Accuracy of preoperative staging for a priori resectable esophageal cancer

Accuracy of preoperative staging for a priori resectable esophageal cancer Summary This study assessed the accuracy of preoperative staging in patients undergoing oncological esophagectomy for adenocarcinoma and squamous cell carcinoma. All patients undergoing surgery for resectable esophageal cancer in a university hospital from 2005 to 2016 were identified from our institutional database. Patients with neoadjuvant treatment were excluded to avoid bias from down-staging effects. Routinely, all patients had an upper endoscopy with biopsy, a thoracoabdominal CT scan, an 18-FEG PET-CT, and endoscopic ultrasound. Preoperative staging was compared to histopathological staging of surgical specimen that was considered as gold standard. There were 51 patients with a median age of 65 years (IQR: 59.3–73 years) having 21 squamous cell carcinoma and 30 adenocarcinoma, respectively. T- and N-stages were correctly predicted in 26 (51%) and 37 patients (72%), respectively. Overall, 18 patients (35%) were preoperatively diagnosed with a correct T- and N-stage. There was no difference between adenocarcinoma and squamous cell carcinoma. Accuracy of the T-stage was not influenced by the smoking status. The N-stage was not correct in 7/22 smoking patients (32%) and 6/29 nonsmoking patients (21%).The N-stage was underestimated in smoking patients as 6/22 patients (27%) had a histologically confirmed N+ who were preoperatively classified as N0. In conclusion, only 35% of patients had a correct assessment. Separate T- and N-stage prediction was improved with 51% and 72%, respectively. Major efforts are needed for improvement. INTRODUCTION Successful management of esophageal cancer represents an ongoing challenge and long-term outcome has only slightly improved in recent years. It remains among the most lethal gastrointestinal malignancies, and its reported overall 5-year survival reaches barely 20%.1,2 The incidences are increasing in many countries; whereby a marked shift from squamous cell carcinoma to adenocarcinoma is observed in Western countries where obesity and chronic gastroesophageal reflux have become major risk factors.1–4 In order to provide a tailored treatment to individual patients as well as a cost-effective use of health care resources, a precise staging is crucial. Patients with early tumor stages without regional lymph node metastasis should be distinguished, as those patients can be treated solely by an endoscopic or surgical intervention. Similarly, patients with advanced primary tumors or affected lymph nodes must be identified in a timely manner as multimodality treatments are considered most promising for advanced tumors.1,2,5 Once distant metastases have been excluded, the pretreatment staging is focused on the T- and N-stage in patients who should be treated with a curative intent. Basically, the T-stage means assessing the esophageal wall thickness and possible infiltrations into adjacent structures defining a T4 tumor that normally precludes any curative resection. Locoregional lymph node involvement is suspected based on imaging criteria such as size, form, and cluster formation.6 The latter two criteria are used to overcome the limitation of size, as even normally sized lymph nodes may be affected. Most commonly, different modalities are combined, such as CT, PET/CT, upper endoscopy with tissue biopsy and endosonography (EUS).7–9 There are several studies that reported separately the accuracy of CT, PET/CT, or EUS, but information of the overall accuracy using different modalities—as this is the case in clinical practice—are scarce. The aim of this study is to assess the global accuracy of the initial tumor staging of patients with resectable esophageal cancer without distant metastases who did not undergo neoadjuvant treatment. To this end, findings of the pretreatment staging were compared to the pathological assessment of the resected specimen. MATERIAL AND METHODS Patients More than 120 single items for each patient undergoing surgery for esophageal cancer at our university hospital are registered in an electronical data base since 2000. Data have been collected retrospectively until 2006, since then, data are prospectively recorded. A national quality control program has been introduced by law in 2014 with regular benchmarking. All patients with a priori resectable esophageal cancer, i.e., adenocarcinoma and squamous cell carcinoma, undergoing oncological esophagectomy from 2005 to 2016 have been identified. A priori resectable cancers have been defined as preoperative tumor stages that were not considered to require neoadjuvant treatments. Higher staged tumors (>uT2 and/or ≥uN+) were treated with neoadjuvant chemoradiotherapy prior to surgery. This restriction aimed to avoid possible bias from downstaging effects (Fig. 1). Patients who presented the following criteria were excluded: tumor recurrences after endoscopic or definitive chemoradiation, proximal esophageal cancer, tumors of the gastroesophageal junction (Siewert III) necessitating an enlarged total gastrectomy, and laryngeal tumors that required gastric pull up. And of course, patients who underwent esophagectomy for benign disease were also not considered. Fig. 1 View largeDownload slide Graph shows the flowchart of patient selection. NAT: neoadjuvant therapy. Fig. 1 View largeDownload slide Graph shows the flowchart of patient selection. NAT: neoadjuvant therapy. Preoperative staging Standardized preoperative staging included an upper gastrointestinal endoscopy and EUS, a thoracoabdominal CT scan, and a whole body 18-FDG-PET/CT scan. The more advanced stage was considered true, if the respective staging tool revealed different tumor stages. While the T-stage was predominantly determined by endoscopy and EUS, lymph nodes were assessed using radiological criteria and/or cytological assessment. Lymph nodes were considered to be tumor positive if their size was ≥1 cm, a central hypodensity was present, peripheral rim enhancement, and conglomeration of ≥3 lymph nodes despite the normal size.6 Cytological assessment of lymph nodes was rarely performed; only if the respective radiological studies were inconclusive or they were potentially considered as distant metastases. EUS was performed by different operators, but all of them are board certified gastroenterologists. Most patients were discussed at the weekly multidisciplinary tumor board, and tumor stages were determined after a thorough review of all staging examinations. Some few patients were directly admitted for surgery. The policy of neoadjuvant therapies was not uniformly accepted during the early study period. Tumor staging was performed according to the 7th edition of the TNM classification.10 According to current guidelines, at least six lymph nodes were required to define a pN0 status. If necessary, reclassification was performed of patients who were initially classified using previous editions of the TNM classification. Surgery The standard operation technique for patients with tumors located in the middle or distal third of the esophagus, meaning all tumors located below the level of the tracheal bifurcation, was a combined thoraco-abdominal approach as described by Lewis.11,12 Transhiatal resections or three-field resections were performed for proximal and middle intrathoracic tumors, or if a thoracotomy was