1. Introduction
Kidney cancer is one of the most common types of cancer with an estimated 138,600 new cases and 54,000 deaths in Europe in 2020 [
1]. International guidelines consider partial nephrectomy (PN) as the gold standard technique for localized kidney tumor treatment [
2,
3]. Different approaches are described, such as open partial nephrectomy (OPN), laparoscopic partial nephrectomy (LPN) or robot-assisted laparoscopic partial nephrectomy (RALPN). Compared to both OPN and LPN, RALPN provides decreased intraoperative blood loss, shorter hospitalization time, fewer complications and shorter ischemia times [
4,
5]. As a result, RALPN has become the new gold standard technique for mini-invasive PN [
6]. LPN can be complex, with a prolonged learning curve, due to limited ergonomics and to technical challenges such as laparoscopic suture [
7]. Gallucci et al. have described a zero-ischemia technique of PN after selective embolization of tumor vessels. The main disadvantages of their technique were prolonged delays between the embolization step and the surgical step (up to 24 h), resulting in peri-lesioned edema increasing the tumor dissection difficulty [
8,
9,
10]. Next-generation hybrid operating rooms combine traditional surgical equipment with advanced imaging technology and allow both procedures to be performed together while optimizing the time between the embolization and the surgical steps [
11]. LPN after hyperselective embolization of tumor vessels (HETV) in a hybrid operating room (HOR) is a non-clamping approach which has been performed in our center since 2015 [
12]. The first short-term results were encouraging regarding operative times, bleeding, postoperative renal function, and the risk of arterial pseudoaneurysm [
13,
14,
15]. The aim of this study was to evaluate the long-term oncological and functional results of LPN after HETV in HOR for localized kidney tumors.
4. Discussion
In this study, we reported a series of LPN with HETV in HOR with a median follow-up of 27 months. The demographic data of our study were similar to those already published, in terms of age, BMI, ASA scores, tumor size, and R.E.N.A.L. complexity [
4,
5,
6,
20].
The median total operative time of 168 min, including HETV, patient repositioning and LPN, appears to be shorter compared to other LPN series [
6,
21]. Our operative times seem similar to those described in the RALPN series [
4,
5]. The median laparoscopic time alone was reduced to 75 min. This could be explained by the absence of renal pedicle dissection and the possibility of suturless PN in most of the cases. In addition, blue dye embolization facilitated macroscopic tumor localization without the use of intra-body ultrasound [
19]. In our study, only one (0.4%) laparoconversion was necessary due to the difficulty in tumor localization. Simmons et al. described six (1%) cases of conversion [
22], while Masson-Lecomte et al. had seven (3.18%) conversions in the robot-assisted laparoscopy group and five (11.1%) in the laparoscopy group [
23]. Dissection was always possible without difficulty related to perilesional edema; a limiting factor described when embolization was performed remotely from the surgical procedure [
10].
Intraoperative bleeding is the most important and severe complication of PN. In contemporary RALPN series, bleeding occurred in 6% of the cases. Despite the off-clamp PN, intraoperative bleeding was 100 mL which is lower than bleeding reported in RALPN (150 to 300 mL) [
4,
5,
24]. Hemostasis was achieved using hemostatic agents and in the majority of the cases did not require cortical suturing (96.7%), which could contribute to the preservation of healthy renal parenchyma [
25]. Furthermore, only five (2.1%) major hemorrhages occurred, and four of them required intraoperative blood transfusions. These results are lower than those reported in the literature [
4,
6,
21,
26].
The low rate of intraoperative hemorrhagic complications, preventing organ hypoperfusion, could also contribute to nephron preservation. However, we did have three (1.2%) postoperative renal bleeding episodes and one (0.4%) hepatic bleeding episode, which required additional postoperative embolization. For one patient the bleeding occurred because of an embolization failure. For two other patients, bleeding occurred because of a large tumor excision with non-embolized healthy renal parenchyma removal. Simone et al. described two cases (1%) of secondary bleeding requiring additional embolization in their series of LPN after HETV [
9]. However, these complications remain lower than those reported by George et al. who performed 16 (5.54%) embolizations for postoperative bleeding [
26]. The use of glue for embolization appears to be a safe and effective technique [
15].
