Article Text

Download PDFPDF

Technetium-99m-indocyanine green versus technetium-99m-methylene blue for sentinel lymph node biopsy in early-stage endometrial cancer
  1. Silvia Cabrera1,2,
  2. Vicente Bebia1,
  3. Silvia Franco-Camps1,
  4. Cristina Forcada1,
  5. Diego Villasboas-Rosciolesi3,
  6. Ignacio Navales3,
  7. Assumpció Pérez-Benavente1 and
  8. Antonio Gil-Moreno1,2
  1. 1 Gynecologic Oncology Unit, Gynecology Department, Hospital Universitari Vall d’Hebron, Vall d’Hebron Barcelona Hospital Campus, Barcelona, Spain
  2. 2 Universitat Autònoma de Barcelona, Barcelona, Spain
  3. 3 Nuclear Medicine Department, Institut de Diagnòstic per Ia Imatge, Vall d’Hebron Barcelona Hospital Campus, Barcelona, Spain
  1. Correspondence to Dr Silvia Cabrera, Gynecologic Oncology Unit, Gynecology Department, Hospital Universitari Vall d’Hebron, Vall d'Hebron Barcelona Hospital Campus, Barcelona, Spain; scabrera{at}


Introduction The low accuracy of pre-operative imaging techniques for prediction of nodal status strengthens the relevance of sentinel lymph node (SLN) biopsy in endometrial cancer, although the optimal method for its detection is still under investigation. The increasing use of indocyanine green (ICG) has aroused concern about its enhanced visualization of lymphatic channels, which could lead to a specimen that is thought to be nodal tissue not subsequently yielding a lymph node on pathologic analysis ('empty node packet'). Our main objective was to compare the overall and bilateral detection rates for SLN biopsy using two combined techniques: technetium-99m-ICG (Tc-99m-ICG) versus technetium-99m-methylene blue (Tc-99m-MB). Our secondary aim was to compare the ‘empty node packet’ rates between the two cohorts.

Methods A prospective, non-randomized, single-center trial including patients diagnosed with endometrial cancer (any grade or histology) in pre-operative early stage, and operated on between February 2017 and July 2019. All tracers were injected intracervically. Pelvic and aortic lymphadenectomy were performed on patients at intermediate or high risk of recurrence pre-operatively. All SLNs were sent for intra-operative frozen section and afterwards processed following an ultrastaging protocol.

Results Eighty-four patients were included, 58% (n=49) in the Tc-99m-MB group and 42% (n=35) in the Tc-99m-ICG group. Overall detection rate was 93% and was not statistically different between the two groups. A better bilateral detection rate was observed among Tc-99m-ICG patients (69% vs 41%, p=0.012). The 'empty node packet' rate was 4% in the Tc-99m-ICG cohort and 0% in the Tc-99m-MB cohort (p=0.032).

Discussion Tc-99m-ICG is a feasible, safe technique for SLN biopsy in early-stage endometrial cancer, and appears to be superior in terms of bilateral detection to Tc-99m-MB. The addition of Tc-99m to ICG could decrease the rate of 'empty node packets' and better define the anatomic location of SLNs in patients with endometrial cancer.

  • sentinel lymph node
  • surgical oncology
  • endometrial neoplasms

Statistics from

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.


  • Overall detection rate is similar between technetium-99m-indocyanine green (Tc-99m-ICG) and technetium-99m-methylene blue.

  • Combined Tc-99m-ICG improves bilateral pelvic sentinel lymph node detection rates.

  • Adding Tc-99m to ICG may decrease the 'empty node packets' described when using ICG alone.


Endometrial cancer is the most common gynecologic cancer in Europe,1 and lymph node status remains one of the main prognostic factors. However, two randomized trials and one meta-analysis found no evidence as to whether lymphadenectomy decreased the risk of death or recurrence in women with pre-operative stage I endometrial cancer. Furthermore, women who underwent lymphadenectomy were more likely to suffer from systemic morbidity, lymphedema, or lymphocyst.2–4 In this context, sentinel lymph node (SLN) biopsy has become of increasing interest in endometrial cancer patients,5 who are at low risk of having lymph node disease,6 and are usually obese and affected by multiple comorbidities.7 The low accuracy of pre-operative imaging techniques for prediction of nodal status strengthens the potential relevance of SLNs in endometrial cancer.8

According to some practice guidelines, SLN biopsy can be considered for surgical staging of patients with pre-operative initial-stage endometrial cancer.9 10 The accurate application of the Sloan-Kettering Cancer Center algorithm reduces the false-negative rate to less than 2%.11 The use of ultrastaging protocols increases the detection rate of low-volume disease (micrometastases and isolated tumor cells) even when its clinical impact is not clear.12 Patients with unknown nodal status are more frequently subjected to adjuvant therapies when compared with patients with nodal study.6

