Prognostic Significance of a New Grading System of Lymph Node Morphology After Neoadjuvant Radiochemotherapy for Esophageal Cancer

Transcription

1 Prognostic Significance of a New Grading System of Lymph Node Morphology After Neoadjuvant Radiochemotherapy for Esophageal Cancer Elfriede Bollschweiler, MD, PhD, Arnulf H. Hölscher, MD, Ralf Metzger, MD, Sarah Besch, MD, Stefan P. Mönig, MD, Stephan E. Baldus, MD, and Uta Drebber, MD Department of General, Visceral and Cancer Surgery, University of Cologne, Cologne; Institute of Pathology, University of Düsseldorf, Düsseldorf and Institute of Pathology, University of Cologne, Cologne, Germany Background. Along with primary tumor response, lymph node (LN) status after radiochemotherapy is one of the most important prognostic factors for advanced esophageal carcinoma. We investigated the influence of neoadjuvant radiochemotherapy on histomorphologic parameters of LNs. Methods. One hundred ninety-two patients with esophageal carcinoma underwent surgery after preoperative radiochemotherapy. Response of primary tumor was graded as minor or major. Two matched subgroups were chosen: 20 patients with minor response and 20 patients with major response. Histomorphologic criteria of LNs underwent univariate and multivariate analyses and correlated with tumor response and prognosis statistics. Results. The LNs from 40 patients (N 1276) were examined (median number of LNs per patient, 31). Of patients with minor response, 65% showed LN metastasis; of those with major response, 20% did so (p 0.011). Major responders had significantly lower rates of capsular and central fibrosis and vascular transformation and had more sarcoidlike lesions. Logistic regression analysis did not distinguish these parameters between major and minor responders. The 5-year survival rate was 55% for major responders and 10% for minor responders (p 0.025), 47% for patients with LN metastasis (LNM) and 18% for patients with LNM (p 0.041). An optimal prognostic factor, LN morphologic grading, was defined as follows: low risk, no LNM and less than 3 LNs with central fibrosis; medium risk, no LNM and central fibrosis in 3 or more LNs or LNM with an LN ratio of less than 0.05; high risk, all other cases. The 5-year survival rate was 56%, 25%, and 0% for patients considered to have low, medium, and high risk, respectively, according to LN morphologic grading (p < 0.003). With the inclusion of this classification in the Cox regression analysis, no other factors showed prognostic relevance. Conclusions. Grading of LN morphology after neoadjuvant radiochemotherapy is the most important prognostic factor for patients with esophageal cancer. (Ann Thorac Surg 2011;92:2020 7) 2011 by The Society of Thoracic Surgeons Neoadjuvant treatment strategies for esophageal cancer are matters of ongoing debate with controversial results from phase II and III trials [1 3]. However, there is consensus that patients with histopathologic major response of the primary tumor after neoadjuvant radiochemotherapy definitively benefit from these treatment modalities, with 5-year survival rates up to 60% [1, 4, 5]. Histopathologic effects of neoadjuvant radiochemotherapy have also been observed in lymph node (LN) metastases (LNMs) [6, 7]. The presence of LNM is a strong indicator for poor outcome in esophageal cancer. Studies have reported higher frequency of ypn0 categories in patients receiving neoadjuvant radiochemotherapy [8, 9]. Schneider and associates [5] found a significant correlation between the frequency of patients without LNM after neoadjuvant therapy (ypn0) (p 0.003) and major histopathologic regression, which indicates a positive response to neoadjuvant treatment. This association Accepted for publication June 29, Address correspondence to Dr Bollschweiler, Department of General, Visceral and Cancer Surgery, University of Cologne, Kerpener Str 62, Köln, Germany; elfriede.bollschweiler@uk-koeln.de. has also been reported for other cancer entities (eg, rectal [10] and lung [11] cancers). Multivariate survival analysis in patients with R0 resections identified ypn category as a prognostic indicator second only to the grade of histomorphologic regression of the primary esophageal tumor. Miyata et al [12] found that tumor budding in the invasive border of the tumor correlated significantly with clinical response and prognosis of patients with esophageal squamous cell carcinomas (SCCs) receiving neoadjuvant chemotherapy. To date, limited data are available regarding the effects of radiochemotherapy on the histopathologic structure of LNs in esophageal cancer. An earlier study by our group found that radiochemotherapy does not necessarily reduce the number of resected and assessed LNs compared to surgery alone. However, the size of metastatic LNs generally decreases after radiochemotherapy [13]. For other tumor entities like rectal cancer, an LN regression grade for mesorectal LNs after neoadjuvant radiochemotherapy has been defined [14]. Newman and coworkers [15] published a study examining the histopathologic evidence for tumor regression in axillary LNs of patients with breast cancer treated with preoperative chemotherapy. This study demonstrated that specific patterns of 2011 by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc doi: /j.athoracsur

