SUMMARY
There is debate surrounding the appropriate threshold for lymph node harvest during esophagectomy in patients with esophageal cancer, specifically for those receiving preoperative radiation. The purpose of this study was to determine the impact of lymph node yield on survival in patients receiving preoperative chemoradiation for esophageal cancer. The National Cancer Database (NCDB) was utilized to identify patients with esophageal cancer that received preoperative radiation. The cohort was divided into patients undergoing minimal (<9) or extensive (≥9) lymph node yield. Demographic, operative, and postoperative outcomes were compared between the groups. Kaplan–Meier analysis with the log rank test was used to compare survival between the yield groups. Cox proportional hazards model was used to determine the association between lymph node yield and survival. In total, 886 cases were included: 349 (39%) belonging to the minimal node group and 537 (61%) to the extensive group. Unadjusted 5-year survival was similar between the minimal and extensive groups, respectively (37.3% vs. 38.8%; P > 0.05). After adjustment using Cox regression, extensive lymph node yield was associated with survival (hazard ratio 0.80, confidence interval 0.66–0.98, P = 0.03). This study suggests that extensive lymph node yield is advantageous for patients with esophageal cancer undergoing esophagectomy following induction therapy. This most likely reflects improved diagnosis and staging with extensive yield.
Keywords: esophageal cancer, esophagectomy, induction radiation, lymph node, preoperative radiation
INTRODUCTION
There is contention surrounding the appropriate number of lymph nodes to retrieve during esophagectomy in patients with esophageal cancer. It remains unclear if the number of lymph nodes yielded significantly impacts survival as small single-centre studies have produced conflicting results to date.1–3
Specifically, the literature has not adequately addressed the impact of lymph node yield in patients receiving radiation prior to esophagectomy. Both single-centre and population-based studies have concluded that extent of lymphadenectomy was not associated with overall or disease-specific survival, but these studies have grouped both nonirradiated and irradiated esophagectomy patients together in their analyses, thus potentially masking the impact of lymph node yield for the irradiated cohort.2,4
This is a compelling area for investigation as there is additional debate surrounding the impact of lymph node yield in esophageal cancer patients that have received induction therapies. A secondary analysis of the ChemoRadiotherapy for Oesophageal cancer followed by Surgery Study (CROSS) trial, a randomized control trial comparing induction chemoradiotherapy versus surgery alone for esophageal cancer, concluded that lymph node count only matters for patients who received an esophagectomy without induction therapy.5 Conversely, an analysis of the Surveillance, Epidemiology, and End Results (SEER) database suggests that lymph node count is associated with survival in patients that have received induction radiotherapy, though this study was unable to account for chemotherapy, patient comorbidities, and pathologic stage of the tumour.6
By utilizing the National Cancer Database (NCDB), this study aims to determine the impact of lymph node yield on survival in esophageal cancer patients that have received chemoradiation prior to esophagectomy. An analysis from this national sample will expand on the knowledge drawn from existing publications from single-centre studies, especially considering NCDB is a nationwide, facility-based comprehensive cancer registry.
MATERIALS AND METHODS
National Cancer Database
The NCDB, a joint programme of the American Cancer Society and the Commission on Cancer (CoC) of the American College of Surgeons (ACS), is recognized as the largest clinical registry in the world and provides a comprehensive view of over 70% of newly diagnosed malignancies nationwide.7 The ACS has executed a business associate agreement that includes a data use agreement with each of its CoC-accredited hospitals. The NCDB Esophageal Participant User File uses surgical codes from the CoC’s Facility Oncology Registry Data Standards (FORDS), and all of our variables, including partial and total esophagectomy, are defined by FORDS. Partial esophagectomy was defined as esophagectomy requiring an intrathoracic anastomosis. Total esophagectomy was defined as either a McKeown or transhiatal esophagectomy.