contraindicated in distal tumors. In all cases, a two-field lymphadenectomy (abdominal and mediastinal) was performed and a gastric transplant was used as a neo-esophagus. Since 2006, the abdominal part of operations was performed laparoscopically by default. Always, a thoracotomy was performed. Only three senior surgeons with large experience in upper GI surgery were involved in all operations. Statistical analysis Preoperative staging of the T-stage, N-stage, and the combination of both was compared to final histopathological staging of surgical specimen that was considered as gold standard. In addition, it was compared whether preoperative staging could distinguish early versus advanced tumor stages to reliably indicate the need for a neoadjuvant treatment. Smoking habits were assessed as particular risk factor that could possibly confound staging results. Categorical data were presented as numbers with percentages, while continuous data were shown as median and ranges. Accuracy of preoperative testing was described as sensitivity, specificity, positive predictive values (PPV), and negative predictive values (NPV); they were expressed as percentages and 95% confidence intervals for ease interpretation. These parameters were calculated using 2 × 2 tables. Chi-square and Fisher tests were used for categorical data and p values < 0.05 were considered as statistically significant. Correct staging was defined as the preoperative T, N, or both were confirmed by the postoperative histological staging. Understaging was considered if the preoperative T, or N were lower as the postoperative values (e.g. preoperative T < postoperative T), and overstaging was defined if the preoperative T or N were overestimated (e.g. preoperative T > postoperative T). Data analysis was performed with MedCalc Software, Version 12.4.0 (B-8400 Ostend, Belgium). RESULTS Patient characteristics (Table 1, Fig. 1) There were 51 patients (38 male, 13 female patients) with a median age 65 years (IQR: 59.3–73 years). Regarding the tumor type, 21 patients had a SCC and 30 patients had an AC. Twenty-two of the 51 patients (43%) were active smokers or had a history of smoking. After the staging, all patients underwent oncological esophagectomy without any neoadjuvant treatment. While 40 patients (78%) underwent a thoracoabdominal resection with an intrathoracic anastomosis, the remaining 11 patients underwent either a thoracoabdominal resection with a cervical anastomosis or a transhiatal resection. The median number of resected lymph nodes in the overall group was 19 (range: 3–79). There were 17 lymph nodes (range: 3–79) for SCC and 19.5 lymph nodes (range: 8–40) for AC, respectively. Prediction of T-stage Overall, the T-stage was correctly predicted in 26 patients (51%), while it was understaged in 12 patients (23.5%) and overstaged in 13 patients (25.5%), respectively. While correct prediction of T-stages was only slightly different between AC and SCC, an almost significant higher rate of understaging was observed for AC (AC 33.3% vs. SCC 9.5%, P = 0.051). SCC and AC had a similar rate of overstaging (SCC 28.5% vs. AC 23.3%, P = 0.939). Differences between AC vs. SCC were not statistically significant (P = 0.139). (Table 2). The differentiation of early T-stages (T1 and 2) versus advanced T-stages (T3 and 4) was precise, as 35 out of 37 pT1/T2 tumors were correctly staged; the sensitivity, specificity, PPV and NPV for detection of early tumor stages were 94.6% (95% CI 81.8–99.3%), 71.4% (95% CI 41.9–91.6%), 89.7% (95% CI 79.2–95.3%) and 83.3% (95% CI 55.5–95.3%), respectively. Similarly, 10 out of 14 T3/T4 tumors (71.4%) were correctly identified (Table 3). Table 1 Patients characteristic   Characteristic  N  51  Male/female  38/13  Median age (IQR range) (years)  65 (59.3–73)  Median body mass index (range) (kg/m2)  25.3 (16.8–33.2)  Active smoking and/or history of smoking  22 (43%)  Squamous cell carcinoma  21 (41%)  Adenocarcinoma  30 (59%)  Type of operation  Transthoracic approach  40 (78%)  Triple approach/transhiatal approach  11 (22%)  Median number of resected lymph nodes (range)  Overall group  19 (3–79)  Squamous cell carcinoma  17 (3–79)  Adenocarcinoma  19.5 (8–40)    Characteristic  N  51  Male/female  38/13  Median age (IQR range) (years)  65 (59.3–73)  Median body mass index (range) (kg/m2)  25.3 (16.8–33.2)  Active smoking and/or history of smoking  22 (43%)  Squamous cell carcinoma  21 (41%)  Adenocarcinoma  30 (59%)  Type of operation  Transthoracic approach  40 (78%)  Triple approach/transhiatal approach  11 (22%)  Median number of resected lymph nodes (range)  Overall group  19 (3–79)  Squamous cell carcinoma  17 (3–79)  Adenocarcinoma  19.5 (8–40)  View Large Table 2 Accuracy of T-, N-, and overall staging   T correct†  T understaged†  T overstaged†  Overall group (n = 51)  26 (51%)  12 (23.5%)  13 (25.5%)  SCC  13 (61.9%)  2 (9.5%)  6 (28.5%)  AC  13 (43.3%)  10 (33.3%)  7 (23.3%)    N correct‡  N understaged‡  N overstaged‡  Overall group  37 (72.6%)  9 (17.6%)  5 (9.8%)  SCC  16 (76.2%)  4 (19%)  1 (4.8%)  AC  21 (70%)  5 (16.7%)  4 (13.3%)    T and N correct§  T and/or N not correct§    Overall group  18 (35.3%)  33 (64.7%)    SCC  9 (42.9%)  12 (57.1%)    AC  9 (30%)  21 (70%)      T correct†  T understaged†  T overstaged†  Overall group (n = 51)  26 (51%)  12 (23.5%)  13 (25.5%)  SCC  13 (61.9%)  2 (9.5%)  6 (28.5%)  AC  13 (43.3%)  10 (33.3%)  7 (23.3%)    N correct‡  N understaged‡  N overstaged‡  Overall group  37 (72.6%)  9 (17.6%)  5 (9.8%)  SCC  16 (76.2%)  4 (19%)  1 (4.8%)  AC  21 (70%)  5 (16.7%)  4 (13.3%)    T and N correct§  T and/or N not correct§    Overall group  18 (35.3%)  33 (64.7%)    SCC  9 (42.9%)  12 (57.1%)    AC  9 (30%)  21 (70%)    †understaged: preoperative T < postoperative T overstaged: preoperative T > postoperative T P = 0. 1391 SCC vs. AC ‡understaged: preoperative N < postoperative N overstaged: preoperative N > postoperative N P = 0.4738 SCC vs. AC §P = 0.3491 SCC vs. AC SCC, squamous cell carcinoma; AC: adenocarcinoma. View Large Table 3 Preoperative staging to predict early vs. advanced T-stage   pT1 and 2  pT3 and 4  T1 and 2  35  4  T3 and 4  2  10    pT1 and 2  pT3 and 4  T1 and 2  35  4  T3 and 4  2  10  Early: T1 and 2    Advanced: T3 and 4. View Large As T1 tumors may potentially benefit for local treatment by endoscopic mucosal or submucosal resection, it was assessed whether T1 tumor can be reliably distinguished from T2 tumors. Only 18 out of 27 T1 tumors (66.6%) were preoperatively correctly predicted, whereas the remaining eight tumors were misinterpreted as T2 tumors. Moreover, four tumors that were preoperatively classified as T2 were finally pT3. Another six tumors were considered to be a T1, but in fact, there were real pT2 tumors. The sensitivity, specificity, PPV, and NPV for detection of early tumor stages were 66.6% (95% CI 46–83.5%), 14.3% (95% CI 0.4–57.9%), 75% (95% CI 66.7–81.8), and 10% (95% CI 1.7–42.4%), respectively (Table 4). Table 4 Preoperative staging to predict pT1- vs. pT2-stage   pT1  pT2  T1  18  6  T2  9  1    pT1  pT2  T1  18  6  T2  9  1  View Large Table 5a Preoperative staging to predict tumor-positive pN-stage   pN0  pN+  N0  29  11  N+  4  7    pN0  pN+  N0  29  11  N+  4  7  Early: N0    Advanced: N+ N+: positive lymph nodes. View Large Prediction of N-stage (Tables 2 and 5a,b) Overall, the N-stage was correctly predicted in 37 patients (72.6%), while it was understaged in nine patients (17.6%) and overstaged in five patients (9.8%), respectively. Prediction of N-stages was not different between AC and SCC (P = 0.473). There, 29 out of 33 patients who were preoperatively correctly identified as N0 (87.9%), but only 7 out of 18 patients (38.9%) were correctly assessed as N+. The sensitivity, specificity, positive and negative predictive values for the correct identification of tumor-free lymph nodes were 87.9% (95% CI 71.