Abstract
Background:
Novel studies are helping to consider minimally invasive surgery for treating patients with renal cell carcinoma and venous tumor thrombus. Evidence regarding its feasibility and safety is still sparse and does not include a subclassification for level III thrombi. We aim to compare the safety of laparoscopic vs open surgery in patients with levels I–IIIa thrombus.
Materials and Methods:
This is a cross-sectional comparative study using single-institutional data on adult patients treated surgically between June 2008 and June 2022. Participants were categorized into open and laparoscopic surgery groups. Primary outcome was difference in the incidence of 30-day major postoperative complications (Clavien–Dindo III–V) between groups. Secondary outcomes were differences in operative time, length of hospital stay, intraoperative blood transfusions, delta hemoglobin level, 30-day minor complications (Clavien–Dindo I–II), estimated overall survival, and progression-free survival between groups. A logistic regression model was performed including adjustment for confounding variables.
Results:
Overall, 15 patients in the laparoscopic group and 25 patients in the open group were included. Major complications occurred in 24.0% of patients within the open group and 6.7% of patients were treated laparoscopically (p = 0.120). Minor complications arose in 32.0% of patients treated with open surgery and in 13.3% of patients treated in the laparoscopic group (p = 0.162). Although not significant, there was a higher perioperative death rate within open surgery cases. The laparoscopic approach presented a crude odds ratio for major complications of 0.22 (95% confidence interval 0.02–2.1, p = 0.191) compared with open surgery. No differences were found between groups regarding oncologic outcomes.
Conclusion:
Laparoscopic approach for patients with venous thrombus levels I–IIIa seems to be as safe as open surgery.
Introduction
As a characteristic hallmark of renal cell carcinoma (RCC), up to 10% of patients with this disease have venous tumor thrombus (VTT) invading the renal vein and/or inferior vena cava (IVC), 1,2 and aggressive surgery is widely accepted and recommended for patients with good performance status. 3,4
Minimally invasive surgery (MIS) has excellent results in selected cases of locally advanced RCC, with the advantage of lower morbidity and shorter convalescence time compared with open procedures. 5 –7 Nonetheless, the open approach is still considered the gold standard of treatment for this subset of patients. 8,9 Currently, even though the role of MIS in this scenario is yet controversial, 10 novel studies are helping to redefine this paradigm. However, evidence regarding feasibility and short-term perioperative outcomes is still sparse and of low quality. 9
Although the Mayo classification system is probably the most frequently used for surgical planning, 11 level III thrombus may imply different challenges regarding its cephalad extension. In line with this, Ciancio and coworkers. proposed in 2002 a model that divides this level into four subgroups, defining more precisely the complexity of the surgery. 12 Nevertheless, this classification has not been considered in later series reported. 1
This study aims to compare the safety of open vs laparoscopic radical nephrectomy with tumor thrombectomy (RNTT) in patients with VTT levels I–IIIa of Ciancio classification system.
Materials and Methods
This is a cross-sectional comparative study using single-institutional data retrieved from the Clinical Research Office of the Endourological Society (CROES) database.
Adult patients treated with RNTT in our institution, between January 2010 and June 2022, were considered eligible, and the Ciancio classification system was used to describe VTT extension. 12
In our center, we are confident that up to level IIIa the procedure can be safely approached with MIS. Therefore, since 2016, we intend to perform these cases laparoscopically. Nevertheless, for thrombus level IIIb or higher, as well as in the case of contraindications for laparoscopy (previous abdominal surgeries or poor cardiorespiratory performance), we perform open surgeries. For this reason, patients with Ciancio stages IIIb–IV were excluded. We also excluded patients with VTT level 0 (since they are managed without much difficulty than a standard radical nephrectomy 1 ) and patients with incomplete clinicopathologic data.
Participants were categorized into two study groups: Open RNTT and Lap RNTT. In the former, we included patients with Ciancio levels I–IIIa treated with open surgery. In the latter, we included patients with the same thrombus levels and the intention to be treated by laparoscopy, regardless of whether conversion was necessary.
