- Email: [email protected]
Indocyanine green and fluorescence lymphangiography for sentinel node identification in patients with melanoma
The American Journal of Surgery xxx (2018) 1e4
Contents lists available at ScienceDirect
The American Journal of Surgery journal homepage: www.americanjournalofsurgery.com
Indocyanine green and fluorescence lymphangiography for sentinel node identification in patients with melanoma Colette R. Pameijer a, *, Anna Leung a, 1, Rogerio I. Neves a, Junjia Zhu b a b
Department of Surgery, The Pennsylvania State University, College of Medicine, Hershey, PA, USA Department of Public Health Sciences, The Pennsylvania State University, College of Medicine, Hershey, PA, USA
a r t i c l e i n f o
a b s t r a c t
Article history: Received 10 October 2017 Received in revised form 29 November 2017 Accepted 4 January 2018
Background: Infrared fluorescence imaging with indocyanine green (ICG) is a novel method for sentinel node localization. Our objective was to assess ICG and fluorescence imaging for preoperative and intraoperative utility. Methods: 87 eligible patients participated in this prospective study. All patients received injection of ICG dye in addition to both methylene blue and 99mTc. Each sentinel node was assessed for the presence of each dye. Results: ICG was visible prior to incision in 44% of subjects. 99mTc identified a mean of 1.89 SLN per patient. ICG identified a mean of 1.87 SLN while methylene blue (MB) dye identified a mean of 0.71 SLN. 99m Tc and ICG identified the same number of sentinel nodes per patient (P ¼ .73) while methylene blue was inferior in its ability to localize sentinel nodes (P < .001). Conclusion: Our findings indicate that ICG/fluorescence imaging has limited ability to identify the nodal basin preoperatively, but is equivalent to 99mTc for intraoperative identification of sentinel nodes and superior to MB. © 2018 Elsevier Inc. All rights reserved.
Keywords: Melanoma Sentinel node biopsy Indocyanine green Lymphangiography
1. Introduction Sentinel lymph node biopsy has become a standard method of staging lymph node basins for multiple cancers, including melanoma. The status of the sentinel lymph node provides valuable prognostic information, and helps patients and physicians make decisions about further treatment. The current standard method involves radio-labeled colloid (99mTc), and often a visible blue dye, injected at the site of cancer prior to incision. Mapping with lymphoscintigraphy identifies the location of the sentinel node, which can be variable with tumors located on the trunk. The radioactive signal from the sentinel node is localized with a gamma probe in the operating room, and the blue dye is a visual aid during surgery. When both 99mTc and blue dye are used, the success rate of sentinel node localization is 96e99%.1 Despite this high reported
* Corresponding author. The Pennsylvania State University, College of Medicine, Department of Surgery, 500 University Drive, H149, P. O. Box 850, Hershey, PA, 17033-0850, USA. E-mail address: [email protected] (C.R. Pameijer). 1 Present address: Kaiser e LAMC, 4760 Sunset Medical Offices, 4760 W. Sunset Blvd., Los Angeles, CA 90027, USA.
success rate, both 99mTc and blue dye have drawbacks. Blue dye stains surrounding tissues, which can obscure tissue planes, and methylene blue in particular may lead to increased infectious complications and skin necrosis.2 Coordinating scheduling between the operating room and radiology can also be problematic, with inconvenience to the surgeon, the patient or both. Another concern is that the parent substance of 99mTc is produced by a small number of facilities worldwide, and the possibility of a shortage of 99m Tc from these ageing facilities has prompted interest in developing alternative methods of imaging. Indocyanine green (ICG) is a green dye with near infrared fluorescent properties. After dermal injection of ICG, real time lymphangiography can be performed in the operating room using a laser assisted imaging device (SPY Machine, Novadaq Corp., Bonita Springs, FL). Sentinel node dissection can be performed concurrent with imaging. We and others (Jain, Korn) have reported preliminary studies suggesting that ICG and fluorescence imaging is feasible in patients with melanoma.3,4 Here we report a more comprehensive study addressing its clinical utility in comparison to 99mTc and methylene blue.
https://doi.org/10.1016/j.amjsurg.2018.01.009 0002-9610/© 2018 Elsevier Inc. All rights reserved.
