Journal of Surgical Oncology 2015;111:1016–1020
Accuracy of EUS and CT Imaging in Preoperative Gastric Cancer Staging MARK FAIRWEATHER, MD,1 KUNAL JAJOO, MD,2 NISHA SAINANI, MD,3 MONICA M. BERTAGNOLLI, MD,1 AND JIPING WANG, MD, PhD1* 1
Department of Surgery, Division of Surgical Oncology, Brigham and Women’s Hospital, Boston, Massachusetts 2 Department of Gastroenterology, Brigham and Women’s Hospital, Boston, Massachusetts 3 Department of Radiology, Brigham and Women’s Hospital, Boston, Massachusetts
Background: Neoadjuvant therapy is recommended for locally advanced gastric cancer patients (stage IB–IIIC). The objective of this study is to evaluate the accuracy of endoscopic ultrasound (EUS) and computed tomography (CT) in identifying patients with locally advanced gastric cancer. Methods: Patients with gastric adenocarcinoma who underwent gastrectomy at our institution were reviewed. Preoperative EUS and CT staging were compared with pathologic staging to determine accuracy. Results: Between 2001 and 2013, 280 patients were identified. Forty-nine patients (18%) who underwent preoperative staging by EUS were analyzed. The accuracy of EUS in identifying individual T stage and N stage was 41.0% and 42.9%, respectively. EUS had moderate accuracy in differentiating locally advanced from early (stage 0–IA) disease (75.5%, area under curve (AUC) 0.772). For individual T stage and N stage, the accuracy of CT was 4.0% and 56%, respectively. CT had relatively poor accuracy in differentiating locally advanced from early disease (60.0%, AUC 0.649). Conclusions: EUS and CT had poor performance in identifying individual T and N stage although EUS demonstrated moderate accuracy in identifying patients with locally advanced disease. A combined staging approach, in addition to further optimization of staging modalities, is required for accurate staging for patients with gastric cancer.
J. Surg. Oncol. 2015;111:1016–1020. ß 2015 Wiley Periodicals, Inc.
KEY WORDS: endoscopic ultrasound; computed tomography; T stage; N stage; neoadjuvant
INTRODUCTION While the incidence of gastric cancer continues to decrease in the United States, approximately 21,000 new cases of gastric cancer are estimated to have been diagnosed in 2013 [1]. The 5-year survival following R0 resection remains poor (23–49%), especially for patients with locally advanced disease (14–55%) [2–4]. For patients with resectable gastric cancer, the Intergroup 0116 trial demonstrated that adjuvant chemoradiation therapy significantly improved 3-year overall survival (50% vs. 41%, P ¼ 0.005) and lowered local recurrence rates (19% vs. 29%) when compared to patients undergoing surgery alone [5]. Results from the MAGIC trial in 2006, and later from the FNCLCC/FFCD phase III trial in 2011, shifted the focus to the efficacy of neoadjuvant chemotherapy, showing improved disease-free survival and overall survival compared to surgery alone in a similar patient population [2,3]. Early results reported from the randomized, phase II COMPASS trial investigating the optimal length of neoadjuvant chemotherapy further support these results, as they found 10% of patients undergoing four cycles of neoadjuvant chemotherapy had a complete pathologic response, with limited chemotherapy-related toxicities in patients with stage IIA–IV disease [6]. As a result, the National Comprehensive Cancer Network (NCCN) recommends the consideration of neoadjuvant chemotherapy for patients with locally advanced disease (stage IB–IIIC) [7]. Previous studies investigating the accuracy of endoscopic ultrasound (EUS) and computed tomography (CT) imaging in preoperative gastric cancer staging that preceded the most recent 7th edition of the American Joint Committee on Cancer (AJCC) Cancer Staging Manual have demonstrated inconsistent results [8– 11]. In addition, there have been modifications of the T stage definitions including upstaging of subserosal involvement from T2b in the 6th edition to T3 in the newest edition and serosal
ß 2015 Wiley Periodicals, Inc.
involvement upstaging to T4 from T3 in the older edition. There are an increased number of permutations of TNM groupings with two additional stages (IIB and IIIC). These changes have resulted in increased specificity of the gastric wall layers that may lead to decreased accuracy of preoperative staging with EUS and CT imaging. The objective of this study was to evaluate the accuracy of EUS and CT imaging in identifying patients with locally advanced gastric cancer, utilizing the TNM definitions from the 7th edition AJCC Staging Manual.
