Objective A retrospective cohort study comparing survival and perioperative outcomes of patients with early vulvar cancer who underwent sentinel lymph node biopsy versus standard lymphadenectomy
Methods Patients diagnosed between January 2012 and December 2015 with vulvar squamous cell carcinoma of less than 4 cm in size, with invasion of at least 1 mm, who underwent sentinel lymph node biopsy, lymphadenectomy, or both were identified from the National Cancer Database. Overall survival was evaluated following generation of Kaplan-Meier curves and compared with the log-rank test for patients who had at least 1 month of follow-up. A Cox model was constructed to control for confounders.
Results A total of 1583 patients were identified; 304 patients (19.2%) underwent sentinel lymph node biopsy alone. Sentinel lymph node biopsy utilization increased 13.9% between 2012 and 2015. Patients who underwent sentinel node biopsy alone were less likely to have comorbidities compared with those undergoing lymphadenectomy only or sentinel node biopsy with lymphadenectomy (25.3% vs 32.9% vs 31.9%, p=0.042), had smaller tumors (median 1.6 vs 2.0 vs 2.0 cm, p<0.001), and were less likely to have positive lymph nodes (11% vs 19.6% vs 28.1%, p<0.001). There was no difference in 3 year overall survival between the three groups (86.3% vs 82.1% vs 77.9%, p=0.26). After controlling for age, race, insurance, comorbidities, lymph node metastases, and tumor size, sentinel lymph node biopsy alone was not associated with worse overall survival compared with lymphadenectomy (HR 0.86, 95% CI 0.57 to 1.32). The sentinel node only group had shorter inpatient stays compared with lymphadenectomy only (median 1 vs 2 days, p<0.001) and a lower rate of unplanned readmission (1.7% vs 5.0%, p=0.010).
Conclusions The utilization of sentinel lymph node biopsy is increasing in the management of vulvar cancer and is associated with superior perioperative outcomes without impacting overall survival.
- vulvar neoplasms
- sentinel lymph node
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Utilization of sentinel lymph node biopsy for early stage vulvar cancer is increasing in the United States.
Sentinel lymph node biopsy is associated with superior perioperative outcomes without impacting overall survival.
The implementation of sentinel node biopsy in real world practice has yielded outcomes similar to those described in prospective trials.
Vulvar cancer is a rare disease, with an incidence of 2.5 per 100 000, with rates rising 0.6% per year for the past 10 years.1 According to the National Comprehensive Cancer Network guidelines, assessment of nodal disease is important in early stage vulvar cancer to determine whether patients should receive adjuvant treatment with radiation and chemotherapy.2 As management of gynecological cancers has trended towards less invasive surgeries, the use of sentinel lymph node biopsy for assessment of nodal disease has become an area of interest.
Vulvar cancer has been no exception; the GROINSS-V trial in 2008 and GOG-173 in 2012 are two landmark trials that established the safety and efficacy of sentinel node biopsy for staging of early vulvar cancer. Based on these trials, adoption of the procedure has been advocated for patients with unifocal tumors less than 4 cm in the absence of clinically suspicious lymph nodes.3 4 Since then, other institutional and literature reviews have shown similar outcomes, however there have been few population-based data regarding the rate of utilization of the sentinel node technique.5–7 There have also been relatively limited data comparing sentinel node biopsy to standard lymphadenectomy with regard to overall survival and perioperative outcomes in real world practice. The aim of this study was to evaluate the uptake of sentinel lymph node biopsy and compare survival and perioperative outcomes to standard lymphadenectomy for early stage vulvar cancer using a large database.
The National Cancer Database was accessed and patients diagnosed with pathologically confirmed vulvar squamous cell carcinoma (based on ICD-O-3 histology codes as grouped by the International Agency for Research on Cancer) between January 2012 and December 2015 were identified. The National Cancer Database, established jointly by the American Cancer Society and Commission on Cancer of the American College of Surgeons, is a hospital-based database capturing approximately 70% of malignancies diagnosed in the United States. Patient data are prospectively collected from commission-accredited cancer programs and are frequently audited to ensure their quality. All data are de-identified and available for research purposes. The American College of Surgeons and the Commission on Cancer have not verified and are not responsible for the analytical or statistical methodology employed, nor the conclusions drawn from these data. Penn Medicine deemed this study exempt from Institutional Board Review.
