Article Text

Download PDFPDF

Sentinel lymph node mapping in endometrial cancer: performance of hysteroscopic injection of tracers
  1. Fabio Martinelli,
  2. Antonino Ditto,
  3. Giorgio Bogani,
  4. Umberto Leone Roberti Maggiore,
  5. Mauro Signorelli,
  6. Valentina Chiappa and
  7. Francesco Raspagliesi
  1. Gynecologic Oncology, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy
  1. Correspondence to Dr Fabio Martinelli, Fondazione IRCCS Istituto Nazionale dei Tumori, Milano, Italy; Fabio.Martinelli{at}istitutotumori.mi.it

Abstract

Objective To report on the performance of hysteroscopic injection of tracers (indocyanine green (ICG) and technetium-99m (Tc-99m)) for sentinel lymph node (SLN) mapping in endometrial cancer.

Methods Single-center retrospective evaluation of consecutive patients who underwent SLN mapping following hysteroscopic peritumoral injection of tracer. Detection rate (overall/bilateral/aortic) diagnostic accuracy, and oncologic outcomes were evaluated.

Results A total of 221 procedures met the inclusion criteria. Mean patient age was 60 (range 28–84) years and mean body mass index was 26.9 (range 15–47) kg/m2 . In 164 cases (70.9%) mapping was performed laparoscopically. The overall detection rate of the technique was 94.1% (208/221 patients). Bilateral pelvic mapping was found in 62.5% of cases with at least one SLN detected and was more frequent using ICG than with Tc-99m (73.8% vs 53.3%; p<0.001). In 47.6% of cases SLNs mapped in both pelvic and aortic nodes, and in five cases (2.4%) only in the aortic area. In eight patients (3.8%) SLNs were found in aberrant (parametrial/presacral) areas. Mean number of detected SLNs was 3.7 (range 1–8). In 51.9% of cases at least one node other than SLNs was removed. Twenty-six patients (12.5%) had nodal involvement: 12 (46.2%) macrometastases, six (23.1%) micrometastases, and eight (30.7%) isolated tumor cells. In 12 cases (46.8%) the aortic area was involved. Overall, 6/221 (2.7%) patients had isolated para-aortic nodes. Three false-negative results were found, all in the Tc-99m group. All had isolated aortic metastases. Overall sensitivity was 88.5% (95% CI 71.7 to 100.0) and overall negative predictive value was 96.5% (95% CI 86.8 to 100.0). There were 10 (4.8%) recurrences: five abdominal/distant, four vaginal, and one nodal (in the aortic area following a unilateral mapping plus side-specific pelvic lymphadenectomy). Most recurrences (9/10 cases) were patients in whom a completion lymphadenectomy was performed. No deaths were reported after a mean follow-up of 47.7 months.

Conclusions Hysteroscopic injection of tracers for SLN mapping in endometrial cancer is as accurate as cervical injection with a higher detection rate in the aortic area. ICG improves the bilateral detection rate. Adding lymphadenectomy to SLN mapping does not reduce the risk of relapse.

  • endometrial neoplasms
  • sentinel lymph node
  • hysteroscopes
  • surgery
  • SLN and lympadenectomy

Statistics from Altmetric.com

Request Permissions

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.

HIGHLIGHTS

  • Sentinel lymph node detection rate and accuracy after hysteroscopic injection of tracers are similar to cervical injection.

  • Hysteroscopic peritumoral injection leads to 50% mapping in the aortic area.

  • In nearly half the patients with nodal metastases the aortic area was involved.

Introduction

Uterine cancer incidence is increasing and endometrial cancer represents the most common gynecologic neoplasm in Western countries.1–3 Surgical staging encompasses nodal evaluation.4 Sentinel lymph node (SLN) mapping has been widely used in recent years for nodal staging5 and is largely accepted as a compromise between no lymphadenectomy and full lymphadenectomy.6 7 The majority of published series report on cervical injection of tracers with a detection rate of 81%, bilateral detection rate of 50%, and aortic detection rate of 17%.5 Among the tracers used, indocyanine green (ICG) proved to be superior to blue dyes with a detection rate comparable to radiocolloid.8–10

Since 2004, our group has adopted hysteroscopic injection for SLN mapping in endometrial cancer.11–14 Here we report on the performance of hysteroscopic peritumoral injection of tracer for SLN mapping in endometrial cancer staging, focusing on detection rate, diagnostic accuracy, and patients’ outcomes.

Methods

Data from the National Cancer Institute of Milan Institutional Review Board approved and prospectively maintained database on SLN mapping in endometrial cancer were evaluated.