–96.6%), 38.9% (95% CI 17.3–64.3%), 72.5% (95% CI 64.1–79.6%), and 63.6% (95% CI 37.1–83.8%), respectively. Among the patients who finally had tumor-positive lymph nodes, the precise N-stage (N1, N2, and N3) could not be precisely be predicted (Table 5b). Table 5b Preoperative staging to predict tumor-positive pN1–3-stages   pN0  pN1  pN2  pN3  N0  29  6  2  3  N+  4  3  4  0    pN0  pN1  pN2  pN3  N0  29  6  2  3  N+  4  3  4  0  View Large Table 6 Preoperative staging to predict early vs. advanced stage   pT1 and 2, pN0  pT1 and 2, pN+      pT3 and 4, pN0 or pN+  T1 and 2, N0  27  9  T1 and 2, N+  3  12  T 3 and 4, N0 or N+        pT1 and 2, pN0  pT1 and 2, pN+      pT3 and 4, pN0 or pN+  T1 and 2, N0  27  9  T1 and 2, N+  3  12  T 3 and 4, N0 or N+      Early: T1 and 2, N0 Advanced: T1 and 2, N+/T3 and 4, N0 or N+ N+: positive lymph nodes. View Large Table 7 Impact of smoking on the accuracy of preoperative staging   T staging  N staging    not correct  not correct  smoking, n = 22  11/22, 50%  7/22, 31.8%      but 6/22 N0 to N+ missed  No smoking, n = 29  11/29, 37.9%  6/29, 20.7%      but 3/26 N0 to N+ missed    T staging  N staging    not correct  not correct  smoking, n = 22  11/22, 50%  7/22, 31.8%      but 6/22 N0 to N+ missed  No smoking, n = 29  11/29, 37.9%  6/29, 20.7%      but 3/26 N0 to N+ missed  View Large Combined prediction of T-stage and N-stage (Table 2) Overall, 18 patients (35.3%) were preoperatively diagnosed with a correct T- and N-stage; whereas in 33 patients (64.7%) either the T- and/or the N-stage was not correct. Prediction of combined T-, and N-stage was not different between AC and SCC (P = 0.349). Prediction of early versus advanced tumor stages (Table 6) Early tumor stages are characterized as T1 and T2 tumors without lymph node metastasis, and its precise identification is relevant as such early tumor stages can be scheduled for surgery without the need of any neoadjuvant treatment. In contrast, advanced tumor stages are characterized as T3 and T4 tumors, with or without positive lymph nodes. There were 30 patients (58.8%) with early disease and 21 patients (41.2%) with advanced disease, respectively. The sensitivity and specificity of preoperative staging to correctly predict early tumor stages were 90% (95% CI 73.5–97.9%) and 57.1% (95% CI 34–78.2%), while the positive and negative predictive values were 75% (95% CI 64.3–83.3%) and 80% (95% CI 56.2–92.6%), respectively. Impact of smoking on the accuracy of preoperative staging (Table 7) In the group of 22 patients who were smoking, in 11 patients (50%) the T-stage was not correct, and in 7 patients (1.8%) the N-stage was not correct. Of note, 6 out of these 7 patients were falsely classified as N0, but in fact, the lymph nodes were infiltrated by tumor cells (≥N1). Within the group of 29 nonsmoking patients, in 11 patients (37.9%) the T-stage was not correctly predicted, as well as 6 patients (20.7%) had no correct N-stage. Understaging was observed in 3 of these 6 patients (≥N1). Overall, 6 patients (27.3%) who were smoking had a correct T- and N-stage, and 12 patients (41.4%) who were not smoking had a correct T- and N-stage, respectively. DISCUSSION A meaningful decision-making and management of esophageal cancer requires a precise tumor staging; in particular, as different treatment modalities according to the tumor stage are available. To this end, once distant metastases are excluded, early T-stages and the presence of suspected lymph nodes are of particular interest, as its treatments modalities are critically related to these two findings.1,2 Standard imaging of resectable esophageal cancer includes a computed tomography (CT) of the chest and abdomen scan and a EUS.13 A PET–CT is most helpful to exclude distant organ metastases including distant lymph node metastases.13,14 Assessing the T-stage by CT, the thickness of the esophageal wall is considered pathological if larger >5 mm and a major asymmetry is present.15 However, these two signs are unspecific, and not surprisingly, the reported accuracy to reliably differentiate T1, T2, and T3 tumors is only 49%–59%.16 The CT is most useful to detect tumor invasion into adjacent structures (T4 tumors), with reported sensitivity and specificity rates of mediastinal invasion of 85%–100%.16 The EUS is more precise to define T1, T2, and T3 tumors, with reported accuracy rates ranging from 85% to 95%, and even T1a and T1b tumors can be predicted with an 85% accuracy.17,18 Nevertheless, EUS has several important limitations; such as it is operator dependent and incomplete visualization of the cancerous lesion can occur in case of esophageal strictures and stenosis. EUS assessment of cervical tumors has also an increased risk of aspiration pneumonia.7 And finally, T-stages cannot reliably be assessed after neoadjuvant chemoradiotherapy.19 In our series, we used CT and EUS together to define T- and N-stages. We found that the T-stage was correctly predicted in 51% of all patients, while 23% were understaged and 26% were overstaged. The identification of early T-stages (T1 and T2) was much better with a sensitivity of 90%. In contrast, only two third of T1 tumors were correctly identified among the early T-stages. Overall our results are within the reported ranges in the literature, but rather at the inferior border. To this end, this series probably represent the ‘real life’, where not only very well trained specialists performed the EUS. Noteworthy, there was no difference between the two tumor types (SCC vs. AC). Tumor-free lymph nodes were identified with a high accuracy of 88%, whereas the correct identification of tumor-affected lymph nodes was much more difficult, and only 39% were preoperatively correctly identified. An interesting finding was the high failure rate to identify tumor-affected lymph nodes in patients who were actively smoking or had a positive history of smoking. The reported sensitivity, specificity, and overall accuracy to detect lymph node metastases by CT are rather low with 30%–60%, 60%–80%, and 46%–58%, respectively.16 As CT scans are only able to depict morphological changes, therefore, small tumor deposits can easily be hidden in ‘normally’ sized lymph nodes, as well as enlarged lymph nodes may occur in case of inflammation, smoking or infectious diseases. Accuracy of EUS to detect pathological lymph nodes based on morphological criteria as size, clear borders, circular form and homogenously hypoechoic texture has been reported to range from 72% to 80%.16 More recently, accuracy can be improved in case of fine needle aspiration of suspected malignancy.16,17 Separate assessments of T- and N-stages are important regarding the indications for neoadjuvant treatment as well as local resections for superficial tumors, while the combined correct prediction of the T- and N-stage in individual patients is rarely reported. Not surprisingly, as CT and EUS have its limitations, only in 18 out of 51 patients (35%), both, the T- and N-stage were correctly predicted preoperatively. Some shortcomings need to be addressed impacting the results of this actual series. First, the number of included patients is rather small, and data were analyzed retrospectively. But only by excluding patients with neoadjuvant therapies, preoperative assessment can be reliably compared to pathological findings. Most patients reveal an advanced tumor stage at diagnosis. Despite the analysis was retrospectively done, all data those from 2005 were prospectively collected, and all patients were treated in a single center. Second, EUS was performed solely by gastroenterologists, of whom; the level of experience was different; and a routine biopsy policy was not adopted until recently. In most centers, EUS are performed by different investigators, and thus, this mixture of different levels of experience represents the real clinical situation. Third, EUS-guided biopsies were underused for several reasons: If radiological criteria for malignancy were fulfilled, lymph nodes were not biopsied. During early years, EUS was performed by using a radial EUS, while linear EUS devices were only introduced later. Finally, patients who were just referred for surgery underwent staging externally, for cost reasons and limited availability of biopsy facilities, only patients with unclear lymph node findings underwent biopsies. In conclusion, members of multidisciplinary tumor boards must be aware of the limited accuracy of preoperative staging, i.e. T- and N-stages. Both, under- and overstaging are common; and patients are at potential risk to be under- and/or overtreated. PET-CT, MRI and EUS-guided FNA should probably be more generously used to improve initial tumor staging. Notes Specific author contributions: Winiker M and Schafer M contributed equally to this work; Winiker M and Schafer M collected, and analyzed the data, and drafted the manuscript; Allemann P and Mantziari S provided analytical oversight; Schafer M designed and supervised the study; Allemann P, Mantziari S, Figueiredo AG, and Demartines N revised the manuscript for important intellectual content; all authors have read and approved the final version to be published. Supported by: no support. Institutional review board statement: The study was reviewed and approved by the Institutional Review Board of the Department of Surgery, University Hospital of Lausanne. Biostatistics statement: The statistical methods of this study were reviewed by Professor Markus Schäfer, Department of Visceral Surgery, University Hospital CHUV, CH-1011 Lausanne, Switzerland. Informed consent statement: Preoperatively, all patients provided informed written consent to be included into the institutional data base. Since 2014, esophageal cancer treatment has been centralized by a federal law; and according to this new law, all patients are obligated to provide their data for quality control and research. Conflicts of interest: The authors declare that they have no conflict of interest. References 1 Pennathur A, Gibson M K, Jobe B A, Luketich J D. Oesophageal cancer. Lancet  2013; 381: 400– 12. Google Scholar CrossRef Search ADS PubMed  2 Rustgi A K, El-Serag H B. Esophageal carcinoma. N Engl J Med  2014; 371: 2499– 509. Google Scholar CrossRef Search ADS PubMed  3 Joliat G R, Hahnloser D, Demartines N, Schafer M. Future development of gastrointestinal cancer incidence and mortality rates in Switzerland: a tumor registry and population-based projection up to 2030. Swiss Med Wkly  2015; 145: w14188 [PMID: 26376322 doi:10.4414/smw.2015.14188]. Google Scholar PubMed  4 Thrift A P. The epidemic of oesophageal carcinoma: where are we now? Cancer Epidemiol  2016; 41: 88– 95. Google Scholar CrossRef Search ADS PubMed  5 Jang R, Darling G, Wong R K. Multimodality approaches for the curative treatment of esophageal cancer. J Natl Compr Canc Netw  2015; 13: 229– 38. Google Scholar CrossRef Search ADS PubMed  6 Sarvanan K, Bapuraj J R, Sharma S C et al.   Computed tomography and ultrasonographic evaluation of metastatic cervical lymph nodes with surgicoclinicopathologic correlation. J Laryngol Otol  2002; 116: 194– 9. Google Scholar CrossRef Search ADS PubMed  7 Cho J W. The role of endosonography in the staging of gastrointestinal cancers. Clin Endosc  2015; 48: 297– 301. Google Scholar CrossRef Search ADS PubMed  8 Tirumani H, Rosenthal M H, Tirumani S H, Shinagare A B, Krajewski K M, Ramaiya N H. Esophageal carcinoma: current concepts in the role of imaging in staging and management. Can Assoc Radiol J  2015; 66: 130– 9. Google Scholar CrossRef Search ADS PubMed  9 Schmidt T, Lordick F, Hermann K, Ott K. Value of functional imaging by PET in esophageal cancer. J Natl Compr Canc Netw  2015; 13: 239– 47. Google Scholar CrossRef Search ADS PubMed  10 Blackshaw G, Lewis W G, Hopper A N et al.   Prospective comparison of endosonography, computed tomography, and histopathological stage of junctional oesophagogastric cancer. Clin Radiol  2008; 63: 1092– 8. Google Scholar CrossRef Search ADS PubMed  11 Lewis I. Carcinoma of the oesophagus. Radical resection with oesophagogastrostomy for a midthoracic growth by a right transpleural approach. Proc R Soc Med  1945; 38: 483– 4. 12 Holscher A H, Schneider P M, Gutschow C, Schroeder W. Laparoscopic ischemic conditioning of the stomach for esophageal replacement. Ann Surg  2007; 245: 241– 6. Google Scholar CrossRef Search ADS PubMed  13 Stahl M, Mariette C, Haustermans K, Cervantes A, Arnold D. Oesophageal cancer: ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol  2013; 24( Suppl 6): vi51– 6. Google Scholar CrossRef Search ADS PubMed  14 Findlay J M, Bradley K M, Maile E J et al.   Pragmatic staging of oesophageal cancer using decision theory involving selective endoscopic ultrasonography, PET, and laparoscopy. Br J Surg  2015; 102: 1488– 99. Google Scholar CrossRef Search ADS PubMed  15 Noh H M, Fishman E K, Forastiere A A, Bliss D F, Calhoun P S. CT of the esophagus: spectrum of disease with emphasis on esophageal carcinoma. Radiographics  1995; 15: 1113– 34. Google Scholar CrossRef Search ADS PubMed  16 Napier K J, Scheerer M, Misra S. Esophageal cancer: a review of epidemiology, pathogenesis, staging workup and treatment modalities. World J Gastrointest Oncol  2014; 6: 112– 20. Google Scholar CrossRef Search ADS PubMed  17 Puli S R, Reddy J B, Bechtold M L, Antillon D, Ibdah J A, Antillon M R. Staging accuracy of esophageal cancer by endoscopic ultrasound: a meta-analysis and systematic review. World J Gastroenterol  2008; 14: 1479– 90. Google Scholar CrossRef Search ADS PubMed  18 Thosani N, Singh H, Kapadia A et al.   Diagnostic accuracy of EUS in differentiating mucosal versus submucosal invasion of superficial esophageal cancers: a systematic review and meta-analysis. Gastrointest Endosc  2012; 75: 242– 53. Google Scholar CrossRef Search ADS PubMed  19 Bohle W, Kasper M, Zoller W G. Different accuracy of endosonographic tumor staging after neoadjuvant chemotherapy and chemoradiotherapy in esophageal cancer. Surg Endosc  2016; 30: 2922– 8. Google Scholar CrossRef Search ADS PubMed  © The Authors 2017. Published by Oxford University Press on behalf of International Society for Diseases of the Esophagus. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Diseases of the Esophagus Oxford University Press

Accuracy of preoperative staging for a priori resectable esophageal cancer

Loading next page...