Using the collaborative staging schema in the National Cancer Database, patients with squamous cell carcinoma who had at least 1 mm depth of invasion, no evidence of metastatic disease, tumors less than 4 cm in size (T1b in the staging schema) and who underwent pathological evaluation of inguinofemoral nodes were identified. Patients who underwent sentinel node biopsy as specified by the database (codes 2, 6 and 7) were captured. The sentinel node group was further delineated by code type into sentinel node biopsy only and sentinel node biopsy with lymphadenectomy (performed either concomitant or as a separate procedure). The third comparator group included patients who underwent lymphadenectomy (code 5), defined by those having had at least four lymph nodes removed not utilizing sentinel node technique. We excluded patients with neoadjuvant radiation and those whose radiotherapy to surgery interval was unknown.
Demographic, clinical-pathological, and treatment variables were extracted from the dataset. Patient age was recoded as age less than 65 and age 65 years and over. Race was recoded as white, black, or other/unknown. Insurance status was recoded as private, government (including Medicaid and Medicare), and uninsured/unknown. The presence of comorbidities was assessed using the Charlson-Deyo Comorbidity index. It was classified as absent (score 0) or present (score ≥1). For analysis purposes, we assessed sentinel lymph node biopsy as two separate groups (either alone or with lymphadenectomy) with regard to demographics and survival. For perioperative outcomes, only the sentinel node biopsy alone group was compared with the lymphadenectomy group; those who underwent sentinel node biopsy with lymphadenectomy were excluded from analysis since we could not specify whether surgical outcomes were related to the sentinel node dissection or lymphadenectomy portions of the case.
Statistical analyses were performed as follows. Distribution frequencies of categorical variables were compared using the chi-square test or Fisher’s exact test, and continuous variables were compared using the Mann Whitney U test. Overall survival was evaluated following generation of Kaplan-Meier curves and compared using the log-rank test for patients who had at least 1 month of follow-up. A Cox model was constructed to control for a priori selected confounders and variables that were significantly different between comparator groups. All statistical analyses were performed with the Statistical Package for the Social Sciences v.24 statistical package (International Business Machines Corp., Armonk, New York) and the alpha level of statistical significance was set at 0.05. In accordance with the journal’s guidelines, we will provide our data for the reproducibility of this study in other centers if requested.
A total of 1583 patients were identified; 304 patients (19.2%) underwent sentinel lymph node biopsy alone, 138 (8.7%) underwent sentinel node biopsy plus lymphadenectomy, and 1141 (72.1%) underwent lymphadenectomy. Combining the sentinel node only and sentinel node with lymphadenectomy groups, a 68% relative increase in the use of sentinel node biopsy per year of diagnosis was noted; 20.3% in 2012, 27.5% in 2013, 30.2% in 2014 and 34.2% in 2015. There was no difference in utilization of sentinel lymph node biopsy alone and with lymphadenectomy between academic and non-academic facilities (52.7% vs 47.3%, p=0.12).
Patients in the sentinel node only group were less likely to have comorbidities based on the Charlson index and had smaller size tumors. Otherwise demographic data were not significantly different between the three groups (Table 1). Status of pathologically examined inguinal lymph nodes was available for 1565 patients among the three comparison groups. Patients who had sentinel lymph node biopsy plus lymphadenectomy were more likely to have positive lymph nodes compared with sentinel node biopsy only and lymphadenectomy only (28.1% vs 11% vs 19.6%, p<0.001) and to receive adjuvant radiotherapy (26.1% vs 10.9% vs 19.4%, p<0.001). When the detection of positive lymph nodes was further stratified by tumor size, there was no difference between the three groups for tumors less than 2 cm (Table 2).
