Assessment of Carbon Nanoparticle Suspension Lymphography–Guided Distal Gastrectomy for Gastric Cancer

Key Points

Question  Is use of carbon nanoparticle suspension injection (CNSI) in distal gastrectomy associated with improved detection of lymph nodes (LNs), including metastatic LNs, compared with conventional lymphadenectomy?

Findings  In this cohort study of 312 patients with gastric cancer, significantly more LNs were detected in the CNSI group than in the conventional group. CNSI was associated with accurately identifying metastatic LNs.

Meaning  These findings suggest that CNSI in distal gastrectomy was associated with improved identification of lymph nodes and was diagnostically useful.

Importance  Carbon nanoparticle suspension injection (CNSI) can be used to visualize lymph node (LN) drainage in gastric cancer. The tracing and diagnostic value of carbon nanoparticle suspension lymphography-guided distal gastrectomy for gastric cancer has not been thoroughly reported.

Objective  To compare the number of lymph nodes identified in patients with gastric cancer receiving a CNSI vs no injection.

Design, Setting, and Participants  This is a retrospective cohort study including patients with clinical T1 to T4 disease who underwent laparoscopic or robotic distal gastrectomy. Data from a cohort of 1225 patients at the Fourth Hospital of Hebei Medical University (Shijiazhuang, China) from November 2019 to February 2021 were analyzed. Patients were divided into the CNSI group and conventional group after 1:1 propensity matching analysis. The mean number of LNs detected was compared between groups, and the diagnostic role of CNSI was analyzed in the CNSI group. Statistical analysis was performed from May to July 2021.

Exposure  CNSI was peritumorally injected under an endoscope 1 day before surgery in the CNSI group, and the conventional group did not receive any treatment before surgery.

Main Outcomes and Measures  The main outcome was the number of LNs detected. Gastrectomy with systematic D1+ (ie, stations 1, 3, 4sb, 4d, 5, 6, and 7) or D2 (ie, all D1 stations, plus 8a, 9, 11p, and 12a) lymphadenectomy was performed. Black-stained LNs and nonblack-stained LNs were examined separately in the CNSI group.

Results  A total of 312 consecutive patients (mean [SD] age, 56.7 [10.4] years; 216 [69.2%] men) who underwent distal gastrectomy were enrolled, including 78 patients in the CNSI group, and another 78 patients determined from 1:1 propensity score matching, making an overall cohort size of 156 patients. The mean (SD) number of LNs detected in the CNSI group was 59.6 (21.4), which was significantly higher than that in the conventional group (30.0 [11.3] LNs; P < .001). In the CNSI group, the mean (SD) number of LNs detected at black-stained LN stations was significantly higher than that at nonstained LN stations (9.2 [6.1] LNs per station vs 3.5 [3.2] LNs per station; P < .001). For black-stained LN stations, the sensitivity was 97.8% (95% CI, 91.6%-99.6%), specificity was 38.1% (95% CI, 34.2%-42.3%), positive predictive value was 20.1% (95% CI, 16.6%-24.2%), and negative predictive value was 99.1% (95% CI, 96.4%-99.8%); for the black-stained LNs, sensitivity was 97.6% (95% CI, 95.3%-98.8%), specificity was 35.4% (95% CI, 33.9%-36.8%), positive predictive value was 11.6% (95% CI, 10.5%-12.8%), and negative predictive value was 99.4% (95% CI, 98.8%-99.7%).

Conclusions and Relevance  These findings suggest that CNSI was associated with facilitating the dissection of all positive LNs, which could improve surgical quality. Carbon nanoparticle suspension-guided lymphography may be an alternative to conventional systematic lymphadenectomy for distal gastrectomy.