Data from consecutive patients who underwent SLN mapping after hysteroscopic injection of technetium-99m (Tc-99m) or ICG from January 2005 to January 2019 were analyzed. Included patients were: women aged ≥18 years, with no concomitant and/or previous cancer, and signed informed consent; women with apparent early-stage endometrial cancer who underwent: hysterectomy ± bilateral salpingo-oophorectomy and SLN mapping ± lymphadenectomy. Tumor histologic classification and staging were performed according to International Federation of Gynecology and Obstetrics (FIGO) 2009 criteria.15

Mapping Technique and Pathologic Assessment of SLNs

Detailed descriptions have been reported in previous papers.11–13 16 Briefly, Tc-99m (111 MBq eluted in 5 mL) was used until November 2013; thereafter ICG (4 mL of 1.25 mg/mL eluted ICG) became the preferred tracer. The hysteroscopic procedure was performed without cervical dilation; no more than 6 hours before surgery, without local and/or general anesthesia when Tc-99m was used; following laparoscopic evaluation under general anesthesia when ICG was the tracer. Saline was used for uterine distension. Hysteroscopic peritumoral injections were performed subendometrially: only around the clock (at 3, 6, 9, and 12) in the case of a single lesion, regardless of its location (fundus, corunal os, lower segment); while, if the entire cavity was involved by the tumor, anterior, posterior, lateral walls, and fundus were injected. To avoid the possible risk of spillage, an intracavitary pressure lower than 40 mm Hg was maintained during the hysteroscopy.17 Since the introduction of ICG, tubal occlusion at the cornual os was also carried out to avoid any peritoneal spillage during the hysteroscopic procedure. This also prevented any ICG spillage that could have masked the operating field if it all had become fluorescent. Caution was taken not to occlude the utero-ovarian/infundibolopelvic pathway. Mapping was auditory-driven in the Tc-99m group (either laparotomic or laparoscopic, since 2010) and real-time visual-driven in the ICG group (all laparoscopic).

Hysterectomy was performed before mapping in the Tc-99m group, to avoid background radioactivity disturbance, and after mapping in the ICG group to reduce surgical dissection that could impair mapping. The uterus was sent for frozen section. Completion lymphadenectomy (pelvic ± aortic) was performed according to the “diagnostic accuracy 2005–2010” study protocol.12 Patients with a stage IBG2 (FIGO 1988) or higher were submitted to sentinel node detection followed by systematic pelvic and para-aortic lymphadenectomy. Therafter lymphadenectomy was based on identification of tumor risk factors and clinical judgement. Patients with poorly differentiated EC and deeply invasive tumors underwent a completion pelvic and aortic lymphadenectomy. Patients with well-differentiated or moderately differentiated cancers confined to the inner half of the myometrium underwent SLN biopsy only. Since April 2017 only SLN mapping has been performed.

Frozen sectioning of SLNs was not routinely performed. There were only a few cases during the study period in which there was a change of tracers (transition from Tc-99m to ICG). Frozen sectioning of SLNs was done to reduce the risk of empty packets. In cases with macroscopic SLN involvement completion lymphadenectomy was performed.

Pathologic evaluation included hematoxylin & eosin (H&E) and immunohistochemistry for cytokeratins in SLNs. In detail, lymph nodes were isolated from the fatty tissue without freezing or preservation and were examined by standard techniques. Nodes of major axis >0.5 cm were dissected; nodes <0.5 cm were entirely fixed and embedded. Three sections were obtained from each node at different levels (100–500 µm apart), and stained with H&E. Immunohistochemistry for cytokeratins (AE1–AE3; Dako Corporation Glostrup, Denmark) was assessed in the same number of SLN sections that gave negative results for metastasis at H&E evaluation.

SLNs were considered positive if they contained macrometastasis or micrometastasis, or isolated tumor cells (ITCs), according to the combined evaluation using H&E and immunohistochemistry.18 In case of positive SLNs at permanent and no completion lymphadenectomy already performed, no other surgeries were carried out to complete nodal dissection.

During the study period seven surgeons were involved.

