European Urology

1. Introduction

Laparoscopic partial nephrectomy (LPN) is gaining popularity as a minimally invasive nephron-sparing treatment for selective renal tumors [1], [2], and [3]. It is now well recognized that patients experience less severe chronic kidney disease and fewer cardiovascular events in the long term after nephron-sparing surgery compared to radical nephrectomy [4] and [5]. Warm ischemic (WI) injury during PN is one of the important issues influencing short- and long-term renal function, and some novel techniques have emerged to decrease this injury, such as segmental renal artery clamping [6], zero ischemia [7], zero ischemia with vascular microdissection technique [8] and [9], and a nonclamping technique [10]. Segmental renal artery clamping is a promising method and provides improved functional outcomes by decreasing WI injury [6]. However, in our previous study, segmental renal artery clamping was not very precise, and missed or inappropriate clamping might occur during the procedure [6]. Thus, a more refined surgical technique and higher quality of radiologic imaging are needed to obtain more satisfactory clinical outcomes. We present our experience using the precise segmental clamping technique under the guidance of dual-source computed tomography (DSCT) angiography in the past 2 yr.

2. Patients and methods

From December 2009 to November 2011, a total of 125 patients with renal tumors underwent LPN using the precise clamping technique. All operations were performed by the same laparoscopic surgeon (C.Y.). All patients had unilateral kidney tumor with diameters ranging from 1.4 to 7.0 cm; the contralateral kidney was normal. A glomerular filtration rate (GFR) study was done to evaluate the split renal function by renal scintigraphy before and 3 min after the operation. Inclusion criteria for LPN with precise segmental artery clamping included all patients with renal tumor deemed to be candidates for PN.

All the patients underwent computed tomography (CT) and CT angiography examination preoperatively. We applied DSCT with 1.2-mm slice thickness for the three-dimensional (3D) reconstructive angiography. Radiologic data were then collected and analyzed. The information included tumor size, position (polar, anterior, posterior, striding), growth pattern (exophytic, mesophytic, endophytic), and renal arterial segmentation and its anatomic relation with the tumor. We defined striding tumor as the lesion located on both anterior and posterior sides of the kidney. We defined exophytic tumor as the lesion extending >60% from the natural surface of the kidney, endophytic tumor as the lesion extending <40% from the kidney surface, and mesophytic tumor as a lesion extending between 40% and 60% from the surface of the natural border of the kidney [11]. Based on this information, we could precisely orient the segmental arteries feeding the tumor. The target of our precise clamping was where the feeding segmental artery had no further branches before entering the parenchyma. So the presegmental artery, the common stem of two or more branches, was not our superselective target. According to the position of origin from the intra- or extrahilar stem, the branches to be clamped were classified into three types: extrahilar, intrahilar, and mixed (Fig. 1).

Fig. 1 Three anatomic variants of target branches: (A) extrahilum type, (B) intrahilum type, and (C) mixed type. Top row: sketches of the anatomic variants. Bottom row: correspondent computed tomography angiography. Orange arrow points to the target branch.

All the procedures were performed using a retroperitoneal approach. The first port for the laparoscope was made 2 cm above the iliac crest. A 12-mm port was made below the tip of twelfth rib, and the other two 5-mm ports were placed at the anterior auxiliary line. We slightly modified port positioning for anterior tumors or when the segmental arteries were to be clamped from the anterior hilum. Briefly, all four ports were moved 2–3 cm toward the midline in those situations to facilitate manipulation from the anterior side of the kidney. After entering the retroperitoneal space, the perirenal fat surrounding the tumor was removed. The main renal artery was not routinely isolated. For intrahilar branches, dissection of segmental arteries was close to the parenchyma from the anterior or posterior hilum (Fig 2, Fig 3, and Fig 4).

Fig. 2 Dissection approach from the posterior hilum. A left-kidney, 3.2-cm-diameter, endophytic tumor located on the posterior side and fed by a single branch. Left top: cross-section of computed tomography (CT) scan. Left bottom: dissection approach from the posterior side. Right top: dual-source CT (DSCT) angiography with three-dimensional reconstruction. Right bottom: DSCT angiography overview from posterior side. Orange arrow points to the target branch, yellow arrow points to the presegmental artery/stem of target branch.

Fig. 3 Dissection approach from the anterior hilum. A right-kidney, 3.8-cm-diameter, mesophytic tumor located on the anterior side and fed by two branches. Left top: cross-section of computed tomography (CT) scan. Left bottom: dissection approach from the anterior hilum. Right top: dual-source CT (DSCT) angiography with three-dimensional reconstruction. Right bottom: DSCT angiography overview from the anterior side. Orange arrow points to the target branch, yellow arrow points to the presegmental artery/stem of target branch.