The incidence and mortality of gastric cancer (GC) have decreased; however, it remains one of the most common cancers in the world, ranking as the fifth most commonly diagnosed cancer after lung cancer, breast cancer, colorectal cancer, and prostate cancer and ranking second for mortality.^1-3

Gastrectomy is the only possible cure for advanced GC. Adequate lymph node (LN) dissection plays an important role in gastrectomy, and adequate lymphadenectomy is an important factor for accurate postoperative pathological staging and improved prognosis.^4-6 Unvisualized LNs are difficult to identify in adipose tissue, and standard lymphadenectomy bears the risk of incomplete resection of potentially metastatic LNs.⁷ Moreover, sufficient dissection of LNs from the resected specimen is critical for the evaluation of nodal status.^8,9

In recent years, various dyes and tracers have been used clinically to observe LN drainage from primary tumors. Carbon nanoparticle suspension injection (CNSI) has been found to provide surgeons with effective visualization of the lymphatic anatomy and sentinel LNs.^10-12

CNSI^13,14 is different from other tracers and has a diameter of 150 nm. After being injected into the surrounding tissue of the tumor, CNSI is swallowed by macrophages and quickly enters lymphatic vessels but does not enter blood vessels. It is retained in LNs and stains the LNs black, thus achieving the living staining of drainage LNs in the tumor region. In addition, CNSI can be observed in vivo after approximately 3 to 4 months,¹³ allowing sufficient time to observe the stained LNs for adequate intraoperative dissection and postoperative LN detection.

However, the pattern of LN staining of carbon nanoparticles and whether stained lymphography detects all potentially metastatic LNs are unknown. This study aimed to evaluate LN dissection quality and the pattern of LN staining associated with carbon nanoparticle suspension lymphography–guided distal gastrectomy for GC, analyze the diagnostic value associated with peritumoral CNSI injection 1 day before surgery, and analyze the sensitivity associated with CNSI in detecting metastatic LN stations and LNs during distal gastrectomy.

This cohort study was approved by the Medical Ethics Committee of the Fourth Hospital of Hebei Medical University, and informed consent was obtained from patients who underwent relevant data analysis. This study is reported following the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

We retrospectively reviewed a prospectively collected GC database to identify patients who underwent robotic or laparoscopic gastrectomy between May 2019 and December 2020. The inclusion criteria were age 18 years or older, histologically confirmed adenocarcinoma of the stomach or esophagogastric junction, distal subtotal gastrectomy performed according to the tumor location, and clinical stage T1 to T4a, N0 to N3, and M0. The exclusion criteria were a previous history of other cancers, having received previous curative resection (such as endoscopic submucosal dissection or endoscopic mucosal resection), and having received preoperative chemotherapy or radiation therapy.

CNSI (50 mg per dose, Chongqing Lesmei Pharmaceutical) was marked in the endoscopy division 1 day before surgery. CNSI was injected submucosally at 4 points (proximal side, distal side, and left and right sides) 0.5 to 1 cm from the tumor edge under endoscopy. The test dose for each point was approximately 0.25 mL (Figure 1).

All patients who were enrolled underwent radical distal subtotal gastrectomy using the da Vinci Surgical Xi system (Intuitive Surgical) or laparoscopy. D1+ (ie, stations 1, 3, 4sb, 4d, 5, 6, and 7) lymphadenectomy was performed for patients with clinically early GC without suspicion of nodal metastases, and D2 (ie, all D1 stations, plus 8a, 9, 11p, and 12a) lymphadenectomy was performed for patients with advanced GC or any suspicion of nodal metastases. When residual stained LNs were identified in the range of D1+ or D2 LN dissection, these LNs were additionally removed (Video; eFigure 1 in the Supplement). Although we occasionally detected stained LNs outside the planned dissection area (eg, stations 4sa, 11d, and 14v), excessive dissection beyond the scope of D2 lymphadenectomy was not performed.