Study population
Patients with clinical T2–3 N0–3 M0 esophageal cancer were identified using International Classification of Diseases for Oncology, 3rd Edition (ICD-O-3) histology and topography codes.8 Those patients with either squamous cell carcinoma or adenocarcinoma histology from 2006 to 2013 were selected for analysis. Patients were excluded if they did not undergo simultaneous neoadjuvant chemoradiation. Patients who underwent an associated gastrectomy or laryngectomy with their esophagectomy were excluded. Patients with tumours recorded to be greater than 20 cm or lymph node yields greater than 60 lymph nodes were excluded for likelihood of miscoding. Finally, patients were also excluded if they had missing information regarding the number of lymph nodes yielded or missing information regarding their overall survival.
Statistical analysis
The primary exposure of interest was the amount of lymph nodes yielded during esophageal resection. The primary endpoint was overall survival defined as the date of diagnosis to date of death from any cause and censored date of most recent alive follow-up, which are recorded in the NCDB. Secondary endpoints included 30-day mortality, 90-day mortality, 30-day readmission, and index hospitalization length of stay.
In order to model the relationship between our primary exposure and outcomes of interest, we developed a multivariable logistic regression model incorporating restricted cubic splines (RCS) functions. The RCS is a piecewise, polynomial function that can flexibly examine the association between a predictor and an outcome without assuming any relationship a priori. Using RCS allows for the evaluation of a possible nonlinear relationship and also allows the ability to detect natural cut-points in the primary exposure. Covariates in the model included age, gender, race (white, black, other), Charlson–Deyo comorbidity score (0,1, ≥2), insurance status (private, Medicaid/Medicare, other governmental, uninsured, unknown), income quartile, education level quartile, facility type (Comprehensive Community Cancer Program, Community Cancer Program, Academic Comprehensive Cancer Program), histology type (SCC, AC), pathologic T and N stages, and extent of surgery. The derived cut-point or threshold was then used to dichotomize the study cohort into extensive and minimal lymphadenectomy groups defined as patients who had a retrieved lymph node count more or less than the threshold, respectively.
Descriptive statistics were compared between the minimal and extensive groups using Pearson chi-square test for categorical variables and Welch’s t-test for continuous variables. Unadjusted Kaplan–Meier curves with the log rank test were run for both groups to determine the impact of extent of lymph node yield on survival. A Cox proportional hazards model was run to determine the impact of extent of lymph node yield on overall survival while accounting for potential confounding variables. This multivariable model included: dichotomous variable for minimal versus extensive lymph node yield, patient age, gender, race (white, black, other), Charlson–Deyo comorbidity score (0,1, ≥2), insurance status (private, Medicaid/Medicare, other governmental, uninsured, unknown), income quartile, education level quartile, facility type (Comprehensive Community Cancer Program, Community Cancer Program, Academic Comprehensive Cancer Program), histology type (SCC, AC), clinical T and N stages, pathologic T and N stages, and extent of surgery. Both clinical and pathologic staging was included in the regression analysis as these were both thought to be relevant factors that influence oncologic survival. Missing data in noncritical fields were uncommon, with average missingness of model covariates reaching 1.5%. Therefore, simple imputation for missing covariate data was used in the multivariable model where the mode was imputed for categorical covariate and the median for continuous covariates. All P-values reported are two-sided with the significance level set to 0.05. All analyses were performed using R 3.0.1 (The R Foundation for Statistical Computing, Vienna, Austria).
RESULTS
During the study period, 70,321 cases of esophageal cancer were reported, and a total of 886 patients were included in the analysis after application of our inclusion criteria (Fig. 1). Of those, 724 (81.7%) patients were diagnosed with adenocarcinoma and 162 (18.3%) with squamous cell carcinoma. Graphical visualization of the multivariable adjusted logistic regression models using RCS demonstrated the curve crossed a hazard ratio of 1 at a lymph node count of 9 (Fig. 2). Based on this model-determined threshold, the study cohort was stratified into two groups: patients having minimal lymphadenectomy (<9 lymph nodes) and patients having extensive lymphadenectomy (≥9 lymph nodes) yield. Based on this cutoff, 349 (39%) of recipients were categorized to the minimal node group and 537 (61%) to the extensive node group. Demographic comparisons between the groups (Table 1) revealed that patients in the extensive node group were more educated compared to the minimal node group. Patients in the extensive node group were also more likely to be treated at academic research centres. There was a statistically significant difference in distribution of pT and pN stage elements as well as cT stage elements between the minimal and extensive node groups, with the extensive node group accounting for more advanced cases of esophageal cancer. The extensive node group was more likely to have nodal upstaging (14.0% vs. 4.3%, P < 0.001) and more likely to receive adjuvant chemotherapy compared to the minimal node group (10.6% vs. 6.0%, P = 0.025). For all secondary endpoints examining postoperative outcomes, there was no statistically significant difference between the minimal and extensive node groups (Table 2).
