Navigating Nodal Metrics for Node-Positive Gastric Cancer in the United States: An NCDB-Based Study and Validation of AJCC Guidelines

View More View Less
  • 1 Department of Surgical Oncology,
  • | 2 Department of Gastrointestinal Medical Oncology,
  • | 3 Department of Pathology, and
  • | 4 Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas.

Background: The optimal number of examined lymph nodes (ELNs) and the positive lymph node ratio (LNR) for potentially curable gastric cancer are not established. We sought to determine clinical benchmarks for these values using a large national database. Methods: Demographic, clinicopathologic, and treatment-related data from patients treated using an R0, curative-intent gastrectomy registered in the National Cancer Database during 2004 to 2016 were evaluated. Patients with node-positive (pTxN+M0) disease were considered for analysis. Results: A total of 22,018 patients met the inclusion criteria, with a median follow-up of 2.2 years. Mean age at diagnosis was 65.6 years, 66% were male, 68% were White, 33% of tumors were located near the gastroesophageal junction, and 29% of patients had undergone preoperative therapy. Most primary tumors (62%) were category pT3–4, 67% had a poor or anaplastic grade, and 19% had signet features. Clinical nodal staging was inaccurate compared with staging at final pathology. The mean [SD] number of nodes examined was 19 [11]. On multivariable analysis, the pN category, ELNs, and LNR were independently associated with survival (all P<.0001). Using receiver operating characteristic (ROC) analysis, an optimal ELN threshold of ≥30 was established for patients with pN3b disease and was applied to the entire cohort. Node positivity and LNR had minimal change beyond 30 examined nodes. Stage-specific LNR thresholds calculated by ROC analysis were 11% for pN1, 28% for pN2, 58% for pN3a, 64% for pN3b, 30% for total combined. By using an ELN threshold of ≥30, prognostically advantageous stage-specific LNR values could be determined for 96% of evaluated patients. Conclusions: Using a large national cancer registry, we determined that an ELN threshold of ≥30 allowed for prognostically advantageous LNRs to be achieved in 96% of patients. Therefore, ≥30 examined nodes should be considered a clinical benchmark for practice in the United States.

Background

Despite decreasing incidence, gastric cancer remains a prevalent and deadly condition in the United States.1 Five-year survival is among the lowest of commonly diagnosed malignancies.2 Fortunately, there is a modest trend toward earlier detection, and approximately half of patients with newly diagnosed disease present with localized or locoregional disease potentially amenable to curative resection.3 Among these patients, nodal metastases are the most important prognostic determinant of survival. However, debate remains regarding the optimal management of lymph nodes, which can be assessed by the number of stations sampled and by the total number of examined lymph nodes (ELNs).4 Both metrics address different and important aspects of oncologic adequacy with regard to lymphadenectomy. Increasingly, there is consensus among Western surgeons that an extended lymphadenectomy (D1+/D2) can be performed safely and confers a survival advantage.57 However, the determination of optimal ELNs and subsequent contextualization by lymph node ratio (LNR) is underutilized and not well established.8

One of the challenges with calculating ELN and LNR thresholds involves patient selection and data input. Low-burden nodal disease appropriately requires a lower ELN count and LNR for discrimination, whereas the converse is true for high-burden disease. Given the inaccuracy of clinical staging in gastric cancer, it is prudent to apply to all patients a more stringent threshold predicated on input from the highest-burden disease. Given this context, we sought to establish ELN and LNR clinical benchmarks in a stage-specific manner for gastric cancer surgery in the United States using the National Cancer Database (NCDB).

Methods

Patient Selection

We queried the NCDB spanning the years 2004 to 2016 for patients with gastric cancer (primary site codes: C16.0–C16.6, C16.8, and C16.9). The following histologic codes were subsequently applied: 8000–8152, 8154–8231, 8243–8245, 8247–8248, 8250–8499, 8560, and 8574. From this initial capture, the following histologic codes were excluded: 8000, 8003, 8004, 8005, 8010, 8012, 8013, 8014, 8022, 8030, 8032, 8033, 8040, 8041, 8046, 8050, 8082, 8153, 8154, 8220, 8240, 8241, 8246, 8249, 8310, and 8390. Any patient with clinical or pathologic M-stage positive disease, patients lacking pT or pN information, and patients who did not undergo a definitive surgical resection were excluded. Patients with missing or inaccurate survival data were also excluded, as were those with final pathologic node-negative disease or a positive surgical margin. We did not exclude patients based on administration or timing of administration of systemic therapy.

Covariates for Evaluation

Demographic covariates included sex, age, and race. Clinicopathologic covariates included tumor location (gastroesophageal vs gastric), clinical and pathologic TNM stage, margin status, tumor grade (well/moderate vs poor differentiation), and presence of signet cells, linitis, or lymphovascular invasion (LVI). The type of gastrectomy was documented, but information on the type of lymphadenectomy (D1 vs D2) is not recorded in the NCDB. Other treatment-related covariates included preoperative therapy (chemotherapy vs chemoradiation), number of nodes removed, number of positive nodes, and hospital length of stay after definitive surgical procedure.