Morbidity and mortality were assessed within 30 days after surgery. Postoperative complications were graded according to the Clavien-Dindo classification.¹⁵

After carbon nanoparticle suspension lymphography–guided lymphadenectomy, each LN station was first separated ex vivo from the resected specimen in accordance with the definitions of the Japanese classification of gastric carcinoma.¹⁶ LNs were classified as black-stained if they were stained with CNSI and nonstained if they were not stained with CNSI. Stations containing stained LNs were classified as black-stained stations, while those without stained LNs were classified as nonstained stations. Each LN was measured with a Vernier caliper, and LNs smaller than 5 mm were defined as micro-LNs (Figure 1). The presence and absence of CNSI staining were subsequently matched with the results of histopathological analysis. For true-positive (TP) stations, LNs were defined as stained stations and contained metastatic LNs on histopathological examination. For false-negative (FN) stations, LNs were defined as nonstained stations but contained metastatic LNs on histopathological examination. For false-positive (FP) stations, LNs were defined as black-stained but were tumor-free. For true-negative (TN) stations, LNs were defined as nonstained and tumor-free. The sensitivity of carbon nanoparticle suspension lymphography-guided lymphography for detecting LN metastasis in stations or LNs was calculated as TP / (TP + FN). The specificity was calculated as TN / (TN + FP); the PPV was calculated as TP / (TP + FP); and the NPV was calculated as TN / (TN + FN).

Continuous variables are reported as medians with IQRs, and the t test was used to detect differences. Categorical variables are reported as numbers and proportions, and the χ² test or Fisher exact test was used to compare the differences in proportions depending on the application criteria. To overcome biases from the different distributions of covariates among patients in the 2 study groups, a propensity score analysis was performed. The model was used to obtain a 1:1 match using the nearest-neighbor matching method. Patients for whom the propensity score was not applicable were excluded from further analysis. Two-sided P < .05 was considered statistically significant. SPSS statistical software version 24 (IBM) was used for all statistical analyses. Statistical analysis was performed from May to July 2021.

A total of 312 patients (mean [SD] age, 56.7 [10.4] years; 216 [69.2%] men) underwent laparoscopic and robotic distal subtotal gastrectomy from May 2019 to December 2020. Three patients had a previous history of other cancers, 40 patients had received preoperative chemotherapy or chemoradiotherapy, and 12 patients had received previous curative resection. The remaining 257 patients were included in the analysis, of whom 78 received preoperative endoscopic injections of CNSI (CNSI group), and 179 underwent surgery without CNSI (conventional group). After 1:1 propensity matching analysis, there were 78 patients each in the CNSI group and conventional group (eFigure 2 in Supplement). The baseline characteristics of patients in both groups were balanced after matching (Table 1).

A total of 669 LN stations were detected in the CNSI group, including 4647 LNs, and 704 LN stations were detected in the conventional group, including 2343 LNs. A mean (SD) of 59.6 (21.4) LNs per patient were detected in the CNSI group, which was significantly higher than that in the conventional group (30.0 [11.3] LNs per patient; P < .001). There was no difference in number of LN stations detected between the CNSI and conventional groups (mean [SD], 8.6 [1.7] LN stations per patient vs 9.0 [2.0] LN stations per patient; P = .14).

A mean (SD) of 6.9 (4.1) LNs per station were detected in the CNSI group, which was higher than the rate in the conventional group (3.3 [2.5] LNs per station). In particular, there were statistically significant differences in stations 3, 4d, 5, 6, 7, 8a, 9, and 12a (Figure 2). In the CNSI group, 447 black-stained stations (66.8%) and 3126 black-stained LNs (67.3%) were detected. There was mean (SD) of 40.1 (16.5) black-stained LNs detected per patient and a mean (SD) of 5.5 (1.4) black-stained stations detected per patient in the CNSI group. Furthermore, in the CNSI group, the mean (SD) number of LNs detected in black-stained stations was significantly higher than that detected in nonstained stations (9.2 [6.1] LNs per station vs 3.5 [3.2] LNs per station; P < .001), except station 12a. All LNs detected in station 3 were black-stained, and there were significant differences in the numbers of LNs detected in stained vs unstained stations in station 1 (mean [SD], 5.0 [1.0] LNs vs 1.0 [0.0] LNs; P = .002), station 5 (mean [SD], 5.1 [2.1] LNs vs 1.7 [0.5] LNs; P = .01), and station 11p (mean [SD], 5.2 [2.7] LNs vs 2.3 [0.9] LNs; P = .003) (Figure 3).

