Abstract
Importance
Fluorescent imaging with indocyanine green can be used to visualize lymphatics. Peritumoral injection of indocyanine green may allow for visualization of every draining lymph node from a primary lesion on near-infrared imaging.
Objectives
To evaluate the role of fluorescent lymphography using near-infrared imaging as an intraoperative tool for achieving complete lymph node dissection and compare the number of lymph nodes retrieved with the use of near-infrared imaging and the number of lymph nodes retrieved without the use of near-infrared imaging.
Design, Setting, and Participants
This prospective single-arm study was conducted among 40 patients who underwent robotic gastrectomy between August 30, 2013, and July 21, 2014, at a single-center, tertiary referral teaching hospital. After propensity score matching, the results of these 40 patients were compared with the results of 40 historical control patients who underwent robotic gastrectomy without indocyanine green injection between January 1, 2012, and August 31, 2013. Statistical analysis was performed from January 1, 2015, to July 31, 2016.
Interventions
Robotic gastrectomy with systemic lymphadenectomy and retrieval of lymph nodes under near-infrared imaging after peritumoral injection of indocyanine green to the submucosal layer 1 day before surgery.
Main Outcomes and Measures
The primary outcome was the number of retrieved lymph nodes in each nodal station.
Results
Among the 40 patients in the study (19 women and 21 men; mean [SD] age, 52.2 [11.7] years), no complications related to indocyanine green injection or near-infrared imaging were observed. On completion of the lymphadenectomy, the absence of fluorescent lymph nodes in the dissected area was confirmed. A mean (SD) total of 23.9 (9.0) fluorescent lymph nodes were recorded among a mean (SD) total of 48.9 (14.6) overall lymph nodes retrieved. The mean number of overall lymph nodes retrieved was larger in the near-infrared group than in the historical controls (48.9 vs 35.2; P < .001), with a significantly greater number of lymph nodes retrieved at stations 2, 6, 7, 8, and 9. In the near-infrared group, 5 patients exhibited lymph node metastases, and all metastatic lymph nodes were fluorescent.
Conclusions and Relevance
This study’s findings suggest that fluorescent lymphography may be useful intraoperatively for identifying and retrieving all necessary lymph nodes for a complete and thorough lymphadenectomy.
This single-arm study evaluates the role of fluorescent lymphography using near-infrared imaging as an intraoperative tool for achieving complete lymph node dissection and compares the number of lymph nodes retrieved with the use of near-infrared imaging and the number of lymph nodes retrieved without the use of near-infrared imaging.
Key Points
Question
Is it possible to complete lymphadenectomy for gastric cancer by application of fluorescent image–guided lymphadenectomy?
Findings
In this single-arm study, peritumoral indocyanine green injection as a fluorescent lymphatic tracer given 1 day before radical gastrectomy successfully visualized all lymph nodes draining from the primary lesion on near-infrared imaging. More lymph nodes were retrieved with low noncompliance after robotic gastrectomy for gastric cancer with near-infrared imaging than without near-infrared imaging.
Meaning
Fluorescent lymphography may be a useful tool for performing complete lymphadenectomy, assessing the adequacy of lymphadenectomy intraoperatively, and retrieving lymph nodes thoroughly from resected specimens.
Introduction
Surgery is the mainstay of treatment for gastric cancer. Radical oncologic surgery for gastric cancer involves performing gastric resection with negative margins and adequate systemic lymph node dissection (LND) for patients without distant metastasis; for advanced gastric cancer, D2 LND (ie, lymph nodes surrounding the main arteries supplying the stomach are also removed) is standard, while D1+ LND, which is less extensive (ie, only perigastric lymph nodes are removed), is recommended for early gastric cancer.1,2 Adequate LND facilitates both proper staging and achievement of better survival outcomes.3,4,5
In the current TNM staging system, nodal classification can be determined only when the total number of lymph nodes examined exceeds 15.6,7 Therefore, sufficient dissection of lymph nodes from the resected specimen is critical to the evaluation of nodal status.8,9 Insufficient LND and inadequate lymph node evaluation have been found to result in poorer survival after gastrectomy among patients with gastric cancer.10,11 We suspect that a single technique that allows for intraoperative assessment of LND and for thorough examination of lymph nodes from the resected specimen would aid in classification of the nodal stage. Near-infrared (NIR) fluorescent imaging after peritumoral injection of indocyanine green (ICG) has been found to provide surgeons with effective visualization of the lymphatic anatomy.12,13 The technique shows high sensitivity and good depth penetration with even a small amount of ICG. Moreover, ICG remains detectable by NIR imaging for several days after its injection into the lymphatic system, allowing a sufficient time interval for visualizing all draining lymph nodes.14,15
In the present study, we tested the hypothesis that peritumoral injection of ICG at a sufficient time before surgery would allow for visualization of every draining lymph node from the primary tumor under NIR imaging, facilitating more complete lymphadenectomy of regional lymph nodes during gastrectomy. Moreover, we suspected that the use of NIR imaging when retrieving lymph nodes from resected specimens would facilitate the harvesting of small fluorescent lymph nodes that would otherwise prove difficult to identify by conventional, manual methods of lymph node retrieval. Altogether, by facilitating more complete LND and more thorough lymph node retrieval from resected specimens, the use of fluorescent lymphography under NIR imaging with ICG would increase the number of lymph nodes harvested compared with conventional gastrectomy without lymphography.