With regard to perioperative outcomes within the sentinel node only and lymphadenectomy groups, patients who had sentinel node biopsy had shorter inpatient stay (median 1 vs 2 days, p<0.001) and lower rates of unplanned readmission within 30 days from surgery (1.7% vs 5%, p=0.010).
Regarding survival outcomes, the median follow-up of the lymphadenectomy, sentinel node only, and sentinel node with lymphadenectomy groups were 33.5, 31.4, and 32 months, respectively. There was no difference in overall survival between patients undergoing sentinel node biopsy only, sentinel node biopsy with lymphadenectomy, or lymphadenectomy (Figure 1). The 3 year overall survival rates from time of diagnosis were 86.3%, 77.9%, and 82.1%, respectively (p=0.26). After controlling for age, race, insurance, comorbidities, tumor size, and lymph node metastases, sentinel node biopsy only was not associated with worse survival compared with lymphadenectomy (HR 0.86, 95% CI 0.57 to 1.32), nor was sentinel node biopsy with lymphadenectomy (HR 1.21, 95% CI 0.74 to 1.98, Table 3).
We performed a sensitivity survival analysis excluding the sentinel lymph node biopsy plus lymphadenectomy group. In this secondary analysis, there was no difference in overall survival between patients undergoing sentinel node biopsy only and lymphadenectomy; 3 year overall survival rates from time of diagnosis were 86.3% and 82.1% respectively (p=0.18). We additionally assessed another variation of the Cox model using the composite sentinel node group (sentinel node biopsy alone and with lymphadenectomy). After controlling for the above variables, sentinel lymph node biopsy either alone or with lymphadenectomy was similarly not associated with worse survival compared with lymphadenectomy (HR 0.98, 95% CI 0.70 to 1.4). We included comorbidities and positive lymph nodes in the models because they were significantly different between comparator groups, and included age, race, tumor size, and insurance because they are generally clinically relevant.8–10
Our results indicate that the utilization of sentinel lymph node biopsy is increasing in the management of early vulvar cancer in the United States without impacting overall survival. This is consistent with the limited data from other population-based studies since the publication of GROINSS-V and GOG-173. Additionally, we found shorter hospital stays and decreased readmission rates with use of sentinel node biopsy.
Our results are similar to other retrospective studies demonstrating increased utilization of the sentinel node technique over the last two decades. Zigras and colleagues performed a retrospective cohort study evaluating 732 women in Ontario with vulvar cancer who had groin node assessment.11 They found an increase in the proportion of patients undergoing sentinel node biopsy, with a 35.1% increase between 2008 and 2016. While this increase is higher than we observed, the difference may be explained by the regional nature of that study, which is less likely to capture a variety of practice settings and geographical differences in utilization. The rate of sentinel node biopsy in the Zigras study was also significantly different based on institution, although there was no difference between academic and community hospitals. Similarly, Cham and colleagues published a retrospective cohort study evaluating 2273 women from the Perspective database to determine factors associated with sentinel node biopsy, although they included all patients with vulvar cancer and did not restrict patients based on tumor size or depth of invasion. They also noted a slightly higher increase in sentinel node biopsies with a 22% rise between 2006 and 2015, with women who received treatment in urban medical centers and those diagnosed more recently being more likely to have undergone the procedure.12 However, the inclusion of lower-stage vulvar cancer and the use of a smaller database likely contributed to the higher rate of utilization.
With respect to prior publications that contain cancer-specific pathology and staging information, our study similarly demonstrated no difference in overall survival outcomes between sentinel node biopsy alone, sentinel node biopsy with lymphadenectomy, and lymphadenectomy. For example, Zhou and colleagues evaluated 1475 women who underwent sentinel node biopsy or lymphadenectomy using the Surveillance, Epidemiology, and End Results cancer registry database from 2003 to 2013, and noted an increase in the number of patients receiving sentinel node biopsy between 2008 and 2013 compared with 2003 to 2007 (13.9% vs 3.7%, p<0.001).13 Similar to our demographics, the sentinel node group was more likely to have smaller tumor size. The 5 year overall survival was not significantly different between the two groups, with 77.5% in the lymphadenectomy group and 82.5% in the sentinel node group. These overall survivals are consistent with those identified in our study.