Statistical Analysis

Covariates including patient demographics, tumor characteristics, treatments, and outcomes were analyzed. Categorical covariates were compared between cohorts using the Pearson chi-square test, and continuous covariates were compared between cohorts in the same manner using the Student t test. Survival curves were plotted using the Kaplan-Meier method, with 95% confidence intervals calculated by using Greenwood’s formula. Unadjusted and multivariable analyses to identify prognostic markers of overall survival (OS) were performed using Cox proportional hazard models. All tests performed were 2-sided, and a significance level of P<.05 indicated statistical significance. All statistical analyses were performed using Intercooled STATA, version 16.1 (StataCorp LLC), and graphical images were created using GraphPad Prism, version 8.0d (GraphPad Software, Inc).

Results

Patient, Tumor, and Treatment-Related Details

Based on the exclusion criteria, a starting cohort of 222,389 patients was filtered to a final cohort of 22,018 patients (10%) who met the inclusion criteria with a median follow-up of 2.2 years (supplemental eTable 1, available with this article at JNCCN.org). Mean [SD] age at diagnosis was 65.6 [12.7] years, 66% were male, and 68% were White (Table 1). The primary tumor was located in the gastroesophageal junction in 33% of patients, and 33% had no evidence of nodal disease on clinical staging despite positive final pathology. Among the primary tumors analyzed, 62% were category pT3–4, 67% had a poor or anaplastic grade, 19% had signet features, and <1% had linitis. Only 53% of patients had documented LVI status, and of these, 33% were LVI-positive; among patients with pN0 disease, however, only 9.5% were LVI-positive (P<.0001; data not shown). A total of 29% of patients were treated using systemic preoperative therapy, and 46% underwent total gastrectomy. The mean [SD] number of nodes examined was 19 [11], and 11% of patients had documented resection of an additional organ. Finally, 27% of patients were admitted for >14 days after their definitive oncologic resection.

Table 1.

Cohort Details

Table 1.

Association of Node Burden, ELN, and LNR With Survival

Node burden, measured by pN category, ELNs, and LNR, was evaluated for prognostic value by multivariable analysis using the Cox proportional hazards model (Table 2). An increasing pN category was associated with reduced survival, and the greatest increase in risk of death (47%) occurred between the pN3a and pN3b categories. Patients with category pN3b disease accordingly had the greatest risk of death (hazard ratio [HR], 1.87; 95% CI, 1.69–2.07; P<.0001). ELN count was protective, with each additional 10 nodes examined conferring an approximate 10% reduction in the risk of death. An ELN count of ≥30 was associated with the lowest risk of death (HR, 0.66; 95% CI, 0.61–0.72; P<.0001). Finally, LNR was the most discriminant marker for OS of the 3 nodal metrics. Regardless of pN category, an LNR of ≥75% was associated with a >70% increased risk of death (Figure 1).

Table 2.

Overall Survival Analysis for Nodal Metrics

Table 2.
Figure 1.
Figure 1.

Kaplan-Meier survival curves for OS stratified by (A) pN category, (B) LNR, and (C) ELNs.

Abbreviations: ELN, examined lymph node; LNR, lymph node ratio; OS, overall survival.

Citation: Journal of the National Comprehensive Cancer Network 2021; 10.6004/jnccn.2021.7038

LNR and ELN Prognostic Value Accounting for Node Stage

We next evaluated ELNs and LNR stratified by pN category to address the influence of node burden. Across pN categories, a low LNR (<25%) was associated with a similar risk of death (Table 3). However, an increasing LNR beyond 25% was concordantly associated with an increased risk of death, and the magnitude of this increase correlated with pN category (an LNR of ≥75% was associated with the greatest risk: HRs of 1.78 for pN1, 2.72 for pN2, 2.96 for pN3a, and 3.32 for pN3b). Thus, although LNR partly addressed stage migration, stratifying by node burden provided more accurate prognostic categories for this metric. In contrast, increasing ELNs was protective across all pN categories, although this effect diminished with increasing node burden (ELNs ≥30: HRs of 0.48 for pN1, 0.63 for pN2, 0.86 for pN3a, and 1.61 for pN3b). For very high node burden disease (category pN3b), an increasing ELN count maintained a protective effect for ≥30 nodes examined, but overall this subgroup had a worse prognosis regardless of node retrieval.

Table 3.

OS Analysis for LNR and ELNs Stratified by pN Category

Table 3.

LNR and ELN Thresholds Adjusted for Node Stage

Given the prognostic value of LNR and ELN, we sought to determine threshold values for clinical use. We performed receiver operating characteristic (ROC) analyses for each continuous variable stratified by pN category to adjust for node burden (Table 4). Threshold values were determined by maximal sensitivity and specificity using the Youden’s J statistic. The C index for LNR was uniformly larger than that for ELNs, which was expected based on the multivariable analysis. Threshold values for LNR increased by 17% between pN1 and pN2 and by 30% between pN2 and pN3a. Only a 6% increase in LNR was observed between pN3a and pN3b. The increase in discriminant LNR value with increasing pN category was a result of an increasing positive node burden with a diminishing increase in node retrieval. In contrast, the calculated ELN threshold increased concordantly with node category, and the greatest increase occurred between the pN3a and pN3b categories (from 19 to 27). Based on this analysis, we chose an ELN threshold of 30 for the following reasons: (1) this value captured all calculated thresholds for pN categories, (2) 30 nodes was in line with the recommendations of the 8th edition of the AJCC Cancer Staging Manual, and (3) rounding up from a calculated threshold of 27 for pN3b disease may partly address issues of stage migration.9

Table 4.