Methods
Patients and Eligibility
We conducted a prospective single-arm study allowing for comparison with historical controls (eFigure 1 in the Supplement). We enrolled patients who were scheduled for robotic gastrectomy between August 30, 2013, and July 21, 2014. We included patients between 20 and 80 years of age regardless of sex. Patients had histologically proven gastric adenocarcinoma, tumor sizes of less than 4 cm in diameter, and clinical stage I disease (T1N0M0, T1N1M0, and T2N0M0) according to the seventh staging system from the American Joint Committee on Cancer.7 We excluded patients with an Eastern Cooperative Oncology Group performance status greater than 1, an American Society of Anesthesiologists score higher than 3, a requirement of additional organ resection, a history of chemotherapy or radiotherapy for current gastric cancer, a history of prior gastric resection regardless of reason, and the presence or history of other malignant neoplasms. This study was approved by the Institutional Review Board of Severance Hospital, Yonsei University College of Medicine, Seoul, Korea (No. 4-2013-0414). The study was performed according to ethical standards. Written informed consent was obtained from patients after a full explanation of the study. This study is registered at ClinicalTrials.gov (NCT01926743).
Experimental Procedures
Endoscopic ICG Injection
As a fluorescent contrast agent, ICG (Dongindang Pharmaceutical) was prepared as a 1.25-mg/mL solution in sterile water (Figure 1A). During endoscopy for tumor localization on the day before surgery,16 we injected 0.6 mL of the prepared solution, containing 0.75 mg of ICG, along the submucosal layer of the stomach at 4 points around the primary tumor. The total volume of injected ICG solution was 2.4 mL, for a total ICG amount of 3 mg.
Figure 1. Schema of the Study Procedures and Illustration to Define Fluorescent Lymph Nodes (LNs) and Fluorescent Stations.
A, Procedure defining fluorescent LNs. A fluorescent LN is a node visualized as green under near-infrared imaging (NIR). B, Fluorescent stations contain fluorescence-emitting components. Fluorescent stations can contain both fluorescent LNs and nonfluorescent LNs. ICG indicates indocyanine green.
Surgery
We used the da Vinci Si Surgical System (Intuitive) equipped with the Firefly mode to acquire NIR fluorescent images for all patients. Because the surgical system contains a module for fluorescence imaging, a simple finger click can change between visible light and NIR imaging (green spots under a dim background) without the need to change any equipment. During surgery, the surgeon frequently turned on the NIR mode before and after dissection at each lymph node station (Video 1).
Video 1. D2 Dissection With Near-Infrared Imaging.
We determined the extent of gastric resection for distal gastrectomy or total gastrectomy according to the tumor location. We performed D1+ LND for early gastric cancer without evidence of lymph node metastasis and D2 LND if there was any suspicion of advanced gastric cancer or lymph node metastasis. The LND procedures have been described in detail previously.17,18,19 When residual lymph nodes containing fluorescence were identified in the dissected area, these lymph nodes were additionally removed (Video 2). Although we occasionally detected fluorescent lymph nodes outside the planned dissection area (stations 10, 11d, 12a, and 14v), excessive dissection beyond the scope of D2 LND was not performed. After LND, the surgeon rechecked the surgical field under the NIR mode to confirm a lack of residual fluorescence, ensuring complete LND.
Video 2. Additional Lymph Node Removal.
Lymph Node Harvest From the Resected Specimen
We separated lymph node–bearing soft tissue from resected specimens ex vivo according to the definition of lymph node stations based on the Japanese classification.20 We then retrieved fluorescence-containing nodes from each station under direct visualization using NIR imaging (Video 3). Lymph node stations that emitted fluorescence were classified as fluorescent stations, and fluorescence-emitting lymph nodes at the fluorescent stations were classified as fluorescent lymph nodes (Figure 1B). We reexamined paraffin blocks containing the lymph nodes using the NIR imaging system to confirm the fluorescence status and to match the fluorescence-containing lymph nodes and the results of pathologic testing.