2 Ann Thorac Surg BOLLSCHWEILER ET AL 2011;92: RADIOCHEMOTHERAPY AND HISTOMORPHOLOGIC PATTERN IN LYMPH NODES Abbreviations and Acronyms AC adenocarcinoma HR hazard ratio LN lymph node LNG lymph node group according to the Japanese Classification of Gastric Cancer LNM lymph node metastasis LN-MG lymph node morphology grading LQ lower quartile MaHR major histopathathologic response of the primary tumor MiHR minor histopathologic response of the primary tumor p p symbol as a descriptor of the TNM classification for histopathologic staging RTx/CTx radiochemotherapy SCC squamous cell carcinoma TNM tumor-nodal-metastases (TNM) Classification Sytem of Malignant Tumors UQ upper quartile y y symbol as a descriptor of the TNMclassification for preoperative multimodality therapy pathologic findings are reliable indicators of treatment response. The aim of the current study was to analyze the effects of radiochemotherapy on LNs in patients with locally advanced esophageal cancer, evaluating various histomorphologic parameters, along with prognostic relevance. Patients and Methods Patients Five hundred twenty-three patients with esophageal cancer were treated at the Department of Surgery between January 1, 1997, and December 31, Sixty-four patients received nonsurgical palliative treatment due to advanced disease or functional inoperability. The remaining 459 patients underwent surgical resection for curative treatment. This study included only patients with clinically staged T3 or T4 carcinomas. One hundred ninety-two patients (65%) received preoperative radiochemotherapy according to a standardized protocol, which is described in detail elsewhere [5]. Of these 192 patients, 40 were randomly selected or matched for this study as follows: group I consisted of 20 patients with minor response (10 with adenocarcinoma [AC] and 10 with SCC) and group II consisted of 20 patients with major response (10 with AC and 10 with SCC). The scientific protocol was approved by the local ethics commission. Staging TNM (tumor-node-metastasis classification system of malignant tumors) staging was performed according to the 2021 criteria of the International Union Against Cancer [16]. Clinical staging was based on results from barium swallow, endoscopy, endoscopic ultrasound, and computed tomography examinations of the chest and abdomen (4-mm sections). Endoscopy and endoscopic ultrasound were performed in a standardized manner for all patients. Neoadjuvant Radiochemotherapy On days 1 to 5, cisplatin (20 mg/m 2 /d) was administered as a short-term infusion and 5-fluorouracil (1000 mg/m 2 /d) was administered as a continuous infusion over 24 hours. Radiation therapy was administered with the use of linear accelerators with 10-MV to 15-MV photons. Radiation therapy was planned to encompass tumor volume with 5-cm cephalocaudal and 2-cm radial margins, and the planning target volume included enlarged regional nodes based on computer tomography evaluation and endoscopic ultrasound. Radiation was delivered in daily fractions of 1.8 Gy (days 1 to 5, 8 to 12, 15 to 19, and 22 to 26) to a total dose of 36 Gy using a multiple field technique. Surgical resection was performed 4 to 5 weeks following completion of radiochemotherapy, after clinical restaging. Surgery Treatment of choice was subtotal en bloc esophagectomy via a right transthoracic approach including 2-field lymphadenectomy of the mediastinal and abdominal LNs [17]. The specimens were removed en bloc, including nodes. To ensure primary tumor integrity, LNs were dissected partially in the operating theater and partially by pathologists according to a standardized protocol. Pathology Histopathologic examination of all resected specimens consisted of thorough evaluations of tumor stage, residual tumor (R) category, grading, and number of examined and involved LNs. The LN ratio was defined as the number of involved LNs divided by the number of resected LNs. A series of sections from each node were selected and stained with hematoxylin and eosin as well as with periodic acid Schiff (PAS). All dissected LNs were microscopically analyzed for metastatic disease. According to the sixth edition of the UICC TNM Classification of Malignant Tumors, section on tumors of the lower thoracic esophagus, metastasis in the celiac LN group (LNG 9 [lymph node group according to the Japanese Classification of Gastric Cancer]) was classified as M1a and in other nonregional locations as M1b [16]. After neoadjuvant therapy, the pathologic assessment is more complex because of possible tumor regression. Thus, such a classification is identified with the prefix y to indicate decreased reliability compared to the ptnm classification after surgery alone. Histomorphologic Evaluation of Lymph Nodes The LNs of the 40 patients of this study were reexamined. A total of 1276 LNs were assessed. More than 1 section per centimeter was reviewed from each tumor nodule. All tumors measuring less than 1 cm and/or exhibiting complete pathologic response were submitted in their entirety for microscopic