Cervical injection has emerged as a useful technique for bilateral pelvic SLN identification in endometrial cancer. The cervix is easily accessible and rarely distorted by anatomic variations or bulky tumors in these patients. However, cervical injection is associated with a significant lower rate of aortic detection compared with uterine injection.13

The traditional tracers used for SLN mapping are blue dyes and radiolabeled tracers. The main problem with blue dyes is that they quickly travel further within the lymphatic system and stain second-level nodes.5 Conversely, technetium-99m (Tc-99m) tends to remain in the SLNs.5 Radiolabeled tracers allow pre-operative imaging with planar scintigraphy and single-photon emission computed tomography (SPECT) combined with computed tomography (CT).14 The pre-operative SPECT/CT increases the number of detected SLNs and provides anatomic information about their location and relation to other abdominal structures.15 More recently, fluorescing dyes such as indocyanine green (ICG), detected by the use of near infrared imaging technologies, have emerged as feasible alternatives to the traditional methods for SLN mapping in endometrial cancer.16 17 One of the main disadvantages of using ICG is its enhanced visualization of lymphatic channels and its rapid extravasation when the surgeon starts the dissection. This could lead to a specimen that is thought to be nodal tissue not subsequently yielding a lymph node on pathologic analysis ('empty node packet').18

The aim of this study was to compare the overall and bilateral detection rate for SLN biopsy in early-stage endometrial cancer using two combined techniques: Tc-99m-ICG versus Tc-99m-methylene blue (Tc-99m-MB). The secondary objective was to compare the ‘empty node packet’ rates between cohorts.


Patient Population

From February 2017 to July 2019, all consecutive patients aged over 18 years and diagnosed with pre-operative initial-stage endometrial cancer (International Federation of Gynecology and Obstetrics (FIGO) 2009 stage I-II, any grade or histology)19 or with endometrial atypical hyperplasia were enrolled in this study. Patients were excluded if they had evidence of extrauterine disease, previous pelvic lymphadenectomy or radiotherapy, or were not suitable to undergo surgery with curative intention.

Patients were assigned to the Tc-99m-ICG or Tc-99m-MB group depending on the availability of ICG-detector cameras or the patient's medical characteristics (thyroidal hyperfunction or tumors, allergic to iodine or ICG were excluded for ICG detection, allergic to methylene blue were excluded for this tracer). Data were collected prospectively and maintained in an Excel (Microsoft, Redmond, USA) based electronic database. The present study was approved by the Clinical Research Ethics Committee of Hospital Universitari Vall d’Hebron. Written informed consent was obtained from all patients.

Tracer Injection and Detection Techniques

Tc-99m-nanocolloid (Nanocoll, Gipharma, Saluggia, Italy) was injected intracervically with four injections of 0.2 mL at 3 and 9 o’clock, deep (1–1.5 cm) and superficially (1–2 mm), the day before the surgery (mean 148±45 MBq). Pre-operative imaging with SPECT/CT was performed at 2 hours post-injection with a hybrid SPECT/CT device (Discovery 670 Pro, GE Healthcare, Chicago, IL, USA). Image reconstruction and fusion were performed using a Xeleris 3.0 software package (GE Healthcare, Chicago, IL, USA). SLNs were localized intra-operatively using a gamma probe (Europrobe, Model 3.2, Eurorad, Eckbolsheim, France).

ICG (four injections of 1 mL, ICG-Pulsion 5 mg/mL, Pulsion Medical Systems, Feldkirchen, Germany) and MB (four injections of 1 mL at 0.25%, MB 10 mg/mL, Xalabarder Farma, Barcelona, Spain) were also injected intracervically at 3 and 9 o’clock, deep and superficially, after the anesthetic induction. SLNs were detected intra-operatively by laparoscopic or robotic fluorescence cameras in the ICG cohort, or visually in the MB cohort.

Surgical Procedures

All patients were classified pre-operatively into three recurrence risk groups according to their imaging tests and pre-operative biopsy. Low-risk patients (endometrioid grade 1–2, stage IA) underwent hysterectomy, bilateral salpingo-oophorectomy, peritoneal and retroperitoneal exploration, and SLN biopsy. Intermediate-risk (endometrioid grade 1–2 stage IB-II or endometrioid grade 3 stage IA) and high-risk (endometrioid grade 3 stage IB-II or non-endometrioid histology) also underwent a comprehensive pelvic and aortic lymphadenectomy. In some cases, intra-operative frozen section of the uterus was performed for more accurate indication of lymphadenectomy. Some patients in the intermediate-risk or high-risk groups were not candidates for aortic lymphadenectomy due to advanced age or comorbidities, and in these cases a bilateral pelvic lymphadenectomy was performed.