ROC Analysis of LNR and ELN Threshold

Table 4.

Interpreting the Utility of LNR and ELN Thresholds Among Node-Positive Patients

We next sought to contextualize our determined LNR and ELN thresholds among all patients with node-positive gastric cancer. For this cohort of 22,018 patients with stage I–III gastric cancer treated using an R0, curative-intent resection, only 6% of individuals had ≥16 positive nodes, and <3% had >20 positive nodes (Figure 2A). These percentages were likely higher in actuality due to the lack of inadequate oncologic resection in a subset of patients; nonetheless, they provided context for the scope of node burden.

Figure 2.
Figure 2.

(A) Distribution of positive node burden (blue) and associated cumulative incidence (orange) (n=22,018). (B) Scatterplot displaying LNR by ELNs. Blue curve represents a modeled semilog curve, indicating diminishing reduction in LNR with increasing ELNs. Orange-hashed line represents capture among the entire cohort with calculated ELN threshold ≥30. (C) Line graph of mean LNR and positive node count with SD error stratified by ELNs.

Abbreviations: ELN, examined lymph node; LNR, lymph node ratio.

Citation: Journal of the National Comprehensive Cancer Network 2021; 10.6004/jnccn.2021.7038

When we compared LNR to ELNs for each patient, we found that LNR decreased as ELNs increased in a semilogarithmic manner (Figure 2B). Our ELN threshold of 30 represented a cutoff at which the rate of reduction in LNR diminished thereafter. Similarly, when the average LNR and node burden were compared by ELNs, it was again observed that by approximately 30 ELN, the average LNR became asymptotic with a simultaneous minimal increase in total positive node burden (Figure 2C). The model became unstable after approximately 50 nodes due to low statistical power despite a large starting number. Finally, if we applied the calculated LNR thresholds for each pN category (Table 4) and used an ELN threshold of 30 for all patients, we observed that 96% of patients were captured with a prognostically advantageous LNR (meaning a ratio less than the stage-specific threshold value) (Table 5).

Table 5.

LNR Determination Stratified by Nodal Burden Using 30 Examined Nodes

Table 5.

Temporal Trends for Nodal Metrics and Survival

We sought to characterize temporal trends in ELNs, node burden, LNR, and OS in the United States. From 2004 to 2014, the average 3-year OS rate increased by 10%, from 37.7% to 47.4% (Figure 3). During the same time interval, the average number of ELNs increased from 15.9 to 22.4, which was associated with a concomitant decrease in LNR, from 40.9% to 27.8%. There was minimal reduction in the mean positive lymph node burden. Taken together, the overall trend was toward improved oncologic resections; however, national data indicate that more work is needed, because only 23% of gastrectomies in 2016 had ≥30 lymph nodes examined (data not shown).

Figure 3.
Figure 3.

Line graph representing mean 3-year OS, ELNs, positive node count, and LNR over time.

Abbreviations: ELN, examined lymph node; LNR, lymph node ratio; OS, overall survival.

Citation: Journal of the National Comprehensive Cancer Network 2021; 10.6004/jnccn.2021.7038

Discussion

In this retrospective analysis focused on nodal metrics for patients with stage I–III gastric cancer treated using an R0, curative-intent gastrectomy, there were 5 main findings. First, the pN category, LNR, and ELNs were independently associated with OS on multivariable analysis. Second, when adjusting for node burden, LNR maintained excellent prognostic value and was a useful resource beyond pN staging. Third, with the use of ROC analysis, we identified an ELN threshold of 30 to be prognostically discriminant for a high positive node burden disease (pN3b); this ELN threshold thus served as a useful clinical benchmark for all oncologic resections. Fourth, and in concordance with our ROC analysis, nodal positivity had a diminishing increase beyond 30 examined nodes, associated with the stabilization of LNR. Fifth, when applying an ELN threshold of ≥30 nodes, 96% of patients were captured with a prognostically advantageous, stage-adjusted LNR.

The recommendation for an ELN threshold ≥30 is supported by the following points. First, low pN stage is inherently confounded by oncologic inadequacy and stage migration, evidenced by the inappropriately high average LNR observed in this study. A better approach is to determine an ELN threshold using high-burden disease, in which there is less difference between documented positive nodes and the actual node burden for a given patient. Second, optimal node yield stratified by stage has only post hoc utility. This claim is supported by the inaccuracy of tumor and node clinical staging for gastric cancer. For example, in our cohort of patients with node-positive gastric cancer, 33% were clinically staged as node-negative. Of these patients, 55% ultimately had ≥pN2 nodal disease, and 25% had pN3a/b disease. Among patients with cN1 category disease, 50% were upstaged to ≥pN2, and among patients with cN2 disease, 34% were upstaged to pN3a/b (data not shown). Clinical T staging is similarly unreliable for predicting node burden. In our cohort, pN3a/b disease was present in 18% of cT1 lesions, 23% of cT2 lesions, 28% of cT3 lesions, and 41% of cT4 lesions (data not shown). Taken together, when approaching the surgical management of a patient with gastric cancer, it is unsafe to assume a low node burden by clinical staging metrics, and we are obliged to act accordingly with sound oncologic principles. An ELN threshold ≥30, which was shown to be protective for high-burden disease in this large cohort, is a reasonable standard of care for all patients.