Video 3. Lymph Node Retrieval From Soft Tissue Ex Vivo.
Measurement of LND Efficiency
We categorized the total number of overall lymph nodes retrieved with reference numbers of 16 and 30 because the minimum number of lymph nodes for proper nodal staging is 16 according to current staging guidelines.21 The eighth edition of the American Joint Committee on Cancer staging system suggests that the removal of 30 or more nodes is desirable.21
We defined noncompliance of LND per station as containing no lymph nodes from the dissected station. We defined noncompliance per patient as the absence of lymph nodes from 2 or more lymph node stations that were supposed to be harvested. Contamination was defined as the presence of lymph nodess from 2 or more stations that were not supposed to be harvested.3,22
Historical Controls
We identified 132 consecutive robotic gastrectomies for gastric cancer performed between January 1, 2012, and August 31, 2013, in a prospectively collected gastric cancer database at our institution. To make the eligibility criteria equal between the control group and the lymphography group, we excluded 21 patients with disease beyond clinical stage I and 10 patients with combined resection of another organ. Then, we used propensity score matching to compensate for differences in baseline characteristics between the prospectively enrolled patients and the historical controls. The propensity scores were calculated with covariates of age, sex, depth of tumor invasion, type of gastric resection, and extent of LND. All patients in the prospective study group were matched 1:1 with a control with the nearest available score without replacement. Finally, of the 101 patients in the historical control group, 40 were matched by propensity score matching. The 2 groups were well balanced with regard to baseline characteristics and pathologic features after matching (Table 1), and no significant differences between the 2 groups in operative outcomes were detected (Table 2).
Table 1. Clinicopathologic Characteristics of the NIR Fluorescence Imaging Group and Historical Controls After Propensity Score Matching.
| Characteristic | Patients, No. (%) | P Value | |
|---|---|---|---|
| NIR Imaging Group (n = 40) | Control Group (n = 40) | ||
| Age, mean (SD), y | 52.2 (11.7) | 52.1 (11.3) | .99 |
| Sex | |||
| Male | 21 (53) | 19 (48) | .66 |
| Female | 19 (48) | 21 (53) | |
| BMI, mean (SD) | 23.3 (2.6) | 23.3 (3.4) | .93 |
| Tumor size, mean (SD), mm | 25.4 (17.7) | 21.8 (13.3) | .32 |
| Tumor location | |||
| Lower third | 20 (50) | 19 (48) | .43 |
| Middle third | 16 (40) | 13 (33) | |
| Upper third | 4 (10) | 8 (20) | |
| Histologic type (Lauren classification) | |||
| Intestinal | 11 (28) | 16 (40) | .36 |
| Diffuse | 27 (68) | 20 (50) | |
| Mixed | 2 (5) | 4 (10) | |
| Depth of invasion | |||
| T1a | 24 (60) | 25 (63) | >.99 |
| T1b | 11 (28) | 11 (28) | |
| T2 | 5 (13) | 4 (10) | |
| LN metastasis | |||
| N0 | 35 (88) | 36 (90) | >.99 |
| N1 | 5 (13) | 4 (10) | |
Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); LN, lymph node; NIR, near-infrared.
Table 2. Operative Outcomes of the NIR Fluorescence Imaging Group and Historical Controls After Propensity Score Matching.
| Characteristic | Patients, No. (%) | P Value | |
|---|---|---|---|
| NIR Imaging Group (n = 40) | Control Group (n = 40) | ||
| Extent of gastrectomy | |||
| Distal | 34 (85) | 32 (80) | .56 |
| Total | 6 (15) | 8 (20) | |
| Extent of LNDa | |||
| D1+ | 24 (60) | 25 (63) | .82 |
| D2 | 16 (40) | 15 (38) | |
| Anastomosis | |||
| Gastroduodenostomy | 21 (53) | 21 (53) | .79 |
| Gastrojejunostomy | 13 (33) | 11 (28) | |
| Roux-en-Y esophagojejunostomy | 6 (15) | 8 (20) | |
| Operation time, mean (SD), min | 191.0 (39.5) | 209.0 (55.3) | .10 |
| Estimated blood loss, mean (SD), mL | 46.8 (35.3) | 47.9 (42.1) | .90 |
| Length of hospital stay, mean (SD), d | 6.3 (3.1) | 7.2 (9.7) | .61 |
| Complication | |||
| No | 35 (98) | 36 (90) | >.99 |
| Yes | 5 (13) | 4 (10) | |
Abbreviations: LN, lymph node; LND, LN dissection; NIR, near-infrared.