3 2022 BOLLSCHWEILER ET AL Ann Thorac Surg RADIOCHEMOTHERAPY AND HISTOMORPHOLOGIC PATTERN IN LYMPH NODES 2011;92: examination. All slides were evaluated by 1 pathologist and controlled by another experienced pathologist. The following histomorphologic parameters were evaluated qualitatively and semiquantitatively (no, few, medium, and florid) for the presence of metastasis, extracapsular invasion, lymphangiosis carcinomatosis, tumor necrosis, central or capsular fibrotic changes, sinus histiocytosis, eosinophilic and neutrophilic granulocytes, giant cells, foam cells, desmoplasia, sarcoidlike lesions, and vascular transformation. Histopathologic Response Classification of the Primary Tumor The degree of histomorphologic regression of the primary tumor was classified into 4 categories (Cologne Regression Scale): grade I, greater than 50% vital residual tumor cells (VRTCs); grade II, 10% to 50% VRTCs; grade III, nearly complete response with less than 10% VRTCs; and grade IV, complete response [5, 18]. This analysis was performed by 2 independent staff pathologists blinded to all other clinical data. Because of prognostic implications, regression grades III and IV were classified as major histomorphologic response compared to grades I and II, which were categorized as minor histopathologic response. Histopathologic tumor regression is the most significant independent prognostic indicator [5]. Statistics Beginning in 1997, clinical data were collected prospectively according to a standardized protocol for all patients with esophageal cancer. Matching Procedure To determine independent prognostic factors, it is necessary to include the relevant prognostic factors of the UICC/TNM classification such as T, N, and M categories; type of histology; and response classification of the primary tumor after neoadjuvant therapy. In addition, patient age and gender and year of surgical procedure may be of prognostic relevance. In the current study, only patients with clinically staged T3 (n 190) or T4 (n 2) carcinomas without distant metastases (cm0) after restaging procedure were included. The clinical N category was not a criterion for therapeutic decision, because there is no valid diagnostic method available. As a consequence, the cn category was not a matching criterion. For the matching procedure, we first randomly selected 10 patients from the 27 AC patients showing major response after neoadjuvant radiochemotherapy. For the remaining groups, propensity score analysis was used to select 10 patients from the 51 SCC patients showing major response after neoadjuvant radiochemotherapy, 10 patients from the 65 AC patients showing minor response, and 10 patients from the 49 SCC patients showing minor response. The propensity score is the most widely used balancing score. For each patient, it estimates the predisposition for belonging to one group versus another. We used the nearest neighbor matching procedure described by John Painter, a routine implemented in SPSS software. The matching criteria were patient age, patient gender, and year of operation. Table 1. Demographic Data of Patients With Advanced Esophageal Cancer After Neoadjuvant Radiochemotherapy and Esophagectomy Showing Comparison Between Total Group and Randomly Selected Subgroups of Patients Data Analysis The data regarding the histomorphologic criteria of the LNs were reevaluated and added to the existing patient file. Frequencies of categorical variables, the median, with the lower (LQ) and upper (UQ) quartiles, or minimum and maximum, were used for descriptive statistics. Differences in patient, tumor, and treatment characteristics were assessed using the Mann-Whitney test for continuous variables and the 2 test or Fisher s exact test for categorical variables with a significance level of p Multivariate analysis was performed using logistic regression analysis and classification tree-based modeling. The median follow-up time of the surviving patients was 5.5 years (range, 4.2 to 7.6 years). Kaplan-Meier plots were used to depict survival distribution. The log-rank test was used to evaluate for survival differences. Multivariate analysis of survival used Cox regression analysis to identify independent prognostic variables. The level of significance was set to p All statistical analyses were performed using the statistic program SPSS for Windows version For graphic presentation of the results, MedCalc for Windows, version (MedCalc Software, Mariakerke, Belgium) was used. Results All Patients (N 192) Subgroup of Patients (n 40) Significance Age, y, median p Gender, No. male/ 152/40 35/5 p female Histology, n p Squamous cell carcinoma Adrencarcinoma No. of resected lymph nodes, median (lower quartile upper quartile) 28 ( ) 31 ( ) p The demographic characteristics of all patients with advanced esophageal carcinoma undergoing neoadjuvant therapy (n 192) and of the randomly selected patients (n 40) are shown in Table 1. There are no significant differences regarding age, gender, histology of tumor, and/or median number of resected LNs among all patients or among the subgroups of randomly selected patients. Histomorphologic Lymph Node Patterns in Patients With Minor Versus Major Tumor Response From the 40 patients, a total of 1276 LNs were reevaluated. Approximately two thirds (13 of 20) of the patients with minor response to radiochemotherapy had LNMs.