With respect to the literature, there is an important distinction between node yield for adequate staging and node yield for theoretical oncologic value. The former functions primarily to address issues of stage migration and the associated improvements in subsequent prognostication.1018 However, it would be misleading to conclude from this type of analysis that an ELN threshold designed for addressing stage migration should serve as the oncologic standard of care. For example, a commonly used threshold is ≥16 examined nodes, which is the current recommendation by NCCN.7 Although it is true that patients with <16 examined nodes typically have poorer outcomes, which is likely a product of both stage migration and oncologic inadequacy, this value does not capture the potential therapeutic value of extended lymphadenectomy for most patients.

Others have attempted to identify a node threshold for theoretical oncologic benefit and survival advantage. Siewert et al19 evaluated 1,654 patients with gastric cancer who were treated using R0 resection. They observed that resection of >25 nodes was associated with increased 10-year survival for patients with stage II disease, even when accounting for insufficient node dissection. Similarly, Brenkman et al20 evaluated a Dutch cohort of 3,764 patients with gastric cancer treated using curative-intent gastrectomy and observed that a lymph node yield of >25 nodes was associated with prolonged survival on multivariable analysis. Smith et al21 evaluated 3,814 patients with resectable gastric cancer from the SEER database (1973–1999) and observed that each increase in 10 examined nodes resulted in an approximately 7% increase in survival. Furthermore, this survival advantage persisted for up to a threshold of 40 lymph nodes. Schwarz and Smith,22 using the SEER database, also evaluated the impact of extended lymphadenectomy in patients with potentially curable advanced gastric cancer (stage III–IV M0). Using a cutpoint analysis, they ascertained that the optimal ELN threshold was 30 nodes for pN2 disease and 40 nodes for N3 disease. Ichikura et al23 evaluated 926 patients treated using curative gastrectomy and observed that ≥30 examined nodes was associated with improved survival for patients with advanced-stage disease. Finally, Mirkin et al,24 using the NCDB (2003–2011), investigated 1,036 patients with node-negative gastric cancer treated with neoadjuvant therapy and curative-intent gastrectomy and observed a survival benefit with >30 nodes examined.

In contrast, our study focused on individuals with node-positive gastric cancer, because ELN is particularly important in this cohort of patients. The current recommendation by the 8th edition of the AJCC Cancer Staging Manual suggests a minimum of 16 nodes, but highlights that the removal of ≥30 nodes is desirable.9 However, the NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines) for Gastric Cancer (Version 1.2021) only recommend removal of at least 16 nodes.7 The current study helps support the higher ELN threshold and can provide justification for performing a second pathologic assessment for additional nodes if this threshold is not met.

The variability in recommended thresholds for node yield partly relates to methodology. Lower thresholds (eg, 20–25 nodes) are often the result of arbitrary or experiential distinctions. Conversely, thresholds derived from statistical modeling are dependent on the input. For example, thresholds predicated on a combination of low- and high-burden nodal disease tend to overestimate the threshold for low burden and underestimate the threshold for high burden. For this reason, studies that included subgroup analyses based on stage are likely more accurate, and tend to support a higher threshold for advanced-stage disease (eg, in the 30- to 40-node range). It may be that ≥40 examined nodes provide additional prognostic value beyond our recommendation of 30 nodes. However, only 5% of patients had ≥40 nodes examined in our cohort, which influenced the ROC subanalysis for pN3b disease. Thus, as the average number of retrieved nodes increases, so too will the ELN threshold.25 However, note that in our dataset, there was a minimal increase in the average number of positive nodes beyond 30 ELNs, and therefore LNR changed only marginally, as well.

Note also that in the modern era of gastric cancer treatment, neoadjuvant chemotherapy has become the standard of care for resectable and locally advanced disease. Neoadjuvant therapy has the potential to downstage tumor burden, and multiple randomized controlled trials have shown an associated improvement in survival outcome measures. The MAGIC trial, which was published in 2006, confirmed a survival benefit with perioperative fluorouracil or capecitabine plus cisplatin and epirubicin (ECF), and this trial provided supporting evidence that tumor downstaging with preoperative cytotoxic therapy improves rates of R0 resection.26 More recently, the FLOT4 trial evaluated perioperative treatment using fluorouracil plus leucovorin, oxaliplatin, and docetaxel compared with ECF. This treatment was associated with a pathologic complete response in 17% of patients, a ypN0 category was observed in 49%, and an R0 resection was achieved in 85%, all significantly higher responses than with ECF treatment.27,28 Finally, the CROSS trial for esophageal junction cancers, published in 2012, was notable for a 29% rate of complete pathologic response with neoadjuvant chemoradiotherapy, which was associated with significantly improved OS.29 Taken together, there is the potential for neoadjuvant therapy to reduce the burden of nodal metastases among a treatment cohort and thus influence the calculated threshold for examined nodes.