 
/lp/ou_press/accuracy-of-preoperative-staging-for-a-priori-resectable-esophageal-a0Z82KcxL7
Publisher
The International Society for Diseases of the Esophagus
Copyright
© The Authors 2017. Published by Oxford University Press on behalf of International Society for Diseases of the Esophagus. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
1120-8694
eISSN
1442-2050
D.O.I.
10.1093/dote/dox113
Publisher site
See Article on Publisher Site

Abstract

Summary This study assessed the accuracy of preoperative staging in patients undergoing oncological esophagectomy for adenocarcinoma and squamous cell carcinoma. All patients undergoing surgery for resectable esophageal cancer in a university hospital from 2005 to 2016 were identified from our institutional database. Patients with neoadjuvant treatment were excluded to avoid bias from down-staging effects. Routinely, all patients had an upper endoscopy with biopsy, a thoracoabdominal CT scan, an 18-FEG PET-CT, and endoscopic ultrasound. Preoperative staging was compared to histopathological staging of surgical specimen that was considered as gold standard. There were 51 patients with a median age of 65 years (IQR: 59.3–73 years) having 21 squamous cell carcinoma and 30 adenocarcinoma, respectively. T- and N-stages were correctly predicted in 26 (51%) and 37 patients (72%), respectively. Overall, 18 patients (35%) were preoperatively diagnosed with a correct T- and N-stage. There was no difference between adenocarcinoma and squamous cell carcinoma. Accuracy of the T-stage was not influenced by the smoking status. The N-stage was not correct in 7/22 smoking patients (32%) and 6/29 nonsmoking patients (21%).The N-stage was underestimated in smoking patients as 6/22 patients (27%) had a histologically confirmed N+ who were preoperatively classified as N0. In conclusion, only 35% of patients had a correct assessment. Separate T- and N-stage prediction was improved with 51% and 72%, respectively. Major efforts are needed for improvement. INTRODUCTION Successful management of esophageal cancer represents an ongoing challenge and long-term outcome has only slightly improved in recent years. It remains among the most lethal gastrointestinal malignancies, and its reported overall 5-year survival reaches barely 20%.1,2 The incidences are increasing in many countries; whereby a marked shift from squamous cell carcinoma to adenocarcinoma is observed in Western countries where obesity and chronic gastroesophageal reflux have become major risk factors.1–4 In order to provide a tailored treatment to individual patients as well as a cost-effective use of health care resources, a precise staging is crucial. Patients with early tumor stages without regional lymph node metastasis should be distinguished, as those patients can be treated solely by an endoscopic or surgical intervention. Similarly, patients with advanced primary tumors or affected lymph nodes must be identified in a timely manner as multimodality treatments are considered most promising for advanced tumors.1,2,5 Once distant metastases have been excluded, the pretreatment staging is focused on the T- and N-stage in patients who should be treated with a curative intent. Basically, the T-stage means assessing the esophageal wall thickness and possible infiltrations into adjacent structures defining a T4 tumor that normally precludes any curative resection. Locoregional lymph node involvement is suspected based on imaging criteria such as size, form, and cluster formation.6 The latter two criteria are used to overcome the limitation of size, as even normally sized lymph nodes may be affected. Most commonly, different modalities are combined, such as CT, PET/CT, upper endoscopy with tissue biopsy and endosonography (EUS).7–9 There are several studies that reported separately the accuracy of CT, PET/CT, or EUS, but information of the overall accuracy using different modalities—as this is the case in clinical practice—are scarce. The aim of this study is to assess the global accuracy of the initial tumor staging of patients with resectable esophageal cancer without distant metastases who did not undergo neoadjuvant treatment. To this end, findings of the pretreatment staging were compared to the pathological assessment of the resected specimen. MATERIAL AND METHODS Patients More than 120 single items for each patient undergoing surgery for esophageal cancer at our university hospital are registered in an electronical data base since 2000. Data have been collected retrospectively until 2006, since then, data are prospectively recorded. A national quality control program has been introduced by law in 2014 with regular benchmarking. All patients with a priori resectable esophageal cancer, i.e., adenocarcinoma and squamous cell carcinoma, undergoing oncological esophagectomy from 2005 to 2016 have been identified. A priori resectable cancers have been defined as preoperative tumor stages that were not considered to require neoadjuvant treatments. Higher staged tumors (>uT2 and/or ≥uN+) were treated with neoadjuvant chemoradiotherapy prior to surgery. This restriction aimed to avoid possible bias from downstaging effects (Fig. 1). Patients who presented the following criteria were excluded: tumor recurrences after endoscopic or definitive chemoradiation, proximal esophageal cancer, tumors of the gastroesophageal junction (Siewert III) necessitating an enlarged total gastrectomy, and laryngeal tumors that required gastric pull up. And of course, patients who underwent esophagectomy for benign disease were also not considered. Fig. 1 View largeDownload slide Graph shows the flowchart of patient selection. NAT: neoadjuvant therapy. Fig. 1 View largeDownload slide Graph shows the flowchart of patient selection. NAT: neoadjuvant therapy. Preoperative staging Standardized preoperative staging included an upper gastrointestinal endoscopy and EUS, a thoracoabdominal CT scan, and a whole body 18-FDG-PET/CT scan. The more advanced stage was considered true, if the respective staging tool revealed different tumor stages. While the T-stage was predominantly determined by endoscopy and EUS, lymph nodes were assessed using radiological criteria and/or cytological assessment. Lymph nodes were considered to be tumor positive if their size was ≥1 cm, a central hypodensity was present, peripheral rim enhancement, and conglomeration of ≥3 lymph nodes despite the normal size.6 Cytological assessment of lymph nodes was rarely performed; only if the respective radiological studies were inconclusive or they were potentially considered as distant metastases. EUS was performed by different operators, but all of them are board certified gastroenterologists. Most patients were discussed at the weekly multidisciplinary tumor board, and tumor stages were determined after a thorough review of all staging examinations. Some few patients were directly admitted for surgery. The policy of neoadjuvant therapies was not uniformly accepted during the early study period. Tumor staging was performed according to the 7th edition of the TNM classification.10 According to current guidelines, at least six lymph nodes were required to define a pN0 status. If necessary, reclassification was performed of patients who were initially classified using previous editions of the TNM classification. Surgery The standard operation technique for patients with tumors located in the middle or distal third of the esophagus, meaning all tumors located below the level of the