Fig. 1.
Flow diagram of study cohort determination.
Fig. 2.
Restricted cubic splines curve for lymph node yield as a continuous variable. The results of a multivariable model utilizing restricted cubic splines for hazard of overall survival are depicted. The line of best fit (solid line) and 95% confidence intervals (CI) (hashed line) are shown.
Table 1.
Descriptive statistics for minimal versus extensive lymph node yield
| Variable | Overall | Minimal yield (<9 lymph nodes) | Extensive yield (≥9 lymph nodes) | P-value |
|---|---|---|---|---|
| Patient characteristics | ||||
| Age, years (IQR) | 61 (54, 68) | 62 (54, 68) | 61 (54, 68) | 0.644 |
| Female | 141 (15.9%) | 63 (18.1%) | 78 (14.5%) | 0.191 |
| Race | 0.137 | |||
| White | 829 (93.6%) | 321 (92%) | 508 (94.6%) | |
| Black | 32 (3.6%) | 18 (5.2%) | 14 (2.6%) | |
| Other | 25 (2.8%) | 10 (2.9%) | 15 (2.8%) | |
| Charlson comorbidity score | 0.085 | |||
| 0 | 688 (77.7%) | 259 (74.2%) | 429 (79.9%) | |
| 1 | 160 (18.1%) | 70 (20.1%) | 90 (16.8%) | |
| 2 | 38 (4.3%) | 20 (5.7%) | 18 (3.4%) | |
| Education quartile | 0.041 | |||
| Bottom | 102 (11.7%) | 43 (12.6%) | 59 (11.2%) | |
| Second | 223 (25.6%) | 95 (27.8%) | 128 (24.2%) | |
| Third | 321 (36.9%) | 134 (39.2%) | 187 (35.4%) | |
| Top | 224 (25.7%) | 70 (20.5%) | 154 (29.2%) | |
| Income quartile | 0.148 | |||
| Bottom | 125 (14.4%) | 57 (16.7%) | 68 (12.9%) | |
| Second | 200 (23%) | 85 (24.9%) | 115 (21.8%) | |
| Third | 271 (31.2%) | 104 (30.5%) | 167 (31.6%) | |
| Top | 273 (31.4%) | 95 (27.9%) | 178 (33.7%) | |
| Distance to cancer center (IQR) | 17 (7, 39.5) | 15.4 (6.2, 39.5) | 18.9 (7.6, 39.4) | 0.84 |
| Treatment facility | <0.001 | |||
| Community Program | 67 (7.8%) | 39 (11.3%) | 28 (5.4%) | |
| Comprehensive Community Program | 320 (37%) | 149 (43.3%) | 171 (32.9%) | |
| Integrated Network Cancer Program | 54 (6.2%) | 20 (5.8%) | 34 (6.5%) | |
| Research/Academic Program | 423 (49%) | 136 (39.5%) | 287 (55.2%) | |
| Insurance | 0.209 | |||
| Private | 475 (53.6%) | 181 (51.9%) | 294 (54.7%) | |
| Medicare/Medicaid | 362 (40.9%) | 141 (40.4%) | 221 (41.2%) | |
| Other government | 15 (1.7%) | 9 (2.6%) | 6 (1.1%) | |
| Unknown | 12 (1.4%) | 7 (2%) | 5 (0.9%) | |
| Uninsured | 22 (2.5%) | 11 (3.2%) | 11 (2%) | |
| Tumour characteristics | ||||
| Tumour size | 40 (25, 53) | 40 (23, 50) | 40 (27, 55) | 0.447 |
| cT stage element | 0.03 | |||
| 2 | 223 (25.2%) | 102 (29.2%) | 121 (22.5%) | |
| 3 | 663 (74.8%) | 247 (70.8%) | 416 (77.5%) | |
| pT stage element | <0.001 | |||
| 0 | 132 (15.8%) | 60 (18.3%) | 72 (14.2%) | |
| 0-IS | 6 (0.7%) | 1 (0.3%) | 5 (1%) | |
| 1 | 100 (12%) | 30 (9.1%) | 70 (13.8%) | |
| 2 | 165 (19.8%) | 70 (21.3%) | 95 (18.8%) | |