Given this context, our calculated threshold of 30 examined nodes was modeled using a population spanning a decade in whom the use of neoadjuvant therapy markedly increased during that time period. For example, in 2004, only 15% of patients with gastric cancer received neoadjuvant therapy, whereas in 2017, this number was approximately 60%.30 Although these rates accurately capture the practice patterns in the United States, our calculated threshold will be higher than for a population in whom all patients received neoadjuvant therapy given the anticipated lower average node burden in the latter.

The prognostic utility of increasing ELNs has been attributed to stage migration and locoregional disease extirpation. At lower thresholds, stage migration is the critical determinant; however, stage migration diminishes with increasing yield, and there may be a survival advantage from increased disease clearance at higher thresholds. It is important to contextualize node yield by the associated LNR. For example, an ELN count of 20 would theoretically be associated with an LNR of 10% for pN1 disease, 25% for pN2 disease, 75% for pN3a disease, and up to 100% for pN3b disease. Regardless of ELNs, a high LNR is the most significant adverse prognostic indicator for gastric cancer, because it provides a surrogate measure for both nodal burden and the likelihood of missed locoregional disease. To highlight this relationship, patients in our cohort with category pN1–2 disease and an LNR of <50% (n=182) had a lower risk of death than did all other patients with category pN disease and an LNR of ≥50%. Decreasing LNR is concordantly associated with improved survival, assuming that perigastric nodal stations are appropriately sampled.31 On ROC analysis, we calculated a threshold LNR of 11% for pN1 disease to serve as a lower limit. This threshold would not be realistically achievable for pN3b disease because it would require the examination of 166 nodes. Thus, for high-burden disease (pN3a/b) we chose to accept the stage-specific calculated threshold. The takeaway from these findings is that the ELN threshold is not a goal in and of itself, but rather it serves to maintain the lowest possible LNR for a patient population.

Factors influencing lymph node retrieval are multiple and include patient habitus, use of preoperative therapy, perigastric lymph node quantity, patient comorbidities, extent of lymphadenectomy, and surgical technique.32 In this study, we observed a trend of improved lymph node yield over time. In 2016, the final year of this analysis, 70% of patients had ≥16 nodes examined, but only 23% had ≥30 nodes removed. However, these numbers were 42% and 10%, respectively, in 2004. Morgan et al33 evaluated 3,321 patients with gastric cancer from the SEER Cancer Registry of Greater California and the California Cancer Registry and observed that in 2015, only 45.5% of those treated at a dedicated cancer center had ≥15 nodes removed during gastrectomy. This number is lower than what we observed, but both studies highlight the need for improved node retrieval for gastric cancer surgery in the United States.

This study had several limitations. As a retrospective analysis, this study was subject to the inherent limitations of that study design. The NCDB is a large national database that may be limited by missing or inaccurate information for certain patients. Verification of information cannot be performed by chart review. We did not differentiate node yield by operative setting, because previous studies have shown that yields are higher in dedicated cancer centers.33 Furthermore, the calculated ELN and LNR thresholds were based on a patient population that on average was shown to have inadequate node retrievals, and this may have artificially lowered the calculated cutoffs. We did not distinguish between patients who had undergone preoperative systemic therapy (29%). It remains to be determined whether preoperative therapy impacts node yield in gastric cancer. We have attempted to take these biases into consideration in our analysis and presentation of the data.

Conclusions

An ELN threshold of ≥30 lymph nodes has been shown to support a sufficiently low LNR for most patients with gastric cancer who were treated in the United States. This serves to mitigate understaging, increase the likelihood of locoregional disease eradication, and potentially improve survival.

References

  • 1.

    Anderson WF, Camargo MC, Fraumeni JF Jr, et al. Age-specific trends in incidence of noncardiac gastric cancer in US adults. JAMA 2010;303:17231728.

  • 2.

    Bray F, Ferlay J, Soerjomataram I, et al. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2018;68:394424.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 3.

    Thrift AP, El-Serag HB. Burden of gastric cancer. Clin Gastroenterol Hepatol 2020;18:534542.

  • 4.

    Seevaratnam R, Bocicariu A, Cardoso R, et al. How many lymph nodes should be assessed in patients with gastric cancer? A systematic review. Gastric Cancer 2012;15(Suppl 1):S7088.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 5.

    Songun I, Putter H, Kranenbarg EM, et al. Surgical treatment of gastric cancer: 15-year follow-up results of the randomised nationwide Dutch D1D2 trial. Lancet Oncol 2010;11: 439449.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 6.

    Seevaratnam R, Bocicariu A, Cardoso R, et al. A meta-analysis of D1 versus D2 lymph node dissection. Gastric Cancer 2012;15(Suppl 1):S6069.

  • 7.

    Ajani JA, D’Amico TA, Bentrm DJ, et al. NCCN Clinical Practice Guidelines in Oncology: Gastric Cancer. Version 1.2021. Accessed February 22, 2021. To view the most recent version, visit NCCN.org

    • Search Google Scholar
    • Export Citation
  • 8.

    Marchet A, Mocellin S, Ambrosi A, et al. The ratio between metastatic and examined lymph nodes (N ratio) is an independent prognostic factor in gastric cancer regardless of the type of lymphadenectomy: results from an Italian multicentric study in 1853 patients. Ann Surg 2007;245:543552.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 9.