tracheal bifurcation, was a combined thoraco-abdominal approach as described by Lewis.11,12 Transhiatal resections or three-field resections were performed for proximal and middle intrathoracic tumors, or if a thoracotomy was contraindicated in distal tumors. In all cases, a two-field lymphadenectomy (abdominal and mediastinal) was performed and a gastric transplant was used as a neo-esophagus. Since 2006, the abdominal part of operations was performed laparoscopically by default. Always, a thoracotomy was performed. Only three senior surgeons with large experience in upper GI surgery were involved in all operations. Statistical analysis Preoperative staging of the T-stage, N-stage, and the combination of both was compared to final histopathological staging of surgical specimen that was considered as gold standard. In addition, it was compared whether preoperative staging could distinguish early versus advanced tumor stages to reliably indicate the need for a neoadjuvant treatment. Smoking habits were assessed as particular risk factor that could possibly confound staging results. Categorical data were presented as numbers with percentages, while continuous data were shown as median and ranges. Accuracy of preoperative testing was described as sensitivity, specificity, positive predictive values (PPV), and negative predictive values (NPV); they were expressed as percentages and 95% confidence intervals for ease interpretation. These parameters were calculated using 2 × 2 tables. Chi-square and Fisher tests were used for categorical data and p values < 0.05 were considered as statistically significant. Correct staging was defined as the preoperative T, N, or both were confirmed by the postoperative histological staging. Understaging was considered if the preoperative T, or N were lower as the postoperative values (e.g. preoperative T < postoperative T), and overstaging was defined if the preoperative T or N were overestimated (e.g. preoperative T > postoperative T). Data analysis was performed with MedCalc Software, Version 12.4.0 (B-8400 Ostend, Belgium). RESULTS Patient characteristics (Table 1, Fig. 1) There were 51 patients (38 male, 13 female patients) with a median age 65 years (IQR: 59.3–73 years). Regarding the tumor type, 21 patients had a SCC and 30 patients had an AC. Twenty-two of the 51 patients (43%) were active smokers or had a history of smoking. After the staging, all patients underwent oncological esophagectomy without any neoadjuvant treatment. While 40 patients (78%) underwent a thoracoabdominal resection with an intrathoracic anastomosis, the remaining 11 patients underwent either a thoracoabdominal resection with a cervical anastomosis or a transhiatal resection. The median number of resected lymph nodes in the overall group was 19 (range: 3–79). There were 17 lymph nodes (range: 3–79) for SCC and 19.5 lymph nodes (range: 8–40) for AC, respectively. Prediction of T-stage Overall, the T-stage was correctly predicted in 26 patients (51%), while it was understaged in 12 patients (23.5%) and overstaged in 13 patients (25.5%), respectively. While correct prediction of T-stages was only slightly different between AC and SCC, an almost significant higher rate of understaging was observed for AC (AC 33.3% vs. SCC 9.5%, P = 0.051). SCC and AC had a similar rate of overstaging (SCC 28.5% vs. AC 23.3%, P = 0.939). Differences between AC vs. SCC were not statistically significant (P = 0.139). (Table 2). The differentiation of early T-stages (T1 and 2) versus advanced T-stages (T3 and 4) was precise, as 35 out of 37 pT1/T2 tumors were correctly staged; the sensitivity, specificity, PPV and NPV for detection of early tumor stages were 94.6% (95% CI 81.8–99.3%), 71.4% (95% CI 41.9–91.6%), 89.7% (95% CI 79.2–95.3%) and 83.3% (95% CI 55.5–95.3%), respectively. Similarly, 10 out of 14 T3/T4 tumors (71.4%) were correctly identified (Table 3). Table 1 Patients characteristic   Characteristic  N  51  Male/female  38/13  Median age (IQR range) (years)  65 (59.3–73)  Median body mass index (range) (kg/m2)  25.3 (16.8–33.2)  Active smoking and/or history of smoking  22 (43%)  Squamous cell carcinoma  21 (41%)  Adenocarcinoma  30 (59%)  Type of operation  Transthoracic approach  40 (78%)  Triple approach/transhiatal approach  11 (22%)  Median number of resected lymph nodes (range)  Overall group  19 (3–79)  Squamous cell carcinoma  17 (3–79)  Adenocarcinoma  19.5 (8–40)    Characteristic  N  51  Male/female  38/13  Median age (IQR range) (years)  65 (59.3–73)  Median body mass index (range) (kg/m2)  25.3 (16.8–33.2)  Active smoking and/or history of smoking  22 (43%)  Squamous cell carcinoma  21 (41%)  Adenocarcinoma  30 (59%)  Type of operation  Transthoracic approach  40 (78%)  Triple approach/transhiatal approach  11 (22%)  Median number of resected lymph nodes (range)  Overall group  19 (3–79)  Squamous cell carcinoma  17 (3–79)  Adenocarcinoma  19.5 (8–40)  View Large Table 2 Accuracy of T-, N-, and overall staging   T correct†  T understaged†  T overstaged†  Overall group (n = 51)  26 (51%)  12 (23.5%)  13 (25.5%)  SCC  13 (61.9%)  2 (9.5%)  6 (28.5%)  AC  13 (43.3%)  10 (33.3%)  7 (23.3%)    N correct‡  N understaged‡  N overstaged‡  Overall group  37 (72.6%)  9 (17.6%)  5 (9.8%)  SCC  16 (76.2%)  4 (19%)  1 (4.8%)  AC  21 (70%)  5 (16.7%)  4 (13.3%)    T and N correct§  T and/or N not correct§    Overall group  18 (35.3%)  33 (64.7%)    SCC  9 (42.9%)  12 (57.1%)    AC  9 (30%)  21 (70%)      T correct†  T understaged†  T overstaged†  Overall group (n = 51)  26 (51%)  12 (23.5%)  13 (25.5%)  SCC  13 (61.9%)  2 (9.5%)  6 (28.5%)  AC  13 (43.3%)  10 (33.3%)  7 (23.3%)    N correct‡  N understaged‡  N overstaged‡  Overall group  37 (72.6%)  9 (17.6%)  5 (9.8%)  SCC  16 (76.2%)  4 (19%)  1 (4.8%)  AC  21 (70%)  5 (16.7%)  4 (13.3%)    T and N correct§  T and/or N not correct§    Overall group  18 (35.3%)  33 (64.7%)    SCC  9 (42.9%)  12 (57.1%)    AC  9 (30%)  21 (70%)    †understaged: preoperative T < postoperative T overstaged: preoperative T > postoperative T P = 0. 1391 SCC vs. AC ‡understaged: preoperative N < postoperative N overstaged: preoperative N > postoperative N P = 0.4738 SCC vs. AC §P = 0.3491 SCC vs. AC SCC, squamous cell carcinoma; AC: adenocarcinoma. View Large Table 3 Preoperative staging to predict early vs. advanced T-stage   pT1 and 2  pT3 and 4  T1 and 2  35  4  T3 and 4  2  10    pT1 and 2  pT3 and 4  T1 and 2  35  4  T3 and 4  2  10  Early: T1 and 2    Advanced: T3 and 4. View Large As T1 tumors may potentially benefit for local treatment by endoscopic mucosal or submucosal resection, it was assessed whether T1 tumor can be reliably distinguished from T2 tumors. Only 18 out of 27 T1 tumors (66.6%) were preoperatively correctly predicted, whereas the remaining eight tumors were misinterpreted as T2 tumors. Moreover, four tumors that were preoperatively classified as T2 were finally pT3. Another six tumors were considered to be a T1, but in fact, there were real pT2 tumors. The sensitivity, specificity, PPV, and NPV for detection of early tumor stages were 66.6% (95% CI 46–83.5%), 14.3% (95% CI 0.4–57.9%), 75% (95% CI 66.7–81.8), and 10% (95% CI 1.7–42.4%), respectively (Table 4). Table 4 Preoperative staging to predict pT1- vs. pT2-stage   pT1  pT2  T1  18  6  T2  9  1    pT1  pT2  T1  18  6  T2  9  1  View Large Table 5a Preoperative staging to predict tumor-positive pN-stage   pN0  pN+  N0  29  11  N+  4  7    pN0  pN+  N0  29  11  N+  4  7  Early: N0    Advanced: N+ N+: positive lymph nodes. View Large Prediction of N-stage (Tables 2 and 5a,b) Overall, the N-stage was correctly predicted in 37 patients (72.6%), while it was understaged in nine patients (17.6%) and overstaged in five patients (9.8%), respectively. Prediction of N-stages was not different between AC and SCC (P = 0.473). There, 29 out of 33 patients who were preoperatively correctly identified as N0 (87.9%), but only 7 out of 18 patients (38.9%) were correctly assessed as N+. The sensitivity, specificity, positive and negative predictive values for the correct identification of tumor-free lymph nodes were 87.9% (95% CI 71.