4 Ann Thorac Surg BOLLSCHWEILER ET AL 2011;92: RADIOCHEMOTHERAPY AND HISTOMORPHOLOGIC PATTERN IN LYMPH NODES In contrast, only 20% (4 of 20) of patients with major response had lymphatic invasion (p 0.01). The frequency of LNM categories differed significantly between the 2 groups (Table 2). Compared to minor responders, major responders showed significantly less capsular fibrosis and less vascular transformation in LNs without metastasis (Table 2). This histomorphologic pattern was the same for both SCCs and ACs (data not shown). In the multivariate logistic regression analysis including all histopathologic variables, only the presence of metastasis, capsular fibrosis, vascular transformation, and sarcoidlike lesions significantly affected the designations of primary tumor minor or major response. Using these 4 parameters, only 57% of the studied patients were correctly classified. Lymph Node Morphology Classification and Long-Term Survival All living patients survived longer than 4 years. Consequently, patients living longer than 4 years were defined as long-term survivors (n 15). Using this parameter as the dependent variable and the histologic parameters of the LNs as independent variables, the multivariate classification tree identified 3 prognostically relevant categories of LN regression grades. Lymph Node Morphologic Grading Low-risk lymph node morphologic grading (LNMG) was defined as no LNM and fewer than 3 LNs with central fibrosis, medium-risk LNMG as no LNM and 3 or more LNs with central fibrosis or LNM with an LN ratio less than 0.05; and high-risk LNMG as all other cases. The classification tree is provided in Figure 1. Prognostic Relevance of Lymph Node Morphologic Grading Univariate analysis of prognosis confirmed the known prognostic factors of tumor response, ypn category, LN ratio, and capsular invasion (Table 3). In addition, patients without LNM and central fibrosis in fewer than 3 LNs had significantly better prognoses than those with ypn0 and 3 or more LNs with central fibrosis. The LNMG classification showed the most significant prognostic relevance with 5-year survival rates of 56% for patients with low-risk LNMG, 25% for patients with medium-risk LNMG, and 0% for patients with high-risk LNMG (p 0.003). Figure 2 shows the Kaplan-Meier survival curve for LNMG. In the multivariate analysis including the prognostically relevant variables, only Lymph Node Morphology Grade retained prognostic relevance with hazard ratio (HR) values as follows: low-risk Lymph Node Morphologic Grade, mediumrisk Lymph Node Morphologic Grade, 2.2 (95% confidence interval [CI], 0.98 to 5.4; p 0.063) and low-risk Lymph Node Morphologic Grade high-risk Lymph Node Morphologic Grade, 6.3 (95% CI, 2.2 to 18.1, p 0.001). Comment Our study demonstrates that specific pathologic findings within the LNs of patients with advanced esophageal cancer undergoing preoperative radiochemotherapy are reliable indicators of treatment response. These features include central fibrosis within nonmetastatic LNs as well as low rates of metastasis compared to the total number of resected LNs. As expected, stratification of these patients according to the presence of these features identified patient subsets with varying prognoses. Patients with the best outcomes were those without LNM but with signs of treatment response such as central fibrosis present in only few LNs at the time of surgery. Figure 3 shows an LN with central fibrosis after neoadjuvant ra- Table 2. Frequency of the Different Histomorphologic Characteristics of Analyzed Lymph Nodes Comparing Patients With Primary Tumor Minor and Major Response After Neoadjuvant Radiochemotherapy and Esophagectomy Minor Response (n 20) Major Response (n 20) Significance Lymph nodes, n p LNM no LNM yes 62 5 Capsular fibrosis, % LNM no p LNM yes p Central fibrosis, % LNM no p LNM yes p Eosinophilic granulocytes, % LNM no p LNM yes p Neutrophilic granulocytes, % LNM no p LNM yes 0 0 p Hyaline degeneration, % LNM no p 1.0 Sarcoidlike lesion, % LNM no p Vascular transformation, % LNM no p Necrosis, % LNM no p Giant cells, % LNM no p Foam cells, % LNM no p Extracapsular invasion, n 13 n 0 % LNM yes p Tumor necrosis, % n 18 n 2 LNM yes p Desmoplasmosia, % n 38 n 5 LNM yes p Lymphangiosis n 15 n 1 carcinomatosis, % LNM yes p LNM lymph node metastasis. 2023