Demographic information collected included gender, age, and comorbidities. Preoperative variables recorded were the side and maximum size of the renal tumor and thrombus level. Intraoperative variables recorded were surgical technique for control of tumor thrombus, number of intraoperative red blood cell (RBC) units transfused, and operative time. Postoperative complications were evaluated at 30 days using the Clavien–Dindo classification. Variation of hemoglobin (Hb) level (“Delta Hb”) was estimated using the last preoperative Hb level and the lowest postoperative Hb level obtained during the hospital stay. Hospitalization time was also recorded. For the Lap RNTT group, the need for hand-assisted or open surgery conversion was stated.
To further demonstrate the safety of the laparoscopic approach, pathologic data and oncologic outcomes were also included in the analysis.
Institutional review board approval was obtained.
Primary outcomes
The primary outcomes were difference in incidence of postoperative Clavien–Dindo III–V complications between groups.
Secondary outcomes
Secondary outcomes were differences in operative time, length of hospital stay, intraoperative blood transfusions, delta Hb, and minor complications (Clavien–Dindo I–II), estimated overall survival (eOS), and progression-free survival (ePFS) between groups.
Surgical technique
In the open technique, the retroperitoneum is accessed through a Chevron incision with a vertical supraumbilical extension. In the laparoscopic technique, we choose a transperitoneal approach.
In right-sided tumors, the infrarenal IVC and the left renal vein are encircled with a tourniquet. In laparoscopic surgery, we used a Rummel tourniquet 13 and secured it in place using a clip.
For level III thrombi, the short hepatic veins are ligated with Hem-o-locks® and released with an ultrasonic scalpel to gain access to retrohepatic IVC, by retracting the caudate lobe. Then, retrocaval dissection of the intrahepatic IVC is carried out until it can be encircled with a Rummel tourniquet in this proximal location. The right renal vein is dissected and exposed. The right renal artery is exposed and clipped in the interaortocaval region. Next, the distal IVC tourniquet is cinched. Once it is ensured that the patient can tolerate IVC cross-clamping, the left renal vein and the proximal IVC tourniquet are also cinched. Then, a cavotomy is performed on its right edge and the thrombus is dissected from its lumen.
The right renal vein ostium is excised en bloc with the thrombus. Finally, after IVC reconstruction with a 4/0 polypropylene running suture, tourniquets are released in the following order: left renal vein, suprarenal IVC, and infrarenal IVC. When assuring hemostatic control, the nephrectomy is completed, and the specimen was extracted through a Pfannenstiel incision.
In the case of left-sided tumors, blood flow to the right kidney needs to be excluded with Bulldog clamps to properly control the thrombus-bearing IVC. In open surgery, both sides are accessible through the Chevron incision, but in laparoscopy, the patient needs to be repositioned. The surgery starts with the nephrectomy in right lateral decubitus, leaving the renal vein anchored or resected with a mechanical suture. Next, we extract the specimen (through the left iliac fossa) and continue with the management of the IVC in a left decubitus, requiring temporary clamping of the right renal hilum during the cavotomy. The clamping sequence is caudal IVC first, then right artery, right vein, and cephalic IVC. After IVC reconstruction, its flow is restored first and the right kidney afterward.
Statistical analysis
Demographic and clinical characteristics were summarized using descriptive statistics. Continuous variables are expressed as mean plus confidence intervals (CIs) or medians plus interquartile range (IQR) as needed and compared by a t-test or Mann–Whitney U test, as appropriate. Categorical variables are expressed as absolute values and percentages (%) and compared with chi-squared tests.
A logistic regression model was used to calculate the odds ratio (OR) for the laparoscopic approach. Then, a stepwise adjustment including confounding variables with >20% crude coefficient modification was performed. Aiming to inform adjusted OR, confounding variables were introduced in order of modification magnitude.
Assessment of eOS and ePFS was performed by using Kaplan–Meier curves and log-rank tests.
We considered p < 0.05 as significant and used IBM SPSS Statistics V22.0 for analysis.
Results
Baseline characteristics
In total, 15 patients in the Lap RNTT group and 25 patients in the Open RNTT group were included (Fig. 1). Mean age at diagnosis was 61.7 years (95% CI 57.9–65.4) and 75% (30/40) of patients were men.
Study flowchart. IVC = inferior vena cava; Lap = laparoscopic; RCC = renal cell carcinoma; RNTT = radical nephrectomy with tumor thrombectomy.
Overall, no differences were found neither for tumor maximum diameter nor for VTT level. Moreover, most patients were treated with thrombectomy, and no differences were found between groups.