Please cite this article in press as: Pameijer CR, et al., Indocyanine green and fluorescence lymphangiography for sentinel node identification in patients with melanoma, The American Journal of Surgery (2018), https://doi.org/10.1016/j.amjsurg.2018.01.009
2
C.R. Pameijer et al. / The American Journal of Surgery xxx (2018) 1e4
2. Methods Patients with melanoma of the trunk or extremities were prospectively enrolled in this trial. Patients with head and neck melanoma were excluded from this study based on our pilot experience. ICG is poorly visualized through muscle, and there seems to be high background uptake in the parotid gland; both of these features limit the utility of ICG and fluorescence imaging. This study was approved by the institutional review board, and the dermal injection of ICG was approved under an IND from the FDA. Novadaq provided material support for the trial and was made aware of the findings prior to submission of this manuscript, but the company was not involved in the analysis or reporting of the results. All subjects had T1b or greater melanoma, and met current National Comprehensive Cancer Network (NCCN) guidelines for sentinel node biopsy. All subjects had injection of 99mTc sulfur colloid and lymphangiography prior to surgery, and injection of indocyanine green (ICG) and methylene blue dye (MB) in the operating room. Up to 1 ml of MB was injected, and 0.9 ml ICG (2.5 mg/ml). The ICG and MB were injected into the dermis separately, after induction of anesthesia and prior to prepping the patient. We chose methylene blue for the lower cost and lower risk of anaphylactic reaction compared to other blue dyes. Each lymph node basin was examined with the gamma probe and imaged with the SPY Elite (Novadaq) prior to making skin incision, with assessment of the visibility of ICG through the skin (yes or no). After incision was made, the basin was periodically imaged with SPY, or dissection was carried out with real time lymphangiography until the sentinel node was identified. The camera of the SPY machine was positioned over the basin, with the image projected onto a monitor. Similar to laparoscopy, the surgeon can watch the image and operate at the same time. Each sentinel node was assessed for the presence of each dye. A lymph node with any one of the dyes present was considered a sentinel node, and was resected. Fluorescence and blue dye were measured as yes or no, and not quantitated. Each basin was palpated intraoperatively for any abnormal nodes. If no sentinel node could be identified by any method, no nodes were removed and the incision closed. These patients were offered imaging surveillance of the nodal basin, typically with ultrasound. Additional data collected included demographic data, the dose of ICG and MB given, adverse effects from any dye, the number of lymph node basins dissected, the total number of sentinel nodes identified in each basin, and the number of malignant sentinel nodes. Pair-wise comparisons were made between the three dyes regarding the number of detected sentinel nodes, and the number of detected metastatic nodes. Based on the results of our pilot study, in which ICG identified 37 of 41 sentinel nodes, the identification rate of ICG was estimated at 90.24% with 95% Blyth-Still-Casella confidence interval (77.88%, 96.6%). We predicted that ICG would identify at least 95% of sentinel nodes. Based on one-sided binomial test, 239 sentinel nodes are needed to achieve 95% power to detect a 5% difference. We estimated that 89 patients were needed in this study. Descriptive statistics were used to summarize the demographic characteristics of the patients in this study. The comparison of the number of sentinel nodes identified by each method was made by the nonparametric signed-rank test for paired-samples. The identification rates were reported using both point estimate and 95% confidence intervals. All analyses were done using statistical software SAS version 9.4 (SAS Institute, Cary, NC, USA). The statistical significance level used was 0.05. 3. Results A total of 89 subjects were enrolled between 2012 and 2016. One
subject retracted consent prior to surgery and another subject was deemed ineligible after enrollment, leaving 87 evaluable subjects. The demographic characteristics of the subjects are outlined in Table 1. Subjects received a median of 0.8 ml MB (range 0.2e1.4 ml) and 0.9 ml ICG (range 0.5e1). There were no adverse reactions to the MB or the ICG. ICG was visible prior to skin incision in 44% of patients. There was no significant difference in age (65 vs 61), BMI (28.5 vs 30) or dose of ICG (0.8 for both) between patients who had visible nodes compared to those in whom ICG was not visible through the skin. The visibility was similar in the axillary and inguinal basins, with 42% visible in the axilla and 47% in the groin. A total of 181 sentinel nodes were resected, for a median of 2 nodes per patient. Seventy-four patients localized to a single lymph node basin, 12 to two basins and 1 patient to 3 basins. Table 2A demonstrates the number of sentinel nodes identified by each dye in each lymph node basin dissected. ICG and fluorescence imaging was effective in the second and third nodal basins, with 10/13 sentinel nodes identified compared to 11/13 nodes identified by 99m Tc and gamma probe, and only 3/13 identified by MB. Twenty sentinel nodes were positive for metastatic disease in 16 patients. 99m Tc identified a mean of 1.89 sentinel nodes per patient, and ICG a mean of 1.87 sentinel nodes, shown in Table 2B. MB identified a mean of 0.71 sentinel nodes per patient. The 99mTc and ICG were not significantly different in their ability to localize sentinel nodes, P ¼ .73. Both 99mTc and ICG were significantly better than MB, with P < .001 in each case. Of the 16 patients with metastases in the sentinel node, 99mTc identified 16/16 (100%) patients, ICG identified 15/16 (93.8%, 95% CI: 69.8e99.8) and MB identified 10/16 (62.5%, 95% CI: 35.4e84.8). 99mTc and ICG were equivalent in identifying patients with metastatic sentinel nodes (P ¼ 1), and both were significantly better than MB, P ¼ .0078 for 99mTc and P ¼ .039 for ICG. ICG failed to localize any sentinel node in three patients, one of whom had metastasis in the sentinel node. All three had trunk lesions of at least intermediate depth, all localized to the axilla and all occurred early in the trial, within the first 30 subjects. All three subjects received 0.9 ml of ICG. In 2 of the 3 subjects the 99mTc and MB did not localize a sentinel node either, thus all 3 methods failed in these two patients and we relied on identifying palpably abnormal nodes. None contained melanoma, and at 8 month follow-up one patient was well but had no further visits. The other patient was found at one year follow-up to have an axillary recurrence as well as metastatic disease. She died of disease 2 months later. The positive sentinel node not detected by ICG was the only node (1 of 181) detected by MB but not ICG. MB failed to localize a sentinel node in 35 patients. Table 3 outlines the number of sentinel nodes positive for each method relative to the other methods. While 99mTc and ICG were largely in agreement with respect to detection of sentinel nodes, the overlap was not complete, with each method identifying a small number of additional nodes compared to the other. MB frequently did not identify sentinel
Table 1 Demographic characteristics of the patients. Median (range) or % Age Gender Depth of melanoma (mm) BMI Number of SN/subject Location melanoma Arm Leg Trunk
65 (26e94) Male 49% 1.18 (0.4e7.8) 28 (17e51.5) 2.1 (1e5) 24.1% 27.6% 48.3%
Please cite this article in press as: Pameijer CR, et al., Indocyanine green and fluorescence lymphangiography for sentinel node identification in patients with melanoma, The American Journal of Surgery (2018), https://doi.org/10.1016/j.amjsurg.2018.01.009
C.R. Pameijer et al. / The American Journal of Surgery xxx (2018) 1e4 Table 2A Number of sentinel nodes identified in 1st, 2nd and 3rd nodal basin by each method.