5 2024 BOLLSCHWEILER ET AL Ann Thorac Surg RADIOCHEMOTHERAPY AND HISTOMORPHOLOGIC PATTERN IN LYMPH NODES 2011;92: Fig 1. Graphical presentation of the classification tree analysis for the definition of prognostically relevant groups of histomorphologic signs of regression in lymph nodes (LN) after neoadjuvant radiochemotherapy. diochemotherapy. These patients were probably node negative or had a low burden of metastatic disease at the time of diagnosis. At the other end of the spectrum, patients with residual viable LNMs in more than 5% of the resected LNs had the worst outcomes. These patients probably had the most advanced disease at diagnosis, with metastases resistant to systemic treatment. In the middle of the spectrum are the patients with or without residual LNMs at the time of surgery showing evidence of treatment response; these patients had outcomes somewhere between those of the first group of patients and those of the node-positive patients. Therefore, the morphologic pattern of response to neoadjuvant therapy as measured with LNMG has particular significance. First, it identifies the group of patients whose tumor has not yet or has just begun to metastasize to the LNs. Second, it identifies those patients whose tumors responded well to the radiochemotherapy. This latter group wins out in the end, since these patients receive proven prognostic benefits from the neoadjuvant therapy compared to those who undergo surgery alone. The group of patients without LNM, showing only minimal signs of regression, would probably have the same prognosis after undergoing esophagectomy and lymphadenectomy regardless of pretreatment. Patients with LNM and minimal regression did not benefit from the neoadjuvant therapy. To date, numerous studies have examined the prognostic relevance of primary tumor response. Various classifications have been used, which are differentiated primarily by the magnitude of regression [19]. There is consensus that the best prognoses are attained for patients showing no tumor cells on the histopathologic examination of the resected specimen, a complete response, after pretreatment. Nevertheless, there is tumor recurrence even after complete response, which again underscores the importance of lymphatic metastasis. Leers and colleagues [20] found that patients with ypn0 esophageal cancer after neoadjuvant therapy had significantly Fig 2. Kaplan-Meier survival curve comparing patients with different grades of lymph node morphology (LNMG) after neoadjuvant radiochemotherapy and esophagectomy.

6 Ann Thorac Surg BOLLSCHWEILER ET AL 2011;92: RADIOCHEMOTHERAPY AND HISTOMORPHOLOGIC PATTERN IN LYMPH NODES Table 3. Univariate Analysis of Prognosis for Patients With Advanced Esophageal Cancer After Neoadjuvant Radiochemotherapy and Esophagectomy No, of Patients 5-y Survival Rate, % Significance 2025 All patients Gender p Male Female 5 40 Histology p Squamous cell carcinoma Adrenocarcinoma Tumor Response, p Cologne regression grade [5, 17] Major Minor Tumor response, p Schneider et al [5] Major pn Major pn Minor pn Minor pn ypn category p N N LN morphology grade p Low risk Medium risk High risk 8 0 LN ratio p ( 5%) ( 5%) 11 0 Capsular fibrosis p LNs LNs Central fibrosis p LNs LNs Sarcoidlike lesion No p Yes Vascular transformation p No Yes 8 12 Capsular invasion p ypn ypn1 without ypn1 with 6 0 Tumor necrosis p ypn1 without ypn1 with 8 0 Lymphangiosis p carcinomatosis ypn ypn1without ypn1with 4 0 LN lymph node. Fig 3. Histomorphologic sign of tumor regression: Lymph node with central fibrosis from a patient with esophageal cancer. worse survival rates than patients receiving surgery alone (49% versus 85%, p 0.005). The relevance of LN status post esophagectomy regarding prognosis is well known. This is true for patients undergoing primary surgery [21-24] and perhaps even more for those receiving neoadjuvant pretreatment [6, 25]. Multivariate survival analysis in patients with R0 resections has identified ypn category as a prognostic indicator second only to the grade of histomorphologic regression of the primary tumor [5]. Consequently, Schneider and associates [5] included ypn category in the regression classification for patients with neoadjuvant treated esophageal cancer. The present study verified these results. Patients whose response to radiochemotherapy was absent or minimal (minor response of primary tumor) often showed a ypn1 category. In contrast, patients responding well to neoadjuvant therapy (major response of primary tumor) had a significantly decreased rate of ypn1 cases, and of these only 1 or 2 metastatic nodes per patient were evident. Although the number of patients in each subgroup was small, the different prognoses yielded by the combined classification system were significant (p 0.039). The pathologic evaluation and