Median operative time was 282.5 minutes (IQR 210–359.5) and was similar across groups (p = 0.208), as well as the median number of RBC units transfused intraoperatively (p 0.081).
Three laparoscopically treated patients needed conversion to hand-assisted surgery (20%) and two of these patients were then converted to open surgery (13.3% of the Lap RNTT cohort).
More data regarding baseline characteristics and operative variables are given in Tables 1 and 2.
Baseline Characteristics for the Overall Cohort
BMI = body mass index; CI = confidence interval; IQR = interquartile range; Lap = laparoscopic; RNTT = radical nephrectomy with tumor thrombectomy; SD = standard deviation; VTT = venous tumor thrombus.
Operative and Postoperative Variables
Hb = hemoglobin; IQR = interquartile range; RBCs = red blood cells.
Postoperative complications
Of patients treated with open surgery, 24.0% of patients had major complications, whereas one patient (6.7%) in the Lap RNTT group had such an event (p = 0.120).
With regard to minor complications, 32.0% of patients in the Open RNTT group and 13.3% in the MIS group presented Clavien–Dindo I–II events (p = 0.162).
Although not reaching significance, when considering only Clavien–Dindo V complications, a higher perioperative death rate was found within the open surgery cases, with three deaths reported:
Case 1: 55-year-old man with a 46-mm right renal tumor with level II VTT. Comorbidities: severe obesity (body mass index [BMI] 37.4), arterial hypertension (AHT), diabetes, and current smoking habits. He had an acute ischemic abdomen on postoperative day 4 that could not be resolved.
Case 2: 61-year-old noncomorbid man with a 90-mm right renal mass with level II VTT. During the 43 days of in-hospital stay, he had multiple complications: hemorrhagic shock, acute renal failure treated with hemodialysis, biliary fistula, and septic shock with multiorgan failure. He died in hospital.
Case 3: 69-year-old woman with AHT history and a 90-mm right renal tumor with level II VTT. Intraoperative finding of VTT level IIIc. Death was because of massive pulmonary embolism while dissecting IVC.
Relating to logistic regression analysis, the laparoscopic approach presented a crude OR for Clavien–Dindo III–V complications of 0.22 (95% CI 0.02–2.1, p = 0.191) when compared with open surgery. After adjusting for confounders, the OR for statistically proven confounders (BMI and tumor size) for laparoscopic surgery is 0.12 (95% CI 0.01–1.77, p = 0.122), being unable to demonstrate that approach is a risk factor for these complications.
Oncologic outcomes
Overall, 32.5% (13/40) of cases had metastatic disease at diagnosis (p 0.542) and 20.0% (8/40) had nodal disease (p 0.734). No differences were found relating to positive surgical margins (p 0.696) as well as lymph node count (p 0.532). More data on pathologic features are given in Table 3.
Pathologic Variables
ISUP = International Society of Urological Pathology; SD = standard deviation.
Regarding disease progression, 40.0% of laparoscopically treated patients (6/15) and 32.5% of the patients treated with open surgery (13/40) suffered from either local relapse or metastatic disease at follow-up. The 1-year ePFS for Lap RNTT and Open RNTT patients was 40.5% (95% CI 5.5–75.5) and 73.6% (95% CI 53.2–94.0), respectively (p 0.394).
In terms of overall survival, 26.7% (4/15) of patients from the Lap RNTT group and 36.0% (9/25) of patients from the Open RNTT group died at a median 13.5 months (IQR 4–36) follow-up. The 1-year eOS for Lap RNTT and Open RNTT patients was 51.4% (95% CI 12.4–91.4) and 74.4% (95% CI 56.6–92.2), respectively (p 0.851).
Discussion
This study was an attempt to evaluate the safety of the laparoscopic approach for the treatment of RCC+VTT up to level IIIa, using Clavien–Dindo complications as a surrogate for this outcome.