| 3 | 279 (33.5%) | 82 (25%) | 197 (38.9%) | |
| 4 | 2 (0.2%) | 1 (0.3%) | 1 (0.2%) | |
| X | 150 (18%) | 84 (25.6%) | 66 (13%) | |
| Upstage T status | 37 (4.2%) | 14 (4.0%) | 23 (4.3%) | 0.843 |
| pN stage element | <0.001 | |||
| 0 | 409 (49%) | 172 (52.4%) | 237 (46.7%) | |
| 1 | 235 (28.1%) | 67 (20.4%) | 168 (33.1%) | |
| 2 | 39 (4.7%) | 6 (1.8%) | 33 (6.5%) | |
| 3 | 13 (1.6%) | 1 (0.3%) | 12 (2.4%) | |
| X | 139 (16.6%) | 82 (25%) | 57 (11.2%) | |
| Upstage N status | 90 (10.2%) | 15 (4.3%) | 75 (14.0%) | <0.001 |
| Treatment specifics | ||||
| Days to definitive surgery (IQR) | 129 (111, 153) | 133 (111, 156.8) | 128 (111, 148) | 0.24 |
| Adjuvant chemo | 0.025 | |||
| Yes | 78 (8.8%) | 21 (6%) | 57 (10.6%) | |
| No | 808 (91.2%) | 328 (94%) | 480 (89.4%) | |
| Surgical endpoints | ||||
| Surgical approach (2010–2012) | 0.233 | |||
| VATS | 92 (24.7%) | 26 (20.6%) | 66 (26.8%) | |
| Converted | 17 (4.6%) | 4 (3.2%) | 13 (5.3%) | |
| Open | 263 (70.7%) | 96 (76.2%) | 167 (67.9%) | |
| Nodes removed (IQR) | 11 (5, 18) | 4 (0, 6) | 16 (12, 22) | <0.001 |
| Positive margins | 46 (5.3%) | 21 (6.2%) | 25 (4.7%) | 0.408 |
| Surgical margins | 0.298 | |||
| R0 | 823 (94.7%) | 316 (93.8%) | 507 (95.3%) | |
| R1 | 29 (3.3%) | 12 (3.6%) | 17 (3.2%) | |
| R2 | 2 (0.2%) | 2 (0.6%) | 0 (0%) | |
| Positive NOS | 15 (1.7%) | 7 (2.1%) | 8 (1.5%) | |
IQR, Interquartile range; NOS, not otherwise specified; LOS, length of stay.
Table 2.
Postoperative outcomes comparisons for minimal versus extensive lymph node yield
| Variable | Overall | Minimal yield (<9 lymph nodes) | Extensive yield (≥9 lymph nodes) | P-value |
|---|---|---|---|---|
| 30-day mortality | 19 (2.2%) | 8 (2.3%) | 11 (2.1%) | 0.999 |
| 90-day mortality | 52 (6%) | 21 (6.1%) | 31 (5.9%) | 0.999 |
| 30-day readmission | 36 (4.2%) | 17 (5.1%) | 19 (3.6%) | 0.395 |
| Hospital LOS (IQR) | 9 (8, 14) | 9 (8, 13.2) | 10 (7.8, 14) | 0.137 |
Similarly, an unadjusted Kaplan–Meier curve with the log rank test comparing survival of the minimal node yield and the extensive node yield groups did not demonstrate a statistically significant difference (5 year KM: 37.3% vs. 38.8%, respectively, log rank P = 0.196) (Fig. 3). The Cox proportional hazards model for the overall cohort demonstrated that the following factors are independently associated with survival: having extensive lymph node dissection (HR 0.801, CI 0.654–0.982, P = 0.033), age of 80 or greater (HR 2.431, CI 1.148–5.147, P = 0.020), being in the top education quartile (HR 0.642, CI 0.421–0.978, P = 0.039), other governmental insurance (HR 1.998, CI 1.053–3.789, P = 0.034), pathologic N1 status (HR 2.072, CI 1.635–2.626, P < 0.001), and pathologic N3 status (HR 2.715, CI 1.35–5.462, P = 0.005) (Table 3).
