The SENTIREC-endo study – Risks and benefits of a national adoption of sentinel node mapping in low and intermediate risk endometrial cancer

Objective. The SENTIREC-endo study aims to investigate risks and bene ﬁ ts of a national protocolled adoption ofsentinellymphnode(SLN)mappinginwomenwithearly-stagelow-gradeendometrialcancer(EC)withlow-(LR) and intermediate-risk (IR) of lymph node metastases. Methods. We performed a national multicenter prospective study of SLN-mapping in women with LR and IR EC from March 2017 – February 2022. Postoperative complications were classi ﬁ ed according to Clavien-Dindo.

• The change in clinical practice contributed to a more correct treatment allocation independent of myometrial invasion.
• SLN mapping carries a very low risk of perioperative complications in women with early-stage low grade endometrial cancer.• The risk of early lymphedema after SLN mapping is low, with an incidence of swelling in one or both legs of 5.2%.• The national adoption of SLN mapping for early-stage endometrial cancer prompted significant benefits and negligible harm.

G R A P H I C A L A B S T R A C T
a b s t r a c t a r t i c l e i n f o

Introduction
Surgical staging of endometrial cancer (EC) serves to allocate women with lymph node metastases to adjuvant therapy.Sentinel lymph node (SLN) mapping represents a sensitive staging procedure that enables the identification of small-volume metastases [1].It remains unknown whether a national adoption of SLN mapping with several centers and surgeons involved, to all women with low-grade EC, prompts more benefits than harms.Prospective trials on SLN mapping in women with low-grade EC have shown high sensitivity and low false-negative rates [2,3] and a recent meta-analysis reported a negative predictive value of 100% [4].Preceded by a surgeon proficiency study, the Danish prospective SENTIREC-endo study undertook a national protocolled adoption of the SLN mapping technique for women with low grade EC.To vouch for the implementation of a new surgical approach, the associated risks should be minimal.
Most women with EC are diagnosed with early-stage, low-grade endometrioid adenocarcinoma.The majority of these women have <50% myometrial invasion and represent a low-risk (LR) group regarding metastases and recurrence.Previous international guidelines did not recommend surgical lymph node staging for women with LR EC, and a systematic implementation of SLN mapping entails extended surgery for this large group.On the other hand, implementation of SLN mapping to women with LR EC is likely to identify lymph node metastasis with high accuracy and has an expected low risk of surgical harm [4].
The depth of myometrial invasion influences the risk of lymph node metastases [5].No imaging modality has proven safe to accurately identify women with ≥50% myometrial invasion [6], so-called intermediate risk (IR) EC.Two randomized trials did not find any survival benefit of routine systematic pelvic lymph node dissection (PLD) in this group [7,8].Identification of lymph node metastasis after PLD for IR EC is estimated to be 10% [9].Applying the SLN mapping technique with ultrastaging of SLNs allows the diagnosis of micro-metastases and isolated tumor cells while the meticulous mapping allows the identification of metastases outside the usual areas of PLD.This improves the allocation to adjuvant treatment compared to PLD [10].
The majority of women with EC are elderly, with a high BMI and significant comorbidities [11].These are all well-known risk factors for surgical complications and postoperative lymphedema.The major risk factors for lymphedema are the number of lymph nodes removed and BMI [11].After PLD, the risk of lymphedema is reported to vary from 18 to 49% [12,13].The wide range is likely due to non-standardized assessment of lymphedema.Only a few series have reported on the risk of lymphedema after SLN mapping in women with EC [12][13][14].As for other cancer diagnoses, the risk of lymphedema is expected to be low after SLN mapping only.
This study investigates the benefits and potential side effects of a national adoption of the SLN technique in women with early-stage low-grade LR and IR EC.We hypothesize that the risk of surgical complications is comparable to the risk of hysterectomy alone and that the incidence of lower leg lymphedema after SLN mapping does not exceed 10%.