PATIENTS AND METHODS Patients With Institutional Review Board approval, records of all patients with gastric adenocarcinoma undergoing preoperative assessment of TNM staging with EUS and CT imaging followed by gastrectomy with curative intent between 2001 to 2013 at the Brigham and Women’s Hospital were identified from an institutional gastric cancer database. Patients with tumors originating at the gastroesophageal junction or patients that received preoperative chemotherapy or radiation therapy were excluded.
* Correspondence to: Jiping Wang, MD, PhD, Department of Surgery, Division of Surgical Oncology, Brigham and Women’s Hospital, 75 Francis Street, Boston, MA. Fax: 617-582-6177. E-mail: [email protected] Received 27 january 2015; Accepted 19 March 2015 DOI 10.1002/jso.23919 Published online 14 April 2015 in Wiley Online Library (wileyonlinelibrary.com).
Preoperative Gastric Cancer Staging Pathologic Staging Surgical pathology reports were reviewed and TNM staging was classified according to the 7th edition of the AJCC Cancer Staging Manual [12]. A T1 tumor invades the lamina propria, muscularis mucosa, or the submucosa and T2 tumors invade the muscularis propria. T3 tumors penetrate the subserosal connective tissue and T4 tumors invade the serosa or adjacent structures. The presence of metastatic lymph nodes were staged N1 for 1–2 lymph nodes, N2 for 3–6 lymph nodes, and N3 for 7 or more lymph nodes.
which were correctly identified. Positive predicted value was defined as the proportion of locally advanced cases identified by EUS and CT imaging which were true pathological proven stage IB–IIIC diseases. Negative predicted value was defined as proportion of early disease cases identified by EUS and CT imaging which were true pathological proven stage 0–IA disease. Receiver operating characteristic (ROC) curve analysis was used to evaluate the overall concordance between EUS/CT imaging and pathologic staging. Statistical analyses were performed using Statistical Analysis System (SAS) software (Cary, NC).
Endoscopic Ultrasound Technique
RESULTS
EUS was performed with a 5 to 10 MHz radial echoendoscope (GFUE160-AL5, Olympus America, Center Valley, PA). EUS reports were reviewed by an experienced gastroenterologist for determination of TNM stages. Node positive disease (N1þ) was recorded if a lymph node was 8 mm in the short-axis diameter.
Patient and Tumor Characteristics
CT Image Analysis Preoperative CT images were retrospectively reviewed by an expert abdominal radiologist who was blinded to the EUS and surgical pathology staging. Only those patients’ CT scans with adequate stomach distention and image quality were staged by the radiologist. TNM staging was assigned using protocols previously described in the literature [13,14]. T1 tumors are those that have focal thickening of the inner layer with a visible outer layer and a clear fat plane around the tumor. T2 tumors exhibit focal or diffuse thickening of the gastric wall with transmural involvement and smooth outer border or have only a few small linear strands of soft tissue extending into the fat plane involving less than one third of the tumor. T3 tumors are transmural with blurring of at least one third of the tumor or have wide reticular strands surrounding the outer border of the tumor. T4 tumors have obliteration of the fat plane between tumor and adjacent organs or invasion of adjacent organs. N1 nodes are >8 mm in short-axis diameter and within 3 cm of the primary lesion and nodes >3 cm away are considered N2.
Data Analysis The preoperative T and N stages determined by EUS or CT were compared with the results from surgical pathology report. Accuracy was defined as the percentage of correctly identified pathologic T and N stages by EUS and/or CT imaging. Early stage cancer was defined as stage 0–IA and locally advanced gastric cancer was defined as patients with either T2–T4 tumor or positive lymph nodes (stage IB– IIIC). Sensitivity was defined as the proportion of locally advanced (stage IB–IIIC) cases which were correctly identified by EUS and CT imaging. Specificity was defined as the proportion of early stage cases
1017
Two-hundred and eighty patients with gastric adenocarcinoma underwent gastrectomy with curative intent, of which, 88 patients (31%) underwent preoperative EUS staging. Of the 88 patients who underwent preoperative EUS, 39 patients received preoperative chemotherapy and/or radiotherapy and were excluded, with the remaining 49 patients (18%) included in the analysis. Median age was 67 years (range 31–90 years) and the majority of patients were male (n ¼ 31, 63%). The most common tumor locations were the antrum (n ¼ 15, 31%) and cardia (n ¼ 10, 20%).