D1 dissection entails gastrectomy and the resection of perigastric lymph nodes; D2 dissection is a D1 plus all the nodes along the left gastric artery, common hepatic artery, celiac artery, and splenic artery; and D1+ dissection is midpoint between D1 and D2.1
Statistical Analysis
Statistical analysis was performed from January 1, 2015, to July 31, 2016. We report all continuous variables as mean values with SDs and ranges. Continuous variables were analyzed with the Mann-Whitney test for the prospective study data set and with the t test for comparison with the historical controls. Categorical variables were compared with the χ2 test or the Fisher exact probability test for comparison with the historical controls, as appropriate. All P values were from 2-sided tests and results were deemed statistically significant at P < .05. All statistical analyses, including propensity score matching, were performed with IBM SPSS Statistics, version 23 for Windows (IBM Corp).
Results
Between August 30, 2013, and July 21, 2014, we enrolled 41 eligible patients. One patient was excluded owing to having lymph nodes that were indistinguishable from ICG-stained soft tissue under NIR imaging because of accidental injection of ICG into the peritoneal space during endoscopy. Thus, 40 patients were included in the analyses. The mean (SD) interval between the endoscopic injection of ICG and the operation was 22.4 (3.3) hours (range, 14.4-28.0 hours). We did not observe any complications attributable to endoscopic injection of ICG. No intraoperative events related to the application of NIR imaging or delayed complications related to injection of ICG were observed. Although 5 patients experienced complications during their hospital stay, none was related to injection of ICG or NIR imaging. During surgery, we confirmed that no residual fluorescent lymph nodes were left in the planned dissection area after completion of LND for all patients.
Lymph Node Examination
We dissected 432 lymph node stations containing 1956 lymph nodes from the 40 patients in the NIR group, for a mean (SD) of 10.8 (1.4) stations per patient (range, 9-14 stations per patient) (eFigure 2 in the Supplement). The mean (SD) number of overall harvested lymph nodes per patient was 48.9 (14.6) (range, 25-92), and the mean (SD) number of lymph nodes per station was 4.5 (3.9) (range, 0-30). For all patients, more than 15 lymph nodes were retrieved, and for 37 patients (92.5%), more than 30 lymph nodes were retrieved.
We identified 289 fluorescent stations for a mean (SD) of 7.2 (1.6) stations per patient (range, 3-10 stations per patient). Not all fluorescent stations contained lymph nodes because, occasionally, soft tissues without lymph nodes were stained with ICG. Of the fluorescent stations, 254 contained 1 or more fluorescent lymph nodes; the other 35 stations had no lymph nodes. Among the 1956 lymph nodes retrieved, 289 fluorescent stations contained 1590 lymph nodes, and 143 nonfluorescent stations contained 366 lymph nodes. Because not all lymph nodes retrieved from fluorescent stations were fluorescent, 956 of the 1590 lymph nodes in the fluorescent stations were fluorescent, for a mean (SD) of 23.9 (9.0) fluorescent lymph nodes per patient and 3.3 (3.0) fluorescent lymph nodes per station. The other 634 lymph nodes in the fluorescent stations were nonfluorescent. We additionally dissected ICG-positive 11p station in 4 patients and 14v station in 2 patients among 26 patients planned to undergo D1+ LND during distal subtotal gastrectomy.
We retrieved significantly more lymph nodes from fluorescent stations than from nonfluorescent stations (mean [SD] number, 5.5 [4.1] fluorescent stations vs 2.5 [2.7] nonfluorescent stations; P < .001). In fact, at the same anatomical lymph node station, more lymph nodes were retrieved when the fluorescent station was stained than not, except at stations 2 and 4sa. However, statistically significant differences in retrieved lymph nodes per station were observed only for stations 4sb (2.7 vs 1.2; P = .046) and 6 (7.5 vs 4.3; P = .01) (Figure 2).
Figure 2. Mean Number of Fluorescent Lymph Nodes (LNs), Nonfluorescent LNs in Fluorescent Stations, and LNs in Nonfluorescent Stations According to LN Station.
aP < .05.
bOverall number of LN stations examined (289 fluorescent stations that emitted fluorescence and 143 nonfluorescent stations that did not emit flurescence).
Of 432 stations dissected, 54 (12.5%) did not contain lymph nodes. The rate of noncompliance of LND per station was 5.2% for fluorescent stations (15 of 289), whereas a higher rate of noncompliance of 27.3% was recorded for nonfluorescent stations (39 of 143; P < .001). Noncompliance per patient, which was defined as the absence of lymph nodes from 2 or more lymph node stations that were supposed to be harvested, was observed for 14 patients (35.0%), whereas contamination of lymph nodes that were not supposed to be harvested was noted in 3 patients (7.5%).
Five patients had lymph node metastases, and each patient had 1 metastatic lymph node. All 5 metastatic lymph nodes were fluorescent.