Fig. 3.
Unadjusted Kaplan–Meier curve for minimal and extensive lymph node yield.
Table 3.
Cox proportional hazards model for minimal versus extensive lymph node yield
| Variables | Hazard ratio | Lower 95% CI | Upper 95% CI | P-value |
|---|---|---|---|---|
| Extensive yield (≥9 lymph nodes) | 0.801 | 0.654 | 0.982 | 0.033 |
| Age (<50 yo) | Ref | |||
| Age (50–59 yo) | 1.164 | 0.803 | 1.687 | 0.424 |
| Age (60–69 yo) | 1.258 | 0.858 | 1.844 | 0.24 |
| Age (70–79 yo) | 1.287 | 0.828 | 2.002 | 0.262 |
| Age (>79 yo) | 2.431 | 1.148 | 5.147 | 0.02 |
| Female gender | 0.764 | 0.572 | 1.02 | 0.068 |
| Race: white | Ref | |||
| Race: black | 0.731 | 0.424 | 1.26 | 0.259 |
| Race: other | 1.026 | 0.599 | 1.757 | 0.926 |
| Charlson–Deyo score | 1.182 | 0.995 | 1.404 | 0.057 |
| Facility type: Community Program | Ref | |||
| Facility type: Comprehensive Community Program | 1.167 | 0.813 | 1.674 | 0.402 |
| Facility type: Integrated Network Cancer Program | 0.743 | 0.44 | 1.257 | 0.269 |
| Facility type: Research/Academic Program | 0.911 | 0.634 | 1.311 | 0.617 |
| Insurance: private | Ref | |||
| Insurance: Medicare/Medicaid | 1.185 | 0.931 | 1.509 | 0.168 |
| Insurance: other government | 1.998 | 1.053 | 3.789 | 0.034 |
| Insurance: unknown | 1.011 | 0.405 | 2.52 | 0.982 |
| Insurance: uninsured | 1.264 | 0.67 | 2.387 | 0.469 |
| Clinical T stage: 2 | Ref | |||
| Clinical T stage: 3 | 1.173 | 0.931 | 1.479 | 0.176 |
| Clinical N stage: 0 | Ref | |||
| Clinical N stage: 1 | 1.063 | 0.857 | 1.32 | 0.576 |
| Clinical N stage: 2 | 1.529 | 0.997 | 2.345 | 0.052 |
| Clinical N stage: 3 | 2.704 | 0.958 | 7.636 | 0.06 |
| Clinical N stage: X | 1.295 | 0.674 | 2.486 | 0.438 |
| Pathologic T stage: 0 | Ref | |||
| Pathologic T stage: IS | 1.501 | 0.46 | 4.898 | 0.501 |
| Pathologic T stage: 1 | 1.019 | 0.679 | 1.527 | 0.929 |
| Pathologic T stage: 2 | 0.906 | 0.634 | 1.295 | 0.589 |
| Pathologic T stage: 3 | 1.056 | 0.759 | 1.469 | 0.745 |
| Pathologic T stage: 4 | 3.754 | 0.881 | 16 | 0.074 |
| Pathologic T stage: X | 0.894 | 0.535 | 1.494 | 0.668 |
| Pathologic N stage: 0 | Ref | |||
| Pathologic N stage: 1 | 2.072 | 1.635 | 2.626 | <0.001 |
| Pathologic N stage: 2 | 1.597 | 0.962 | 2.651 | 0.07 |
| Pathologic N stage: 3 | 2.715 | 1.35 | 5.462 | 0.005 |
| Pathologic N stage: X | 1.097 | 0.681 | 1.766 | 0.704 |
| Surgery extent: total esophagectomy | Ref | |||
| Surgery extent: partial esophagectomy | 1.025 | 0.846 | 1.243 | 0.798 |
| Histology: AC | Ref | |||
| Histology: SCC | 1.286 | 0.976 | 1.695 | 0.074 |
| Income quartile: first | Ref | |||
| Income quartile: second | 0.796 | 0.562 | 1.127 | 0.199 |
| Income quartile: third | 0.849 | 0.603 | 1.196 | 0.348 |
| Income quartile: top | 1.043 | 0.717 | 1.515 | 0.827 |
| Education quartile: first | Ref | |||
| Education quartile: second | 0.943 | 0.662 | 1.343 | 0.743 |
| Education quartile: third | 0.896 | 0.622 | 1.291 | 0.555 |
| Education quartile: top | 0.642 | 0.421 | 0.978 | 0.039 |
CI, confidence intervals
DISCUSSION