Pathologic Evaluation

All SLNs were sent for intra-operative frozen section. Negative nodes were paraffin-embedded for ultrastaging protocol. Stains with hematoxylin and eosin (H&E) and immunohistochemistry using AE1/AE3 (Dako, Glostrup, Denmark) were performed. Non-SLNs were examined by routine H&E. Disease staging after obtaining definitive histology results was in accordance with the 2009 FIGO classification.19

Data Analysis

Quantitative variable distributions were tested for normality by using a goodness-of-fit test (Shapiro-Wilk), and described by their mean and SD or their median and IQR depending on whether they were normally or non-normally distributed, respectively. Qualitative variable distributions were described based on percentage and size. Multivariate regression analysis was performed in order to control the interaction and confusion of the significant variables submitted to analysis. Statistical significance at p<0.05 was considered for a risk α=5%. Statistical analysis was performed using Stata statistical software v14.2 (StataCorp, College Station, TX, USA).


From February 2017 to July 2019, 84 consecutive patients diagnosed with endometrial cancer (any grade or histology) or endometrial atypical hyperplasia were enrolled in this study, 49 patients (58%) in the Tc-99m-MB group and 35 patients (42%) in the Tc-99m-ICG group.

Patient Characteristics

The mean age of the general cohort was 63.1 years, and mean body mass index (BMI) was 30.7 kg/m2. The majority (86%) of the patients were diagnosed with endometrioid adenocarcinoma, and 81% were grade 1–2. After pre-operative studies, 58% of the patients were classified as pre-operative low-risk, 24% as intermediate-risk, and 13% as high-risk. No differences between the two cohorts were found in terms of age, BMI, tumor diameter, lymphovascular invasion, or other histologic characteristics.

A minimally invasive approach was preferred in most (96%) cases. Both groups were comparable regarding surgical variables, except for a larger proportion of patients undergoing robotic surgery in the Tc-99m-ICG group (71% vs 33%, p=0.001) (Table 1). There were no cases of allergic reactions to any of the tracers.

Table 1

Demographic and clinical variables of the patients included in the analysis

After the surgery nine patients (11%) were upstaged to stage III, five of them because of nodal invasion. Two patients (50%) with endometrial atypical hyperplasia were diagnosed with endometrial cancer.

SLN Mapping Accuracy

According to the final pathology reports, the overall detection rate was 93%, and was not statistically different between groups (Tc-99m-MB 94% vs Tc-99m-ICG 91%, p=0.69). We found six patients (7%) with absence of drainage for both detection techniques, three patients in each group (p=0.09).

Bilateral pelvic detection was obtained in 69% patients in the Tc-99m-ICG cohort and in 41% patients in the Tc-99m-MB cohort, this difference being statistically significant (p=0.012). We observed a higher rate of aortic mapping in the Tc-99m-MB group (18% vs 6%), although this difference did not reach statistical significance (p=0.11).

Four surgical specimens believed to be SLNs were found not to contain any lymphatic tissue in definitive histology. All of these cases were reported in the TC-99m-ICG group, and all of them were excised by ICG detection alone and were intra-operatively 'cold' for Tc-99m. The incidence of ‘empty node packets’ in the Tc-99m-ICG group was 4% and it was significantly higher than in the Tc-99m-MB cohort (0%, p=0.032) (Table 2).

Table 2

Sentinel lymph node biopsy variables

When assessing which factors could be associated with a better bilateral detection rate, only Tc-99m-ICG detection method and younger age remained statistically significant after performing multivariate analysis (Table 3).

Table 3

Univariate and multivariate analysis of factors associated with bilateral detection in our cohort

We detected seven positive SLNs in five patients (6%), all in the Tc-99m-MB cohort. All of them were detected with Tc-99m, either intra-operatively or pre-operatively with SPECT/CT. MB detected six out of the seven positive SLNs.

Only 23 patients (27%) received a complete pelvic and aortic lymphadenectomy, therefore the sensitivity and negative predictive value of the technique cannot be calculated in our study.