Comparison With Historical Controls
A total of 422 lymph node stations containing 1408 lymph nodes were dissected from 40 historical control patients. The mean (SD) number of overall lymph nodes harvested per patient was 35.2 (11.2) (range, 18-78), and the mean (SD) number of lymph nodes per station was 3.3 (3.2) (range, 0-22). The mean (SD) number of lymph nodes retrieved per patient among the historical controls (35.2 [11.2]) was significantly lower than that for patients in the NIR group (48.9 [14.6]; P < .001). Although at least 16 lymph nodes were retrieved for all patients in either group, a total of 30 or more lymph nodes were retrieved from 37 patients in the NIR group (92.5%) and from only 25 patients in the historical control group (62.5%; P = .001).
When we divided lymph nodes into perigastric (stations 1, 2, 3, 4, 5, and 6) and extraperigastric (stations 7, 8, 9, 11, and 12a) regions, the mean (SD) number of retrieved lymph nodes was significantly higher in the NIR group than in the historical control group for both the perigastric (28.8 [11.6] vs 23.5 [8.7]; P = .02) and extraperigastric (19.9 [8.4] vs 11.3 [6.0]; P < .001) regions. In subgroup analysis according to the extent of gastrectomy and LND, patients who underwent distal gastrectomy in the NIR group showed significantly more lymph nodes retrieved than did their counterparts in the historical control group. The differences in the number of harvested lymph nodes were more obvious in the extraperigastric area (Figure 3A). The lymph nodes in the stations of D1+ LND for distal gastrectomy (stations 1, 3, 4sb, 4d, 5, 6, 7, 8, and 9), which are always dissected during gastrectomy for gastric cancer, were harvested more often in the NIR group than in the historical control (mean [SD] number, 43.1 [13.1] vs 31.3 [9.8]; P < .001). Although the overall mean (SD) number of lymph nodes per station was higher in the NIR group than in the historical controls (4.5 [3.9] vs 3.3 [3.2]; P < .001), significant differences were observed only for stations 2, 6, 7, 8, and 9 (Figure 3B).
Figure 3. Total Harvested Lymph Nodes (LNs) and Total Retrieved LNs According to LN Station.
A, Total harvested LNs in the near-infrared (NIR) group and historical controls. B, Total retrieved LNs in the NIR group and historical controls according to LN station. Patients were divided according to the extent of gastrectomy and LN dissection (LND; perigastric stations, 1, 2, 3, 4, 5, and 6; extraperigastric stations, 7, 8, 9, 11, and 12a). Fluorescent LNs are those emitting fluoresence, whereas nonfluorescent LNs do not. DG indicates distal gastrectomy; TG, total gastrectomy.
aP < .05.
bOverall number of lymph node stations examined (432 fluorescent and nonfluorescent stations for LNs in the NIR group and 422 stations for LNs in historical controls).
The rate of noncompliance per station in the historical controls (78 of 422 [18.5%]) was significantly higher than that in the NIR group (54 of 432 [12.5%]; P = .02). This rate was significantly higher than the rate of noncompliance for fluorescent stations in the NIR group (15 of 289 [5.2%]; P < .001) but lower than that for nonfluorescent stations in the NIR group (39 of 143 [27.3%]; P = .03). Noncompliance per patient was observed for 14 patients (35.0%) in the NIR group and 23 patients (57.5%) in the historical control group (P = .04). Contamination of lymph nodes that were not supposed to be harvested was similar between the 2 groups: 1 patient (2.5%) among the historical controls and 3 patients (7.5%) among the NIR group (P = .62).
Discussion
Fluorescent lymphography with endoscopic peritumoral injection of ICG given 1 day before robotic gastrectomy effectively facilitated the visualization of lymph nodes that drained from the primary tumor. Using fluorescent lymphography with NIR imaging allowed us to perform complete lymphadenectomy for all patients and to assess the thoroughness of the lymphadenectomy intraoperatively. Consequently, we were able to retrieve more lymph nodes with low noncompliance after robotic gastrectomy for gastric cancer with NIR imaging than without NIR imaging.
Although lymphography has previously been used in surgery for gastric cancer, the reported techniques were used only to identify lymphatic drainage patterns or sentinel lymph nodes.15,23,24 As the longer wavelengths of NIR allow for greater penetration and identification of lymph nodes inside thick fatty tissues, ICG under NIR imaging has been found to offer better visualization of lymph nodes than other dyes observable by the naked eye under visible light.12 Better visualization of lymph vessels and lymph nodes during surgery enables the surgeon to achieve en bloc resection of the lymphatics and lymph nodes without breakage of lymphatic structures, thereby preventing tumor cell spillage during resection, thus potentially providing oncologic benefit. Moreover, fluorescent lymphography can help surgeons perform safer surgical procedures by preventing injuries when dissecting lymph node–bearing soft tissue around blood vessels or the pancreas. In the present study, we found that intraoperative fluorescent lymphography guided the surgeon for the complete removal of all lymph nodes containing ICG within the extent of the planned LND. In addition, the technique allowed for intraoperative assessment of the thoroughness of the lymphadenectomy because any fluorescent lymph nodes left in the dissected area would have been easily detected. Thus, a study evaluating fluorescent lymph nodes left behind by NIR imaging after completing LND without NIR imaging would confirm lymph node stations additionally dissected with NIR visualization.