99m
Tc
MB ICG
1st Basin n ¼ 167
2nd Basin n ¼ 13
3rd Basin n ¼ 1
153 56 149
10 2 9
1 1 1
4. Discussion Indocyanine green has long standing FDA approval for use in perfusion imaging, and recently has been explored for use in localizing sentinel nodes. Prior studies have demonstrated the feasibility of its use for sentinel node localization, but many of these studies are small case series or retrospective in nature. The only prospective trial to date of fluorescence imaging for sentinel node localization found that 96% of sentinel nodes were identified by ICG after skin excision, although the rate of localization prior to incision was reported as only 21%.5 To our knowledge, this is the first independent comparison of 99mTc, MB and ICG, with further exploration of the utility and limitations of this technology. ICG had limited utility for preoperative localization, as less than half the patients had visible fluorescence prior to skin incision. We did not identify any obvious patient factors that would help predict the transcutaneous visibility of ICG, although we did not measure the depth of the sentinel node from the skin. BMI may not be the best reflection of visibility, as it does not predict the location of the sentinel node in the basin. The distributor (Novadaq Corp.) has stated that the fluorescence imaging is visible through 5 mm of tissue, and the sentinel node is usually quickly visible once the superficial fascia is incised. For melanomas located on the trunk or with any potential for drainage to multiple basins, our experience suggests that ICG would serve as an excellent complement to 99mTc, but correct localization of the basin(s) still requires 99mTc with its capacity for preoperative mapping. Once a target basin has been identified by such mapping, however, the intra-operative localization of sentinel nodes by ICG is equivalent to 99mTc, and ICG provided significantly better localization than methylene blue. Ballardini et al.6 found that ICG and 99mTc were equivalent in identifying sentinel nodes in patients with breast cancer, and they suggest that ICG can be used alone. The nodal basin is predictable in this patient population, although internal mammary nodes would not be visualized. In the present study, the number of nodes identified by 99mTc but not ICG was small, and involved both trunk and extremity tumors. While our study provides evidence of efficacy and comparative utility, it is not large enough to have detected low frequency events, and no common attribute could be identified in those patients for whom ICG failed to detect a sentinel node. Systemic
Tc
MB ICG
For all patients (n ¼ 87)
For sentinel node þ patients only (n ¼ 16)
Mean # nodes
p-value*
Mean # nodes
P- value *
1.89 0.71 1.87
(reference) <.001 .73
1.25 0.75 1.19
(reference) 0.0078 1
*P - value represents significance compared to
99m
Tc.
Tech99 -
ICG -
MB -
e 7 1
8 e 1
103 102 e
8 nodes were hot but not fluorescent, 7 were fluorescent but not hot. Over 100 nodes were hot or fluorescent but not blue, while only one node was blue but not hot or fluorescent.
absorption of ICG occurs after 45e60 min, but ICG performed well in the second or third basin, meaning there is enough time after injection to dissect multiple regions. MB on the other hand performed poorly in the second and third basin, with only 3 of 13 sentinel nodes identified. Methylene blue has previously been shown to be related to increased wound complications,2 however no such detrimental relationship has been noted with ICG. Wound complications were not specifically tracked in this study. Lymphazurin is sometimes preferred over MB, but several recent studies find that MB and lymphazurin are equivalent in melanoma,7 breast cancer8 and GI malignancy.9 Some authors10,11 have suggested that a higher number of sentinel nodes may be resected per patient with ICG, compared to 99m Tc. ICG has a low molecular weight, and in theory could travel to non-sentinel nodes leading to unnecessary node excision. This concern is balanced by the risk of false negative sentinel node biopsy due to poor localization of the sentinel nodes. Table 4 lists several studies that used ICG to identify sentinel nodes, and the mean number of sentinel nodes per patient. Earlier studies used higher doses of ICG, although most of the studies find a similar number of sentinel nodes, both from one author to the next and comparing ICG to 99mTc. False negative sentinel node biopsy occurs for multiple reasons,12 some of which are technical. Lee et al.13 recently reviewed the John Wayne melanoma population, and found an overall false negative rate of 14.5%. The rate dropped to 8.5% if they excluded patients who developed in-transit metastases, suggesting the biology behind at least some false negative results. Based on our experience and findings in this study, ICG identified the same number of sentinel nodes as 99mTc and did not lead to an excessive number of sentinel nodes being resected. This study was not powered to detect a difference in rates of false negative sentinel node biopsy among the three methods. The issue of cost is very relevant, although a direct comparison of the cost of 99mTc, ICG and MB is difficult. The cost of 99mTc, depending on an institution's routine, will include the 99mTc itself, a professional fee for the injection and the interpretation of the images, and the gamma probe device. In our institution, the professional fee of lymphangiography is paid at approximately 28% of cost, and the technical fees at 35e74% of cost. The cost of using ICG includes the dye itself, the cost of the SPY machine and the drape for sterile use. The cost of methylene blue would include the cost of purchase, storage and administration, which are minimal.
Table 4 Mean number of sentinel nodes identified in several studies using ICG, with or without Tech99.
Table 2B Number of sentinel nodes overall identified by each method.
99m
Table 3 Correlation of Tech99, ICG and MB positivity in all sentinel nodes.