The mean (SD) number of LN metastases was not significantly different between the CNSI and conventional groups (4.8 [9.3] LN metastases vs 4.1 [6.8] LN metastases; P = .62). The mean (SD) number of micro-LNs in the CNSI group was significantly higher than that in the conventional group (12.5 [8.9] micro-LNs vs 4.5 [3.6] micro-LNs; P < .001).

Except for nonmetastatic LNs in stations 1 and 12a, the metastasis rate of black-stained LNs in other stations was higher than that of nonstained LNs, and there were significant differences in stations 4 sb, 4d, 6, 7, 8a, and 9 (eTable 2 in the Supplement).

Of 78 patients in the CNSI group, 36 (46.2%) had LN metastasis. In 34 of these patients with LN metastases, all metastases were observed in black-stained stations only. In 1 patient with LN metastases, metastases were detected in black-stained and nonstained stations, and 1 patient displayed metastases exclusively in nonstained stations. The sensitivity for the detection of metastatic stations of carbon nanoparticle suspension lymphography was 97.8% (95% CI, 91.6%-99.6%), the specificity was 38.1% (95% CI, 34.2%-42.3%), the positive predictive value (PPV) was 20.1% (95% CI, 16.6%-24.2%), and the negative predictive value (NPV) was 99.1% (95% CI, 96.4%-99.8%). The sensitivity for the detection of metastatic LNs of carbon nanoparticle suspension lymphography was 97.6% (95% CI, 95.3%-98.8%), the specificity was 35.4% (95% CI, 33.9%-36.8%), the PPV was 11.6% (95% CI, 10.5%-12.8%), and the NPV was 99.4% (95% CI, 98.8%-99.7%) (Table 2). This analysis applied only to the CNSI group, and the goal was to determine if black staining could estimate LN metastasis.

No significant differences between the CNSI and conventional groups were observed in intraoperative blood loss (mean [SD], 55.5 [29.0] mL vs 57.2 [31.0] mL; P = .70) and operative time (mean [SD], 180.2 [29.1] minutes vs 184.5 [25.6] minutes; P = .33) (eTable 1 in the Supplement). The postoperative recovery process was comparable for both groups. There were no significant differences in time to first flatus, time to ambulation, time to first liquid intake, and postoperative hospital stay between the 2 groups. No significant differences were found between the CNSI and conventional groups regarding the incidence of postoperative complications within 30 days after surgery(6 of 78 patients [7.7%] vs 7 of 78 patients [8.9%]; P = .77), nor regarding severity of postoperative complications.

In this cohort study, the total number of LNs detected and the number of LNs detected per station in the CNSI group were significantly higher than in the conventional group, suggesting that carbon nanoparticle suspension lymphography-guided distal gastrectomy achieved sufficient LN detection. In previous GC studies, carbon nanoparticles were used to identify only sentinel LNs in early GC.⁹ The number of LNs detected and the diagnostic value of black-stained LNs have not been discussed, to our knowledge. Compared with India ink, CNSI is relatively safe.¹⁷ Compared with indocyanine green fluorescence, CNSI does not require complicated surgical procedures and equipment, and making its use simpler than that of indocyanine green fluorescence.¹⁸ Methylene blue, indigo carmine and other dyes have shorter half-lives than CNSI.¹⁹ The CNSI black stain contrasts well with the color of adipose tissue; thus, LNs can be easily identified during surgery. The surgeon can achieve en bloc resection of LNs without breakage of lymphatic structures, reducing the risk of intraoperative tumor cell spillage and preventing vascular injury in the surrounding tissue. Moreover, the quality control of lymphadenectomy can be carried out well by determining whether there are residual stained LNs within the scope of the scheduled LN dissection.