    Amin MB, Greene FL, Edge SB, et al. The Eighth Edition AJCC Cancer Staging Manual: continuing to build a bridge from a population-based to a more “personalized” approach to cancer staging. CA Cancer J Clin 2017;67:9399.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 10.

    Gu P, Deng J, Wang W, et al. Impact of the number of examined lymph nodes on stage migration in node-negative gastric cancer patients: a Chinese multi-institutional analysis with propensity score matching. Ann Transl Med 2020;8:938.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 11.

    Allen CJ, Vreeland TJ, Newhook TE, et al. Prognostic value of lymph node yield after neoadjuvant chemoradiation for gastric cancer. Ann Surg Oncol 2020;27:534542.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 12.

    Deng J, Yamashita H, Seto Y, et al. Increasing the number of examined lymph nodes is a prerequisite for improvement in the accurate evaluation of overall survival of node-negative gastric cancer patients. Ann Surg Oncol 2017;24:745753.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 13.

    Bunt AM, Hermans J, Smit VT, et al. Surgical/pathologic-stage migration confounds comparisons of gastric cancer survival rates between Japan and Western countries. J Clin Oncol 1995;13:1925.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 14.

    Vuong B, Graff-Baker AN, Dehal A, et al. Survival analysis with extended lymphadenectomy for gastric cancer: removing stage migration from the equation. Am Surg 2017;83:10741079.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 15.

    Seeruttun SR, Xu L, Wang F, et al. A homogenized approach to classify advanced gastric cancer patients with limited and adequate number of pathologically examined lymph nodes. Cancer Commun (Lond) 2019;39:111.

    • Search Google Scholar
    • Export Citation
  • 16.

    Wang PL, Xiao FT, Gong BC, et al. A nomogram for predicting overall survival of gastric cancer patients with insufficient lymph nodes examined. J Gastrointest Surg 2017;21:947956.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 17.

    MingHua Z, KeCheng Z, ZhenYu C, et al. Impact of lymph nodes examined on survival in ypN0 gastric cancer patients: a population-based study. J Gastrointest Surg 2021;25:919925.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 18.

    Li SS, Costantino CL, Rattner DW, et al. Outcomes of extended lymphadenectomy for gastroesophageal carcinoma: a large Western series. J Am Coll Surg 2019;228:879891.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 19.

    Siewert JR, Böttcher K, Stein HJ, et al. Relevant prognostic factors in gastric cancer: ten-year results of the German Gastric Cancer Study. Ann Surg 1998;228:449461.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 20.

    Brenkman HJF, Goense L, Brosens LA, et al. A high lymph node yield is associated with prolonged survival in elderly patients undergoing curative gastrectomy for cancer: a Dutch population-based cohort study. Ann Surg Oncol 2017;24:22132223.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 21.

    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;23:71147124.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 22.

    Schwarz RE, Smith DD. Clinical impact of lymphadenectomy extent in resectable gastric cancer of advanced stage. Ann Surg Oncol 2007;14: 317328.

  • 23.

    Ichikura T, Ogawa T, Chochi K, et al. Minimum number of lymph nodes that should be examined for the International Union Against Cancer/American Joint Committee on Cancer TNM classification of gastric carcinoma. World J Surg 2003;27:330333.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 24.

    Mirkin KA, Hollenbeak CS, Wong J. Greater lymph node retrieval improves survival in node-negative resected gastric cancer in the United States. J Gastric Cancer 2017;17:306318.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 25.

    Macalindong SS, Kim KH, Nam BH, et al. Effect of total number of harvested lymph nodes on survival outcomes after curative resection for gastric adenocarcinoma: findings from an Eastern high-volume gastric cancer center. BMC Cancer 2018;18:73.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 26.

    Cunningham D, Allum WH, Stenning SP, et al. Perioperative chemotherapy versus surgery alone for resectable gastroesophageal cancer. N Engl J Med 2006;355:1120.

  • 27.

    Al-Batran SE, Hofheinz RD, Pauligk C, et al. Histopathological regression after neoadjuvant docetaxel, oxaliplatin, fluorouracil, and leucovorin versus epirubicin, cisplatin, and fluorouracil or capecitabine in patients with resectable gastric or gastro-oesophageal junction adenocarcinoma (FLOT4-AIO): results from the phase 2 part of a multicentre, open-label, randomised phase 2/3 trial. Lancet Oncol 2016;17:16971708.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 28.

    Al-Batran SE, Homann N, Pauligk C, et al. Perioperative chemotherapy with fluorouracil plus leucovorin, oxaliplatin, and docetaxel versus fluorouracil or capecitabine plus cisplatin and epirubicin for locally advanced, resectable gastric or gastro-oesophageal junction adenocarcinoma (FLOT4): a randomised, phase 2/3 trial. Lancet 2019;393:19481957.

    • Search Google Scholar
    • Export Citation
  • 29.

    van Hagen P, Hulshof MC, van Lanschot JJ, et al. Preoperative chemoradiotherapy for esophageal or junctional cancer. N Engl J Med 2012;366:20742084.

  • 30.

    Ramdas S, Lowe K. Trends in perioperative treatment of gastric cancer 2004–2017: an NCDB analysis. Accessed April 5, 2021. Available at: https://bulletin.facs.org/2020/10/trends-in-perioperative-treatment-of-gastric-cancer-2004-2017-an-ncdb-analysis/

    • Search Google Scholar
    • Export Citation
  • 31.