Tech99þ ICGþ MBþ
nodes, and only identified one node that the other methods did not. Of the sentinel nodes that were 99mTc þ and ICG, 3 originated from extremity primaries, and 5 from trunk.
3
Author
Dose ICG
Tagaya8 Hirche11 Fujisawa7 Jain3 Korn4
5 mg 11 mg 2 mg 2 mg 1.25e3.75 mg
Mean # nodes ICG
Tech99
5.5 2 2.18 2.8 1.86
e e 1.76 2.6 1.86
Please cite this article in press as: Pameijer CR, et al., Indocyanine green and fluorescence lymphangiography for sentinel node identification in patients with melanoma, The American Journal of Surgery (2018), https://doi.org/10.1016/j.amjsurg.2018.01.009
4
C.R. Pameijer et al. / The American Journal of Surgery xxx (2018) 1e4
Significant savings are unlikely to come from the agents themselves, with cost estimates of $60/gram 99mTc, $40/10 ml ICG and $2/5 ml MB. Professional charges are generated for injection of dye and interpretation of images, however most insurers will not pay professional fees to two specialists, radiologist and surgeon, for the same type of service on the same occasion of service, in this case sentinel node mapping. These considerations suggest that savings to the system as a whole would occur if we can identify a patient population for whom lymphangiography by radiology is unnecessary, thus eliminating several professional, technical and hospital charges. In summary, ICG-based fluorescence imaging provides significantly better localization of sentinel nodes than MB, and ICG identifies the same number of sentinel nodes as 99mTc. Based on our findings, ICG and fluorescence imaging can replace MB, although for straightforward cases with a focused gamma signal a second method may not be necessary at all. ICG and fluorescence imaging is not yet suitable for independent use due to poor visibility prior to incision, although it may be possible for extremity tumors in which the nodal basin is predictable. A future study involving extremity tumors and tracking additional information such as depth of node, absolute fluorescence and time to sentinel node identification as well as additional cost and reimbursement information would further clarify the utility of this method. Disclosure The authors have no financial interests to disclose. This trial received material support from Novadaq. Acknowledgment This trial received material support from Novadaq including ICG and sterile drapes for the SPY machine. Appendix A. Supplementary data Supplementary data related to this article can be found at
https://doi.org/10.1016/j.amjsurg.2018.01.009.
References 1. Gershenwald JE, Tseng CH, Thompson W, et al. Improved sentinel lymph node localization in patients with primary melanoma with the use of radiolabeled colloid. Surgery. 1998;124:203e210. 2. Neves RI, Reynolds BQ, Hazard SW, et al. Increased post-operative complications with methylene blue versus lymphazurin in sentinel lymph node biopsies for skin cancers. J Surg Oncol. 2011;103:421e425. 3. Jain V, Phillips BT, Conkling N, et al. Sentinel lymph node detection using laserassisted indocyanine green dye lymphangiography in patients with melanoma. Int J Surg Oncol. 2013;2013, 904214. 4. Korn JM, Tellez-Diaz A, Bartz-Kurycki M, et al. Indocyanine green SPY eliteassisted sentinel lymph node biopsy in cutaneous melanoma. Plast Reconstr Surg. 2014;133:914e922. 5. Stoffels I, Dissemond J, Poppel T, et al. Intraoperative fluorescence imaging for sentinel lymph node detection: prospective clinical trial to compare the usefulness of indocyanine green vs technetium tc 99m for identification of sentinel lymph nodes. JAMA Surg. 2015;150:617e623. 6. Ballardini B, Santoro L, Sangalli C, et al. The indocyanine green method is equivalent to the 99mTc-labeled radiotracer method for identifying the sentinel node in breast cancer: a concordance and validation study. Eur J Surg Oncol. 2013;39:1332e1336. 7. Liu Y, Truini C, Ariyan S. A randomized study comparing the effectiveness of methylene blue dye with lymphazurin blue dye in sentinel lymph node biopsy for the treatment of cutaneous melanoma. Ann Surg Oncol. 2008;15: 2412e2417. 8. Blessing WD, Stolier AJ, Teng SC, Bolton JS, Fuhrman GM. A comparison of methylene blue and lymphazurin in breast cancer sentinel node mapping. Am J Surg. 2002;184:341e345. 9. Soni M, Saha S, Korant A, et al. A prospective trial comparing 1% lymphazurin vs 1% methylene blue in sentinel lymph node mapping of gastrointestinal tumors. Ann Surg Oncol. 2009;16:2224e2230. 10. Fujisawa Y, Nakamura Y, Kawachi Y, et al. Indocyanine green fluorescencenavigated sentinel node biopsy showed higher sensitivity than the radioisotope or blue dye method, which may help to reduce false-negative cases in skin cancer. J Surg Oncol. 2012;106:41e45. 11. Tagaya N, Yamazaki R, Nakagawa A, et al. Intraoperative identification of sentinel lymph nodes by near-infrared fluorescence imaging in patients with breast cancer. Am J Surg. 2008;195:850e853. 12. Scoggins CR, Martin RC, Ross MI, et al. Factors associated with false-negative sentinel lymph node biopsy in melanoma patients. Ann Surg Oncol. 2010;17: 709e717. 13. Lee DY, Huynh KT, Teng A, et al. Predictors and survival impact of falsenegative sentinel nodes in melanoma. Ann Surg Oncol. 2016;23:1012e1018.