Fig. 4 Dissection approach from both anterior and posterior hila. A left-kidney, 5.5-cm-diameter, mesophytic tumor striding both anterior and posterior sides and fed by two branches. Left top: cross-section of computed tomography (CT) scan. Left middle: dissection approach from the posterior hilum. Left bottom: dissection approach from the anterior hilum. Right top: Dual-source CT (DSCT) angiography with three-dimensional reconstruction. Right middle: DSCT angiography overview from posterior side. Right bottom: DSCT angiography overview from the anterior side. Orange arrow points to the target branch, yellow arrow points to the presegmental artery/stem of target branch.

After exposing the target branches manifested by DSCT angiography, we introduced bulldog clamps and clamped the feeding segmental arteries. The tumor could be controlled by single- or multiple-branch clamping. The parenchyma was incised and the tumor was excised closely around its capsule with margin of 1- to 2-mm normal parenchyma (Fig. 5). Last, the renorrhaphy was performed using our previous technique [6].

Fig. 5 Excision of the tumor, maximally preserving the normal parenchyma. Left: the tumor was excised closely around its pseudocapsule, leaving a 1- to 2-mm margin of normal parenchyma. Right: The artery feeding normal parenchyma was peeled off the pseudocapsule to preserve the blood supply to adjacent parenchyma.

Surgical complications were classified according to Clavien's method [12]. Complication data were collected according to Martin's criteria [13]. All patients underwent GFR examination 3 min postoperatively and CT scan at 6 and 18 mo to evaluate the oncologic outcomes.

Univariable and multivariable logistic regression analyses were applied to assess the correlations of tumor characteristics (size, location, and growth pattern) with the number of clamped branches. Univariable and multivariable linear regression analyses were applied to assess correlations of tumor characteristics, clamped time, and preoperative GFR level with postoperative GFR reduction and estimated blood loss (EBL). The null hypothesis was rejected at p < 0.05.

3. Results

Patients’ general information and preoperative tumor characteristics are listed in Table 1. The median tumor diameter was 3.4 cm (range: 1.4–7.0 cm). The intrahilar feeding branch occupied 83.2% of all three anatomic variants. Surgical outcomes are listed in Table 2. The mean operating time was 87 min, and the median segmental clamping time was 24 min (range: 12–40 min). The median intraoperative blood loss was 200 ml (range: 50–800 ml). Target branches orientated by DSCT angiography could be isolated and clamped in all cases. Satisfactory ischemic areas were obtained by precise clamping in 117 of 125 cases. Insufficient regional ischemia due to missed clamping of feeding arteries was seen in eight patients. The accuracy of DSCT angiography for orienting the feeding artery reached 93.6%. The median hospitalization stay was 7 d. The total surgical complication rate was 14.4%.

Nine patients had arterial or venous bleeding from the parenchymal defect and required blood transfusion. Blood transfusion was indicated when the intraoperative hemoglobin level was <10 g/dl. The hemorrhage was controlled with clips or suturing, leaving the main renal artery unclamped. None of these nine patients was converted to open surgery or radical nephrectomy. One patient had postoperative hemorrhage and received branch embolization intervention 2 d after surgery. Eight patients had gross hematuria that was resolved after 2 wk of bed rest without further intervention. During the follow-up, no patient had renal dysfunction, arteriovenous fistula, aneurysm, or tumor recurrence.

Univariable and multivariable analyses showed that tumor size, location, and growth pattern were three independent factors affecting the number of clamped branches (Table 3). Comprehensively, large tumor size, striding location, and endophytic pattern led to a greater number of clamped branches. Anatomic characteristics and intraoperative variables were integrated into the analysis of GFR reduction and EBL (Table 4 and Table 5). Tumor characteristics and clamped time were not independent factors affecting GFR (size: β = 0.015, p = 0.612; location: β = 0.011, p = 0.769; pattern: β = 0.047, p = 0.267; clamped time: β = 0.022, p = 0.421). Only the number of clamped arteries was significantly correlated with reduced renal function. GFR reduction was greater in cases in which two arterial branches were clamped than in cases of single-branch clamping (β = 0.111; p = 0.001), and greater in cases in which three arterial branches were clamped than in cases of single-branch clamping (β = 0.181; p = 0.001). Only the number of clamped branches was an independent predictor of EBL in multivariable analysis (two branches vs one branch: β = 69.71, p = 0.024; three branches vs one branch: β = 123.29, p = 0.011), once adjusted for the other covariates.

4. Discussion

Renal artery clamping during PN can provide good intraoperative visualization and bleeding control, but inevitably causes WI injury. The WI time should be ≤20 min, as recent studies recommend [14], [15], and [16]. Every effort, therefore, should be made to minimize WI intervals and priority should be given to methods that improve renal surgery paradigms to minimize ischemic parenchymal damage [17]. We previously reported our segmental renal artery clamping technique to obtain regional ischemia and avoid global WI, and showed that it provided improved functional outcomes compared to conventional clamping techniques [6]. However, there were still some limitations such as missed clamping or overclamping with unnecessary parenchymal ischemia. This encouraged us to further refine the technique to achieve more satisfactory surgical outcomes.

