Robotic-Assisted Sentinel Lymph Node Mapping With Indocyanine Green in Pelvic Malignancies: A Systematic Review and Meta-Analysis

Objective: Newer technologies such as near-infrared (NIR) imaging of the fluorescent dye indocyanine green (ICG) and daVinci Xi Surgical System have become promising tools for sentinel lymph node (SLN) mapping. This meta-analysis was conducted to comprehensively evaluate the diagnostic value of SLN in assessing lymph nodal metastasis in pelvic malignancies, using ICG with NIR imaging in robotic-assisted surgery. Materials and Methods: A literature search was conducted using PubMed for studies in English before April 2019. The detection rate, sensitivity of SLN detection of metastatic disease, and factors associated with successful mapping (sample size, study design, mean age, mean body mass index, type of cancer) were synthesized for meta-analysis. Results: A total of 17 articles including 1,059 patients were finally included. The reported detection rates of SLN ranged from 76 to 100%, with a pooled average rate of 95% (95% CI: 93–97; 17 studies). The sensitivity of SLN detection of metastatic disease ranged from 50 to 100% and the pooled sensitivity was 86% (95% CI: 75–94; 8 studies). There were no complications related to ICG administration reported. Conclusions: NIR imaging system using ICG in robotic-assisted surgery is a feasible and safe method for SLN mapping. Due to its promising performance, it is considered to be an alternative to a complete pelvic lymph node dissection.


INTRODUCTION
Pelvic lymphadenectomy (PLND), which remains the most accurate procedure for the detection of lymph node metastasis (LNM) in malignant pelvic tumors, plays an important role in surgical management of endometrial and prostate cancers (1,2). However, the data from a prospective study suggests that lymphadenectomy is associated with the increasing operative time, blood loss and risk for surgical morbidity (e.g., blood vessel and nerve damage, lymphedema, and lymphocyst formation) (3,4). Thus, novel nodal assessment techniques should be developed to improve the accuracy of LNM detection with lower surgical morbidity.
The biopsy of SLN which is defined as the first node to receive the drainage from the primary tumor, has been described by Canbanas in 1977 (5). The utility of SLN mapping can avoid the unnecessary LND when the SLN turns out to be negative (6). The different methods used in SLN mapping, such as blue dye, technetium, and ICG with NIR imaging have been investigated, among which ICG has been used clinically for over two decades with an excellent safety profile (7). Also, as one of four fluorochromes approved by US Food and Drug Administration, ICG may be of significant use in pelvic surgery due to its properties (8).
The NIR fluorescence imaging system in daVinci Xi Surgical System (Intuitive Surgical, Sunnyvale, CA, USA) with Firefly technology provides intraoperative ICG near-infrared fluorescence, especially for ICG at low concentrations in lymphatic mapping. While the high concentrations of ICG can be seen directly in green in color on a background of a grayscale image. Moreover, it brings surgeons great convenience to control the scope completely which the infrared and visible light systems are built in.
Nevertheless, although ICG-NIF imaging in SLN detection appears to be superior, there are few studies of meta-analysis on SLN mapping outcomes, and most of them focused on specific one or two types of cancer, especially on endometrial and cervical cancer. Thus, we performed this meta-analysis to evaluate the detection rates and sensitivity of SLN mapping in malignant pelvic tumors, including endometrial, cervical, bladder and prostate cancers.

Search Strategy
The literature search was conducted on PubMed, only English language studies before April 2019 included. The search terms used are as follow: (robotic OR robot) AND (indocyanine green OR ICG) AND lymph. In addition, the references of included studies were reviewed as supplement.

Study Quality Assessment
The quality of enrolled studies was assessed using the QUADAS-2 (Quality Assessment of Diagnostic Accuracy Studies-2) (9) tool by two reviewers independently. The items are shown in the Appendix
The overall detection rate was estimated as the proportion of patients with at least 1 SLN identified among all the patients going though PLND. When assessing sensitivity and specificity, the patients who failed in SLN mapping were excluded, and sensitivity is defined as the percentage of patients with positive SLN divided by all patients with positive lymph nodes. The specificity is defined as the percentage of patients with negative SLN divided by all patients with negative lymph nodes. Studies in which the calculation of sensitivity was based on the number of removed node packets but not on the patients were excluded during assessment.

Statistical Analysis
The Stata 15.0 and meta-disc were used to conduct all data analysis. The overall detection rate was calculated using a random-effects model under meta-analysis. The sensitivity of detection rates of SLN was evaluated using summary receiver operating characteristic curve (SROC). The I 2 index was used to detect the heterogeneity among the studies. The Funnel plots, Egger's regression intercepts were used for the evaluation of publication bias. The univariate meta-regression was applied for the association of SLN detection rate and study characteristics, including sample size, study design, mean age, mean BMI, and type of cancer.

Characteristics of Enrolled Studies
Of the 78 abstracts screened, 17 articles including 1,059 patients with pelvic malignancies were eligible for inclusion as demonstrated in Figure 1. The sample size of each study ranged from 10 to 197. SNL was evaluated for endometrial cancer in 9 articles, cervical cancer in 1, prostate cancer in 3 and bladder cancer in 1, with the other 3 articles dedicated in both endometrial cancer, and cervical cancer (10-26) ( Table 1).
An overall mean of 3.5 SLNs was removed per patient (13 studies). All the studies conducted PLND with or without paraaortic LND, and the mean non-SLNs identified per person was 17.9 (5 studies). The mean age of 1,059 patients was 62 years (17 studies) and mean BMI was 31.1 kg/m 2 (13 studies) ( Table 1).
No complications, among all 17 articles, were described related to ICG administration.