Applying NIR imaging in the present study, we were able to obtain more lymph nodes and to achieve a lower rate of noncompliance. Although the increase in the number of retrieved lymph nodes may have resulted from more precise dissection, it also may have stemmed from more thorough retrieval of lymph nodes from the resected specimen. Retrieving lymph nodes under NIR visualization permitted identification of even tiny lymph nodes owing to the high sensitivity of NIR imaging compared with the conventional method of separating lymph nodes from fatty tissue with manual lymph node retrieval under the naked eye. Near-infrared imaging allowed for proper staging on a thorough examination of lymph nodes retrieved without the need to further apply complicated methods of lymph node harvesting, such as the fat-clearing technique.25 Because harvesting a sufficient number of lymph nodes is essential for proper staging and because inappropriate staging, especially underestimating, may lead to inappropriate management of patients by withholding adjuvant treatment, an increase in the number of lymph nodes retrieved and the potentially more appropriate staging resulting from lymph node harvesting are likely to have a positive effect on survival.6,26 Thus, in light of our results, we suggest that application of NIR imaging to gastric cancer surgery can be of use in ensuring both higher surgical quality and proper staging. These advantages of our techniques would offer greater benefit to surgeons in Western countries with a higher prevalence of obese patients and low incidences of gastric cancer.
Most studies using NIR imaging with ICG for gastric cancer have focused on sentinel lymph node mapping; however, our purpose for the use of NIR imaging with ICG was to visualize every draining lymph node from the primary tumor for complete LND and intraoperative assessment of the thoroughness of LND. To establish an optimized protocol for injection of ICG, we tested various doses, concentrations, and timings of injection of ICG before initiation of this study. Injection of the commonly used 2.5-mg/mL dose of ICG resulted in extensive diffusion of ICG around lymph nodes, and the strong intensity of fluorescence hampered the identification of the lymphatics and the delineation of the lymph nodes from surrounding tissue. Meanwhile, because ICG requires sufficient time to spread to lymph nodes, intraoperative injection of ICG would make the operation time longer. Moreover, intraoperative endoscopy may have rendered the surgical field cluttered by insufflated air in the small intestine. In a previous study, researchers reported that intraoperative subserosal injection resulted in lower accuracy than submucosal injection, even for the detection of sentinel lymph nodes.15 Another study evaluating NIR fluorescent image–guided lymphadenectomy for gastric cancer also used subserosal injection of ICG for sentinel lymph node biopsy.13 Thus, we opted to apply endoscopic submucosal injection of ICG preoperatively to allow for sufficient distribution of the fluorescent agent and to avoid a lengthy surgery. Furthermore, we discovered that this protocol fit nicely with typical clinical practice, as most patients are admitted to the hospital 1 day before surgery, thereby allowing for injection of ICG on the day of admission and reducing additional hospital visits. However, the efficiency of injecting ICG several hours before or just before starting the operation needs to be evaluated to determine the optimal timing of the injection.
Although this study specifically examined the benefits of NIR imaging for gastric cancer surgery, the concept of the study is applicable for several other tumor types that hold the potential for lymph node metastasis. Examples of such tumors are colorectal, cervical, and esophageal cancer. In addition, injection of ICG boasts an advantage for tumor localization. Because ICG is injected around the tumor, it is also helpful for determining gastric resection margins.
Limitations
Our study has several limitations. First, our study was not conducted as a randomized clinical trial. Second, because it was impossible to apply surgeon masking for fluorescent lymphography during surgery, the surgeon may have been biased toward the extent of LND. Therefore, we used propensity score matching to compare the results with an equivalent patient group to overcome the limitations related to the nonrandomized, nonblinded nature of the study. However, a large-sized multi-institutional prospective randomized clinical trial is needed. Third, our enrollment of patients with relatively early-stage disease with low lymph node metastasis is another limitation. Although all metastatic lymph nodes were fluorescent in this study, it would be difficult to determine metastatic lymph node–specific sensitivity or specificity of fluorescent lymph nodes. High sensitivity and specificity for metastatic lymph nodes among fluorescent lymph nodes would enable application of tumor-specific, personalized LND according to individual tumor locations and lymphatic drainage patterns. Moreover, it remains unclear whether our use of fluorescent lymphography to guide LND and retrieval increased the number of positive lymph nodes retrieved in comparison with conventional methods. Thus, whether this technique can change tumor staging and consequently influence clinical practice should be evaluated in further studies including a larger number of patients with lymph node metastasis. Nonetheless, our observation that this technique increased the number of lymph nodes retrieved suggests that fluorescent lymphography is useful for proper staging. Fourth, our study had a relatively small number of cases to analyze the lymphatic drainage patterns according to each tumor location. Further study with a larger number of patients would show the correlation between tumor location and lymphatic drainage patterns.