Major postoperative complications were noted in 17 (6.9%) patients, which is consistent with the literature [
6]. One patient died postoperatively from cardiac arrest secondary to uncontrolled asthma and had a history of multiple cardiovascular comorbidities (high blood pressure and aortic valve disease with mechanical cardiac valve). Minor complications were mostly isolated postoperative hyperthermia, related to post-embolization syndrome [
14]. Hospitalization times were similar to those reported in RALPN trials [
20].
The deterioration of renal function after PN is a multifactorial and complex process related to non-modifiable factors (age, comorbidities and preoperative kidney function) and modifiable factors (duration of ischemia and nephron sacrifice) [
7]. Mir et al. described a preservation of approximately 90% of renal function after PN [
27].
In a previous study, we evaluated renal function by GFR and computed tomography renal volume 6 months after surgery in 137 patients [
15]. We found a 9.3 mL/min decreased in GFR and a median loss of 21 mL of healthy parenchyma on the operated kidney which is consistent with the literature on robot-assisted surgery [
28]. With a larger population and a longer follow-up, we found a 10% loss of renal function, which remained stable over time. We did not identify the risk factor of significant decrease in renal function. However, we did not have data regarding the volume of healthy renal parenchyma loss nor any pre-existing hypertension or proteinuria. This could be a limitation as it appears that these factors may have an impact on the postoperative GFR [
29,
30,
31].
Preoperative embolization of tumor arteries has several advantages: (1) the reduction in intraoperative bleeding; (2) the selective dissection of the tumor using blue coloration of the tumor to optimize the differentiation from the normal parenchyma and therefore resulting in a better preservation; and (3) performing PN without clamping which avoids the risk of renal ischemia lasting more than 25 to 30 min [
32]. This is an important point especially for patients suffering from preoperative chronic renal disease: HETV limited the loss of renal function to 9% at 42 months. These results should be interpreted with caution, as our population of chronic renal insufficiency patients was only 30 patients and only 4 patients were still followed-up at 42 months. Nevertheless, studies specifically focusing on this population would be interesting.
Out of the 245 patients in our series, 34 (13.9%) had benign tumors. These results are consistent with the literature. Simone et al. found 30% benign tumors, Masson-Lecomte et al. had 16% in the robot-assisted series, and Peyronnet et al. had 14.6% [
4,
9,
22,
23]. Of these 34 benign tumors, 22 were oncocytomas. However, the diagnosis is difficult, and it can be tricky to identify chromophobe renal cell carcinoma, with 9 to 25% of patients having a final diagnosis of clear cell carcinoma [
33,
34]. The other benign tumors included five angiomyolipomas, five symptomatic cysts, one hemangioma, and a metanephric adenoma.
We identified two risk factors for recurrence: positive surgical margins and pathological stage ≥ pT3a. The rate of positive margins was 4.9%. These results are similar to the robot-assisted series of Ingels et al. (4.9%), Peyronnet et al. (5.2%), Pignot et al. (5.7%) and Masson-Lecomte et al. (8%) [
4,
5,
20,
23]. Despite a positive surgical margin rate comparable to that observed in the literature, the 5-year disease-free survival rate in our series (84%) appears to be in the low range compared to multicenter series, which reported that 5-year CFS estimate rates ranged from 86.4% to 98.4% [
35]. It is probably related to the aggressiveness and complexity of the tumors treated in our series. Indeed, 8.1% of our patients had a pT3a tumor with invasion of the peri-renal fat. There were also two atypical histologies in our series (metastasis from lung cancer and Bellini’s carcinoma) which recurred in the year following surgery. Moreover, most series reporting long-term oncological results are retrospective studies whereas our data were collected prospectively, and this may constitute a bias. For example, two recent prospective studies reported 5-year CFS estimate rates of 86% and 91%, which is closer to what we report [
36,
37].
All studies comparing RALPN to OPN or LPN have found benefits for the robot-assisted approach regarding postoperative complications, bleeding, transfusions rate and length of hospital stay. However, our results appear encouraging regarding the robot-assisted LPN series [
4,
5,
6,
23].
The costs associated with the use of robot-assisted surgery are expensive. It could be interesting to compare them to those related to the use of an HOR.
Our study is of course not without limitations. It is an observational study, with no control group and the different comparisons with other surgical approaches, particularly robot-assisted, may be debatable. No conclusion can be drawn about the superiority of one technique over another.
However, the main strength of our study is being in a real-life prospective setting. It represents the activity of our academic department, and the surgical procedures were performed by operators with different levels of experience, some of whom were just beginning their learning curve in laparoscopy.