Please cite this article in press as: Pameijer CR, et al., Indocyanine green and fluorescence lymphangiography for sentinel node identification in patients with melanoma, The American Journal of Surgery (2018), https://doi.org/10.1016/j.amjsurg.2018.01.009
Contents lists available at ScienceDirect
The American Journal of Surgery journal homepage: www.americanjournalofsurgery.com
Indocyanine green and fluorescence lymphangiography for sentinel node identification in patients with melanoma Colette R. Pameijer a, *, Anna Leung a, 1, Rogerio I. Neves a, Junjia Zhu b a b
Department of Surgery, The Pennsylvania State University, College of Medicine, Hershey, PA, USA Department of Public Health Sciences, The Pennsylvania State University, College of Medicine, Hershey, PA, USA
a r t i c l e i n f o
a b s t r a c t
Article history: Received 10 October 2017 Received in revised form 29 November 2017 Accepted 4 January 2018
Background: Infrared fluorescence imaging with indocyanine green (ICG) is a novel method for sentinel node localization. Our objective was to assess ICG and fluorescence imaging for preoperative and intraoperative utility. Methods: 87 eligible patients participated in this prospective study. All patients received injection of ICG dye in addition to both methylene blue and 99mTc. Each sentinel node was assessed for the presence of each dye. Results: ICG was visible prior to incision in 44% of subjects. 99mTc identified a mean of 1.89 SLN per patient. ICG identified a mean of 1.87 SLN while methylene blue (MB) dye identified a mean of 0.71 SLN. 99m Tc and ICG identified the same number of sentinel nodes per patient (P ¼ .73) while methylene blue was inferior in its ability to localize sentinel nodes (P < .001). Conclusion: Our findings indicate that ICG/fluorescence imaging has limited ability to identify the nodal basin preoperatively, but is equivalent to 99mTc for intraoperative identification of sentinel nodes and superior to MB. © 2018 Elsevier Inc. All rights reserved.
Keywords: Melanoma Sentinel node biopsy Indocyanine green Lymphangiography
1. Introduction Sentinel lymph node biopsy has become a standard method of staging lymph node basins for multiple cancers, including melanoma. The status of the sentinel lymph node provides valuable prognostic information, and helps patients and physicians make decisions about further treatment. The current standard method involves radio-labeled colloid (99mTc), and often a visible blue dye, injected at the site of cancer prior to incision. Mapping with lymphoscintigraphy identifies the location of the sentinel node, which can be variable with tumors located on the trunk. The radioactive signal from the sentinel node is localized with a gamma probe in the operating room, and the blue dye is a visual aid during surgery. When both 99mTc and blue dye are used, the success rate of sentinel node localization is 96e99%.1 Despite this high reported
* Corresponding author. The Pennsylvania State University, College of Medicine, Department of Surgery, 500 University Drive, H149, P. O. Box 850, Hershey, PA, 17033-0850, USA. E-mail address: [email protected] (C.R. Pameijer). 1 Present address: Kaiser e LAMC, 4760 Sunset Medical Offices, 4760 W. Sunset Blvd., Los Angeles, CA 90027, USA.
success rate, both 99mTc and blue dye have drawbacks. Blue dye stains surrounding tissues, which can obscure tissue planes, and methylene blue in particular may lead to increased infectious complications and skin necrosis.2 Coordinating scheduling between the operating room and radiology can also be problematic, with inconvenience to the surgeon, the patient or both. Another concern is that the parent substance of 99mTc is produced by a small number of facilities worldwide, and the possibility of a shortage of 99m Tc from these ageing facilities has prompted interest in developing alternative methods of imaging. Indocyanine green (ICG) is a green dye with near infrared fluorescent properties. After dermal injection of ICG, real time lymphangiography can be performed in the operating room using a laser assisted imaging device (SPY Machine, Novadaq Corp., Bonita Springs, FL). Sentinel node dissection can be performed concurrent with imaging. We and others (Jain, Korn) have reported preliminary studies suggesting that ICG and fluorescence imaging is feasible in patients with melanoma.3,4 Here we report a more comprehensive study addressing its clinical utility in comparison to 99mTc and methylene blue.
https://doi.org/10.1016/j.amjsurg.2018.01.009 0002-9610/© 2018 Elsevier Inc. All rights reserved.