The concept of SLN biopsy is rapidly emerging in the surgical management of initial-stage endometrial cancer. However, the optimal technique for this procedure is yet to be defined. The superiority of ICG over blue dyes has been widely reported.20–22 In a recent prospective trial published by Backes et al, 204 patients were enrolled to receive either ICG or blue dye. The authors reported an overall detection rate of 90% and a bilateral detection rate of 68%. ICG optimally mapped 83% of hemipelvis and blue dye 64% (p<0.0001).23 When comparing ICG versus combined Tc-99m-MB, a significant increase in bilateral detection rate was reported with ICG. In a retrospective multicenter study that included 342 patients, the overall detection rate was similar between patients undergoing SLN mapping with Tc-99m-MB versus ICG (97%, NS), although bilateral detection was significantly higher for patients mapped with ICG (84% vs 73%, p=0.03).5 These results have also been confirmed in three meta-analyses.13 24 25

The combination of ICG and Tc-99m reported in this trial has been scarcely explored. Recently, Kessous et al reported a series of 157 patients who were randomly assigned to receive ICG and Tc-99m or ICG, Tc-99m, and blue dye. They reported an overall detection rate of 97% and a bilateral detection rate of 81% within the cohort of ICG and Tc-99m, and concluded that the use of a triple combination of tracers does not improve any of the parameters obtained with ICG and Tc-99m.26

We found some advantages in the combined use of Tc-99m and ICG. First, the possibility of obtaining a pre-operative SPECT/CT, that provides the surgeon with accurate anatomic information and facilitates the process of intra-operative SLN resection.27 We found an 82% correlation between pre-operative SPECT/CT and intra-operative detection with a gamma probe, in concordance with other studies.14 28 Due to this acknowledged discordance, careful exploration of both pelvic and aortic regions with the gamma probe is recommended during surgery, especially in those patients in which no drainage is observed in SPECT/CT.

Second, the supposed reduction of false-negative rates. In our study, all the positive SLNs were 'hot' for Tc-99m. Although our study had few cases, other series have reported that Tc-99m has the higher sensitivity in detecting affected SLNs. In the study of How et al, Tc-99m was the only tracer that detected all 13 SLNs containing metastatic disease (ICG detected 77% and MB 53%).29 Similarly, Elisei et al reported a series of 40 stage I endometrial cancer patients who underwent Tc-99m-MB mapping. The rate of lymphatic metastases in their study was 16% (nine pathologic SLNs), and all of them were detected with Tc-99m (seven were 'hot' and blue, two were only 'hot').28 Conversely, in two meta-analyses the pooled analysis of sensitivity and false-negative rates showed no difference between tracers.13 25

Finally, a reduction in the factors associated with bilateral detection failure when using an isolated tracer. In the case of Tc-99m, mapping failure by both pre-operative SPECT/CT imaging and intra-operative gamma probe is mainly associated with incorrect tracer injection,14 30 that leads to major venous drainage or peritoneal extravasation. With regard to ICG, incorrect tracer injection is reported less frequently as a cause of mapping failure since the injection can be repeated if necessary. Prior studies have demonstrated a higher bilateral detection rate failure in patients with increased BMI, who require lysis of adhesions at the beginning of surgery or with clinically enlarged nodes.23 31 32 There is increasing concern about the phenomenon of ‘empty node packet’. Bedyńska et al reported that almost 40% of the detection failure with ICG was due to ‘empty packets’.30 Thomaier et al found that in the first 25 procedures the ‘empty node packet’ rate was 20%, compared with 7% in subsequent procedures.18 The FILM trial reported absence of nodal tissue in 5% of the specimens detected with ICG and 0% of those detected with blue dye.16 We found an ‘empty node packet’ rate of 4% in our Tc-99m-ICG cohort, which is lower than the rates reported by other authors. The combination with Tc-99m could be the cause of the decrease in this rate, when compared to other series that report higher rates using ICG alone. As in the FILM trial, this phenomenon was only associated to the use of ICG in our study.

In conclusion, Tc-99m-ICG is a feasible, oncologically safe technique for SLN biopsy in early-stage endometrial cancer patients, and appears to be superior in terms of bilateral mapping to Tc-99m-MB. The use of Tc-99m in combination with ICG gives pre-operative imaging information and might decrease the ‘empty packets’ rate. In our opinion, this association could be instrumental in improving the bilateral detection rate and decreasing the reported rates of ‘empty packets’ in patients with early-stage endometrial cancer and consequently should be further investigated.



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

  • Contributors SC is the principal investigator, designed the study, reviewed the literature, collected and analyzed data, and wrote the paper. VB, SF, and CF contributed to the collection of data, critical revision of the manuscript, and read and approved the final draft. VB performed the statistical analyses. DVR and IN performed the nuclear medicine studies, reviewed the manuscript, and approved the final draft. APB and AGM reviewed the manuscript for intellectual and scientific content, and approved the final draft.

  • 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.

  • Ethics approval Comité Ético de Investigación Clínica Y Comisión de Proyectos de Investigación del Hospital Universitari Vall d’Hebron, ID: PR(AMI)389/2016.

  • Provenance and peer review Commissioned; externally peer reviewed.

  • Data availability statement All data relevant to the study are included in the article; more information related to this work is available upon reasonable request to the corresponding author.