Although MIS in this scenario is still unstandardized, and it appears it is more the exception than the rule, encouraging results published with these approaches may indicate this scenario begins to change in the medium term. 9,14
To our knowledge, in only three studies, a direct comparison between pure laparoscopic and open surgery was reported, 15 –17 and together with our results, it seems that there is still inconclusive evidence regarding postoperative complication rates. For instance, we found that both study groups had similar outcomes, which is consistent with the findings of Xu et al. 15 With a bigger sample size, Liu and coworkers could demonstrate a significant difference, reporting 19.5% complications in the laparoscopic group against 47.8% complications in the open cohort. 16 Nevertheless, the laparoscopic group had significantly smaller renal tumors and a higher proportion of low-level VTT (Mayo level I). 15,16
In our cohort, we found a higher prevalence of VTT level II in the Open group. Even so, a comparable distribution of high complexity cases (VTT level IIIa) was also found between groups. Moreover, VTT level did not represent a confounding variable in our logistic regression model.
Finally, after 1:1 propensity matching, Zhang and coworkers found a higher complication rate (39.4% against 21.3%) and Clavien–Dindo grade in the open surgery patients, without a difference in baseline characteristics between groups. 17
Besides not being able to demonstrate a difference, a tendency for fewer complications in the laparoscopic group in our study and in Xu et al. study must be highlighted. Of note, these results were regardless of the type of MIS approach, since Xu et al. used a complete retroperitoneal approach while we performed transperitoneal surgeries. 15
With regard to the need for conversion, we report a 13.3% conversion rate. This outcome is extremely variable in the literature since Liu and coworkers described 22.0% conversions, Zhang and coworkers a 25.5% conversion rate, and Xu et al. reported no conversions in 17 cases, but some patients were treated with hybrid procedures (laparoscopic nephrectomy and open thrombectomy). 15 –17
Concerning oncologic outcomes, we did not detect any significant difference neither in ePFS nor in eOS between study groups. Immediate pathologic data support these findings, since no difference was found in either lymph node count or positive surgical margins.
Since we performed the first RNTT in our institution, there has been a shift in surgical approach. In the beginning, we carried out every case with open surgery. But, since 2016, the number of laparoscopic RNTT has gradually increased for patients up to level IIIa thrombus. This is in line with the development of experience in laparoscopic surgery in our department. Before leading off the surgical challenge of performing a laparoscopic RNTT, our surgeons developed skills in other minimally invasive procedures such as radical, partial, and live kidney donor nephrectomies.
Most of the reports on this topic include the Mayo classification system to describe the cephalad extension of the thrombus. 9 But, for a safe approach to the VTT, a subclassification of level III thrombi should be considered. 12 Since optimal exposure and control of hepatic veins in this level are of utmost importance for gaining liver flow control, this system helps to anticipate the sequence of surgical maneuvers needed at each sublevel and to optimize surgical planning. 18,19
More proximal thrombi extending beyond the major hepatic vessels need control of the suprahepatic IVC. This requires laparoscopic liver mobilization and Pringle maneuver, and both are extremely challenging from a technical point of view. 14 For this reason, in our practice, we perform laparoscopic surgeries only in levels I–IIIa.
In the constant evolution of minimally invasive approaches, robotic surgery seems to occupy a dominant role since there are reports of levels III–IV thrombus effectively operated with these techniques. 19,20 Moreover, some studies go further and explore different robotic approaches (e.g., transperitoneal vs retroperitoneal). 21 Nevertheless, despite having a robotic surgery program since 2008, we prefer the laparoscopic approach for three reasons: first, the easiest conversion in case it becomes necessary; second, the possibility to choose the site and number of ports for each patient; and third, because robotic approach may hinder intraoperative assistance from consulting services that are less familiarized with robotic surgery. 10
Although our results are encouraging, some limitations must be stated. First, the retrospective, nonrandomized, and single institutional design with its inherent selection bias. Second, since no established criteria exist to decide between surgical modalities, it still depends on surgeons' preferences and an adequate selection of the case. 20 Third, besides a difference regarding level II thrombi between groups was found, VTT level did not represent a confounding variable in our logistic regression model. Finally, we did not include patient-reported outcomes for the evaluation of surgical safety.
We encourage further research in this area considering multi-institutional registries as well as studies with a prospective design.
Conclusion
Laparoscopic approach for patients with RCC and venous thrombus levels I–IIIa seems to be as safe as open surgery.
Footnotes
Acknowledgment
The authors thank Ignacio Pablo Tobia, MD for serving as a statistics advisor.
Authors' Contributions
All authors have contributed equally to the conception and design of the article, as well as to acquisition, analysis, and interpretation of the data. All authors read and approved the final version of the article.
Author Disclosure Statement
No competing financial interests exist.
Funding Information
No funding was received for this article.