Please cite this article in press as: Pameijer CR, et al., Indocyanine green and fluorescence lymphangiography for sentinel node identification in patients with melanoma, The American Journal of Surgery (2018), https://doi.org/10.1016/j.amjsurg.2018.01.009
2
C.R. Pameijer et al. / The American Journal of Surgery xxx (2018) 1e4
2. Methods Patients with melanoma of the trunk or extremities were prospectively enrolled in this trial. Patients with head and neck melanoma were excluded from this study based on our pilot experience. ICG is poorly visualized through muscle, and there seems to be high background uptake in the parotid gland; both of these features limit the utility of ICG and fluorescence imaging. This study was approved by the institutional review board, and the dermal injection of ICG was approved under an IND from the FDA. Novadaq provided material support for the trial and was made aware of the findings prior to submission of this manuscript, but the company was not involved in the analysis or reporting of the results. All subjects had T1b or greater melanoma, and met current National Comprehensive Cancer Network (NCCN) guidelines for sentinel node biopsy. All subjects had injection of 99mTc sulfur colloid and lymphangiography prior to surgery, and injection of indocyanine green (ICG) and methylene blue dye (MB) in the operating room. Up to 1 ml of MB was injected, and 0.9 ml ICG (2.5 mg/ml). The ICG and MB were injected into the dermis separately, after induction of anesthesia and prior to prepping the patient. We chose methylene blue for the lower cost and lower risk of anaphylactic reaction compared to other blue dyes. Each lymph node basin was examined with the gamma probe and imaged with the SPY Elite (Novadaq) prior to making skin incision, with assessment of the visibility of ICG through the skin (yes or no). After incision was made, the basin was periodically imaged with SPY, or dissection was carried out with real time lymphangiography until the sentinel node was identified. The camera of the SPY machine was positioned over the basin, with the image projected onto a monitor. Similar to laparoscopy, the surgeon can watch the image and operate at the same time. Each sentinel node was assessed for the presence of each dye. A lymph node with any one of the dyes present was considered a sentinel node, and was resected. Fluorescence and blue dye were measured as yes or no, and not quantitated. Each basin was palpated intraoperatively for any abnormal nodes. If no sentinel node could be identified by any method, no nodes were removed and the incision closed. These patients were offered imaging surveillance of the nodal basin, typically with ultrasound. Additional data collected included demographic data, the dose of ICG and MB given, adverse effects from any dye, the number of lymph node basins dissected, the total number of sentinel nodes identified in each basin, and the number of malignant sentinel nodes. Pair-wise comparisons were made between the three dyes regarding the number of detected sentinel nodes, and the number of detected metastatic nodes. Based on the results of our pilot study, in which ICG identified 37 of 41 sentinel nodes, the identification rate of ICG was estimated at 90.24% with 95% Blyth-Still-Casella confidence interval (77.88%, 96.6%). We predicted that ICG would identify at least 95% of sentinel nodes. Based on one-sided binomial test, 239 sentinel nodes are needed to achieve 95% power to detect a 5% difference. We estimated that 89 patients were needed in this study. Descriptive statistics were used to summarize the demographic characteristics of the patients in this study. The comparison of the number of sentinel nodes identified by each method was made by the nonparametric signed-rank test for paired-samples. The identification rates were reported using both point estimate and 95% confidence intervals. All analyses were done using statistical software SAS version 9.4 (SAS Institute, Cary, NC, USA). The statistical significance level used was 0.05. 3. Results A total of 89 subjects were enrolled between 2012 and 2016. One
subject retracted consent prior to surgery and another subject was deemed ineligible after enrollment, leaving 87 evaluable subjects. The demographic characteristics of the subjects are outlined in Table 1. Subjects received a median of 0.8 ml MB (range 0.2e1.4 ml) and 0.9 ml ICG (range 0.5e1). There were no adverse reactions to the MB or the ICG. ICG was visible prior to skin incision in 44% of patients. There was no significant difference in age (65 vs 61), BMI (28.5 vs 30) or dose of ICG (0.8 for both) between patients who had visible nodes compared to those in whom ICG was not visible through the skin. The visibility was similar in the axillary and inguinal basins, with 42% visible in the axilla and 47% in the groin. A total of 181 sentinel nodes were resected, for a median of 2 nodes per patient. Seventy-four patients localized to a single lymph node basin, 12 to two basins and 1 patient to 3 basins. Table 2A demonstrates the number of sentinel nodes identified by each dye in each lymph node basin dissected. ICG and fluorescence imaging was effective in the second and third nodal basins, with 10/13 sentinel nodes identified compared to 11/13 nodes identified by 99m Tc and gamma probe, and only 3/13 identified by MB. Twenty sentinel nodes were positive for metastatic disease in 16 patients. 99m Tc identified a mean of 1.89 sentinel nodes per patient, and ICG a mean of 1.87 sentinel nodes, shown in Table 2B. MB identified a mean of 0.71 sentinel nodes per patient. The 99mTc and ICG were not significantly different in their ability to localize sentinel nodes, P ¼ .73. Both 99mTc and ICG were significantly better than MB, with P < .001 in each case. Of the 16 patients with metastases in the sentinel node, 99mTc identified 16/16 (100%) patients, ICG identified 15/16 (93.8%, 95% CI: 69.8e99.8) and MB identified 10/16 (62.5%, 95% CI: 35.4e84.8). 99mTc and ICG were equivalent in identifying patients with metastatic sentinel nodes (P ¼ 1), and both were significantly better than MB, P ¼ .0078 for 99mTc and P ¼ .039 for ICG. ICG failed to localize any sentinel node in three patients, one of whom had metastasis in the sentinel node. All three had trunk lesions of at least intermediate depth, all localized to the axilla and all occurred early in the trial, within the first 30 subjects. All three subjects received 0.9 ml of ICG. In 2 of the 3 subjects the 99mTc and MB did not localize a sentinel node either, thus all 3 methods failed in these two patients and we relied on identifying palpably abnormal nodes. None contained melanoma, and at 8 month follow-up one patient was well but had no further visits. The other patient was found at one year follow-up to have an axillary recurrence as well as metastatic disease. She died of disease 2 months later. The positive sentinel node not detected by ICG was the only node (1 of 181) detected by MB but not ICG. MB failed to localize a sentinel node in 35 patients. Table 3 outlines the number of sentinel nodes positive for each method relative to the other methods. While 99mTc and ICG were largely in agreement with respect to detection of sentinel nodes, the overlap was not complete, with each method identifying a small number of additional nodes compared to the other. MB frequently did not identify sentinel
Table 1 Demographic characteristics of the patients. Median (range) or % Age Gender Depth of melanoma (mm) BMI Number of SN/subject Location melanoma Arm Leg Trunk
65 (26e94) Male 49% 1.18 (0.4e7.8) 28 (17e51.5) 2.1 (1e5) 24.1% 27.6% 48.3%
Please cite this article in press as: Pameijer CR, et al., Indocyanine green and fluorescence lymphangiography for sentinel node identification in patients with melanoma, The American Journal of Surgery (2018), https://doi.org/10.1016/j.amjsurg.2018.01.009
C.R. Pameijer et al. / The American Journal of Surgery xxx (2018) 1e4 Table 2A Number of sentinel nodes identified in 1st, 2nd and 3rd nodal basin by each method.