This single-center review includes all patients with endometrial cancer who underwent hysteroscopic injection of tracer for SLN mapping. Demographics and pathologic data were analyzed. SLNs were examined for the number and location of the nodes. Only first draining nodes were considered as SLNs, while ‘echelon’ or ‘secondary’ nodes were not considered as SLNs. Standard descriptive statistics were used. Accuracy was calculated only among those patients who underwent completion pelvic and aortic lymphadenectomy. The European Society for Medical Oncology (ESMO), European Society of Gynaecological Oncology (ESGO), and European Society for Radiotherapy & Oncology (ESTRO) ESMO/ESGO/ESTRO risk group classification was used: Low risk: Stage I endometrioid, grade 1–2, <50% myometrial invasion, lymphovascular space invasion (LVSI) negative; Intermediate risk: Stage I endometrioid, grade 1–2, ≥50% myometrial invasion, LVSI negative; High-intermediate risk: Stage I endometrioid, grade 3, <50% myometrial invasion, regardless of LVSI status; Stage I endometrioid, grade 1–2, LVSI unequivocally positive, regardless of depth of invasion; High risk: Stage I endometrioid, grade 3, ≥50% myometrial invasion, regardless of LVSI status; Stage II; Stage III endometrioid, no residual disease; Non-endometrioid (serous or clear cell or undifferentiated carcinoma, or carcinosarcoma). Statistics were performed using SPSS 15.0 (SPSS Inc., Chicago, IL, USA). All calculated p values were two-sided, and p values <0.05 were considered statistically significant.

Results

A total of 231 procedures were performed. Mean patient age was 60 (range 28–84) years and mean body mass index was 26.9 (range 15–47) kg/m2. Laparoscopy was the route of access in 164 cases (70.9%). Three conversions to laparotomy occurred. In 10 cases mapping was aborted (nine technical equipment failure, not related to the hysteroscopic procedure; one vagal reaction). At least one SLN was detected in 208 patients (94.1%). Overall bilateral and overall aortic mapping were 58.8% and 47.0%, respectively. Clinical and pathologic data of patients with at least 1 SLN detected are reported in Table 1. Bilateral pelvic mapping was found in 62.5% of cases (130/208) and was significantly more frequent in the ICG group than in the Tc-99m group (73.8% vs 53.3%; p<0.001). In 47.6% of cases (99/208) SLNs mapped both to pelvic and aortic nodes, and in five cases (2.4%) only in the aortic area. In eight patients (3.8%) SLNs were found in aberrant (parametrial/pre-sacral) areas. Mean number of detected SLNs was 3.7 (range 1–8). In 51.9% of cases (108/208) at least one node other than SLNs was removed. There was a significant reduction of completion lymphadenectomy over time (Figure 1).

Figure 1

Trend in completion lymphadenectomy.

Table 1

Clinical and pathologic data for the 208 patients who underwent sentinel lymph node mapping

Some 26 patients (12.5%) had nodal involvement: 12 (46.2%) macrometastases; six (23.1%) micrometastases; and eight (30.7%) ITCs.

In six cases (23.1%) only aortic nodes were positive; in six cases (23.1%) both pelvic and aortic nodes and in 14 cases (53.8%) only the pelvic area was involved. Among patients with nodal involvement, 22 belonged to the high-intermediate-risk group on uterus according to ESMO-ESGO-ESTRO, three were in the high-risk group, and only one was in the low-risk group (ITCs found on SLNs) (Table 2).

Table 2

Uterine risk factors and node metastases

Three false-negative results were found, all in the Tc-99m group. All these patients belonged to the high-intermediate-risk group on uterus (grade 2 with LVSI) and myometrial infiltration >50%). All had isolated aortic metastases with negative pelvic nodes after full pelvic and aortic dissection: two patients with macrometastases (one also had peritoneal implants) and one with micrometastatisis in one node only.

Overall sensitivity of the technique was 88.5% (95% CI 71.7 to 100.0) and overall negative predictive value (NPV) was 96.5% (95% CI 86.8 to 100.0).

Adjuvant treatment was administered in 57 (27.4%) patients. Fifty patients (87.7%) received radiation therapy ± chemotherapy; in seven cases (12.3%) chemotherapy only was the choice.

There were 10 (4.8%) recurrences: five abdominal/distant, four vaginal, and one nodal (in the aortic area following a unilateral mapping plus side-specific pelvic lymphadenectomy). Most recurrences (9/10 cases) were patients in whom a completion lymphadenectomy was performed. Characteristics of the patients who relapsed are detailed in Table 3.

Table 3

Characteristics of the patients who relapsed

No deaths were reported after a mean follow-up of 47.7 months.

Discussion

SLN mapping is widely used and has gained acceptance for staging purposes in endometrial cancer.5 6 Uterine cervix became the most popular injection site thanks to its ease of access and reproducibility of the technique. Overall the SLN detection rate ranges from 70% to 100%, being superior to 90% in series with ICG. Furthermore, ICG increased the bilateral detection rate (range 65%–85%) compared with other tracers.5 6 8 19 The same figures were found in our series with an overall detection rate of 94.1% and significantly more frequent bilateral mapping in the ICG group than in the Tc-99m group (73.8% vs 53.3%; p<0.001). Recently Faranzestaninan et al20 in a head-to-head comparison of intracervical and fundal injection of tracers found a 97.2% concordance when a successful dual-tracer mapping was achieved for hemipelvis. This finding corroborates the fact that cervical and fundal injections depict the same lymphatic pelvic pathways. The main difference between fundal and cervical injections was the rate of aortic mapping (22.2% vs 4.4%, respectively).