Conclusions
Near-infrared fluorescent lymphography after injection of ICG 1 day before surgery facilitated effective visualization of draining lymph nodes, providing guidance for LND during surgery for gastric cancer. We deemed fluorescent lymphography to be a useful tool for performing a complete lymphadenectomy, assessing the adequacy of lymphadenectomy intraoperatively, and retrieving lymph nodes thoroughly from resected specimens for pathologic staging. Our protocol of florescent lymphography was found to offer increased lymph node retrieval without any discernible disadvantages. Wider adoption of this technique should be considered in clinical practice.
eFigure 1. Schema of the Prospective Study Design for Comparing Study Patients With a Matched Historical Control Group
eFigure 2. Flowchart Detailing the Examination of the Lymph Nodes
References
- 1.Japanese Gastric Cancer Association Japanese gastric cancer treatment guidelines 2014 (ver. 4). Gastric Cancer. 2017;20(1):1-19. doi: 10.1007/s10120-016-0622-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Ajani JA, Bentrem DJ, Besh S, et al. ; National Comprehensive Cancer Network . Gastric cancer, version 2.2013: featured updates to the NCCN guidelines. J Natl Compr Canc Netw. 2013;11(5):531-546. doi: 10.6004/jnccn.2013.0070 [DOI] [PubMed] [Google Scholar]
- 3.Bonenkamp JJ, Hermans J, Sasako M, et al. ; Dutch Gastric Cancer Group . Extended lymph-node dissection for gastric cancer. N Engl J Med. 1999;340(12):908-914. doi: 10.1056/NEJM199903253401202 [DOI] [PubMed] [Google Scholar]
- 4.Songun I, Putter H, Kranenbarg EM, Sasako M, van de Velde CJ. Surgical treatment of gastric cancer: 15-year follow-up results of the randomised nationwide Dutch D1D2 trial. Lancet Oncol. 2010;11(5):439-449. doi: 10.1016/S1470-2045(10)70070-X [DOI] [PubMed] [Google Scholar]
- 5.Wu CW, Hsiung CA, Lo SS, et al. Nodal dissection for patients with gastric cancer: a randomised controlled trial. Lancet Oncol. 2006;7(4):309-315. doi: 10.1016/S1470-2045(06)70623-4 [DOI] [PubMed] [Google Scholar]
- 6.Son T, Hyung WJ, Lee JH, et al. Clinical implication of an insufficient number of examined lymph nodes after curative resection for gastric cancer. Cancer. 2012;118(19):4687-4693. doi: 10.1002/cncr.27426 [DOI] [PubMed] [Google Scholar]
- 7.Edge SB, Byrd DR, Compton CC, Fritz AG, Greene FL, Trotti A, eds. AJCC Cancer Staging Manual. 7th ed New York, NY: Springer; 2010. [Google Scholar]
- 8.Siewert JR, Böttcher K, Stein HJ, Roder JD. Relevant prognostic factors in gastric cancer: ten-year results of the German Gastric Cancer Study. Ann Surg. 1998;228(4):449-461. doi: 10.1097/00000658-199810000-00002 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Wu CW, Hsieh MC, Lo SS, Tsay SH, Lui WY, P’eng FK. Relation of number of positive lymph nodes to the prognosis of patients with primary gastric adenocarcinoma. Gut. 1996;38(4):525-527. doi: 10.1136/gut.38.4.525 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Smith DD, Schwarz RR, Schwarz RE. Impact of total lymph node count on staging and survival after gastrectomy for gastric cancer: data from a large US-population database. J Clin Oncol. 2005;23(28):7114-7124. doi: 10.1200/JCO.2005.14.621 [DOI] [PubMed] [Google Scholar]
- 11.Coburn NG, Swallow CJ, Kiss A, Law C. Significant regional variation in adequacy of lymph node assessment and survival in gastric cancer. Cancer. 2006;107(9):2143-2151. doi: 10.1002/cncr.22229 [DOI] [PubMed] [Google Scholar]