Data Analysis for the Detection Rate and Diagnostic Accuracy of SLN Mapping
The detection rates of SLN ranged from 76 to 100%, with a pooled average of 95% (95% CI: 93-97; 17 studies) with heterogeneity I 2 = 56.2% (Figure 2). The funnel plot of the pooled overall SLN detection rate is shown in Figure 3. The Egger's regression intercept was −4.65 (p = 0.000).
Sensitivity of SLN mapping ranged from 50 to 100%. The pooled sensitivity of SLN detection of metastatic disease was 86% (95% CI: 75-94; 8 studies) (Figure 4). The funnel plot of the pooled sensitivity is shown in Figure 3. The Egger's regression intercept was found out to be

Test of Heterogeneity
Due to the heterogeneity I 2 was found to be 56.2% in detection rate a sub-group analysis was conducted to find the reasons for the observed heterogeneity. Univariate meta-regression of SLN detection rate and study characteristics showed that study size, study method, mean patient age, mean patient BMI, and type of cancer were not significantly associated with detection rates ( Table 2).

DISCUSSION
SLN mapping has been the standard of care for breast cancer and melanoma for a long time (27) and achieved success in many other types of cancer. During the process of assessing the value of SLN mapping, detection rate is taken into consideration in the very first place. Most of meta-analyses on diagnostic efficacy of SLN mapping were focused on uterine cancers, and to our knowledge, this is the first meta-analysis of that in pelvic malignancies.
It should be mentioned that in these previous meta-analyses, studies using both robotic-assisted system and indocyanine green fluorescence tracer in the meantime have not been analyzed statistically before. Compared with prior meta-analyses which studied on tracers including blue dye, ICG and/or 99mTc in  endometrial cancer (28)(29)(30), this meta-analysis with only ICG used as tracer shows higher detection rate of SLN mapping with detection rate of 95% (95% CI: 93-97), and in study of Lin et al. the detection rate was only 76% (95%: 71-81) with blue dye alone. In their study, detection rates were of 93 and 86% in ICG and 99mTc combined with blue dye, respectively, (28). Moreover, in study of Smith et al. detection rates were found higher of 90.3% in ICG vs. 81% in blue dye (31). Hybrid image-guided surgery technologies are increasingly gaining interest, such as combined radio-and fluorescenceguidance. In study of KleinJan et al. use of the hybrid tracer ICG-99mTc-nancolloid was evaluated and the detection rate was over 95% (32). Compared with the conventional radioguided SN approach, the additional cost of ICG-99mTc-nancolloid is negligible (33), and use of ICG involves only minor additional costs (34). According to prior studies, the use of ICG also brings several advantages, such as fewer adverse effects, less pain, and quicker transcutaneous real-time visualization (35,36).
In this meta-analysis, the pooled sensitivity of SLN detection of lymph node metastasis was 86% (95% CI: 75-94). In study of SLN mapping by Rossi et al. there were only two patients with positive lymph nodes in those who had successful mapping  and one of them showed negative SLN, leading to the lowest sensitivity of 50% (10). In other 10 studies with available data, 6 out of them showed the sensitivities of 100% and all of them were over 70% [ (11-14, 17, 21-24, 26); Table 1].
In pilot meta-analysis, robotic-assisted surgery demonstrated higher detection rates than other modalities. The pilot study conducted by Lin et al. showed that robotic-assisted surgery led to 86% detection rate, when laparoscopy and laparotomy got that of 82 and 77%, respectively, in patients with endometrial cancer (28). In the literature before, a higher BMI is linked to lower detection rate of the SLN (13). However, in the study by Rozenholc et al. there was no difference in the detection rate between surgeries that were robotic (mean BMI 44.6) and laparoscopic (mean BMI 29.4) (24). Moreover, when compared with open surgery, robotic-assisted surgery results in fewer blood transfusions and leads to a slightly shorter hospital stay (37).
However, due to the limitations of robotic-assisted surgery, such as higher costs and the lack of haptic feedback, the current robotic technology has not become the standard technique of minimally invasive surgery worldwide. A single robotic surgical system can set a hospital back by about $2 million, and that's just to get started. Some of the instruments are disposable and need to be continually replaced. Additionally, the systems require regular maintenance at rates that can exceed $100,000, which limits the number of hospitals that can buy Da Vinci system (https://www. drugwatch.com/davinci-surgery/). In a survey of complications of ICG angiography in Japan, the results indicated that indocyanine green was as safe as fluorescein (38), which was reported only 0.05% frequency of severe adverse reactions, such as circulatory shock, bronchospasm, laryngospasm, cardiac arrest, myocardial infarction, and tonic seizure (39).
Our systematic review and meta-analysis has limitations as follow. First, we only included English studies, which may lead to a potential language bias. Second, we didn't have individual patient data, such as age, BMI and so on. The results presented in this study were based on unadjusted estimates. Third, the number of studies included is not sufficient enough so that some subgroup analyses cannot be conducted, and it may contribute to a publication bias.
In conclusion, the NIR imaging system in robotic-assisted surgery with ICG dye is quite easy to master, and the present results confirmed that SLN mapping using ICG alone is a reliable and safe approach that performs well diagnostically when assessing lymph nodal metastasis in pelvic malignancies. Although it is considered to be an alternative to a complete pelvic lymph node dissection, studies with larger patient samples are still needed, especially in urology cancers like prostate and bladder cancer.

DATA AVAILABILITY
All datasets generated for this study are included in the manuscript/Supplementary Files.