The aortic detection rate following cervical injection ranges from 7% to 20%, while it is higher (27%–30%) after uterine injection.5 6 Following hysteroscopic peritumoral injection we found an aortic mapping in 50% of patients.

Two arguments deserve attention regarding this point.

First, the question whether or not the so-defined SLNs are real SLNs (first in chain) or echelon nodes. Geppert et al21 studying lymphatic anatomy found a 47% SLN detection rate in the aortic area after fundal injection of ICG (following the infundibolopelvic pathway). However, the same group, in a recent study (SHREC-trial) analyzing pelvic SLN detection in high-risk endometrial cancers after cervical injection of ICG, found no migrations via the infundibolopelvic pathway nor the inframesenteric SLNs (“that would have required the absence of nodes along the pelvic pathways”).22 A clear definition on SLNs is mandatory. If with Tc-99m there is no way to see lymphatics, the introduction of ICG allows one to visualize them. Thanks to real-time visualization, ICG permits one to clearly follow lymphatic channels, thus reducing the risk of considering secondary nodes as SLNs. In case of doubt (echelon vs true SLN) channels were followed back up to the uterus, to make sure that there were no other nodes between the womb and the supposed SLN. Ruiz et al, in aiming to increase the aortic detection rate, used a dual fundal/cervical injection of ICG and found nearly 60% of aortic mapping.23 A Korean group reported an even higher aortic detection rate (86%) following two-step (corunal and cervical) ICG injection.24 Our 50% mapping in the aortic area is in line with reported data, considering that the hysteroscopic injection was peritumoral: not all tumors are on the fundus or coronal os and lower aortic mapping (along infundibolopelvic pathways) was seen for lower uterine lesions.

Second, are there any advantages in detecting more SLNs in the aortic area? The point is not the 2%–3% risk of isolated aortic metastases (2.8% in our series), but whether there is any advantage in knowing whether or not aortic nodes are involved (FIGO stage IIIC2). Recently published data using SLN in high-risk endometrial cancer found cervical injection accurate for pelvic mapping (NPV 99%). However, after aortic lymphadenectomy, more than 40% of node-positive patients also had positive aortic nodes, which were not SLNs.22 25 In our series, almost half the patients with nodal involvement had aortic metastases. Among patients with only positive pelvic nodes, 71.4% (10/14) had successful negative aortic mapping; 21.4% (3/14) without aortic mapping underwent aortic lymphadenectomy without evidence of nodal involvement. Globally the aortic area was adequately evaluated in 92.8% (13/14) of cases. Whether aortic lymphadenectomy could have a therapeutic role is a debatable point, as well as what is the best adjuvant therapy for node-positive patients. Multicenter retrospective series have not identified differences in survival in women with endometrial cancer treated with an SLN mapping procedure or full staging lymphadenectomy.26 27 Undoubtedly having aortic nodal involvement (FIGO stage IIIC2) has a prognostic impact. In a recent Surveillance Epidemiology and End Results (SEER) data analysis of over 3650 women with stage IIIC endometrial cancer, stage IIIC2 was associated with higher all-cause mortality (HR 1.44, 95% CI 1.22 to 1.69) and endometrial cancer-specific mortality (HR 1.49, 95% CI 1.25 to 1.77) compared with stage IIIC1. The authors suggest performing a “selective aortic lymphadenectomy in those women who are most likely to have aortic LN metastasis as an adjunct to sentinel pelvic LN dissection” if no aortic mapping occurred.28

In our study we had three false-negative results, all in the Tc-99m group (prior to December 2013). Whether a false-negative rate of 11.5%, entirely due to isolated aortic metastases, in respect to the high (around 50%) SLN detection rate in this area, might suggest a low accuracy of the method for the aortic area merits further investigation. The progressive implementation of SLN mapping and the introduction of ICG coupled with minimally invasive equipment (that permits real-time mapping) has significantly reduced the number of completion lymphadenectomies (Figure 1). Theoretically a higher ‘true’ SLN detection in the aortic area would enhance staging purposes and reduce the number lymphadenectomies performed.