Several essential factors contribute to the concept of precise segmental clamping. First and foremost is the application of high-resolution 3D angiography using DSCT. Compared to single-source CT, DSCT has higher resolution and significantly improves image quality [18]. Owing to the high-quality and multiangle reconstructive imaging, DSCT can reveal the course of segmental arteries in the sinus and interlobar arteries within the parenchyma until they reach the cortex or the mass [19] and [20]. It serves as guidance to optimize the approach to finding the target branches (the anterior approach, the posterior approach, or both) and to pinpointing the best clamping position. This precise orientation and strictly limited dissection within the hilum decrease the risk of vessel injury compared to our previous method of dissecting all visible branches of the renal artery.

The second factor is a modified retroperitoneal approach with four ports distributed according to tumor location. We developed this modified approach to meet the treatment requirements for both anterior and posterior tumors. It is feasible and convenient for segmental artery dissection at either the anterior or posterior hilum.

Another contributor to precise clamping is our modified method of tumor excision. It has been demonstrated that preserved parenchyma plays an essential role in postoperative renal function [21] and [22], so every effort should be made to maximally preserve the quantity of normal parenchyma. As seen in our study, we excised the tumor closely around its pseudocapsule, leaving only a 1- to 2-mm thickness of normal parenchyma. Safety-margin thickness does not correlate with disease progression [23] and a minimal tumor-free margin is sufficient to avoid local tumor recurrence [24]. Some interlobar arteries supplying adjacent parenchyma might cling to the pseudocapsule and we were able to peel them off to preserve the adjacent parenchymal blood supplying instead of transecting them (Fig. 5). Furthermore, the rate of collecting-system entry and, subsequently, calyceal repair with this method was lower than that of our previous technique.

Our technique highlights the complexity of the segmentation and the variation of renal vascularity. In the precise clamping technique, the optimal target is the feeding branch from the last bifurcation before entering the parenchyma. Any clamping before the last bifurcation will extend the ischemic area, even if it is at a tertiary or quaternary branch from the main renal artery. As a result, precise clamping can produce a minimal ischemic area, yet still be sufficient to control bleeding from the defect.

Reduced renal function may be related to WI injury and the volume of preserved normal parenchyma. With our technique, the ischemic injury has been limited to the tumor area; the blood supply to most normal parenchyma remains unaffected. Additionally, the normal parenchyma has been maximally preserved with our method of tumor excision. The reduction of postoperative renal function, therefore, is mainly affected by tumor characteristics. In univariable analysis, tumor size, location, and growth pattern were all correlated to reduced GFR with statistical significance, but by multivariable analysis, this correlation was lost. Thus a single characteristic could not predict the loss of renal function. These results may be due to interaction among these variables. Similar results occurred in the analysis of factors influencing EBL. Another valuable finding is that three tumor characteristics independently correlate to the number of clamped branches: Large tumor size, striding location, and endophytic pattern are usually accompanied by multiple clamped branches. In other words, we can use the number of clamped branches to represent comprehensive tumor characteristics. That the clamped number was found to be an independent factor in multivariable analysis verifies these results. In short, the clamped number can be a reliable predictor of postoperative renal function and EBL.

Although the outcomes of our precise clamping technique for LPN are promising, several limitations still exist in our study. First, it is retrospective. Second, it is a single-surgeon series. Third, although short-term surgical outcomes are satisfactory, the evaluation of long-term outcomes is still awaited.

5. Conclusions

Our precise segmental clamping technique under the guidance of DSCT angiography is feasible and effective for excising renal tumors and protecting the normal parenchyma. This precise clamping technique further minimizes WI injury during LPN. Tumor size, tumor location, and growth pattern are associated with the number of clamped arterial branches. The number of clamped branches can predict short-term postoperative renal function.

Author contributions: Changjun Yin had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Yin, Shao, Tang.

Acquisition of data: Yin, Shao, Tang, Xu, Ju, Meng, Lv, Zhang, J. Li, Cao.

Analysis and interpretation of data: Shao, P. Li, Tang.

Drafting of the manuscript: Shao, P. Li.

Critical revision of the manuscript for important intellectual content: Yin, Shao, Qin, Ju, J. Li.

Statistical analysis: P. Li, Shao, Cao.

Obtaining funding: None.

Administrative, technical, or material support: Yin, Tang, Xu.

Supervision: Yin.

Other (specify): None.

Financial disclosures: Changjun Yin certifies that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: None.

Funding/Support and role of the sponsor: This study was supported by the Program for Development of Innovative Research Team in the First Affiliated Hospital of Nanjing Medical University, Nanjing, China.