–96.6%), 38.9% (95% CI 17.3–64.3%), 72.5% (95% CI 64.1–79.6%), and 63.6% (95% CI 37.1–83.8%), respectively. Among the patients who finally had tumor-positive lymph nodes, the precise N-stage (N1, N2, and N3) could not be precisely be predicted (Table 5b). Table 5b Preoperative staging to predict tumor-positive pN1–3-stages   pN0  pN1  pN2  pN3  N0  29  6  2  3  N+  4  3  4  0    pN0  pN1  pN2  pN3  N0  29  6  2  3  N+  4  3  4  0  View Large Table 6 Preoperative staging to predict early vs. advanced stage   pT1 and 2, pN0  pT1 and 2, pN+      pT3 and 4, pN0 or pN+  T1 and 2, N0  27  9  T1 and 2, N+  3  12  T 3 and 4, N0 or N+        pT1 and 2, pN0  pT1 and 2, pN+      pT3 and 4, pN0 or pN+  T1 and 2, N0  27  9  T1 and 2, N+  3  12  T 3 and 4, N0 or N+      Early: T1 and 2, N0 Advanced: T1 and 2, N+/T3 and 4, N0 or N+ N+: positive lymph nodes. View Large Table 7 Impact of smoking on the accuracy of preoperative staging   T staging  N staging    not correct  not correct  smoking, n = 22  11/22, 50%  7/22, 31.8%      but 6/22 N0 to N+ missed  No smoking, n = 29  11/29, 37.9%  6/29, 20.7%      but 3/26 N0 to N+ missed    T staging  N staging    not correct  not correct  smoking, n = 22  11/22, 50%  7/22, 31.8%      but 6/22 N0 to N+ missed  No smoking, n = 29  11/29, 37.9%  6/29, 20.7%      but 3/26 N0 to N+ missed  View Large Combined prediction of T-stage and N-stage (Table 2) Overall, 18 patients (35.3%) were preoperatively diagnosed with a correct T- and N-stage; whereas in 33 patients (64.7%) either the T- and/or the N-stage was not correct. Prediction of combined T-, and N-stage was not different between AC and SCC (P = 0.349). Prediction of early versus advanced tumor stages (Table 6) Early tumor stages are characterized as T1 and T2 tumors without lymph node metastasis, and its precise identification is relevant as such early tumor stages can be scheduled for surgery without the need of any neoadjuvant treatment. In contrast, advanced tumor stages are characterized as T3 and T4 tumors, with or without positive lymph nodes. There were 30 patients (58.8%) with early disease and 21 patients (41.2%) with advanced disease, respectively. The sensitivity and specificity of preoperative staging to correctly predict early tumor stages were 90% (95% CI 73.5–97.9%) and 57.1% (95% CI 34–78.2%), while the positive and negative predictive values were 75% (95% CI 64.3–83.3%) and 80% (95% CI 56.2–92.6%), respectively. Impact of smoking on the accuracy of preoperative staging (Table 7) In the group of 22 patients who were smoking, in 11 patients (50%) the T-stage was not correct, and in 7 patients (1.8%) the N-stage was not correct. Of note, 6 out of these 7 patients were falsely classified as N0, but in fact, the lymph nodes were infiltrated by tumor cells (≥N1). Within the group of 29 nonsmoking patients, in 11 patients (37.9%) the T-stage was not correctly predicted, as well as 6 patients (20.7%) had no correct N-stage. Understaging was observed in 3 of these 6 patients (≥N1). Overall, 6 patients (27.3%) who were smoking had a correct T- and N-stage, and 12 patients (41.4%) who were not smoking had a correct T- and N-stage, respectively. DISCUSSION A meaningful decision-making and management of esophageal cancer requires a precise tumor staging; in particular, as different treatment modalities according to the tumor stage are available. To this end, once distant metastases are excluded, early T-stages and the presence of suspected lymph nodes are of particular interest, as its treatments modalities are critically related to these two findings.1,2 Standard imaging of resectable esophageal cancer includes a computed tomography (CT) of the chest and abdomen scan and a EUS.13 A PET–CT is most helpful to exclude distant organ metastases including distant lymph node metastases.13,14 Assessing the T-stage by CT, the thickness of the esophageal wall is considered pathological if larger >5 mm and a major asymmetry is present.15 However, these two signs are unspecific, and not surprisingly, the reported accuracy to reliably differentiate T1, T2, and T3 tumors is only 49%–59%.16 The CT is most useful to detect tumor invasion into adjacent structures (T4 tumors), with reported sensitivity and specificity rates of mediastinal invasion of 85%–100%.16 The EUS is more precise to define T1, T2, and T3 tumors, with reported accuracy rates ranging from 85% to 95%, and even T1a and T1b tumors can be predicted with an 85% accuracy.17,18 Nevertheless, EUS has several important limitations; such as it is operator dependent and incomplete visualization of the cancerous lesion can occur in case of esophageal strictures and stenosis. EUS assessment of cervical tumors has also an increased risk of aspiration pneumonia.7 And finally, T-stages cannot reliably be assessed after neoadjuvant chemoradiotherapy.19 In our series, we used CT and EUS together to define T- and N-stages. We found that the T-stage was correctly predicted in 51% of all patients, while 23% were understaged and 26% were overstaged. The identification of early T-stages (T1 and T2) was much better with a sensitivity of 90%. In contrast, only two third of T1 tumors were correctly identified among the early T-stages. Overall our results are within the reported ranges in the literature, but rather at the inferior border. To this end, this series probably represent the ‘real life’, where not only very well trained specialists performed the EUS. Noteworthy, there was no difference between the two tumor types (SCC vs. AC). Tumor-free lymph nodes were identified with a high accuracy of 88%, whereas the correct identification of tumor-affected lymph nodes was much more difficult, and only 39% were preoperatively correctly identified. An interesting finding was the high failure rate to identify tumor-affected lymph nodes in patients who were actively smoking or had a positive history of smoking. The reported sensitivity, specificity, and overall accuracy to detect lymph node metastases by CT are rather low with 30%–60%, 60%–80%, and 46%–58%, respectively.16 As CT scans are only able to depict morphological changes, therefore, small tumor deposits can easily be hidden in ‘normally’ sized lymph nodes, as well as enlarged lymph nodes may occur in case of inflammation, smoking or infectious diseases. Accuracy of EUS to detect pathological lymph nodes based on morphological criteria as size, clear borders, circular form and homogenously hypoechoic texture has been reported to range from 72% to 80%.16 More recently, accuracy can be improved in case of fine needle aspiration of suspected malignancy.16,17 Separate assessments of T- and N-stages are important regarding the indications for neoadjuvant treatment as well as local resections for superficial tumors, while the combined correct prediction of the T- and N-stage in individual patients is rarely reported. Not