Study design and participants
This is a national prospective multicenter cohort study.Women with presumed FIGO (International Federation of Gynecology and Obstetrics) stage I low-grade histology with LR or IR risk of lymph node metastases (Endometrioid adenocarcinoma grade 1 or grade 2, with </≥50% myometrial invasion) were eligible.Ineligibility criteria were dementia, inability to understand Danish, allergic to iodine, active treatment for another malignancy within the past 5 years, or if PLD was discarded or previously performed.Women were included from March 2017 until March 2022 at four gynecological cancer centres in Denmark: Rigshospitalet, Herlev Hospital, Aarhus University Hospital, and Odense University Hospital.All went through a protocolled surgeon proficiency study with a minimum of 30 SLN procedures until they reached a total SLN detection rate of 80% [15].Surgery of EC in Denmark was centralized to six centers in 2012 and the participating centers in this study undertake the surgical treatment of 76% of EC patients in Denmark [16].
The study was approved by the Regional Committees on Health Research Ethics for Southern Denmark (S-20150207) and the Data Protection Agency (15/52037).Study data were collected and managed using REDCap (Research Electronic Data Capture) [17,18] tools hosted at Odense Explorative Network (OPEN).

Surgical procedure
Surgery was performed as robotic-assisted laparoscopic surgery using the da Vinci Si or Xi Surgical System (Intuitive Surgical, Sunnyvale, CA, USA).SLN mapping followed an international algorithm with the excision of all SLNs, removal of clinically suspicious lymph nodes, and side-specific PLD in cases with no SLN mapping in a hemipelvis [19].However, to acknowledge the low risk of lymph node metastases in women with <50% myometrial invasion, cases with no mapping in one or both hemipelvises underwent pathological intra-operative assessment of myometrial invasion.Women with ≥50% myometrial invasion underwent PLD in the non-mapped hemipelvis while women with <50% myometrial invasion did not have further PLD (Fig. 1).

SLN mapping
Indocyanine Green (ICG) was used as tracer.A vial of 1.25 mg/ml ICG was produced by diluting 25 mg ICG with 20 ml of sterile water.With the woman anesthetized and immediately before surgery, 4 ml of the vial was slowly injected, 1 ml submucosally (<5 mm), and 1 ml deep in the cervical stroma (≥10 mm) on each side of the cervix, at positions 3 and 9 o'clock.The near-infrared (NIR) fluorescence imaging system was used to identify SLNs.SLNs and clinically suspicious lymph nodes were removed in endobags and marked with predefined anatomic locations and sent separately for final pathology.Repeated ICG injection was not allowed as per protocol.

Ultrastaging
All SLNs were examined using a national standardized ultrastaging protocol [15].After excision, the SLNs were fixed in formalin, and then gross sectioned at 2 mm intervals perpendicular to the long axis.Each 2 mm gross section was further cut into microscopic levels, obtained at a 350 μm interval.On each level, three consecutive sections of five μm were obtained.The first section was stained with routine hematoxylin and eosin (HE), the second section was used for immunohistochemical staining for cytokeratin AE1/AE3, and the third section was available for additional analysis.Metastases were categorized according to international standards as macro-metastasis (MAC) >2.0 mm, micrometastasis (MIC) >0.2 to ≤2.0 mm, and isolated tumor cells (ITC) ≤0.2 mm [20].Only the largest metastasis in each lymph node was reported.
If metastases were identified by final pathology in SLNs or non-SLN, a postoperative FDG-PET/CT was performed.If further metastases were suspected, a second surgical procedure was performed to obtain surgical radicallity for stage III-IV disease.During the first years of inclusion, backup PLD was performed for stage II disease since these patients were eligible for another ongoing study protocol.

Intraoperative and postoperative complications
All surgeons completed a standardized surgical report shortly after surgery including a description of any intraoperative complications.The categories of intraoperative complications were; intestinal damage, damage to a larger blood vessel, damage to the urinary tract, or other damage.Postoperative complications were graded according to the Clavien-Dindo classification [21].If a postoperative complication was registered at the local site within 30 days, the medical file was accessed by an experienced medical doctor to correctly classify the event.

Lymphedema
Lymphedema was assessed prospectively by validated patientreported outcome measures (PROMs).The women completed the PROMs before surgery (baseline) and 3,6,12, 24, and 36 months postsurgery.In the present study, we use the lymphedema symptom score from the validated disease-specific Endometrial cancer Module from the European Organisation for Research and Treatment of Cancer (EORTC) QLQ-EN24 [22] to give an overall measure of lymphedema three months postoperatively.This score is based on responses from two items; "Have you had swelling in one or both legs?" and "Have you felt heaviness in one or both legs?"Both items are measured on a 4-point Likert scale: 'Not at all', 'A little', 'Quite a bit', and 'Very much'.A high symptom score corresponds to a high level of symptoms.To calculate an overall incidence of lymphedema we dichotomized the lymphedema symptom score: scores of not at all and a little were set to "no lymphedema" while scores of quite a bit, and very much were set to "lymphedema".The incidence of lymphedema was calculated as the proportion of women who changed from no lymphedema at baseline to lymphedema three months postoperatively.

Statistical analysis
Descriptive statistics were used to summarize patient characteristics and SLN outcomes.To compare patient characteristics between the two defined risk groups we used Wilcoxon rank-sum test for continuous variables and the chi-squared test for categorical variables.
There is no international consensus on the definition of lymphedema [23].For a power calculation, we used the dichotomized lymphedema symptom score.The power calculation was based on the proportion of women with SLN mapping only, who changed from "no lymphedema" to "lymphedema", at 12 months postoperatively.Assuming that the true increase in the incidence of lymphedema is 6%, a sample size of 350 with SLN mapping only is required to obtain the conclusion of a lymphedema incidence of <10%, with a statistical power of at least 80%, using a binomial model.This required 350 women with SLN mapping only and completed PROMS preoperatively and 12 months postoperatively.The inclusion was closed when sufficient PROM data at 12 months was forecasted, with an expected rate of non-responses of 20%.
The lymphedema score was linearly transformed ranging from 0 to 100 according to the EORTC scoring manual [24].A change score was calculated as the difference between the baseline-and the three months postoperative score.The baseline and the three months postoperative scores were compared within each group by a paired t-test.The change score was compared between the two EC risk groups by a two-sample ttest of unequal variance.To evaluate the clinical significance of the lymphedema change score, the recommendations for interpretation by Cocks et al. [25] were applied.In general, for longitudinal symptom scores, Cocks et al. find that a difference below 10 is a small difference and a difference of 5 is generally a trivial or no difference.The difference in lymphedema incidence at three months between the two risk groups was compared by a chi-squared test.

Results
A total of 668 women were enrolled in the study from March 2017 to March 2022.Of these, 627 were included in the final analyses and 506 were included in the analyses of lymphedema (Fig. 2).Of the women with PROM data, 442 underwent SLN mapping only.With an expected rate of nonresponses of 20%, this gave a forecast of at least 353 women at 12 months.Of the included 458 (73%) had <50% myometrial invasion and 169 (27%) had >50% myometrial invasion.The median age was 67 (range 36-88) and the median BMI was 29 (range 17-52).According to the Carlson Comorbidity Index (CCI), 40% of the women had at least one comorbidity (Table 1).

SLN mapping
The total SLN detection rate was 94.3% (591/627) with bilateral detection in 73.2% (459/627).Data on lymph nodes and metastases are presented in Table 2.The surgical lymph node assessment was SLN mapping only for the majority of women 533/627 (85.7%), the assessment was significantly different between the two risk groups (p < 0.001), reflecting the difference in the algorithm for the two groups.Thus, 5.9% (27/458) of women with LR EC had no lymph nodes removed due to failed mapping.

Lymph node metastases
The total rate of metastases in the whole cohort was 9.3% (58/627).In women with LR EC, metastases were found in 20/458 (4.4%): 5 MACs, 9 MICs, and 6 ITCs only.Hence, 75% of women with metastatic lymph nodes were identified with the ultrastaging procedure only.Two women (10%) with metastasis in SLNs, underwent secondary pelvic and paraaortic lymph node dissection due to suspicious nodes on post-operative FDG/PET-CT.They both had paraaortic MAC metastases identified (Table 2).
In women with IR EC, 38/169 (22.5%) had metastases, 17 MACs, 11 MICs, and 10 ITCs.Thus, 55% of metastases were identified by ultrastaging only.One woman with failed unilateral, mapping had metastasis in a non-SLN, identified by the surgical algorithm after ipsilateral PLD.Seven women had metastases in SLN and non-SLNs: Two women had additional metastases in clinically suspicious lymph nodes, one had unilateral failed mapping and metastases in the SLN and in the contralateral hemiPLD, four women underwent additional PLD and paraaortic lymph node dissection due to suspicious lymph nodes on postoperative FDG/PET-CT.All four women had paraaortic MAC metastases.

Intraoperative complications
Table 3 presents perioperative-and early postoperative complications.The rate of perioperative complications was 1.4% among all women (9/627).Four women had intestinal damage, all were serosal lesions and none occurred during the SLN mapping procedure.Three women had blood vessel damage; one during the robotic docking procedure (inferior epigastric vein) and two during the SLN mapping (lateral branches to the external and internal iliac vein).Of these, one was converted to laparotomy.One woman with severe adhesions due to endometriosis, had a small lesion of the bladder, and one woman had minor bleeding from the spleen due to rupture of a benign splenic cyst.Hence, the intraoperative complication rate associated with the SLN mapping procedure was 0.3% (2/627).

Early postoperative complications
The overall incidence of postoperative complications was 8% (50/ 627).Of these, 96% (48/50) were grade I-III complications and 4% (2/ 50) were grade IV-V.All grade I complications (18/50) were per definition mild complications.Of the grade II complications, only 4/20 represented a more severe infection.Of the remaining, 10/20 were urinary tract infections, 5/20 were superficial surgical site infections while one patient was given antibiotics due to a suspected supravaginal haematoma.One woman experienced a Grade IVb complication; she suffered from postoperative apoplexy despite antithrombotic prophylaxis.One woman in the study had a grade V complication.The surgical procedure was complicated by massive adhesions and an undetected intestinal injury.She died 30 days after primary surgery.The complications did not occur during the SLN mapping procedure.Despite the difference in the surgical algorithm, there was no significant difference in perioperative complications, blood loss, and postoperative complications between women with LR and IR EC.

Lymphedema
The compliance with completion of PROMs preoperatively and at three months postoperatively was 80.7% (506/627).The lymphedema baseline score was 11.3 (CI 9.6-13.0)on a 0-100 scale for all women, increasing to 15.7 (CI 13.8-17.7)three months postoperatively.Thus, the change score was 4.5 (CI 2.9-6.0,p < 0.001).The change score did not differ between LR and IR groups (p = 0.72) (Table 4).However, according to Cocks' thresholds of clinically important differences, the change score and the CI corresponds to a trivial to small symptom difference [25].In the total population, the incidence of swelling of one or both legs was 5.2% (CI 3.4-7.5)and the incidence of heaviness in one or both legs was 5.8% (CI 3.9-8.2) three months postoperatively.There was no difference in swelling and heaviness between the LR and IR groups (p = 0.82 and p = 0.40 respectively) (Table 4).

Discussion
To the best of our knowledge this is the first study to report the benefits and potential side-effects associated with a national adoption of Abbreviations: EC: endometrial cancer; SLN: Sentinel lymph node; PLD: Pelvic Lymphadenectomy; MAC: Macrometastasis; MIC: Micrometastasis; ITC: Isolated tumor cells.a Twelve women, involvement of cervical stroma; Two women, positive postoperative PET-CT, PLD at second surgery; One woman, sampling of pelvic lymph nodes from the most common SLN locations due to difficult mapping.Two women, presumed >50% myometrial involvement.b Eight women, with removal of paraaortic lymph nodes, due to positive postoperative PET/CT.Two women with low-and six with intermediate risk EC, six of these women had paraaortic macro metastases.
SLN mapping in women with LR and IR EC.With the SENTIREC-endo study we changed the national clinical practice, and thus present real world clinical data.Women with LR and IR EC profited by the implementation of SLN mapping by increased detection of lymph node metastases compared to previous surgical standards.In women with EC of LR, metastatic lymph nodes were identified in 4.4%.Lymph node metastases were identified in 22% of women with EC of IR thus representing a much higher incidence than previously reported [4,9] and mainly ascribed to the ultrastaging procedure.As the myometrial invasion cannot accurately be assessed preoperatively [6], benefits for both risk groups should counterbalance the potential side effects for the large LR group.Reassuringly, the peri-and postoperative complication rate for all women was low, including a very low risk of early leg lymphedema.
Implementation of SLN mapping relies on a high SLN detection rate and a high sensitivity to identify lymph node metastases.The overall rate of metastases of 9.3% in this study is similar to the incidence of 9.6% reported in a recent meta-analysis [4], including accuracy studies on sentinel node mapping in women with EC of LR and IR.A total SLN detection rate of 94.3% in our national study, is comparable to previous prospective studies [2,3,13,26,27] and superior to the results from a recent study reporting from the SLN implementation at a single center (88.8%) [28].A high detection rate protects against unnecessary lymph node dissection in women at low risk of metastases.Our high detection rate is likely explained by several factors.First, all centers completed a surgeon proficiency study [15].Secondly, although SLN mapping represented a new procedure in 2017 in Denmark, the surgical treatment of EC was centralized in 2012 to six high-volume centers with experienced surgeons who were familiar with retroperitoneal surgery.In our study micrometastases accounted for 34% (20/58) of all lymph node metastases in accordance with previous findings [2].Women with MICs are considered as having metastatic involvement and receive adjuvant treatment whereas the oncological consequence and the best clinical management of patients with ITC remain unsolved [1,29,30].Thus, the national adoption of SLN mapping following the SENTIREC-endo study, prompted a substantial improvement in treatment allocation for women with LR as well as IR EC compared to previous national standards were only women with MAC metastases and IR EC would be identified and allocated to adjuvant treatment.
In total, 1.4% of women experienced a perioperative complication.Two of 627 (0.3%) experienced a blood-vessel damage during SLN mapping, and one of these was converted to laparotomy.In the FIRES study, they had an equal complication rate related to the SLN procedure of 0.3% [2].Persson et al. did not report any complications during the SLN procedure but had an overall perioperative complication rate of 3%.In their setting, only five very experienced surgeons performed the procedure [26].The total postoperative complication rate in our study was low (8%) compared to Danish national data after benign hysterectomy (13.5%) [31] and previous studies comparing complications in women undergoing hysterectomy +/− SLN mapping (8-13%) [13,32,33].Despite the large group of surgeons in our study and the real-life clinical setting with gynecological-oncological specialist apprenticeship, the surgical harm related to the SLN procedure in itself seems almost negligible.
By applying a validated PROM sub-scale, we showed that women with low-grade EC undergoing SLN mapping rarely develop early lymphedema.The mean difference in lymphedema score represents a small-trivial difference according to Cocks et al. [25] and the incidence of swelling and heaviness at three months postoperatively was very low (5-6%).Comparing results from different studies is challenged by the lack of consensus on measuring and quantification of lymphedema [23].In the explorative study by Geppert et al., a low incidence of lymphedema was reported after SLN mapping only; 1.3% in 76 women [13].However, no baseline assessment was available for comparison thus weakening the validity of the result [11].The low incidence of self-reported lymphedema in the present study confirms that the SLN procedure in itself carries a very low risk of early leg lymphedema.Future analyses of the prospectively obtained longitudinal PROM data in the SENTIREC-endo study will supplement with additional valuable data on the risk of chronic lymphedema after the SLN procedure.
A recent Delphi investigation among international experts on SLN mapping found no consensus on the mandatory need for completion lymphadenectomy on a non-mapped hemipelvis for the women with EC of LR or IR [34].We adjusted the SLN algorithm with an intraoperative assessment of myometrial invasion in case of no mapping to obviate the national resistance to include women with LR EC given their favorable prognosis.By applying this adjustment, 5.9% (27/458) of the women with LR EC did not undergo PLD after failed SLN mapping.There is a small risk that these women have non-identified lymph node metastases.However, the incidence of lymph node metastasis of 4.4% in this study is similar to previous studies [35,36].We believe that the benefits of the Danish national adoption of the SLN procedure with a minor adjustment for women with LR EC clearly outweigh potential harms in both groups.

Strengths and limitations
The major strength of this study is its national multicenter, prospective design, with inclusion of a large population investigating benefits and potential harms after SLN mapping in women with EC of LR and IR.We have no loss to follow-up on the early complication data, and a high compliance with completion of PROMs on lymphedema pre-and post-operatively.The sample size calculation was based on the incidence of lymphedema and we have statistical power to conclude on the risk of early leg lymphedema.This study was not an accuracy study and thus we cannot conclude on the negative predictive value of SLN mapping to EC of LR and IR.

Conclusion
The SENTIREC-endo study provides real world data of a national change in clinical practice regarding the surgical staging of women with LR and IR EC.We conclude that SLN mapping carries a very low risk of peri-and postoperative complications including a low risk of lymphedema within three months.Women with LR and IR EC benefit from the procedure in terms of increased detection of lymph node metastases compared to previous standards, thus obtaining a more correct allocation to adjuvant treatment.The results from the SENTIREC-endo study provide high-quality data for both clinicians and patients for the shared decision-making regarding the surgical treatment of earlystage low-grade EC.

Fig. 2 .
Fig.2.Flowchart of the inclusion of women in the SENTIREC-endo study.

Table 1
Demographic and clinical characteristics.

Table 2
Lymph node detection rates and metastases.