    Lin GT, Chen QY, Zhong Q, et al. Intraoperative surrogate indicators of gastric cancer patients’ long-term prognosis: the number of lymph nodes examined relates to the lymph node noncompliance rate. Ann Surg Oncol 2020;27:32813293.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 32.

    Ikoma N, Estrella JS, Hofstetter WL, et al. Surgeon assessment of gastric cancer lymph node specimens with a video of technique. J Gastrointest Surg 2018;22:20132019.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 33.

    Morgan JW, Ji L, Friedman G, et al. The role of the cancer center when using lymph node count as a quality measure for gastric cancer surgery. JAMA Surg 2015;150:3743.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation

Submitted December 15, 2020; final revision received March 24, 2021; accepted for publication March 25, 2021. Published online October 22, 2021.

Author contributions: Data acquisition: Erstad, Blum, Das, Minsky, Ajani, Mansfield, Ikoma, Badgwell. Data analysis: Erstad, Blum, Das, Minsky, Ajani, Mansfield, Ikoma, Badgwell. Data interpretation: Erstad, Estrella. Expert clinical opinion: Blum, Estrella, Das, Minsky, Ajani, Mansfield, Ikoma, Badgwell. Manuscript Preparataion: All authors.

Disclosures: The authors have disclosed that they have not received any financial consideration from any person or organization to support the preparation, analysis, results, or discussion of this article.

Correspondence: Brian D. Badgwell, MD, Department of Surgical Oncology, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Boulevard/Unit 1484, Houston, TX 77030-4009. Email: bbadgwell@mdanderson.org

Supplementary Materials

  • View in gallery

    Kaplan-Meier survival curves for OS stratified by (A) pN category, (B) LNR, and (C) ELNs.

    Abbreviations: ELN, examined lymph node; LNR, lymph node ratio; OS, overall survival.

  • View in gallery

    (A) Distribution of positive node burden (blue) and associated cumulative incidence (orange) (n=22,018). (B) Scatterplot displaying LNR by ELNs. Blue curve represents a modeled semilog curve, indicating diminishing reduction in LNR with increasing ELNs. Orange-hashed line represents capture among the entire cohort with calculated ELN threshold ≥30. (C) Line graph of mean LNR and positive node count with SD error stratified by ELNs.

    Abbreviations: ELN, examined lymph node; LNR, lymph node ratio.

  • View in gallery

    Line graph representing mean 3-year OS, ELNs, positive node count, and LNR over time.

    Abbreviations: ELN, examined lymph node; LNR, lymph node ratio; OS, overall survival.

  • 1.

    Anderson WF, Camargo MC, Fraumeni JF Jr, et al. Age-specific trends in incidence of noncardiac gastric cancer in US adults. JAMA 2010;303:17231728.

  • 2.

    Bray F, Ferlay J, Soerjomataram I, et al. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2018;68:394424.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 3.

    Thrift AP, El-Serag HB. Burden of gastric cancer. Clin Gastroenterol Hepatol 2020;18:534542.

  • 4.

    Seevaratnam R, Bocicariu A, Cardoso R, et al. How many lymph nodes should be assessed in patients with gastric cancer? A systematic review. Gastric Cancer 2012;15(Suppl 1):S7088.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 5.

    Songun I, Putter H, Kranenbarg EM, et al. Surgical treatment of gastric cancer: 15-year follow-up results of the randomised nationwide Dutch D1D2 trial. Lancet Oncol 2010;11: 439449.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 6.

    Seevaratnam R, Bocicariu A, Cardoso R, et al. A meta-analysis of D1 versus D2 lymph node dissection. Gastric Cancer 2012;15(Suppl 1):S6069.

  • 7.

    Ajani JA, D’Amico TA, Bentrm DJ, et al. NCCN Clinical Practice Guidelines in Oncology: Gastric Cancer. Version 1.2021. Accessed February 22, 2021. To view the most recent version, visit NCCN.org

    • Search Google Scholar
    • Export Citation
  • 8.

    Marchet A, Mocellin S, Ambrosi A, et al. The ratio between metastatic and examined lymph nodes (N ratio) is an independent prognostic factor in gastric cancer regardless of the type of lymphadenectomy: results from an Italian multicentric study in 1853 patients. Ann Surg 2007;245:543552.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 9.

    Amin MB, Greene FL, Edge SB, et al. The Eighth Edition AJCC Cancer Staging Manual: continuing to build a bridge from a population-based to a more “personalized” approach to cancer staging. CA Cancer J Clin 2017;67:9399.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 10.

    Gu P, Deng J, Wang W, et al. Impact of the number of examined lymph nodes on stage migration in node-negative gastric cancer patients: a Chinese multi-institutional analysis with propensity score matching. Ann Transl Med 2020;8:938.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 11.

    Allen CJ, Vreeland TJ, Newhook TE, et al. Prognostic value of lymph node yield after neoadjuvant chemoradiation for gastric cancer. Ann Surg Oncol 2020;27:534542.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 12.

    Deng J, Yamashita H, Seto Y, et al. Increasing the number of examined lymph nodes is a prerequisite for improvement in the accurate evaluation of overall survival of node-negative gastric cancer patients. Ann Surg Oncol 2017;24:745753.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 13.

