%) 4 (100) Stage T1a; T1b, n (%) 3 (75); 1 (25) Pathologic tumor size, cm, median (range) 2.7 (1.8 5) Renal function, median (range) Preoperative serum creatinine, mg/dl 0.9 ( ) Discharge serum creatinine, mg/dl 1.1 (0.6 2) Serum creatinine absolute change 0.2 ( 0.1 to 0.4) Preoperative egfr ml/min per 1.73 m ( ) Discharge egfr ml/min per 1.73 m ( ) egfr absolute change ml/min per 1.73 m 2 6( 18 to 22.6) CT = computed tomography; PN = partial nephrectomy; egfr = estimated glomerular filtration rate. * Postoperative heparin-induced thrombocytopenia (Clavien grade 2).

6 216 [(Fig._5)TD$FIG] EUROPEAN UROLOGY 61 (2012) Fig. 5 Case 4: An 81-yr-old woman with a left 2.1-cm central completely intrarenal hilar upper-mid pole mass substantially abutting the renal sinus and collecting system in a broad front manner. Renovascular and tumor three-dimensional reconstruction demonstrated the mass to be supplied by not only an anterior renal artery branch but also an intrarenal tertiary upper branch. [1 3]. Recent advancements in surgical technique now make it possible to eliminate global renal ischemia completely during partial nephrectomy. Zero-ischemia robotic and laparoscopic partial nephrectomy is based on the concept of anatomic renovascular microdissection of tertiary and higher-order renal arterial branches in relationship to the tumor margin and kidney surface [1]. Conventional 3D CT reconstruction techniques typically present all three structures (kidney, tumor, and renal vessels) as opaque. Thus, in the 3D reconstructed image, only the extrarenal vasculature is visualized in adequate anatomic detail; however, the opaque kidney makes it impossible to visualize intrarenal interrelationships of the tumor and adjacent inter-/intralobar arteries. As such, the surgeon s understanding of the relevant intrarenal anatomy during partial nephrectomy currently is wholly based on preoperative CT images combined with intraoperative laparoscopic visualization and twodimensional ultrasound. All these elements must then be mentally correlated by the surgeon in real time to create a working mental image of 3D tumor curvatures, angulations, and length/distribution of the relevant intrarenal arterial tree. Although this approach may be adequate for tumors that are at least partially exophytic (and therefore visible laparoscopically on the renal surface), this approach becomes imprecise for the completely intraparenchymal (and therefore laparoscopically invisible) tumor. Development of new radiologic guidance technology that addresses this issue would benefit the surgeon [4]. To our knowledge, our approach is novel in that it provides a 3D video representation of the opaque tumor and opaque extra- and intrarenal arterial tree in the setting of a semitransparent surface-rendered kidney. Because the kidney itself is rendered semitransparent, it becomes possible to visualize opaquely rendered tumor vasculature interrelationships in exquisite detail. This allows precise anatomy-based surgical planning for targeted vascular microdissection to facilitate zero-ischemia partial nephrectomy, even for intrarenal or central tumors. Without such 3D image guidance, high-precision surgical planning would likely be difficult.

7 EUROPEAN UROLOGY 61 (2012) Selective tumor-specific vascular control offers the considerable and previously unavailable benefit of eliminating global renal ischemic injury during partial nephrectomy [5], even for anatomically complex tumors. The four tumors reported here represent the highest level of technical difficulty, as indicated by their RENAL, C-index, and PADUA scores. The major aim of surgical image navigation is to demonstrate 3D anatomy beyond the laparoscopic surgical view, thus augmenting the reality of the surgical field with the added data derived from radiologic imaging [6 8]. Our navigation system nicely demonstrated the detailed anatomic relationships between the 3D margin of the renal tumor and the 3D course of extra- and intrarenal arteries, visualized through the reconstructed transparent kidney margin. Using our original software, we classified the intrarenal arterial branches by order (eg, secondary, tertiary) and identified specific feeding branches directly into the tumor. Actual intraoperative images (laparoscopic vision, color Doppler ultrasonography) corresponded accurately to the preoperatively reconstructed 3D images of branch arteries and tumor margins (Fig. 2b). Three-dimensional depiction of renal arterial branches provided a means to adapt microdissection strategies, define the appropriate location for radial nephrotomy incision, and more confidently interpret intraoperative laparoscopic and ultrasonic findings. A limitation of this segmentation technique is that it may not be readily adaptable to magnetic resonance imaging (MRI) scans. Our segmentation technique requires highly detailed thin-slice CT cuts of mm thickness. Because MRI scanning takes longer to perform than CT scanning, current MRI scans may not be able to provide such thin (<1 mm) multislice scanning during the patient s short breath-holding period. We present four patients with a completely intrarenal, laparoscopically invisible central hilar tumor undergoing laparoscopic or robotic zero-ischemia partial nephrectomy. Reconstructed 3D renal arteriogram, fused with 3D image of surface-rendered renal tumor and semitransparent kidney, facilitated selective microdissection of tumorspecific arterial branches during zero-ischemia partial nephrectomy. Conflicts of interest: The authors have nothing to disclose. EU-ACME question Please visit to answer the following EU-ACME question online (the EU- ACME credits will be attributed automatically). Question: When performing zero-ischemia partial nephrectomy for a central tumor, three-dimensional (3D) renovascular and tumor reconstruction based on computed tomography scan facilitates: A. Identification of the tumor feeding arterial branch. B. Understanding of the 3D course of the intrarenal arterial branches. C. Recognition of the anatomic interrelationship of tumor and arterial anatomy. D. All of the above. References [1] Gill IS, Eisenberg MS, Aron M, et al. Zero ischemia partial nephrectomy: novel laparoscopic and robotic technique. Eur Urol 2011;59: [2] Gill IS, Aron M, Gervais DA, Jewett MAS. Small renal mass. N Engl J Med 2010;362: [3] Thompson RH, Lane BR, Lohse CM, et al. Every minute counts when the renal hilum is clamped during partial nephrectomy. Eur Urol 2010;58: [4] Ukimura O, Gill IS. Imaging-assisted endoscopic surgery: Cleveland Clinic experience. J Endourol 2008;22: [5] Weld KJ, Bhayani SB, Belani J, et al. Extra-renal vascular anatomy of kidney: assessment of variations and their relevance to partial nephrectomy. Urology 2005;66: [6] Teber D, Guven S, Simpfendörfer T, et al. Augmented reality: a new tool to improve surgical accuracy during laparoscopic partial nephrectomy? Preliminary in vitro and in vivo results. Eur Urol 2009;56: [7] Su LM, Vagvolgyi BP, Agarwal R, Reiley CE, Taylor RH, Hager GD. Augmented reality during robot-assisted laparoscopic partial nephrectomy: toward real-time 3D-CT to stereoscopic video registration. Urology 2009;73: [8] Ukimura O, Gill IS. Image-fusion, augmented reality, and predictive surgical navigation. Urol Clin North Am 2009;36: