Risk factors for failure of bilateral sentinel lymph node mapping in early-stage cervical cancer

Gynecologic Oncology xxx (xxxx) xxx

Contents lists available at ScienceDirect

Gynecologic Oncology journal homepage: www.elsevier.com/locate/ygyno

Risk factors for failure of bilateral sentinel lymph node mapping in early-stage cervical cancer V. Balaya a, b, *, A. Bresset a, b, B. Guani c, L. Magaud d, R. Montero Macias a, b, ^ a, b, A.S. Bats a, b, P. Mathevet c, F. Le curu a, b M. Delomenie a, b, H. Bonsang-Kitzis a, b, C. Ngo a

Gynecologic and Breast Oncologic Surgery Department, Georges Pompidou European Hospital, Paris, France Paris Descartes University, Sorbonne Paris Cit e, Faculty of Medicine, Paris, France Gynecologic Department, University Hospital of Vaud, Lausanne, Switzerland d ^le Sant Hospices Civils de Lyon, Po e Publique, Service Recherche et  epid emiologie cliniques, Lyon, F-69003, France b c

h i g h l i g h t s
Bilateral detection is required to increase sensitivity and decrease false-negative rate of sentinel lymph node biopsy.
Obesity may decrease bilateral detection of sentinel lymph nodes.
Bilateral detection of sentinel lymph nodes is lower in patients aged more than 70 years.
Failure of bilateral detection of sentinel lymph nodes is more frequent with tumor size larger than 20 mm.
Experience and higher surgical skills improved sentinel lymph nodes detection.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 25 June 2019 Received in revised form 21 October 2019 Accepted 23 October 2019 Available online xxx

Background: The objective of this study was to determine clinical, tumoral and surgical factors associated with successful bilateral sentinel lymph node mapping (SBM) in early-stage cervical cancer. Methods: We performed an ancillary work on the data of two prospective trials on SLN biopsy for FIGO IA-IIA cervical cancer (SENTICOL I & II). Patients having Sentinel lymph node (SLN) mapping for earlystage cervical cancer were included between 2005 and 2012 from 28 French oncologic centers. SLN was detected by a combined labeling technique (blue and isotopic). Results: 405 patients were included for analysis: SLNs were identified on at least one side of the pelvis in 381 patients (94.1%) and bilaterally in 326 patients (80.5%). The mean age was 45.4 years [22e85 years]. Most patients had IB1 pathologic FIGO 2018 stage (81.3%) and squamous cell carcinoma (71%). Surgeries were mainly performed by minimally invasive approach (368 patients e 90.9%). By multivariate analysis, lower SBM rate was significantly associated with Age 70 years (ORa ¼ 0.02, 95%CI ¼ [0.001e0.28], p ¼ 0.004), tumor size larger than 20 mm (ORa ¼ 0.46,95%CI ¼ [0.21e0.99], p ¼ 0.048) and Body-mass index higher than 30 kg/m2 (ORa ¼ 0.28, 95%CI ¼ [0.12e0.65], p ¼ 0.003). SBM rate was significantly higher in high skills centers (>5patients/year) (ORa ¼ 8.05, 95%CI ¼ [2.06e31.50], p ¼ 0.003) and in SENTICOL II (2009e2012) compared to SENTICOL I (2005e2007) (ORa ¼ 2.6, 95%CI ¼ [1.23e5.51], p ¼ 0.01). Conclusions: In early-stage cervical cancer, bilateral SLN detection rates is lower in patients aged more than 70years, patients with BMI30 kg/m2 and larger tumor 20 mm whereas stronger experience of SLN biopsy technique improves bilateral SLN detection. © 2019 Elsevier Inc. All rights reserved.

Keywords: Cervical cancer SLN mapping Sentinel lymph node SENTICOL

1. Introduction

rologique, Gyne cologique et * Corresponding author. Service de Chirurgie Cance ^pital Europe en Georges Pompidou, 20, rue Leblanc, 75908, Paris Cedex du Sein Ho 15, France.: E-mail address: [email protected] (V. Balaya).

Lymph node involvement is the main prognosis factor in earlystage cervical cancer. Sentinel lymph node (SLN) biopsy is a diagnostic method widely used in early-stage cervical cancer to assess lymph node status instead of systematic pelvic lymph node

https://doi.org/10.1016/j.ygyno.2019.10.027 0090-8258/© 2019 Elsevier Inc. All rights reserved.

Please cite this article as: V. Balaya et al., Risk factors for failure of bilateral sentinel lymph node mapping in early-stage cervical cancer, Gynecologic Oncology, https://doi.org/10.1016/j.ygyno.2019.10.027

2

V. Balaya et al. / Gynecologic Oncology xxx (xxxx) xxx

dissection [1,2]. The main current challenge of this technique is to achieve maximal sensitivity and lowest false-negative rates [3]. Based on simple criteria such as early FIGO stage (IA2, IB1, IIA primary tumor size < 40 mm), no suspicious pre-, and per-operative lymph nodes, and bilateral negative SLNs after ultra staging, the false-negative rate may be very low at 0.08% in eligible patients [4]. Due to its central topography, historic anatomic work has proven that the uterine cervix has in fact a bilateral lymphatic drainage [5] and therefore a bilateral SLN detection is of paramount importance to be fully reliable [2,6,7]. However, bilateral SLN detection is obtained in 60% of cases only [4]. In endometrial cancer, body-mass index, type of dye injection, clinically suspicious nodes and uterine morphology may have an impact on failure for SLN mapping [8,9] but these risk-factors remain unclear in cervical cancer [10]. SENTICOL I was a prospective multicenter study assessing the diagnostic value of SLN biopsy in early-stage cervical cancer and patients underwent both SLN biopsy and full pelvic lymphadenectomy. Among patients who had SLNs dectected bilaterally, none had false-negative results [2]. SENTICOL II was a prospective randomized multicenter study assessing morbidity and quality of life of SLN biopsy only versus full pelvic lymphadenectomy. SLN biopsy alone induced less surgical morbidity, less lymphedema, and better quality of life than full pelvic lymphadenectomy [11]. These results emphasized that full pelvic lymphadenectomy could be omitted in case of bilateral negative SLN in order to decrease surgical morbidity. In both studies, included patients had early-stage cervical cancer (IAI with Lymphovascular space invasion to IIA1 of the FIGO 2018 classification) and no suspicious nodes at preoperative imaging. The objective of this study was to determine clinical, tumoral and surgical factors associated with successful bilateral sentinel lymph node mapping (SBM) in early-stage cervical cancer through prospective data of SENTICOL I and II. 2. Material and methods 2.1. Patient selection We reviewed the database of two prospective multicentric studies on Sentinel Lymph Node biopsy SENTICOL I [2] and SENTICOL II [11]. In SENTICOL I, patients were included between 2005 and 2007 from 7 French gynecological oncology centers and, in SENTICOL II, between 2009 and 2012 from 28 French gynecological oncology centers (including the seven previous centers). In both studies, all patients had SLN mapping. We included patients with early-stage cervical cancer (IAI with Lymphovascular space invasion to IIA1 of the FIGO 2018 classification) and no suspicious nodes at preoperative imaging. We excluded patients who had suspected lymph node metastasis at preoperative imaging and patients who did not have lymph node staging. This study obtained approval from the Paris Descartes Ethical  de Protection des Personnes HEGP-Broussais). Committee (Comite An informed consent allowing the use of data for secondary analyses was signed by patients. 2.2. SLN detection and ultrastaging Sentinel lymph nodes were detected by a combined labeling technique. The radioactive tracer colloidal rhenium sulfide labeled with technetium [99mTc] was injected using a 25-gauge needle into the four cardinal points of the uterine cervix either on the day before surgery at a dose of 120-MBq or morning of surgery at a dose of 60-MBq. Lymphoscintigraphy was currently performed to guide the surgeons during the surgical procedure. At the beginning of the surgical procedure, 2.5% Patent Blue (2 mL diluted in 2 mL of saline) was injected into the four cardinal points of the cervix.

Intraoperatively, the pelvic and para-aortic areas were firstly explored before opening the retroperitoneal space. Secondly, peritoneum was opened and then the pararectal and paravesical spaces, individualizing the parametrium. Upon the identification of individual blue dyed lymphatic channels and blue colored nodes, we performed radioactivity detection with a gamma probe. SLNs were selectively removed and after the removal of a node, the radioactive intensity of the node was checked ex-vivo once more. Following the extirpation of ‘‘hot’’ and ‘‘blue’’ nodes, residual radioactivity was assessed, and other areas were explored (presacral, supraobturatory, infraobturatory, external iliac and common iliac zones). If no SLNs were retrived, an ipsi-lateral pelvic lymphadenectomy was performed. Frozen section analysis was performed either routinely or only on suspected metastasis nodes depending on the center. According to SENTICOL I protocol, a systematic pelvic lymphadenectomy was performed after SLN biopsy as well as lymphadenectomy of areas containing one or more SLNs whereas in patients included in SENTICOL II, a pelvic lymphadenectomy was performed according to the randomization group (group A, SLN biopsy alone and group B, SLN biopsy with additional pelvic lymphadenectomy). SLNs were analysed after hematoxylineosin-saffron (HES) staining of 200-mm sections. SLNs negative by HES were examined by immunohistochemistry with anti-cytokeratin AE1-AE3 antibodies. Non-SLNs were cut once and were examined by HES only. Isolated tumor cells (ITCs) were defined as < 0.2 mm, micrometastases as between 0.2 and 2 mm, and macrometastases as > 2 mm [12]. 2.3. Data and statistical analysis All data were extracted from two prospective multicentric databases. From these both databases, demographic characteristics, surgical history, and clinical data including FIGO stage were extracted. Body-mass index (BMI) less than 18.5 kg/m2, between 18.5 and 25 kg/m2, between 25 and 30 kg/m2, and higher than 30 kg/m2 defined respectively underweight, normal-weight, overweight and obese patients. Preoperative brachytherapy was not used in all center and was performed in case of tumor larger than 20 mm and/or positive LVSI, according to specific protocol of each center. The usual dose of preoperative brachytherapy was 60 Gy and surgery was performed 6e8 weeks after. Operative records were reviewed, and data about the type of surgical approach and the type of surgery performed (hysterectomy or trachelectomy). Minimally invasive approach was defined as surgery performed by laparoscopy or robotic-assisted laparoscopy. Pathological data was also reviewed (tumor histology, lymphovascular space invasion, parametrial status, vaginal margin status, surgical margin status and tumor size). To assess surgical skills of each center, we defined 3 types of inclusion center according to the number of patients included per year during the inclusion period: type 1 (1 patient/ year, 9 centers), type 2 (1e5 patients/year, 14 centers) and type 3 (>5 patients/year, 5 centers). We performed two analysis. For the first analysis, patients were assessed based on successful bilateral mapping (SBM). Overall detection rate was defined as the proportion of cases in which at least one SLN was identified intraoperatively in one or both hemipelvises. Bilateral detection rate was defined as the proportion of cases in which SLNs were identified intraoperatively in both hemipelvises. For the second analysis, we considered all hemi-pelvises (right and left for each patient) and evaluated the side-specific successful SLN mapping. Side-specific detection rate was defined as the proportion of hemi-pelvises in which at least one SLN was identified intraoperatively. Qualitative variables were expressed as n (%) and were compared by applying chi-square test. Quantitative data were

Please cite this article as: V. Balaya et al., Risk factors for failure of bilateral sentinel lymph node mapping in early-stage cervical cancer, Gynecologic Oncology, https://doi.org/10.1016/j.ygyno.2019.10.027

V. Balaya et al. / Gynecologic Oncology xxx (xxxx) xxx

expressed as mean [range] and were compared by applying the Student's t-test. Variables yielding p values lower than 0.1 by univariate analysis were entered into a multivariate logistic regression model to determine variables independently associated with successful SLN mapping. P values lower than 0.05 were retained as significance set. Data were recorded in an Excel files and all statistical analyses were performed using XLStat Biomed software (AddInsoft V19.4). 3. Results One hundred and forty-five patients were enrolled in the SENTICOL I and 267 patients were enrolled in SENTICOL II. Among these 412 patients, 7 patients were excluded due to final diagnosis of other type of cancer. Finally, 405 patients were included for analysis. The median age was 43 years [22e85 years] and median BMI was 22.9 kg/m2 [14.6e46.9 kg/m2]. Most patients (81.3%) had IB1 pathologic FIGO 2018 stage and squamous cell carcinoma (61.7%). Preoperative brachytherapy was performed in 22.3% of cases. Surgeries were mainly performed by minimally invasive approach (368 patients e 91.9%), including 9 cases by robot-assisted way and whereas 36 patients (8.9%) were operated by laparotomy, including 6 cases of laparoconversion. One hundred patients (24.7%) underwent SLN biopsy alone whereas 305 patients (75.3%) underwent both SLN biopsy and pelvic lymphadenectomy. In 405 patients, 1183 SLNs were detected. SLNs were identified on at least one side of the pelvis in 381 patients (94.1%) and bilaterally in 326 patients (80.5%) while no SLNs were retrievied in 24 patients (5.9%). Two hundred and one SLNs were only blue (17%), 289 SLNs were only hot (24.4%) and 693 were both blue and hot (58.6%). On 810 hemi-pelvises, SLNs were successfully found in 707 cases (87.3%). The median number of SLNs identified per patient was 3 [1e11]. Most of SLNs were found in ilio-obturator area and external iliac area (977 SLNs e 82.6%), followed by common iliac area (109 SLNs e 9.2%), parametrial area (45 SLNs e 3.8%), paraaortic area (23 SLNs e 1.9%) and presacral area (22 SLNs e 1.9%). In patients with SBM, 1104 SLNs were detected: 541 SLNs on the rightside (49%) and 537 on the left-side (48.6%) whereas laterality of 26 SLNs was not precised (2.4%). In patients with unilateral detection (55 patients), 79 SLNs were detected: 36 SLNs on the right-side (45.6%) and 42 SLNs on the left-side (53.2%) whereas laterality of 2 SLNs was not precised (2.5%). In the 381 patients who had at least one SLN identified, 59 patients (15.5%) had at least one positive SLN after ultrastaging: 29 patients had macrometastases, 16 had micrometastases, and 14 had ITCs. SLN metastases were found in 53 of patients with SBM (16.3%) and in 6 patients with unilateral SLN detection (10.9%) (p ¼ 0.31). In the subgroup of patients with SBM and positive SLN (53 patients), most of SLN were positive unilaterally (79.2%). In the subgroup of 305 patients who had both SLN biopsy and full pelvic lymphadenectomy, 228 patients (74.8%) had SBM, 55 patients (18%) had SLNs detected unilaterally and 22 patients (7.2%) had no SLNs found. There were 52 patients who had positive SLN and 61 patients who had at least one positive node. In patients who had at least one SLN detected, the sensitivity was 91.2% and the negative predictive value was 97.2% whereas in patients who had SBM, the sensitivity was 100% and the negative predictive value was 100%.

3

Table 1 Univariate analysis of factors associated with successful bilateral SLN mapping per patients. Predictive variable

p univariate

Total population N ¼ 405

Successful Bilateral SLN mapping N ¼ 326

n

[%]

n

[%]

35.6 64.4

108 218

75.0 83.5

0.04

4 31.9 64.2

9 101 216

56.3 78.3 83.1

0.02

70.9 22.7 6.4

248 68 10

86.4 73.9 38.5

<0.0001

6.9 60.8 19.1 13.2

24 210 57 34 1

85.7 85.7 74.0 64.2

0.001

24.7 75.3

87 239

87.0 78.4

0.06

31.6 68.4

88 238

68.8 85.9

<0.0001

55.8 44.2

180 146

79.6 81.6

0.63

71.0 27.0 2.0

216 87 7 16

86.4 91.6 100.0

0.26

60.8 39.2

193 117 16

90.2 84.8

0.13

22.3 77.7

67 244 15

83.8 87.8

0.35

91.1 8.9

299 26 1

81.2 72.2

0.19

17.1 82.9

53 271 2

76.8 81.1

0.41

73.5 26.5

247 73 7

87.3 71.6

0.0003

27.9 72.1

91 227 8

83.5 80.8

0.54

8.6 81.3 3.8 6.3

30 267 9 19 1

88.2 83.2 60.0 76.0

0.08

Study inclusion SENTICOL I (2005e2007) 144 SENTICOL II (2009e2012) 261 Type of inclusion center Type 1 (1 patient/year) 16 Type 2 (<1e5 patients/year) 129 Type 3 (>5 patients/year) 260 Age [years] < 50 287 50e70 92  70 26 2 BMI (kg/m ) < 18.5 28 18.5e25 245 < 25 - 30 77  30 53 Not specified 2 Parity 0 100 1 305 Menopausal status Post-menopausal 128 Pre-menopausal 277 History of previous pelvic surgery 0 226 1 179 Histology Squamous cell carcinoma 250 Adenocarcinoma 95 Other type 7 Not specified 53 Preoperative LEEP Yes 214 No 138 Not specified 53 Preoperative brachytherapy Yes 80 No 278 Not specified 47 Type of surgical approach Minimal invasive surgery 368 Laparotomy 36 Not specified 1 Node status Patients with 1 positive node Yes 69 No 334 Not specified 2 Tumor size < 20 mm 283  20 mm 102 Not specified 20 LVSI Yes 109 No 281 Not specified 15 FIGO stage (2018) IA1 with emboli - IA2 34 IB1 321 IB2 15 IIA-B 25 Not specified 10

3.1. Patients-specific analysis Results of univariate analysis are shown in Table 1. Successful bilateral mapping (SBM) rate was significantly higher in SENTICOL II cohort (2009e2012) than in SENTICOL I cohort (2005e2007), respectively 83.5% vs 75%, p ¼ 0.04. Centers with low surgical skills (Type 1) had significantly lower SBM rate than type 2 and 3, 56.3%

vs 78.3% and 83.1% respectively (p ¼ 0.02). Patients aged higher than 70 years old had significantly less SBM than patients of 50e70 years and patients younger than 50 years, respectively 38.5% vs 73.9% and 86.4% (p < 0.0001), respectively. Age may be correlated with menopausal status and SBM was significantly lower in

Please cite this article as: V. Balaya et al., Risk factors for failure of bilateral sentinel lymph node mapping in early-stage cervical cancer, Gynecologic Oncology, https://doi.org/10.1016/j.ygyno.2019.10.027

4

V. Balaya et al. / Gynecologic Oncology xxx (xxxx) xxx

menopausal patient (68.8% vs 85.9%, p < 0.0001). Obese patients had significantly lower SBM compared to overweight, normal-BMI and underweight patients, 64.2% vs 74%, 85.7% and 85.7% (p ¼ 0.001) respectively. At final pathologic exam, tumor size larger than 20 mm decreased significantly the SBM rate (71.6% vs 87.3%, p ¼ 0.0003). There was no impact on SBM rate of histologic type, preoperative brachytherapy, preoperative conization and node status. On multivariate analysis, Age >70 years (ORa ¼ 0.02, 95% CI ¼ [0.001e0.28], p ¼ 0.004), obesity (ORa ¼ 0.28, 95% CI ¼ [0.12e0.65], p ¼ 0.003) and tumor size  20 mm (ORa ¼ 0.46, 95%CI ¼ [0.21e0.99], p ¼ 0.048) were independently associated with lower SBM rate (Table 2). Inclusion in SENTICOL II during 2009e2012 period was independently associated with higher SBM rate compared to nclusion in SENTICOL I during 2005e2007 period (ORa ¼ 2.6, 95%CI ¼ [1.23e5.51], p ¼ 0.01). Surgical skill of inclusion center was an independent factor of SBM with higher SBM rate in high skills centers type 2 (ORa ¼ 7.67, 95%CI ¼ [1.94e30.29], p ¼ 0.004) and type 3 (ORa ¼ 8.05, 95%CI ¼ [2.06e31.50], p ¼ 0.003).

3.2. Side-specific analysis As found in patient-specific analysis, Age 70 years, BMI 30 kg/m2, inclusion in low-skill centers, menopause and tumor size 20 mm were also significantly associated with lower successful side-specific SLN mapping (Table 3). Moreover, side-specific SLN mapping failure rate was significantly increased in patients more severe disease (82% in IIA-B vs 89.3% in IB1 and 92.6% in IA, p ¼ 0.02). Minimal invasive approach tended to increase successful side-specific SLN mapping compared to open approach (87.9% vs 80.6%, p ¼ 0.07). Type 2 center with higher skills seemed to have more successful side-specific SLN mapping than type 1 center (81.7% vs 69.2%, p ¼ 0.06). Table 2 Multivariate analysis of factors associated with successful bilateral SLN mapping per patients. Variable Study inclusion SENTICOL I (2005e2007) SENTICOL II (2009e2012) Type of inclusion center Type 1 (1 patient/year) Type 2 (<1e5 patients/year) Type 3 (>5 patients/year) Age [years] < 50 50e70  70 BMI (kg/m2) 18.5e25 < 18.5 < 25 - 30  30 Parity 0 1 Menopausal status Post-Menopausal Pre-Menopausal Tumor size < 20 mm  20 mm FIGO stage (2018) IA1 with LVSI - IA2 IB1 IB2 IIA-B

ORa

IC 95%

P

1 2.6

1.23e5.51

0.01

1 7.67 8.05

1.94e30.29 2.06e31.50

0.004 0.003

1 0.18 0.02

0.02e2.09 0.001e0.28

0.17 0.004

1 1.24 0.88 0.28

0.33e4.73 0.39e2.00 0.12e0.65

0.75 0.76 0.003

1 0.66

0.30e1.46

0.31

1 2.64

0.23e30.44

0.43

1 0.46

0.21e0.99

0.048

1 0.82 0.41 1.94

0.24e2.85 0.06e2.89 0.31e11.95

0.75 0.37 0.48

Table 3 Univariate analysis of factors associated with successful side-specific SLN mapping. Predictive variable

p univariate

Total population N ¼ 810

Successful side-specific SLN mapping N ¼ 707

n

[%]

n

[%]

35.6 64.4

247 460

85.8 88.1

0.33

4.0 31.9 64.2

24 218 465

75.0 84.5 89.4

0.02

70.9 22.7 6.4

522 156 29

90.9 84.8 55.8

<0.0001

6.9 60.8 19.1 13.2

51 448 128 77 3

91.1 91.4 83.1 72.6

<0.0001

24.7 75.3

180 527

90.0 86.4

0.18

31.6 68.4

205 502

80.1 90.6

<0.0001

55.8 44.2

392 315

86.7 88.0

0.59

71.0 27.0 2.0

460 179 14 54

92.0 94.2 100.0

0.35

60.8 39.2

403 250 54

94.2 90.6

0.07

22.3 77.7

145 514 48

90.6 92.4

0.45

91.1 8.9

647 58 2

87.9 80.6

0.07

11.3 88.7

75 628 4

82.4 87.8

0.14

73.5 26.5

516 167 24

91.2 81.9

0.0003

27.9 72.1

194 493 20

89.0 87.7

0.62

8.6 81.3 3.8 6.3

63 573 22 41 8

92.6 89.3 73.3 82.0

0.02

Study inclusion SENTICOL I (2005e2007) 288 SENTICOL II (2009e2012) 522 Type of inclusion center Type 1 (1 patient/year) 32 Type 2 (<1e5 patients/year) 258 Type 3 (>5 patients/year) 520 Age [years] < 50 574 50e70 184  70 52 BMI (kg/m2) < 18.5 56 18.5e25 490 < 25 - 30 154  30 106 Not specified 4 Parity 0 200 1 610 Menopausal status Post-Menopausal 256 Pre-Menopausal 554 History of previous pelvic surgery 0 452 1 358 Histology Squamous cell carcinoma 500 Adenocarcinoma 190 Other type 14 Not specified 106 Preoperative LEEP Yes 428 No 276 Not specified 106 Preoperative brachytherapy Yes 160 No 556 Not specified 94 Type of surgical approach Minimal invasive surgery 736 Laparotomy 72 Not specified 2 Node status Side with ≥1 positive node Yes 91 No 715 Not specified 4 Tumor size < 20 mm 566  20 mm 204 Not specified 40 LVSI Yes 218 No 562 Not specified 30 FIGO stage (2018) IA1 with emboli - IA2 68 IB1 642 IB2 30 IIA-B 50 Not specified 20

The multivariate analysis confirmed that Age 70 years, BMI 30 kg/m2, low-skill centers and tumor size 20 mm were independent factors of side-specific SLN mapping failure (Table 4). Moreover, minimal invasive approach significantly increased successful side-specific SLN mapping compared to open approach (ORa ¼ 5.94, 95%CI ¼ [1.20e29.36], p ¼ 0.03).

Please cite this article as: V. Balaya et al., Risk factors for failure of bilateral sentinel lymph node mapping in early-stage cervical cancer, Gynecologic Oncology, https://doi.org/10.1016/j.ygyno.2019.10.027

V. Balaya et al. / Gynecologic Oncology xxx (xxxx) xxx Table 4 Multivariate analysis of factors associated with successful side-specific SLN mapping. Variable Type of inclusion center Type 1 (1 patient/year) Type 2 (<1e5 patients/year) Type 3 (>5 patients/year) Age [years] < 50 50e70  70 BMI (kg/m2) 18.5e25 < 18.5 < 25 - 30  30 Menopausal status Post-Menopausal Pre-Menopausal Preoperative LEEP No Yes Type of surgical approach Laparotomy Minimal invasive surgery Tumor size < 20 mm  20 mm FIGO stage (2018) IA1 with emboli - IA2 IB1 IB2 IIA-B

ORa

IC 95%

p

1 7.32 2.47

1.25e42.85 0.5e12.25

0.03 0.27

1 0.21 0.02

0.01e3.35 0.002e0.35

0.27 0.01

1 0.93 1.67 0.32

0.29e3.05 0.57e4.90 0.13e0.78

0.91 0.35 0.01

1 2.36

0.15e37.35

0.54

1 0.73

0.35e1.52

0.40

1 5.94

1.20e29.36

0.03

1 0.17

0.07e0.38

<0.0001

1 1.29 0.47 2.81

0.35e4.71 0.05e4.37 0.50e15.73

0.70 0.51 0.24

4. Discussion Bilateral SLN detection is necessary to increase sensitivity and decrease false-negative rates of SLN biopsy. According to the surgical algorithm suggested by Cormier et al. in case of no SLN found on a hemi-pelvis, a side-specific pelvic lymphadenectomy should be performed [13]. In our study, the overall detection rate was 94.1% and the SBM rate was 80.5%. These results are in accordance with those found in the literatture where overall detection and bilateral detection rates were reported respectively at 89.2e93% and 60e86.7% [4,10,14e17]. Considering that uterus is a midline structure, its lymphatic drainage courses bilaterally in the pelvis and SLNs should therefore be identified bilaterally and measuring detection rate by patient may be less accurate than detection rate by hemi-pelvis [18,19]. In this study we found that side-specific detection rate was 87.3% similar to that previously reported of 74e96.9% [4,10,14]. In our cohort, positive SLN rate was 15.5% which was comparable with that already described of 15e20% [1,4,16]. We have shown that age higher than 70 years was an independent risk-factor of failure of SBM by patients and of SLN sidespecific mapping. These results confirmed our previous published data with higher bilateral detection rate in younger women (OR ¼ 1.6 per decades, 95%CI ¼ [1.2e2.2], p ¼ 0.001) [2]. In a cohort of 89 patients, Seong et al. found lower overall detection rates in patients older than 50 years (42.7% vs 57.3%, p ¼ 0.02) [20]. Wuntakal et al. found higher age was associated with lower SLN count (p ¼ 0.005) and also higher unilateral detection (p ¼ 0.04) [21]. This may be explained by the modification with increasing age of lymphatic drainage by sclerosis of lymphatic channels which implies lower concentration of Tc99 m in the SLNs [22]. Obesity, defined as BMI 30 kg/m2, was independently associated with failure of SBM by patients and of SLN side-specific mapping. This factor has been similarly described in breast cancer [23] and endometrial cancer [8,24]. In a cohort of 188 patients with

5

early-stage cervical cancer, Salvo et al. found that BMI 30 kg/m2 decreased significantly bilateral SLN detection (OR ¼ 0.49, 95% CI 0.26e0.94, p ¼ 0.03) [16]. Less SLNs may have been globally detected in obese patients because of thick layer of adipose tissue that does not enable visualization of retroperitoneal lymphatic channels and nodes. A slower lymphatic drainage has also been suggested such as in breast cancer [25]. However, ICG may reverse the negative effect of obesity on SLN mapping as shown in endometrial cancer with an improvement of SLN detection with ICG compared to blue dye [9,24]. Near-infrared imaging with ICG allows deeper visualization of SLN through adipose tissue with a penetration of 1 cm. To improve SLN detection in obese patients, we support the idea that SLN mapping should be performed by fluorescence method rather than combined technique. Tumor size larger than 20 mm decreased significantly the SBM rate per patients and SLN side-specific mapping. Larger tumor size is a well-known risk-factors for failure of bilateral SLN mapping so as NCCN guidelines does not recommend SLN biopsy in tumors 20 mm [4,10,14]. This caution was also justified by higher false negative rates associated with larger tumor. In a review analyzing studies with more 50 patients included, Rob et al. showed a better rate of detection (95.4% vs 80.1%, OR ¼ 0.20, 95% CI ¼ [0.1342e0.3107], p ¼ 0.0001), and a better side-specific detection (84.1% v. 58.8%; OR ¼ 0.27, 95%CI ¼ [0.2116e0.3445], p ¼ 0.0001) in patients with tumor size < 20 mm compared to patients with tumor size  20 mm [14]. In the same way, Kadkhodayan et al. found that detection rate and sensitivity were significantly decreased in patients with tumor size 20 mm, 73.9% vs 93.4% and 81.7% vs 94.7% respectively [10]. In case of larger tumors, tracer injection technique is more difficult due to the frequent presence of a dense and necrotic central core that may inhibit tracer diffusion [26]. In addition to that, larger tumor with exophytic growth may hide usual site of tracer injection [17]. Tumor size may have also an impact on lymphatic drainage because probability of lymphovascular space invasion and lymph nodal involvement are higher in larger tumors which may lead to an occlusion of lymphatic channels that limits tracer spread. Moreover, tumor cells migration may use secondaries lymphatic drainage pathway because of this obstruction and SLNs may be missed due to atypical localizations [27]. However, Dostalek et al. did not find any differences in bilateral detection rate and false negative rate between 3 subgroups of patients with <20 mm, 20e39 mm and 40 mm tumor size [17]. The authors explained that these results may be due to appropriate modification of tracer injection technique. They used longer needle, bigger volume of blue dye and meticulously injected the tracer in the residual stroma around the tumor. We highlighted that minimally invasive approach improved side-specific detection. With a magnified view, this approach permits better exploration of nodes area, especially in the parametrium [18,19,28]. However, this difference is not retrieved in the literature. Kadkhodayan et al. found that laparoscopic and robotic surgery had similar detection rate to open surgery, respectively 93.9%, 90.8% and 88.5% [10]. Furthermore, the results of the recent prospective randomized LACC trial will certainly soon have an impact on surgical practice in favor of open approach in early-stage cervical cancer [29,30]. In our study, patients who were included in SENTICOL II between 2009 and 2012 had a higher SBM rate than patients who were included earlier in SENTICOL I between 2005 and 2007. In addition to that, patients who were included in high-skill centers in SLN biopsy (>5 patients/year) had significantly higher SBM rate than those included in low-skill centers (1 patient/year). These data raise the question of the learning-curve specific to SLN biopsy which has been widely suggested by many authors [16e18,31,32]. In their cohort of 55 patients, Plante et al. noticed an improvement

Please cite this article as: V. Balaya et al., Risk factors for failure of bilateral sentinel lymph node mapping in early-stage cervical cancer, Gynecologic Oncology, https://doi.org/10.1016/j.ygyno.2019.10.027

6

V. Balaya et al. / Gynecologic Oncology xxx (xxxx) xxx

of SLN identification in the last 15 cases compared to their 28 first cases (93% vs 51%, p < 0.01) and concluded that experience in SLN detection increased significantly with time [18]. This acquired experience had an impact in the improvement of SLN detection in subgroup of patients with higher risk of SLN detection failure such as obese patients [16] and patients with larger tumor size [17]. According to Khoury-Collado et al., the cut-off for learning curve in SLN biopsy in endometrial cancer should be fixed at 30 cases [32]. Moreover, it has been demonstrated that patients with early-stage cervical cancer referred at high-volume centers had better functional and oncologic outcomes [33,34]. This ancillary work deserved to assess risk-factors of failure of bilateral SLN mapping in a large cohort of 405 patients from two large prospective multicentric trials on SLN biopsy in early-stage cervical cancer. The main limitation was the use of combined labeling technique (blue dye and radiotracer) and not ICG which was not currently used during inclusion period of SENTICOL I and II. Recently, the FILM trial has shown better SLN detection rates with ICG compared to blue dye in patients with endometrial and cervical cancer [35]. However, in the meta-analysis of Ruscito et al. there were no differences in overall and bilateral detection rates between ICG and association of blue dyes and Technetium-99 m [36]. 5. Conclusion In early-stage cervical cancer, bilateral SLN detection rates are lower in patients aged more than 70 years, patients with BMI 30 kg/m2 and larger tumor 20 mm whereas stronger experience of SLN biopsy technique improves bilateral SLN detection. In absence of these pejorative criterias, SLN biopsy is a reproductive and an efficient technique. Declaration of competing InterestCOI None. CRediT authorship contribution statement V. Balaya: Conceptualization, Project administration, Data curation, Investigation, Formal analysis, Methodology, Writing original draft, Writing - review & editing. A. Bresset: Conceptualization, Project administration, Methodology, Writing - original draft. B. Guani: Data curation, Investigation. L. Magaud: Data curation, Investigation. R. Montero Macias: Formal analysis, Writing - review & editing. M. Delomenie: Formal analysis, Writing - review & editing. H. Bonsang-Kitzis: Formal analysis, Method^ : Formal analysis, Writing ology, Writing - review & editing. C. Ngo review & editing. A.S. Bats: Writing - review & editing. P. Mathevet: Conceptualization, Project administration, Data curation, Investigation, Formal analysis, Writing - original draft, Supervision. curu: Conceptualization, Project administration, Data curation, F. Le Investigation, Formal analysis, Writing - original draft, Writing review & editing, Supervision. References €dt, C. Ko €hler, M. Dürst, A. Schneider, et al., [1] C. Altgassen, H. Hertel, A. Brandsta Multicenter validation study of the sentinel lymph node concept in cervical cancer: AGO Study Group, J. Clin. Oncol. Off. J. Am. Soc. Clin Oncol. 26 (18) (2008 Jun 20) 2943e2951. curu, P. Mathevet, D. Querleu, E. Leblanc, P. Morice, E. Daraï, et al., [2] F. Le Bilateral negative sentinel nodes accurately predict absence of lymph node metastasis in early cervical cancer: results of the SENTICOL study, J. Clin. Oncol. Off. J. Am. Soc. Clin Oncol. 29 (13) (2011 May 1) 1686e1691. [3] D. Cibula, W.G. McCluggage, Sentinel lymph node (SLN) concept in cervical cancer: current limitations and unanswered questions, Gynecol. Oncol. 152 (1) (2019) 202e207.

[4] C. Tax, M.M. Rovers, C. de Graaf, P.L.M. Zusterzeel, R.L.M. Bekkers, The sentinel node procedure in early stage cervical cancer, taking the next step; a diagnostic review, Gynecol. Oncol. 139 (3) (2015 Dec) 559e567. [5] Leveuf J, Godard H. Les lymphatiques de l’uterus. Rev. Chir.. 61:219e248. ^tu, Negative sentinel lymph node [6] I. Popa, M. Plante, M.-C. Renaud, M. Roy, B. Te accurately predicts negative status of pelvic lymph nodes in uterine cervix carcinoma, Gynecol. Oncol. 103 (2) (2006 Nov) 649e653. [7] D. Cibula, N.R. Abu-Rustum, L. Dusek, J. Slama, M. Zik an, A. Zaal, et al., Bilateral ultrastaging of sentinel lymph node in cervical cancer: lowering the falsenegative rate and improving the detection of micrometastasis, Gynecol. Oncol. 127 (3) (2012 Dec) 462e466. [8] E.J. Tanner, A.K. Sinno, R.L. Stone, K.L. Levinson, K.C. Long, A.N. Fader, Factors associated with successful bilateral sentinel lymph node mapping in endometrial cancer, Gynecol. Oncol. 138 (3) (2015 Sep) 542e547. € z, A. Go €kçe, S. Yüksel, et al., Risk factors [9] S. Tas¸kın, M.E. Sarı, D. Altın, C.C. Erso for failure of sentinel lymph node mapping using indocyanine green/nearinfrared fluorescent imaging in endometrial cancer, Arch. Gynecol. Obstet. 299 (6) (2019 Jun) 1667e1672. [10] S. Kadkhodayan, M. Hasanzadeh, G. Treglia, A. Azad, Z. Yousefi, L. Zarifmahmoudi, et al., Sentinel node biopsy for lymph nodal staging of uterine cervix cancer: a systematic review and meta-analysis of the pertinent literature, Eur. J. Surg. Oncol. J. Eur. Soc. Surg. Oncol. Br. Assoc. Surg. Oncol. 41 (1) (2015 Jan) 1e20. [11] P. Mathevet, F. Lecuru, L. Magaud, F. Bouttitie, Sentinel lymph node biopsy for early cervical cancer: results of a randomized prospective, multicenter study (Senticol 2) comparing adding pelvic lymph node dissection vs sentinel node biopsy only, Gynecol. Oncol. 145 (Supplement 1) (2017 Jun) 2e3. [12] P. Hermanek, R.V. Hutter, L.H. Sobin, C. Wittekind, International Union against Cancer. Classification of isolated tumor cells and micrometastasis, Cancer 86 (12) (1999 Dec 15) 2668e2673. [13] B. Cormier, J.P. Diaz, K. Shih, R.M. Sampson, Y. Sonoda, K.J. Park, et al., Establishing a sentinel lymph node mapping algorithm for the treatment of early cervical cancer, Gynecol. Oncol. 122 (2) (2011 Aug) 275e280. [14] L. Rob, R. Lukas, H. Robova, R. Helena, M.J. Halaska, H.M. Jiri, et al., Current status of sentinel lymph node mapping in the management of cervical cancer, Expert Rev. Anticancer Ther. 13 (7) (2013 Jul) 861e870. [15] A.L. Beavis, S. Salazar-Marioni, A.K. Sinno, R.L. Stone, A.N. Fader, A. SantillanGomez, et al., Sentinel lymph node detection rates using indocyanine green in women with early-stage cervical cancer, Gynecol. Oncol. 143 (2) (2016 Nov) 302e306. [16] G. Salvo, P.T. Ramirez, C.F. Levenback, M.F. Munsell, E.D. Euscher, P.T. Soliman, et al., Sensitivity and negative predictive value for sentinel lymph node biopsy in women with early-stage cervical cancer, Gynecol. Oncol. 145 (1) (2017) 96e101. [17] L. Dost alek, M. Zikan, D. Fischerova, R. Kocian, A. Germanova, F. Frühauf, et al., SLN biopsy in cervical cancer patients with tumors larger than 2cm and 4cm, Gynecol. Oncol. 148 (3) (2018) 456e460. ^tu, F. Harel, M. Roy, Laparoscopic sentinel node [18] M. Plante, M.-C. Renaud, B. Te mapping in early-stage cervical cancer, Gynecol. Oncol. 91 (3) (2003 Dec) 494e503. [19] L. Rob, P. Strnad, H. Robova, M. Charvat, M. Pluta, D. Schlegerova, et al., Study of lymphatic mapping and sentinel node identification in early stage cervical cancer, Gynecol. Oncol. 98 (2) (2005 Aug) 281e288. [20] S.J. Seong, H. Park, K.M. Yang, T.J. Kim, K.T. Lim, J.U. Shim, et al., Detection of sentinel lymph nodes in patients with early stage cervical cancer, J. Korean Med. Sci. 22 (1) (2007 Feb) 105e109. [21] R. Wuntakal, A.J. Papadopoulos, S.A. Montalto, M. Perovic, M. Coutts, O. Devaja, Location of sentinel lymph node in cervical carcinoma and factors associated with unilateral detection, Int. J. Gynecol. Canc. Off. J. Int. Gynecol. Canc. Soc. 25 (9) (2015 Nov) 1663e1668. [22] W.C. Conway, M.B. Faries, M.B. Nicholl, A.M. Terando, E.C. Glass, M. Sim, et al., Age-related lymphatic dysfunction in melanoma patients, Ann. Surg. Oncol. 16 (6) (2009 Jun) 1548e1552. [23] A.M. Derossis, J.V. Fey, H.S. Cody, P.I. Borgen, Obesity influences outcome of sentinel lymph node biopsy in early-stage breast cancer, J. Am. Coll. Surg. 197 (6) (2003 Dec) 896e901. [24] A.G.Z. Eriksson, M. Montovano, A. Beavis, R.A. Soslow, Q. Zhou, N.R. AbuRustum, et al., Impact of obesity on sentinel lymph node mapping in patients with newly diagnosed uterine cancer undergoing robotic surgery, Ann. Surg. Oncol. 23 (8) (2016 Aug) 2522e2528. s Olmos, Sentinel node mapping for breast cancer: [25] S. Vidal-Sicart, R. Valde current situation, J. Oncol. 2012 (2012) 361341. [26] D. Cibula, D. Kuzel, J. Sl ama, D. Fischerova, P. Dundr, P. Freitag, et al., Sentinel node (SLN) biopsy in the management of locally advanced cervical cancer, Gynecol. Oncol. 115 (1) (2009 Oct) 46e50. [27] V. Balaya, P. Mathevet, L. Magaud, H. Bonsang-Kitzis, M. Delomenie, R. Montero Macias, et al., Predictive factors of unexpected lymphatic drainage pathways in early-stage cervical cancer, Gynecol. Oncol. 154 (1) (2019 Jul) 102e109. [28] M. Frumovitz, E.D. Euscher, M.T. Deavers, P.T. Soliman, K.M. Schmeler, P.T. Ramirez, et al., “Triple injection” lymphatic mapping technique to determine if parametrial nodes are the true sentinel lymph nodes in women with cervical cancer, Gynecol. Oncol. 127 (3) (2012 Dec) 467e471. [29] P.T. Ramirez, M. Frumovitz, R. Pareja, A. Lopez, M. Vieira, R. Ribeiro, et al., Minimally invasive versus abdominal radical hysterectomy for cervical cancer, N. Engl. J. Med. 379 (20) (2018 Nov 15) 1895e1904.

Please cite this article as: V. Balaya et al., Risk factors for failure of bilateral sentinel lymph node mapping in early-stage cervical cancer, Gynecologic Oncology, https://doi.org/10.1016/j.ygyno.2019.10.027

V. Balaya et al. / Gynecologic Oncology xxx (xxxx) xxx [30] A. Melamed, D.J. Margul, L. Chen, N.L. Keating, M.G. del Carmen, J. Yang, et al., Survival after minimally invasive radical hysterectomy for early-stage cervical cancer, N. Engl. J. Med. 379 (20) (2018 Nov 15) 1905e1914. [31] X. Deng, Y. Zhang, D. Li, X. Zhang, H. Guo, F. Wang, et al., Abdominal radical trachelectomy guided by sentinel lymph node biopsy for stage IB1 cervical cancer with tumors >2 cm, Oncotarget 8 (2) (2017 Jan 10) 3422e3429. [32] F. Khoury-Collado, G.E. Glaser, O. Zivanovic, Y. Sonoda, D.A. Levine, D.S. Chi, et al., Improving sentinel lymph node detection rates in endometrial cancer: how many cases are needed? Gynecol. Oncol. 115 (3) (2009 Dec) 453e455. [33] K. Matsuo, M. Shimada, S. Yamaguchi, M. Matoda, T. Nakanishi, F. Kikkawa, et al., Association of radical hysterectomy surgical volume and survival for earlystage cervical cancer, Obstet. Gynecol. 133 (6) (2019 Jun) 1086e1098.

7

^ , et [34] V. Balaya, P. Mathevet, L. Magaud, M. Delomenie, H. Bonsang-Kitzis, C. Ngo al., Predictive factors of severe perioperative morbidity of radical hysterectomy with lymphadenectomy in early-stage cervical cancer: a French prospective multicentric cohort of 248 patients, Eur. J. Surg. Oncol. J. Eur. Soc. Surg. Oncol. Br. Assoc. Surg. Oncol. 45 (4) (2019) 650e658. [35] M. Frumovitz, M. Plante, P.S. Lee, S. Sandadi, J.F. Lilja, P.F. Escobar, et al., The FILM Trial: a randomized phase III multicenter study assessing near infrared fluorescence in the identification of sentinel lymph nodes (SLN), Gynecol. Oncol. 149 (2018 Jun 1) 7. [36] I. Ruscito, M.L. Gasparri, E.I. Braicu, F. Bellati, L. Raio, J. Sehouli, et al., Sentinel node mapping in cervical and endometrial cancer: indocyanine green versus other conventional dyes-A meta-analysis, Ann. Surg. Oncol. 23 (11) (2016 Oct) 3749e3756.

Please cite this article as: V. Balaya et al., Risk factors for failure of bilateral sentinel lymph node mapping in early-stage cervical cancer, Gynecologic Oncology, https://doi.org/10.1016/j.ygyno.2019.10.027