EUS and Surgical Pathologic Accuracy Determination of EUS staging accuracy for individual surgical pathology T stage is shown in Table I. The overall diagnostic accuracy of EUS for individual T staging was 41.0%. When differentiating between early and locally advanced disease, the accuracy improved to 75.5% (Table II). Ten patients (25.6%) with locally advanced disease were understaged (sensitivity¼74.4%) and 2 (20%) patients with early disease were overstaged by EUS (specificity¼80.0%). The positive and negative predictive values for locally advanced disease were 93.5% and 44.4%, respectively. The overall accuracy for the presence or absence of nodal disease for EUS was 42.9% (Table II). Twenty-five of 27 (93%) patients with node positive disease were understaged (sensitivity ¼ 7.4%) and 3 patients with N0 disease were overstaged as having node positive disease (specificity ¼ 86.4%). The positive and negative predictive values for nodal metastases were 40.0% and 43.2%, respectively. ROC analysis indicated that EUS had moderate accuracy in differentiating early versus locally advanced disease (area under curve (AUC) 0.772) (Fig. 1).
CT Imaging and Surgical Pathology Accuracy Twenty-five of the 49 patients who underwent EUS had good quality preoperative CT imaging which allowed T and N staging to be
TABLE I. Accuracy of Preoperative Endoscopic Ultrasound (EUS) (20/49 ¼ 41.0%) and Computed Tomography (CT) Imaging (1/25 ¼ 4.0%) in Predicting Individual Surgical Pathologic T Staging Pathologic staging EUS Stage (n ¼ 49) T0 (n ¼ 4) T1 (n ¼ 14) T2 (n ¼ 16) T3 (n ¼ 9) T4 (n ¼ 6) Total
Pathologic staging
T0
T1
T2
T3
T4
CT Stage (n ¼ 25)
T0
T1
T2
T3
T4
1 0 0 0 0 1
3 7 2 0 1 13
0 1 4 0 0 5
0 4 6 4 1 15
0 2 4 5 4 15
T0 (n ¼ 11) T1 (n ¼ 3 T2 (n ¼ 9) T3 (n ¼ 2) T4 (n ¼ 0) Total
0 0 1 0 0 1
5 0 2 0 0 7
1 0 0 0 0 1
4 0 5 1 0 10
1 3 1 1 0 6
*The concordance between EUS and pathology stages: sum of the diagonal (shaded)/total ¼ 20/49 ¼ 41.0%; The concordance between CT and pathology stages: 1/ 25 ¼ 4%
Journal of Surgical Oncology
1018
Fairweather et al.
TABLE II. Accuracy of Preoperative Endoscopic Ultrasound (EUS) in Predicting Early and Locally Advanced Disease (37/49 ¼ 77.5) and the Presence or Absence of Nodal Disease (21/49 ¼ 42.9%)
TABLE III. Accuracy of Preoperative Computed Tomography (CT) Imaging in Predicting Early (0–IA) and Locally Advanced (IB–IIIC) Disease (15/ 25 ¼ 60.0%) and the Presence or Absence of Nodal Disease (14/25 ¼ 56.0%)
Pathologic staging EUS Stage 0–IA IB–IIIC Total
N0 N1þ Total
0-IA
IB-IIIC
Pathologic Staging Total
8 10 2 29 10 39 Specificity ¼ 80.0% Sensitivity ¼ 74.4% N0
N1þ
19 3 22 Specificity ¼ 86.4%
25 2 27 Sensitivity ¼ 7.4%
18 31 49
CT Stage NPV ¼ 44.4% PPV ¼ 93.5%
0-IA IB–IIIC Total
Total 44 5 49
NPV¼43.2% PPV¼40.0%
NPV, negative predictive value; PPV, positive predictive value.
N0 N1þ Total
0–IA
IB–IIIC
Total
4 3 7 Specificity ¼ 57.1%
7 11 18 Sensitivity ¼ 61.1%
11 14 25
N0
N1þ
Total
9 3 12 Specificity ¼ 75.0%
8 5 13 Sensitivity ¼ 38.5%
17 8 25
NPV ¼ 36.4% PPV ¼ 78.6%
NPV ¼ 53.0% PPV ¼ 62.5%
NPV, negative predictive value; PPV, positive predictive value.