In the process of postoperative LN sorting, black-stained LNs are more easily obtained from resected specimens without any other methods,²⁰ such as fluorescence laparoscopy, secondary searches, or the Lipodissolve technique. It also was associated with increasing the number of micro-LNs detected compared with conventional treatment. Our results show that with CNSI, the accuracy of negative LNs and LN pathological staging may be improved, and it could provide the guidance for postoperative adjuvant therapy, which in turn may improve the prognosis of patients.^21-25

The results of this study showed that the number of LNs detected in black-stained stations was greater than that detected in nonblack-stained stations, and the metastasis rate of black-stained LNs was higher than that of nonblack-stained LNs. Both black-stained LNs and stations were highly sensitive to pathological diagnosis, and the NPV of unstained LNs was also satisfactory. Our findings of carbon nanoparticle suspension lymphography–guided distal gastrectomy are not equivalent to those of a 2016 study of sentinel LN dissection.²⁵ The purpose of this study was to perform en bloc LN dissection, not only resection of sentinel LNs. A study by Yan et al²⁶ using carbon nanoparticles to show sentinel LNs in early GC reported that in black-stained sentinel LNs in early GC, sensitivity was 90%, specificity was 100%, and accuracy was 98.9%. In contrast, the sensitivities of stained LNs or stations in this study to detect metastasis were higher. This result suggests that carbon nanoparticle suspension lymphography–guided distal gastrectomy was associated with a more thorough LN dissection, reduced number of residual positive LNs, and increased number of negative or micro-LNs.

There are some limitations in this study. First, there is no further analysis on the CNSI method and the optimal injection time before the operation for LN detection. Our injection rules were conducted for peritumoral submucosal injection of CNSI 1 day before surgery. Although this timing may result in an increase in additional endoscopy and be painful for patients, it was associated with a reduction in the probability of surgical field pollution compared with subserosal injection. The submucosal injection 1 day before the operation was associated with an increase in the diffusion time of CNSI in the lymph system. Second, this study did not include patients who underwent total gastrectomy, and we will analyze carbon nanoparticle suspension lymphography–guided total gastrectomy for patients with GC in future studies. Third, carbon nanoparticle–guided distal subtotal gastrectomy LN dissection was associated with an increase in the number of LNs detected. However, in this study, owing to the short follow-up time, no long-term survival analysis was performed. We plan provide survival results after further follow-up.

In this cohort study, carbon nanoparticle suspension lymphography–guided distal gastrectomy was associated with increasing the number of postoperative LNs detected and the accuracy of LN pathological staging. It also was associated with promoting the removal of all potential metastatic LNs, which suggests it may suitable to replace conventional lymphadenectomy for distal gastrectomy.

Accepted for Publication: January 31, 2022.

Published: April 18, 2022. doi:10.1001/jamanetworkopen.2022.7739

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2022 Tian Y et al. JAMA Network Open.

Corresponding Author: Qun Zhao, MD, Third Surgery Department, Fourth Hospital of Hebei Medical University, No.12, Jian-Kang Rd, Shijiazhuang, 050019, China (zhaoqun@hebmu.edu.cn).

Author Contributions: Dr Zhao had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Tian, Liu, Li, Zhao.

Acquisition, analysis, or interpretation of data: Tian, Yang, Lin, Hu, Deng, Ma, Guo, Z. Zhang, Ding, Fan, Z. d. Zhang, Wang, Zhao.

Drafting of the manuscript: Tian, Deng, Z. Zhang, Z. d. Zhang, Wang.

Critical revision of the manuscript for important intellectual content: Yang, Lin, Hu, Ma, Guo, Liu, Ding, Li, Fan, Zhao.

Statistical analysis: Tian, Yang, Lin, Hu, Deng, Liu, Ding, Li, Fan, Z. d. Zhang, Wang, Zhao.

Administrative, technical, or material support: Deng, Ma, Guo.

Conflict of Interest Disclosures: None reported.

Funding/Support: This study was supported by the University Research Project of Hebei Province (grant No. ZD2019139) and the Medical Research Project of Hebei Province (grant No. 20211339).

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.