99m
Tc
MB ICG
1st Basin n ¼ 167
2nd Basin n ¼ 13
3rd Basin n ¼ 1
153 56 149
10 2 9
1 1 1
4. Discussion Indocyanine green has long standing FDA approval for use in perfusion imaging, and recently has been explored for use in localizing sentinel nodes. Prior studies have demonstrated the feasibility of its use for sentinel node localization, but many of these studies are small case series or retrospective in nature. The only prospective trial to date of fluorescence imaging for sentinel node localization found that 96% of sentinel nodes were identified by ICG after skin excision, although the rate of localization prior to incision was reported as only 21%.5 To our knowledge, this is the first independent comparison of 99mTc, MB and ICG, with further exploration of the utility and limitations of this technology. ICG had limited utility for preoperative localization, as less than half the patients had visible fluorescence prior to skin incision. We did not identify any obvious patient factors that would help predict the transcutaneous visibility of ICG, although we did not measure the depth of the sentinel node from the skin. BMI may not be the best reflection of visibility, as it does not predict the location of the sentinel node in the basin. The distributor (Novadaq Corp.) has stated that the fluorescence imaging is visible through 5 mm of tissue, and the sentinel node is usually quickly visible once the superficial fascia is incised. For melanomas located on the trunk or with any potential for drainage to multiple basins, our experience suggests that ICG would serve as an excellent complement to 99mTc, but correct localization of the basin(s) still requires 99mTc with its capacity for preoperative mapping. Once a target basin has been identified by such mapping, however, the intra-operative localization of sentinel nodes by ICG is equivalent to 99mTc, and ICG provided significantly better localization than methylene blue. Ballardini et al.6 found that ICG and 99mTc were equivalent in identifying sentinel nodes in patients with breast cancer, and they suggest that ICG can be used alone. The nodal basin is predictable in this patient population, although internal mammary nodes would not be visualized. In the present study, the number of nodes identified by 99mTc but not ICG was small, and involved both trunk and extremity tumors. While our study provides evidence of efficacy and comparative utility, it is not large enough to have detected low frequency events, and no common attribute could be identified in those patients for whom ICG failed to detect a sentinel node. Systemic
Tc
MB ICG
For all patients (n ¼ 87)
For sentinel node þ patients only (n ¼ 16)
Mean # nodes
p-value*
Mean # nodes
P- value *
1.89 0.71 1.87
(reference) <.001 .73
1.25 0.75 1.19
(reference) 0.0078 1
*P - value represents significance compared to
99m
Tc.
Tech99 -
ICG -
MB -
e 7 1
8 e 1
103 102 e
8 nodes were hot but not fluorescent, 7 were fluorescent but not hot. Over 100 nodes were hot or fluorescent but not blue, while only one node was blue but not hot or fluorescent.
absorption of ICG occurs after 45e60 min, but ICG performed well in the second or third basin, meaning there is enough time after injection to dissect multiple regions. MB on the other hand performed poorly in the second and third basin, with only 3 of 13 sentinel nodes identified. Methylene blue has previously been shown to be related to increased wound complications,2 however no such detrimental relationship has been noted with ICG. Wound complications were not specifically tracked in this study. Lymphazurin is sometimes preferred over MB, but several recent studies find that MB and lymphazurin are equivalent in melanoma,7 breast cancer8 and GI malignancy.9 Some authors10,11 have suggested that a higher number of sentinel nodes may be resected per patient with ICG, compared to 99m Tc. ICG has a low molecular weight, and in theory could travel to non-sentinel nodes leading to unnecessary node excision. This concern is balanced by the risk of false negative sentinel node biopsy due to poor localization of the sentinel nodes. Table 4 lists several studies that used ICG to identify sentinel nodes, and the mean number of sentinel nodes per patient. Earlier studies used higher doses of ICG, although most of the studies find a similar number of sentinel nodes, both from one author to the next and comparing ICG to 99mTc. False negative sentinel node biopsy occurs for multiple reasons,12 some of which are technical. Lee et al.13 recently reviewed the John Wayne melanoma population, and found an overall false negative rate of 14.5%. The rate dropped to 8.5% if they excluded patients who developed in-transit metastases, suggesting the biology behind at least some false negative results. Based on our experience and findings in this study, ICG identified the same number of sentinel nodes as 99mTc and did not lead to an excessive number of sentinel nodes being resected. This study was not powered to detect a difference in rates of false negative sentinel node biopsy among the three methods. The issue of cost is very relevant, although a direct comparison of the cost of 99mTc, ICG and MB is difficult. The cost of 99mTc, depending on an institution's routine, will include the 99mTc itself, a professional fee for the injection and the interpretation of the images, and the gamma probe device. In our institution, the professional fee of lymphangiography is paid at approximately 28% of cost, and the technical fees at 35e74% of cost. The cost of using ICG includes the dye itself, the cost of the SPY machine and the drape for sterile use. The cost of methylene blue would include the cost of purchase, storage and administration, which are minimal.
Table 4 Mean number of sentinel nodes identified in several studies using ICG, with or without Tech99.
Table 2B Number of sentinel nodes overall identified by each method.
99m
Table 3 Correlation of Tech99, ICG and MB positivity in all sentinel nodes.