Increasing the accuracy of nodal staging also leads to increased identification of nodal metastases. Similar to other reports,5 6 more than half of the node-positive patients (53.8%) in our cohort had only low-volume disease (micrometastases and ITCs). This SLN-oriented ‘enhanced staging’ may lead to a better tailoring of subsequent adjuvant treatment prescription, when this is needed. However, the exact role of ITCs has still to be determined.29 It is likely that SLN mapping reduces the risk of recurrences. In a Canadian cohort of 472 patients, women submitted to SLN mapping followed by pelvic lymphadenectomy (n=275) had a lower recurrence rate (30% vs 71.4%) with a significantly higher pelvic recurrence-free survival compared with women who had systematic pelvic lymphadenectomy alone (n=197).30 In our series, 10 patients suffered from recurrences. No pelvic sidewall relapses were recorded and only one nodal relapse (in the aortic area following SLN mapping algorithm application) was identified.

Since the introduction of SLN mapping in endometrial cancer there has been debate as to which is the best injection site. The majority of groups adopted the cervical route which has a high pelvic detection rate and accuracy and is easily reproducible. Fewer groups explored the uterine/fundal injection which has a higher aortic detection rate, but a longer learning curve.6 Here we report on the largest series ever published of SLN mapping following hysteroscopic peritumoral injection of tracers in a ‘real-life’ setting. Changing the tracer (from Tc-99m to ICG), equipment evolution (gamma probes for laparotomy/laparoscopy; near-infrared equipment), and the learning curve of involved medical staff were all included in this series. Despite these drawbacks, detection rates (overall and bilateral) and accuracy of the technique are comparable with data on cervical injection (NPV 96.5% vs 99.7%).5 There were no missed positive pelvic nodes. The hysteroscopic injection leads to a high aortic mapping (nearly 50%). Use of ICG significantly ameliorates bilateral mapping and depiction of lymphatics routes, reducing the risk of echelon node sampling. The majority (96.1%) of patients with nodal involvement were in the high/high-intermediate-risk groups and nearly half of these had aortic metastases. If we assume that SLN mapping is the standard procedure for endometrial cancer staging and look at cervical injection (with one caveat being the low aortic detection rate),6 with the aim of achieving a thorough evaluation of the aortic area, in order to correctly stage patients and drive any further treatments,31 hysteroscopic injection could play a role in this process.

Tips and Tricks for Hysteroscopic ICG Injection and Mapping

  • Start the laparoscopic/robotic procedure to ensure no evidence of peritoneal disease is present, trying to avoid any dissection before injecting tracer.

  • Perform tubal cornual occlusion (bipolar, clips, ultrasound, etc.) without damaging the utero-ovarian/infundibolopelvic pedicles.

  • Start the hysteroscopic procedure, if possible using dedicated equipment for hysteroscopy, so as to be under laparoscopic/robotic visualization.

  • No cervical os dilatation is needed, nor cervical grasping; let the uterus move to avoid false passages or overintroduction/uterine perforation.

  • Use an operative hysteroscope with a working channel that allows introduction of a needle (ie, 40 mm 22G) and has a high workflow of the distension medium.

  • Using saline as distension medium permits adequate lavage of the cavity without risk of overload syndrome.

  • Maintain intrauterine pressure as low as possible to avoid any spillage (this can be also controlled under near infrared vision via laparoscopic/robotic visualization).

  • Wash the uterine cavity as much as possible in order to adequately see tumor diffusion or the site of implants, reducing unnecessary movements (cancer usually bleeds a lot).

  • Make injection peritumoral, subendometrially starting from the highest lesion (ie, fundus/cornual os) and then going back towards the cervix. Myometrium is stiffer than endometrium and cancer. However, observe carefully from the laparoscopic/robotic access to be sure not to make an overintroduction of the needle beyond the uterine serosa.

  • Under near infrared vision, follow the coloration of the lymphatic channels to see in real time where the tracer goes.

  • Undock the hysteroscopic procedure and commence SLN mapping.

  • Overall: be patient!

Acknowledgments

The authors thank Edward Hausler, MD, for his assistance with editing and revising the English in their manuscript.

References

Footnotes

  • Editor's note This paper will feature in a special issue on sentinel lymph node mapping in 2020.

  • Twitter @Fabio Martinelli@DrFMartinelli

  • Contributors FM, AD, FR: conception and design of the work. FM, AD, GB, ULRM, MS, VC: data acquisition. FM: data analysis and interpretation. FM, FR: drafting and revising the paper. FM, FR: final approval of the paper.

  • 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 None declared.

  • Patient consent for publication Not required.

  • Provenance and peer review Commissioned; externally peer reviewed.

  • Data availability statement All data relevant to the study are included in the article or uploaded as supplementary information.