Our study demonstrates that extent of lymph node yield during esophagectomy in patients receiving preoperative chemoradiation is significantly associated with overall survival, with extensive yield (defined as ≥9 nodes) conferring a survival advantage. Specifically, these findings suggest that extensive lymph node yield confers a survival advantage in that it may allow for more accurate nodal staging and, subsequently, delivery of optimal care in the form of adjuvant treatment. In the RCS analysis, the curve did not continue to increase as lymph node count increased once the threshold of nine lymph nodes was reached. This supports the notion that continued lymph node retrieval does not have a therapeutic benefit. These results from a national database serve to clarify the existing literature that extensive lymph node yield is in fact advantageous for this unique population.
At present, the National Comprehensive Cancer Network recommends removal of at least 15 lymph nodes for patients undergoing esophagectomy without induction chemoradiation.9 While recognizing the lack of data regarding lymph node retrieval for patients receiving preoperative chemoradiation, they recommend a similar lymph node yield for this population of patients as well.
Prior to our study, the field lacked an analysis on a national scale that examined patients receiving preoperative chemoradiation for esophageal cancer as a unique cohort. Previously, Shridhar and colleagues found that lymph node yield was not associated with overall survival in patients receiving neoadjuvant chemoradiation for esophageal cancer.3 Conversely, another analysis revealed that lymph node yield was associated with survival for patients with locally advanced esophageal cancer that received induction chemoradiation.1 Notably, both of these studies are single-institution analyses, and their results lack generalizability to other less experienced centres. Moreover, patients being treated at academic research centres were more likely to have an extensive lymph node yield, which again serves to demonstrate the limits of existing single-centre studies of this cohort as lymph node yield may be a surrogate variable for overall quality of care received. Furthermore, these analyses contained small sample sizes that inherently limited their ability to adjust for confounders. Additionally, a single-centre study from the UK and a national analysis from Sweden both revealed that lymph node yield was not associated with overall survival, but these studies failed to separate nonirradiated and irradiated patients in their analyses.2,4 While an analysis of the SEER database was consistent with our findings, this study was unable to account for chemotherapy, patient comorbidities, and pathologic stage of the tumour, all potential confounders that our study was able to address in a statistical model.6
Another analysis by Sampson and colleagues, also using the NCDB, analysed the association between lymphadenectomy and survival in patients undergoing esophageal resection for cancer, concluding that increased lymphadenectomy was advantageous for patients undergoing upfront esophagectomy.10 Our analysis improves on theirs for a couple of reasons. First, their analysis did not treat lymph node yield as a continuous variable. In doing so, we determined the exact lymph node count that is associated with improved survival. Additionally, our manuscript focuses specifically on patients who underwent esophagectomy after induction therapy. This specific cohort merits attention since their lymph nodes are likely affected by induction therapy. Finally, our analysis excluded patients who underwent associated gastrectomies and laryngectomies. Without this exclusion, one could question whether the retrieved lymph nodes were gastric or laryngeal in origin.