surprisingly, as CT and EUS have its limitations, only in 18 out of 51 patients (35%), both, the T- and N-stage were correctly predicted preoperatively. Some shortcomings need to be addressed impacting the results of this actual series. First, the number of included patients is rather small, and data were analyzed retrospectively. But only by excluding patients with neoadjuvant therapies, preoperative assessment can be reliably compared to pathological findings. Most patients reveal an advanced tumor stage at diagnosis. Despite the analysis was retrospectively done, all data those from 2005 were prospectively collected, and all patients were treated in a single center. Second, EUS was performed solely by gastroenterologists, of whom; the level of experience was different; and a routine biopsy policy was not adopted until recently. In most centers, EUS are performed by different investigators, and thus, this mixture of different levels of experience represents the real clinical situation. Third, EUS-guided biopsies were underused for several reasons: If radiological criteria for malignancy were fulfilled, lymph nodes were not biopsied. During early years, EUS was performed by using a radial EUS, while linear EUS devices were only introduced later. Finally, patients who were just referred for surgery underwent staging externally, for cost reasons and limited availability of biopsy facilities, only patients with unclear lymph node findings underwent biopsies. In conclusion, members of multidisciplinary tumor boards must be aware of the limited accuracy of preoperative staging, i.e. T- and N-stages. Both, under- and overstaging are common; and patients are at potential risk to be under- and/or overtreated. PET-CT, MRI and EUS-guided FNA should probably be more generously used to improve initial tumor staging. Notes Specific author contributions: Winiker M and Schafer M contributed equally to this work; Winiker M and Schafer M collected, and analyzed the data, and drafted the manuscript; Allemann P and Mantziari S provided analytical oversight; Schafer M designed and supervised the study; Allemann P, Mantziari S, Figueiredo AG, and Demartines N revised the manuscript for important intellectual content; all authors have read and approved the final version to be published. Supported by: no support. Institutional review board statement: The study was reviewed and approved by the Institutional Review Board of the Department of Surgery, University Hospital of Lausanne. Biostatistics statement: The statistical methods of this study were reviewed by Professor Markus Schäfer, Department of Visceral Surgery, University Hospital CHUV, CH-1011 Lausanne, Switzerland. Informed consent statement: Preoperatively, all patients provided informed written consent to be included into the institutional data base. Since 2014, esophageal cancer treatment has been centralized by a federal law; and according to this new law, all patients are obligated to provide their data for quality control and research. Conflicts of interest: The authors declare that they have no conflict of interest. References 1 Pennathur A, Gibson M K, Jobe B A, Luketich J D. Oesophageal cancer. Lancet  2013; 381: 400– 12. Google Scholar CrossRef Search ADS PubMed  2 Rustgi A K, El-Serag H B. Esophageal carcinoma. N Engl J Med  2014; 371: 2499– 509. Google Scholar CrossRef Search ADS PubMed  3 Joliat G R, Hahnloser D, Demartines N, Schafer M. Future development of gastrointestinal cancer incidence and mortality rates in Switzerland: a tumor registry and population-based projection up to 2030. Swiss Med Wkly  2015; 145: w14188 [PMID: 26376322 doi:10.4414/smw.2015.14188]. Google Scholar PubMed  4 Thrift A P. The epidemic of oesophageal carcinoma: where are we now? Cancer Epidemiol  2016; 41: 88– 95. Google Scholar CrossRef Search ADS PubMed  5 Jang R, Darling G, Wong R K. Multimodality approaches for the curative treatment of esophageal cancer. J Natl Compr Canc Netw  2015; 13: 229– 38. Google Scholar CrossRef Search ADS PubMed  6 Sarvanan K, Bapuraj J R, Sharma S C et al.   Computed tomography and ultrasonographic evaluation of metastatic cervical lymph nodes with surgicoclinicopathologic correlation. J Laryngol Otol  2002; 116: 194– 9. Google Scholar CrossRef Search ADS PubMed  7 Cho J W. The role of endosonography in the staging of gastrointestinal cancers. Clin Endosc  2015; 48: 297– 301. Google Scholar CrossRef Search ADS PubMed  8 Tirumani H, Rosenthal M H, Tirumani S H, Shinagare A B, Krajewski K M, Ramaiya N H. Esophageal carcinoma: current concepts in the role of imaging in staging and management. Can Assoc Radiol J  2015; 66: 130– 9. Google Scholar CrossRef Search ADS PubMed  9 Schmidt T, Lordick F, Hermann K, Ott K. Value of functional imaging by PET in esophageal cancer. J Natl Compr Canc Netw  2015; 13: 239– 47. Google Scholar CrossRef Search ADS PubMed  10 Blackshaw G, Lewis W G, Hopper A N et al.   Prospective comparison of endosonography, computed tomography, and histopathological stage of junctional oesophagogastric cancer. Clin Radiol  2008; 63: 1092– 8. Google Scholar CrossRef Search ADS PubMed  11 Lewis I. Carcinoma of the oesophagus. Radical resection with oesophagogastrostomy for a midthoracic growth by a right transpleural approach. Proc R Soc Med  1945; 38: 483– 4. 12 Holscher A H, Schneider P M, Gutschow C, Schroeder W. Laparoscopic ischemic conditioning of the stomach for esophageal replacement. Ann Surg  2007; 245: 241– 6. Google Scholar CrossRef Search ADS PubMed  13 Stahl M, Mariette C, Haustermans K, Cervantes A, Arnold D. Oesophageal cancer: ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol  2013; 24( Suppl 6): vi51– 6. Google Scholar CrossRef Search ADS PubMed  14 Findlay J M, Bradley K M, Maile E J et al.   Pragmatic staging of oesophageal cancer using decision theory involving selective endoscopic ultrasonography, PET, and laparoscopy. Br J Surg  2015; 102: 1488– 99. Google Scholar CrossRef Search ADS PubMed  15 Noh H M, Fishman E K, Forastiere A A, Bliss D F, Calhoun P S. CT of the esophagus: spectrum of disease with emphasis on esophageal carcinoma. Radiographics  1995; 15: 1113– 34. Google Scholar CrossRef Search ADS PubMed  16 Napier K J, Scheerer M, Misra S. Esophageal cancer: a review of epidemiology, pathogenesis, staging workup and treatment modalities. World J Gastrointest Oncol  2014; 6: 112– 20. Google Scholar CrossRef Search ADS PubMed  17 Puli S R, Reddy J B, Bechtold M L, Antillon D, Ibdah J A, Antillon M R. Staging accuracy of esophageal cancer by endoscopic ultrasound: a meta-analysis and systematic review. World J Gastroenterol  2008; 14: 1479– 90. Google Scholar CrossRef Search ADS PubMed  18 Thosani N, Singh H, Kapadia A et al.   Diagnostic accuracy of EUS in differentiating mucosal versus submucosal invasion of superficial esophageal cancers: a systematic review and meta-analysis. Gastrointest Endosc  2012; 75: 242– 53. Google Scholar CrossRef Search ADS PubMed  19 Bohle W, Kasper M, Zoller W G. Different accuracy of endosonographic tumor staging after neoadjuvant chemotherapy and chemoradiotherapy in esophageal cancer. Surg Endosc  2016; 30: 2922– 8. Google Scholar CrossRef Search ADS PubMed  © The Authors 2017. Published by Oxford University Press on behalf of International Society for Diseases of the Esophagus. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com

Journal

Diseases of the EsophagusOxford University Press

Published: Jan 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off