DISCUSSION
adequately evaluated. The overall accuracy of CT imaging in determination of individual T stage was 4.0% (Table I). In determining early and locally advanced disease, the accuracy improved to 60% (Table III). Seven of 18 (36.8%) patients with locally advanced disease were understaged (sensitivity ¼ 61.1%) and 3 patients with early disease were overstaged by CT imaging (specificity ¼ 57.1). The positive predictive value for locally advanced disease was 78.6% and the negative predictive value was 36.4%. The accuracy of CT imaging in identifying the presence or absence of nodal disease was 56.0% (Table III). Eight of 13 (61.5%) patients with surgical pathology confirmed node positive disease were understaged (sensitivity ¼ 38.5%) and 3 patients with N0 disease were overstaged as having node-positive disease (specificity ¼ 75.0%). The positive predictive value for node positive disease was 62.5% and the negative predictive value was 53%. ROC analysis indicated that CT imaging had relatively poor performance in differentiating early versus locally advanced disease (AUC 0.649) (Fig. 2).
Locally advanced gastric cancer is associated with a high recurrence rate and poor prognosis [15]. Thus, it is important to identify high risk patients preoperatively that may benefit from neoadjuvant therapy. EUS has evolved into a valuable tool in preoperative gastric cancer staging; however, the accuracy and reliability of this modality has been variable. Early studies reported EUS staging accuracies of 57–86% for individual T staging and 50–71.2% for N staging [9,10,16–19]. EUS has been shown to have variable prognostic capability. Smith et al. found that at a median of 13 months, only 1 of 13 patients (7.7%) with EUS stage T1 or T2 had recurred and 17 of 30 (57%) with EUS stage T3 or T4 had recurred (P ¼ 0.03), thus concluding that EUS staging could potentially predict the risk of recurrence and identify patients who may benefit from neoadjuvant therapy [9]. With regards to predicting survival, contrasting data have been reported. Jurgensen et al. reported preoperative EUS T staging and pathologic T staging equally predicted survival while more recently, Spolverato et al., concluded
Fig. 1. Receiver operating characteristic (ROC) curve for endoscopic ultrasound (EUS) in differentiating early versus locally advanced disease.
Fig. 2. Receiver operating characteristic (ROC) curve for computed tomography (CT) imaging in differentiating early versus locally advanced disease.
Journal of Surgical Oncology
Preoperative Gastric Cancer Staging that survival in patients with advanced preoperative EUS T staging was significantly overestimated when compared to surgical pathologic staging [20,21]. Multi-detector row CT imaging is an additional modality for determining preoperative TNM staging. In Asia, this is a common staging modality and radiologic protocols have been established to aid in analysis and optimization of radiographic images [13,14]. These protocols include having patients drink 400–800 cc of tap water to distend the stomach for improved visualization of the stomach wall layers. Accuracy rates in recent studies using multi-detector row CT for T and N staging have been reported to be 71.4–88.9% and 64–78%, respectively [22–25]. In the United States, preoperative CT imaging is primarily used to identify the presence of M1 disease and is not routinely used to categorize T and N stages. In this study, we explored the accuracy of both EUS and CT imaging in the preoperative gastric cancer staging utilizing the most recent AJCC 7th edition guidelines. The results demonstrate a lower overall accuracy of EUS than reported in the literature for individual T stage (47.1% vs. 57–86%) and N stage (42.9% vs. 50–71.2%). One of the largest studies to date included 223 patients from Memorial SloanKettering Cancer Center reported an accuracy of 57% for individual EUS T staging and 50% for identification of nodal disease [10]. Recent data from the US Gastric Cancer Collaborative that included 149 patients from 7 academic institutions demonstrated a comparably poor accuracy with our results for T staging (46.2%), but a more favorable accuracy for N staging (66.7%). The overall accuracy of CT scan is unacceptably low for individual T stage (4%); however, the accuracy for N stage (56%) is comparable with the Asian experience (71.4– 88.9% for T stage and 64–78% for N stage, respectively). Due to the innate operator variability in performing EUS, it could be postulated that high volume geographic areas may report improved accuracy rates; however, this has not been shown to be the case. A recent study from Asia that included 277 patients reported an overall accuracy of 74.7% for T staging and 62.8% for N staging [17]. An additional study from Asia, including 309 patients, reported an overall accuracy of 70.2% for T staging and 71.2% for N staging [19]. In looking at the effect of annual volume, a meta-analysis by Cardoso et al. found no association between EUS performance in T and N staging with regards to EUS annual volume (P ¼ 0.836, 0.99, respectively) [8]. Retrospective studies such as this are becoming increasingly difficult to conduct when patients undergoing neoadjuvant therapy are excluded to avoid overstaging. Our patient numbers in this study are certainly