prognostic significance of response in preoperatively treated LNs were initially defined by Newman and colleagues [15]. In that study, the axillary LNs of 71 patients with locally advanced breast cancer treated with induction chemotherapy were evaluated for histologic evidence of tumor regression. Complete pathologic response in the breast and axilla occurred in 10 patients (14%), and patients without nodal metastases or evidence of tumor regression had the best outcomes [15]. Mandard and coworkers [26] used the Mandard regression classification of primary tumor to characterize the regression seen in mesorectal LNs after neoadjuvant radiochemotherapy for rectal cancer [14]. They found that regression from tumor to fibrosis in the LNs is a quantifiable result, which correlates significantly with primary tumor regression [14]. Gervaz and coworkers [27] used a variation of the Mandard and associates regression system to evaluate the regression of colon carcinoma in the LNs after chemotherapy: TRG1 corre-

7 2026 BOLLSCHWEILER ET AL Ann Thorac Surg RADIOCHEMOTHERAPY AND HISTOMORPHOLOGIC PATTERN IN LYMPH NODES 2011;92: sponded to an absence of tumor cells replaced by abundant fibrosis, TRG2 to rare residual tumor cells scattered throughout abundant fibrosis, TRG3 to more residual tumor cells throughout a predominant fibrosis, TRG4 to a large number of tumor cells predominating over fibrosis, and TRG5 to tumor cells exclusively without fibrosis. Both studies had small numbers of patients and did not demonstrate the prognostic relevance of LN regression. The inclusion of histomorphologic regression criteria with the presence/absence of LN metastasis improved prognostic relevance. In addition, we used the LN ratio, which measures the number of metastatic LNs and the total number of resected LNs. Several studies have shown that the number of LNs that are removed predicts survival in esophageal cancer [22, 28, 29]. Wilson and associates [23] found that the optimal threshold for this survival benefit predicted by Cox regression was removal of a minimum of 23 nodes. These data were taken from patients without neoadjuvant therapy. Mariette and coworkers [6] found that the prognostic role of both the number and the ratio of LNMs were maintained regardless of whether patients received neoadjuvant radiochemotherapy. Moreover, LN ratio was more accurate for inadequately staged patients (15 or fewer examined LNs), whereas the number of LNMs is more pertinent for adequately staged patients (more than 15 examined LNs) [6]. There are several other markers of LNs with prognostic relevance. For example, there is increasing evidence regarding extracapsular LN involvement as a prognostic factor for recurrence and poor prognosis in gastrointestinal malignancies [30-33]. This was identified for patients with and without neoadjuvant therapy. The results of our study are based on a small number of patients but a great number of LNs. This may be a problem for the calculation of prognostic relevance. However, all known relevant prognostic parameters for esophageal cancer were confirmed. To avoid selection bias in our study, we performed a special matching routine (ie, propensity score matching). The goal of this is to balance patient characteristics by incorporating everything recorded that could relate to either systematic bias or simply bad luck. Logistic regression analysis, such as that used for the propensity model, generates a coefficient for each variable. Cases were matched according to the coefficients [34]. In summary, we have defined a grading system of histomorphologic signs in LNs after neoadjuvant radiochemotherapy. This grading is the most important prognostic factor for patients undergoing preoperative radiochemotherapy for advanced esophageal cancer. Patients without metastasis and with isolated signs of tumor regression had the best prognoses. In contrast, patients with LNM and an LNM ratio of greater than 5% had the worst prognoses. This classification may be introduced into a new UICC/TNM version for esophageal cancer after neoadjuvant therapy. Further prospective trials must be performed to confirm the prognostic relevance. References 1. Stahl M, Walz M, Stuschke M, et al. Phase III comparison of preoperative chemotherapy compared with chemoradiotherapy in patients with locally advanced adenocarcinoma of the esophagogastric junction. J Clin Oncol 2009;27: Tepper J, Krasna M, Niedzwiecki D, et al. Phase III trial of trimodality therapy with cisplatin, fluorouracil, radiotherapy, and surgery compared with surgery alone for esophageal