    Bunt AM, Hermans J, Smit VT, et al. Surgical/pathologic-stage migration confounds comparisons of gastric cancer survival rates between Japan and Western countries. J Clin Oncol 1995;13:1925.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 14.

    Vuong B, Graff-Baker AN, Dehal A, et al. Survival analysis with extended lymphadenectomy for gastric cancer: removing stage migration from the equation. Am Surg 2017;83:10741079.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 15.

    Seeruttun SR, Xu L, Wang F, et al. A homogenized approach to classify advanced gastric cancer patients with limited and adequate number of pathologically examined lymph nodes. Cancer Commun (Lond) 2019;39:111.

    • Search Google Scholar
    • Export Citation
  • 16.

    Wang PL, Xiao FT, Gong BC, et al. A nomogram for predicting overall survival of gastric cancer patients with insufficient lymph nodes examined. J Gastrointest Surg 2017;21:947956.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 17.

    MingHua Z, KeCheng Z, ZhenYu C, et al. Impact of lymph nodes examined on survival in ypN0 gastric cancer patients: a population-based study. J Gastrointest Surg 2021;25:919925.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 18.

    Li SS, Costantino CL, Rattner DW, et al. Outcomes of extended lymphadenectomy for gastroesophageal carcinoma: a large Western series. J Am Coll Surg 2019;228:879891.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 19.

    Siewert JR, Böttcher K, Stein HJ, et al. Relevant prognostic factors in gastric cancer: ten-year results of the German Gastric Cancer Study. Ann Surg 1998;228:449461.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 20.

    Brenkman HJF, Goense L, Brosens LA, et al. A high lymph node yield is associated with prolonged survival in elderly patients undergoing curative gastrectomy for cancer: a Dutch population-based cohort study. Ann Surg Oncol 2017;24:22132223.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 21.

    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;23:71147124.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 22.

    Schwarz RE, Smith DD. Clinical impact of lymphadenectomy extent in resectable gastric cancer of advanced stage. Ann Surg Oncol 2007;14: 317328.

  • 23.

    Ichikura T, Ogawa T, Chochi K, et al. Minimum number of lymph nodes that should be examined for the International Union Against Cancer/American Joint Committee on Cancer TNM classification of gastric carcinoma. World J Surg 2003;27:330333.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 24.

    Mirkin KA, Hollenbeak CS, Wong J. Greater lymph node retrieval improves survival in node-negative resected gastric cancer in the United States. J Gastric Cancer 2017;17:306318.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 25.

    Macalindong SS, Kim KH, Nam BH, et al. Effect of total number of harvested lymph nodes on survival outcomes after curative resection for gastric adenocarcinoma: findings from an Eastern high-volume gastric cancer center. BMC Cancer 2018;18:73.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 26.

    Cunningham D, Allum WH, Stenning SP, et al. Perioperative chemotherapy versus surgery alone for resectable gastroesophageal cancer. N Engl J Med 2006;355:1120.

  • 27.

    Al-Batran SE, Hofheinz RD, Pauligk C, et al. Histopathological regression after neoadjuvant docetaxel, oxaliplatin, fluorouracil, and leucovorin versus epirubicin, cisplatin, and fluorouracil or capecitabine in patients with resectable gastric or gastro-oesophageal junction adenocarcinoma (FLOT4-AIO): results from the phase 2 part of a multicentre, open-label, randomised phase 2/3 trial. Lancet Oncol 2016;17:16971708.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 28.

    Al-Batran SE, Homann N, Pauligk C, et al. Perioperative chemotherapy with fluorouracil plus leucovorin, oxaliplatin, and docetaxel versus fluorouracil or capecitabine plus cisplatin and epirubicin for locally advanced, resectable gastric or gastro-oesophageal junction adenocarcinoma (FLOT4): a randomised, phase 2/3 trial. Lancet 2019;393:19481957.

    • Search Google Scholar
    • Export Citation
  • 29.

    van Hagen P, Hulshof MC, van Lanschot JJ, et al. Preoperative chemoradiotherapy for esophageal or junctional cancer. N Engl J Med 2012;366:20742084.

  • 30.

    Ramdas S, Lowe K. Trends in perioperative treatment of gastric cancer 2004–2017: an NCDB analysis. Accessed April 5, 2021. Available at: https://bulletin.facs.org/2020/10/trends-in-perioperative-treatment-of-gastric-cancer-2004-2017-an-ncdb-analysis/

    • Search Google Scholar
    • Export Citation
  • 31.

    Lin GT, Chen QY, Zhong Q, et al. Intraoperative surrogate indicators of gastric cancer patients’ long-term prognosis: the number of lymph nodes examined relates to the lymph node noncompliance rate. Ann Surg Oncol 2020;27:32813293.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 32.

    Ikoma N, Estrella JS, Hofstetter WL, et al. Surgeon assessment of gastric cancer lymph node specimens with a video of technique. J Gastrointest Surg 2018;22:20132019.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 33.

    Morgan JW, Ji L, Friedman G, et al. The role of the cancer center when using lymph node count as a quality measure for gastric cancer surgery. JAMA Surg 2015;150:3743.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 526 526 129
PDF Downloads 295 295 102
EPUB Downloads 0 0 0