Our study also indicated an improvement in perioperative outcomes in the sentinel node only group. There have been only a few studies reporting on perioperative outcomes with regards to sentinel node biopsy. Our findings are consistent with previously published data and add strength to the utility of sentinel lymph node biopsy by way of reporting outcomes from a separate national data source. The Cham study noted a shorter median length of stay in women undergoing sentinel node biopsy compared with lymphadenectomy (2 days vs 3 days)12; our data indicated a median stay of 1 versus 2 days. Along with shorter length of stay, we have also identified improved unplanned readmission rates in the sentinel node group, which in addition to being clinically significant also helps identify key factors to include in future research on perioperative outcomes.
Our results indicated that the sentinel node biopsy with lymphadenectomy group was more likely to have positive lymph nodes and to receive adjuvant treatment despite having the same median tumor size as the lymphadenectomy only group. When we stratified our positive lymph node rates by tumor size, the significantly higher percentage of cases with positive lymph nodes in the sentinel node with lymphadenectomy group was only seen in tumor sizes of at least 2 cm. (Table 2). The fact that larger tumors have higher likelihood of positive lymph nodes has been established; GOG 173 noted only a 2.0% false-negative predictive value for tumors smaller than 4 cm, while the false-negative predictive value was 7.4% for tumors of 4–6 cm.4 While it is unclear if the higher positive node rate in this group is due to patients undergoing a lymphadenectomy as a result of a positive sentinel lymph node, patients with clinically suspicious nodes detected during sentinel node dissection who then undergo lymphadenectomy, or patients undergoing lymphadenectomy following mapping failure during sentinel node biopsy, these results highlight the importance of lymphadenectomy as the next step in the sentinel node algorithm outlined in GROINSS and NCCN guidelines2 and suggest that the guidelines as written are functioning well in real practice.
The primary strengths of our study include its large sample size, use of data from a national population-based database, inclusion of only patients with tumor characteristics favorable for the use of sentinel node biopsy, and relatively high percentage of patients undergoing the procedure. In our study, 27.9% of patients underwent sentinel node biopsy compared with 8.7% in Zhou and colleagues, thus improving our ability to evaluate overall survival. Our study also reports on the real world application of the technique for management of early stage vulvar cancer in the United States and provides reassurance of the positive survival outcomes.
Our study does have the inherent weakness of a retrospective design, as well as limitations related to the database itself. For instance, certain information such as the technique used for sentinel node biopsy, ultra-staging data, pre-operative imaging, clinical findings, and exact tumor location was not available. Additionally, the inability to ensure demographic equality between groups due to the study design resulted in potential confounders, with the sentinel node group having smaller tumor size and fewer comorbidities. However, comorbidities and tumor size were adjusted for in the overall survival model without influencing the results. Finally, median follow-up for survival was relatively short at 2.6 years; it is possible we missed later survival events.
In conclusion, our findings indicate that sentinel lymph node biopsy is increasing in the United States for the management of early vulvar squamous cell carcinoma, and is associated with improved perioperative outcomes and comparable overall survival. Our results suggest that the widespread implementation of sentinel node biopsy in practice has yielded results consistent with those seen in clinical trials. The results of GROINSS-V II, which have not yet been published, will provide additional valuable data on this topic.
Contributors All of the authors listed have made substantial contributions to conceptualization and editing of the manuscript for intellectual content, and all authors have approved the final version submitted for publication.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests RG discloses a grant from AstraZeneca, however there was no relevance to this project. EMK discloses a grant from American Cancer Society and from Tesaro, also with no relevance to this project.
Patient consent for publication Not required.
Ethics approval The study was deemed exempt by the University of Pennsylvania Institutional Review Board (ID #829268).
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement Data are available upon reasonable request. In accordance with the journal’s guidelines, we will provide our data for the reproducibility of this study in other centers if such is requested. The data is in deidentified form from the National Cancer Database.