cancer: CALGB J Clin Oncol 2008;26: Adelstein DJ, Rice TW, Rybicki LA, et al. A phase II trial of accelerated multimodality therapy for locoregionally advanced cancer of the esophagus and gastroesophageal junction: the impact of clinical heterogeneity. Am J Clin Oncol Cancer Clin Trials 2007;30: Bollschweiler E, Metzger R, Drebber U, et al. Histological type of esophageal cancer might affect response to neoadjuvant radiochemotherapy and subsequent prognosis. Ann Oncol 2009;20: Schneider P, Baldus S, Metzger R, et al. Histomorphologic tumor regression and lymph node metastases determine prognosis following neoadjuvant radiochemotherapy for esophageal cancer: implications for response classification. Ann Surg 2005;242: Mariette C, Piessen G, Briez N, et al. The number of metastatic lymph nodes and the ratio between metastatic and examined lymph nodes are independent prognostic factors in esophageal cancer regardless of neoadjuvant chemoradiation or lymphadenectomy extent. Ann Surg 2008; 247: Prenzel K, Konig A, Schneider P, et al. Reduced incidence of nodal micrometastasis after major response to neoadjuvant chemoradiation in locally advanced esophageal cancer. Ann Surg Oncol 2007;14: Walsh TN, Noonan N, Hollywood D, et al. A comparison of multimodal therapy and surgery for esophageal adenocarcinoma. N Engl J Med 1996;335: Rouvelas I, Zeng W, Lindblad M, et al. Survival after neoadjuvant therapy compared with surgery alone for resectable esophageal cancer in a population-based study. World J Surg 2006;30: Read T, Andujar J, Caushaj P, et al. Neoadjuvant therapy for rectal cancer: histologic response of the primary tumor predicts nodal status. Dis Colon Rectum 2004;47: Stupp R, Mayer M, Kann R, et al. Neoadjuvant chemotherapy and radiotherapy followed by surgery in selected patients with stage IIIB non-small-cell lung cancer: a multicentre phase II trial. Lancet Oncol 2009;10: Miyata H, Yoshioka A, Yamasaki M, et al. Tumor budding in tumor invasive front predicts prognosis and survival of patients with esophageal squamous cell carcinomas receiving neoadjuvant chemotherapy. Cancer 2009;115: Bollschweiler E, Besch S, Drebber U, et al. Influence of neoadjuvant chemoradiation on the number and size of analyzed lymph nodes in esophageal cancer. Ann Surg Oncol 2010;17: Caricato M, Ausania F, De Dominicis E, et al. Tumor regression in mesorectal lymphnodes after neoadjuvant chemoradiation for rectal cancer. Eur J Surg Oncol 2007;33: Newman L, Pernick N, Adsay V, et al. Histopathologic evidence of tumor regression in the axillary lymph nodes of patients treated with preoperative chemotherapy correlates with breast cancer outcome. Ann Surg Oncol 2003;10: Sobin L, Wittekind C. UICC TNM Classification of Malignant Tumors. 6th ed. New York, NY: Wiley & Sons; Hölscher A, Schroder W, Bollschweiler E, et al. How safe is high intrathoracic esophagogastrostomy? Chirurg 2003;74: Baldus S, Monig S, Schroder W, et al. [Regression of oesophageal carcinomas after neoadjuvant radiochemotherapy: criteria of the histopathological evaluation]. Pathologe 2004;25:

8 Ann Thorac Surg BOLLSCHWEILER ET AL 2011;92: RADIOCHEMOTHERAPY AND HISTOMORPHOLOGIC PATTERN IN LYMPH NODES 19. Bollschweiler E, Holscher AH, Metzger R. Histologic tumor type and the rate of complete response after neoadjuvant therapy for esophageal cancer. Future Oncol 2010;6: Leers J, Ayazi S, Hagen J, et al. Survival in lymph node negative adenocarcinoma of the esophagus after R0 resection with and without neoadjuvant therapy: evidence for downstaging of N status. J Am Coll Surg 2009;208: Greenstein A, Litle V, Swanson S, et al. Effect of the number of lymph nodes sampled on postoperative survival of lymph node-negative esophageal cancer. Cancer 2008;112: Twine C, Lewis W, Morgan M, et al. The assessment of prognosis of surgically resected oesophageal cancer is dependent on the number of lymph nodes examined pathologically. Histopathology 2009;55: Wilson M, Rosato E, Chojnacki K, et al. Prognostic significance of lymph node metastases and ratio in esophageal cancer. J Surg Res 2008;146: Altorki N, Zhou X, Stiles B, et al. Total number of resected lymph nodes predicts survival in esophageal cancer. Ann Surg 2008;248: Barbour A, Jones M, Gonen M, et al. Refining esophageal cancer staging after neoadjuvant therapy: importance of treatment response. Ann Surg Oncol 2008;15: Mandard A, Dalibard F, Mandard J, et al. Pathologic assessment of tumor regression after preoperative chemoradiotherapy of esophageal carcinoma. Clinicopathologic correlations. Cancer 1994;73: Gervaz