- 12.Vahrmeijer AL, Hutteman M, van der Vorst JR, van de Velde CJ, Frangioni JV. Image-guided cancer surgery using near-infrared fluorescence. Nat Rev Clin Oncol. 2013;10(9):507-518. doi: 10.1038/nrclinonc.2013.123 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Herrera-Almario G, Patane M, Sarkaria I, Strong VE. Initial report of near-infrared fluorescence imaging as an intraoperative adjunct for lymph node harvesting during robot-assisted laparoscopic gastrectomy. J Surg Oncol. 2016;113(7):768-770. doi: 10.1002/jso.24226 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Kim H, Lee SK, Kim YM, et al. Fluorescent iodized emulsion for pre- and intraoperative sentinel lymph node imaging: validation in a preclinical model. Radiology. 2015;275(1):196-204. doi: 10.1148/radiol.14141159 [DOI] [PubMed] [Google Scholar]
- 15.Tajima Y, Yamazaki K, Masuda Y, et al. Sentinel node mapping guided by indocyanine green fluorescence imaging in gastric cancer. Ann Surg. 2009;249(1):58-62. doi: 10.1097/SLA.0b013e3181927267 [DOI] [PubMed] [Google Scholar]
- 16.Kim HI, Hyung WJ, Lee CR, et al. Intraoperative portable abdominal radiograph for tumor localization: a simple and accurate method for laparoscopic gastrectomy. Surg Endosc. 2011;25(3):958-963. doi: 10.1007/s00464-010-1288-3 [DOI] [PubMed] [Google Scholar]
- 17.Song J, Oh SJ, Kang WH, Hyung WJ, Choi SH, Noh SH. Robot-assisted gastrectomy with lymph node dissection for gastric cancer: lessons learned from an initial 100 consecutive procedures. Ann Surg. 2009;249(6):927-932. doi: 10.1097/01.sla.0000351688.64999.73 [DOI] [PubMed] [Google Scholar]
- 18.Kim YM, Son T, Kim HI, Noh SH, Hyung WJ. Robotic D2 lymph node dissection during distal subtotal gastrectomy for gastric cancer: toward procedural standardization. Ann Surg Oncol. 2016;23(8):2409-2410. doi: 10.1245/s10434-016-5166-7 [DOI] [PubMed] [Google Scholar]
- 19.Son T, Lee JH, Kim YM, Kim HI, Noh SH, Hyung WJ. Robotic spleen-preserving total gastrectomy for gastric cancer: comparison with conventional laparoscopic procedure. Surg Endosc. 2014;28(9):2606-2615. doi: 10.1007/s00464-014-3511-0 [DOI] [PubMed] [Google Scholar]
- 20.Japanese Gastric Cancer Association Japanese classification of gastric carcinoma: 3rd English edition. Gastric Cancer. 2011;14(2):101-112. doi: 10.1007/s10120-011-0041-5 [DOI] [PubMed] [Google Scholar]
- 21.Amin MB, Edge S, Greene F, et al, eds. AJCC Cancer Staging Manual. 8th ed New York, NY: Springer; 2017. [Google Scholar]
- 22.Lee JH, Kim YW, Ryu KW, et al. A phase-II clinical trial of laparoscopy-assisted distal gastrectomy with D2 lymph node dissection for gastric cancer patients. Ann Surg Oncol. 2007;14(11):3148-3153. doi: 10.1245/s10434-007-9446-0 [DOI] [PubMed] [Google Scholar]
- 23.Miyashiro I, Hiratsuka M, Kishi K, et al. Intraoperative diagnosis using sentinel node biopsy with indocyanine green dye in gastric cancer surgery: an institutional trial by experienced surgeons. Ann Surg Oncol. 2013;20(2):542-546. doi: 10.1245/s10434-012-2608-8 [DOI] [PubMed] [Google Scholar]
- 24.Becher RD, Shen P, Stewart JH, Geisinger KR, McCarthy LP, Levine EA. Sentinel lymph node mapping for gastric adenocarcinoma. Am Surg. 2009;75(8):710-714. [PubMed] [Google Scholar]
- 25.Candela FC, Urmacher C, Brennan MF. Comparison of the conventional method of lymph node staging with a comprehensive fat-clearing method for gastric adenocarcinoma. Cancer. 1990;66(8):1828-1832. doi: [DOI] [PubMed] [Google Scholar]
- 26.Al-Refaie WB, Gay G, Virnig BA, et al. Variations in gastric cancer care: a trend beyond racial disparities. Cancer. 2010;116(2):465-475. doi: 10.1002/cncr.24772 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eFigure 1. Schema of the Prospective Study Design for Comparing Study Patients With a Matched Historical Control Group
eFigure 2. Flowchart Detailing the Examination of the Lymph Nodes