There are limits to our study that are inherent to utilizing a large national database. First, as a limitation of our dataset, specific approaches to esophagectomy (McKeown, Ivor Lewis, transhiatal) could not be captured. In a single-centre study, Wolff and colleagues compared Ivor Lewis and transhiatal approaches and found that significantly more lymph nodes were collected using Ivor Lewis.11 Further investigation utilizing multi-institutional data is warranted to compare lymph node retrieval in all three approaches and subsequent patient survival. In addition, our analysis was retrospective in nature, and, thus, there were potential confounders we could not account for, including BMI, radiation dose, and previous intra-abdominal or intrathoracic surgery, all of which can complicate nodal dissections and count numbers. Other important variables, such as type of neoadjuvant radiation and radiation fields, were subject to a high degree of missingness (>30%), and were not included in our analysis. Our study was subjected to selection bias. Although we attempted to control for known clinical, socioeconomic, hospital, and tumour covariates through our Cox regression, inherent differences between comparison groups may still occur given the retrospective nature of our analysis and would be best addressed with a prospective study.
In summary, this study serves to solidify that extensive lymph node yield is advantageous for patients undergoing esophagectomy following chemoradiation therapy for treatment of esophageal cancer. Notably, this analysis defines a new threshold of nine extracted lymph nodes for patients undergoing esophagectomy after induction therapy. This likely reflects improved staging and diagnosis that allows for optimal care.
ACKNOWLEDGMENTS
Institutional funding was used for this study. In addition, B.A.Y. is supported by the NIH funded Cardiothoracic Surgical Trials Network 5U01HL088953-05. V.R.E., B.Y.A., and A.Y.C. are supported by the National Center For Advancing Translational Sciences of the National Institutes of Health under Award Number TL1TR001116 and TL1TR002555 (A.Y.C.). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
REFERENCES
- 1. Hanna J M, Erhunmwunsee L, Berry M, D’Amico T, Onaitis M. The prognostic importance of the number of dissected lymph nodes after induction chemoradiotherapy for esophageal cancer. Ann Thorac Surg 2015; 99: 265–8. [DOI] [PubMed] [Google Scholar]
- 2. Lagergren J, Mattsson F, Zylstra J et al. Extent of lymphadenectomy and prognosis after esophageal cancer surgery. JAMA Surg 2015; 1. doi: 10.1001/jamasurg.2015.2611 [DOI] [PubMed] [Google Scholar]
- 3. Shridhar R, Hoffe S E, Almhanna K et al. Lymph node harvest in esophageal cancer after neoadjuvant chemoradiotherapy. Ann Surg Oncol 2013; 20(9): 3038–43. 10.1245/s10434-013-2988-4 [DOI] [PubMed] [Google Scholar]
- 4. Van Der, Johar A, Wijnhoven B, Lagergren P, Lagergren J. Extent of lymph node removal during esophageal cancer surgery and survival. J Natl Cancer Inst 2015; 107(5). 10.1093/jnci/djv043 [DOI] [PubMed] [Google Scholar]
- 5. Koen T A, Shapiro J, Looman C W N et al. Lymph node retrieval during esophagectomy with and without neoadjuvant chemotherapy. Ann Surg 2014; 260(5): 786–93. [DOI] [PubMed] [Google Scholar]
- 6. Wu S, Zhang Z, Liu W, He Z, Li F. Impact of the number of resected lymph nodes on survival after preoperative radiotherapy for esophageal cancer. Oncotarget 2014. (Apr; 7. 10.18632/oncotarget.8113 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Bilimoria K Y, Stewart A K, Winchester D P, Ko C Y. The National Cancer Data Base: a powerful initiative to improve cancer care in the United States. Ann Surg Oncol 2008; 15(3): 683–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Fritz A, Percy C, Jack A. International Classification of Diseases for Oncology, 3rd edn. Geneva: World Health Organization, 2000. [Google Scholar]
- 9. Merkow R P, Bilimoria K Y, Chow W B et al. Variation in lymph node examination after esophagectomy for cancer in the United States. JAMA Surg 2012; 147(6): 505–11. doi: 10.1001/archsurg.2011.2215 [DOI] [PubMed] [Google Scholar]
- 10. Samson P, Crabtree T, Broderick S et al. Quality measures in clinical stage I non-small cell lung cancer: improved performance is associated with improved survival. Ann Thorac Surg 2017; 103(1): 303–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Wolff C S, Castillo S F, Larson D R et al. Ivor Lewis approach is superior to transhiatal approach in retrieval of lymph nodes at esophagectomy. Dis Esophagus 2008; 21(4): 328–33. [DOI] [PubMed] [Google Scholar]