P, Rubbia-Brandt L, Andres A, et al. Neoadjuvant chemotherapy in patients with stage IV colorectal cancer: a 2027 comparison of histological response in liver metastases, primary tumors, and regional lymph nodes. Ann Surg Oncol 2010;17: Peyre C, Hagen J, De Meester S, et al. The number of lymph nodes removed predicts survival in esophageal cancer: an international study on the impact of extent of surgical resection. Ann Surg 2008;248: Bollschweiler E, Baldus S, Schroder W, et al. Staging of esophageal carcinoma: Length of tumor and number of involved regional lymph nodes. Are these independent prognostic factors? J Surg Oncol 2006;94: Wind J, Lagarde S, Ten Kate F, et al. A systematic review on the significance of extracapsular lymph node involvement in gastrointestinal malignancies. Eur J Surg Oncol 2007;33: Metzger R, Drebber U, Baldus SE, et al. Extracapsular lymph node involvement differs between squamous cell and adenocarcinoma of the esophagus. Ann Surg Oncol 2009;16: Metzger R, Bollschweiler E, Drebber U, et al. Neoadjuvant chemoradiotherapy for esophageal cancer: impact on extracapsular lymph node involvement. World J Gastroenterol 2010;16: Alakus H, Holscher AH, Grass G, et al. Extracapsular lymph node spread: a new prognostic factor in gastric cancer. Cancer 2010;116: Blackstone E. Comparing apples and oranges. J Thorac Cardiovasc Surg 2002;123:8 15. INVITED COMMENTARY The prognostic importance of a major response to neoadjuvant therapy in esophageal cancer is accepted. A variety of grading systems for classifying the response of the primary tumor have been developed, including the Cologne Regression Scale described in previous publications by Bollschweiler and colleagues [1]. In their most recent publication titled Prognostic Significance of a New Grading System of Lymph Node Morphology after Neoadjuvant Chemoradiation for Esophageal Cancer, the authors have expanded on their previous work describing a classification scheme for lymph node involvement after neoadjuvant therapy. If confirmed in future studies involving larger numbers of patients, this classification system, which defines 3 groups of patients with distinct survival probabilities, may allow more accurate prediction of outcome than is currently possible. It may also be of therapeutic importance because a high-risk group of patients has been defined with a uniformly fatal outcome. These patients may be candidates for postoperative adjuvant therapy in an attempt to improve survival. The authors results may also shed some light on the true meaning of a response to neoadjuvant therapy. Although there is little doubt that a complete or major response to chemoradiation is associated with a better outcome, it is not clear whether this represents a true benefit in terms of eradication of occult systemic disease or whether a major histopathologic response is simply a marker of a tumor with more favorable biological characteristics. There are data in the present study that would seem to support the latter alternative. The authors have shown that major responders have a 3 times lower likelihood of lymph node involvement, which might be taken to indicate that regression has occurred in the nodes as well as in the primary tumor. The problem with this conclusion is that a major response in the primary tumor was also associated with a significantly lower frequency of signs of tumor response in the lymph nodes, as evidenced by capsular fibrosis and vascular transformation. Further, their observation that long-term survival decreased as the number of nodes showing central fibrosis increased may indicate that the observed survival benefit may be more of a reflection of a lower tumor burden at the time of initial diagnosis. The lack of documentation of the clinical status of the lymph nodes before chemoradiation therapy in this series and the failure to include this factor in their propensity matching makes it difficult to determine with certainty if the groups were truly comparable. Future studies to confirm their findings should include endoscopic ultrasonography and positron emission tomography/computed tomography staging information. Jeffrey A. Hagen, MD Division of Cardiothoracic Surgery University of Southern California 1510 San Pablo St, #514 Los Angeles, CA hagen@usc.edu Reference 1. Bollschweiler E, Hölscher AH, Metzger R, et al. Prognostic significance of a new grading system of lymph node morphology after neoadjuvant radiochemotherapy for esophageal cancer. Ann Thorac Surg 2011;92: by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc doi: /j.athoracsur