Tech99þ ICGþ MBþ
nodes, and only identified one node that the other methods did not. Of the sentinel nodes that were 99mTc þ and ICG, 3 originated from extremity primaries, and 5 from trunk.
3
Author
Dose ICG
Tagaya8 Hirche11 Fujisawa7 Jain3 Korn4
5 mg 11 mg 2 mg 2 mg 1.25e3.75 mg
Mean # nodes ICG
Tech99
5.5 2 2.18 2.8 1.86
e e 1.76 2.6 1.86
Please cite this article in press as: Pameijer CR, et al., Indocyanine green and fluorescence lymphangiography for sentinel node identification in patients with melanoma, The American Journal of Surgery (2018), https://doi.org/10.1016/j.amjsurg.2018.01.009
4
C.R. Pameijer et al. / The American Journal of Surgery xxx (2018) 1e4
Significant savings are unlikely to come from the agents themselves, with cost estimates of $60/gram 99mTc, $40/10 ml ICG and $2/5 ml MB. Professional charges are generated for injection of dye and interpretation of images, however most insurers will not pay professional fees to two specialists, radiologist and surgeon, for the same type of service on the same occasion of service, in this case sentinel node mapping. These considerations suggest that savings to the system as a whole would occur if we can identify a patient population for whom lymphangiography by radiology is unnecessary, thus eliminating several professional, technical and hospital charges. In summary, ICG-based fluorescence imaging provides significantly better localization of sentinel nodes than MB, and ICG identifies the same number of sentinel nodes as 99mTc. Based on our findings, ICG and fluorescence imaging can replace MB, although for straightforward cases with a focused gamma signal a second method may not be necessary at all. ICG and fluorescence imaging is not yet suitable for independent use due to poor visibility prior to incision, although it may be possible for extremity tumors in which the nodal basin is predictable. A future study involving extremity tumors and tracking additional information such as depth of node, absolute fluorescence and time to sentinel node identification as well as additional cost and reimbursement information would further clarify the utility of this method. Disclosure The authors have no financial interests to disclose. This trial received material support from Novadaq. Acknowledgment This trial received material support from Novadaq including ICG and sterile drapes for the SPY machine. Appendix A. Supplementary data Supplementary data related to this article can be found at
https://doi.org/10.1016/j.amjsurg.2018.01.009.
References 1. Gershenwald JE, Tseng CH, Thompson W, et al. Improved sentinel lymph node localization in patients with primary melanoma with the use of radiolabeled colloid. Surgery. 1998;124:203e210. 2. Neves RI, Reynolds BQ, Hazard SW, et al. Increased post-operative complications with methylene blue versus lymphazurin in sentinel lymph node biopsies for skin cancers. J Surg Oncol. 2011;103:421e425. 3. Jain V, Phillips BT, Conkling N, et al. Sentinel lymph node detection using laserassisted indocyanine green dye lymphangiography in patients with melanoma. Int J Surg Oncol. 2013;2013, 904214. 4. Korn JM, Tellez-Diaz A, Bartz-Kurycki M, et al. Indocyanine green SPY eliteassisted sentinel lymph node biopsy in cutaneous melanoma. Plast Reconstr Surg. 2014;133:914e922. 5. Stoffels I, Dissemond J, Poppel T, et al. Intraoperative fluorescence imaging for sentinel lymph node detection: prospective clinical trial to compare the usefulness of indocyanine green vs technetium tc 99m for identification of sentinel lymph nodes. JAMA Surg. 2015;150:617e623. 6. Ballardini B, Santoro L, Sangalli C, et al. The indocyanine green method is equivalent to the 99mTc-labeled radiotracer method for identifying the sentinel node in breast cancer: a concordance and validation study. Eur J Surg Oncol. 2013;39:1332e1336. 7. Liu Y, Truini C, Ariyan S. A randomized study comparing the effectiveness of methylene blue dye with lymphazurin blue dye in sentinel lymph node biopsy for the treatment of cutaneous melanoma. Ann Surg Oncol. 2008;15: 2412e2417. 8. Blessing WD, Stolier AJ, Teng SC, Bolton JS, Fuhrman GM. A comparison of methylene blue and lymphazurin in breast cancer sentinel node mapping. Am J Surg. 2002;184:341e345. 9. Soni M, Saha S, Korant A, et al. A prospective trial comparing 1% lymphazurin vs 1% methylene blue in sentinel lymph node mapping of gastrointestinal tumors. Ann Surg Oncol. 2009;16:2224e2230. 10. Fujisawa Y, Nakamura Y, Kawachi Y, et al. Indocyanine green fluorescencenavigated sentinel node biopsy showed higher sensitivity than the radioisotope or blue dye method, which may help to reduce false-negative cases in skin cancer. J Surg Oncol. 2012;106:41e45. 11. Tagaya N, Yamazaki R, Nakagawa A, et al. Intraoperative identification of sentinel lymph nodes by near-infrared fluorescence imaging in patients with breast cancer. Am J Surg. 2008;195:850e853. 12. Scoggins CR, Martin RC, Ross MI, et al. Factors associated with false-negative sentinel lymph node biopsy in melanoma patients. Ann Surg Oncol. 2010;17: 709e717. 13. Lee DY, Huynh KT, Teng A, et al. Predictors and survival impact of falsenegative sentinel nodes in melanoma. Ann Surg Oncol. 2016;23:1012e1018.
Please cite this article in press as: Pameijer CR, et al., Indocyanine green and fluorescence lymphangiography for sentinel node identification in patients with melanoma, The American Journal of Surgery (2018), https://doi.org/10.1016/j.amjsurg.2018.01.009