limited by this fact. Additionally, despite previous studies that have advocated the use of EUS, this has not been strongly adopted as only a small portion of patients with gastric cancer undergo preoperative EUS staging. In a study by Bentrem et al., only 40% of patients who underwent a gastrectomy underwent preoperative EUS staging [10]. The US Gastric Cancer Collaborative reported rates of EUS utilization of 13.3–45.2% for the 7 participating academic institutions included in their study [21]. At our institution, 31% of patients who underwent gastric resection were staged preoperatively by EUS. Our reported accuracies may be influenced by the recent changes made to the AJCC 7th edition staging system. T categories have been harmonized with those of the esophagus and small and large intestine, whereby T2 tumors invade the muscularis propria, T3 tumors invade the subserosal connective tissue, and T4 tumors invade the serosa or adjacent structures [12]. Our reported accuracy rate of 41.0% for individual T staging may reflect the increased difficulty of delineating the margin between the muscularis propria and the subserosal connective tissue endoscopically. Previously, using the AJCC 6th edition, this differentiation occurred between the muscularis propria or subserosa and the more easily identified serosa. Additionally, T4 tumors are likely to have a lower accuracy rate with the new staging system as they were previously defined as involvement of an adjacent organ, which is intuitively easier to identify, while the 7th edition includes Journal of Surgical Oncology
1019
tumor involving the serosa. Lastly, tumors arising at or within 5 cm of the gastroesophageal junction are staged according to the TNM system for esophageal cancer and were thus not included in this study [12]. While our overall accuracy rate of 41.0% is lower than reported in previous studies, it is important to note that the accuracy of the individual T stage is less important than distinguishing between early and locally advanced disease. By over or understaging T or N stages, patients are potentially being mis-managed by either receiving or not receiving neoadjuvant chemoradiotherapy. We report a low accuracy rate of 42.9% for identifying the presence of nodal disease with EUS. It is concerning that of the 27 patients with node positive disease identified by surgical pathology, 25 patients were reported as not having nodal disease. Potentially contributing to the low accuracy of detecting nodepositive disease, in this study as well as historically, is the presence of metastatic disease in normal-size lymph nodes. Nevertheless, EUS had moderate accuracy in differentiating between early and advanced cancer stages (75.5%). We did find that for T3/T4 tumors, EUS was more likely to understage tumors as T0–T2 (53.3%). This is in contrast to previously reported results in which EUS more commonly resulted in overstaging [19,21,26]. It should be noted, however, that this may be confounded by inclusion of patients who received neoadjuvant therapies. In summary, our results suggest that EUS and CT imaging are not accurate assessment tools for preoperative staging of individual T and N staging in patients with gastric adenocarcinoma. A combination of modalities is likely to improve accuracy in addition to distinguishing between early and locally advanced disease. Optimization of CT imaging protocols is required for improved accuracy of T and N staging. Future investigation whereby patients can be restaged following neoadjuvant therapy to increase the number of patients able to be included in studies such as this is required. Until further data is available, interpretation of EUS and CT imaging should be done with caution with the knowledge that these modalities tend to understage gastric tumors but have reliable accuracies in distinguishing early and locally advanced disease.
ACKNOWLEDGMENT The authors would like to thank Pebbles Ashley and Jacob Barron for their assistance in data collection. No further acknowledgments of assistance or funding.
REFERENCES 1. Cancer Facts & Figures 2013.http://www.cancer.org/research/ cancerfactsstatistics/cancerfactsfigures2013/index Assessed October 17, 2013. 2. Cunningham D,Allum WH, Stenning SP, et al.: Perioperative chemotherapy versus surgery alone for resectable gastroesophageal cancer. N Engl J Med 2006;355:11–20. 3. Ychou M, Boige V, Pignon JP, et al.: Perioperative chemotherapy compared with surgery alone for resectable gastroesophageal adenocarcinoma: An FNCLCC and FFCD multicenter phase III trial. J Clin Oncol 2011;29:1715–1721. 4. Karpeh MS, Leon L, Klimstra D, et al.: Lymph node staging in gastric cancer: Is location more important than number? An analysis of 1,038 patients. Ann Surg 2000;232:362–371. 5. Macdonald JS, Smalley SR, Benedetti J, et al.: Chemoradiotherapy after surgery compared with surgery alone for adenocarcinoma of the stomach or gastroesophageal junction. N Engl J Med 2001;345:725–730. 6. Yoshikawa T, Tanabe K, Nishikawa K, et al.: Induction of a pathological complete response by four courses of neoadjuvant chemotherapy for gastric cancer: Early results of the randomized phase II COMPASS trial. Ann Surg Oncol 2014;21:213–219. 7. Ajani JA, Barthel JS, Bekaii-Saab T, et al.: Gastric cancer. J Natl Compr Canc Netw 2010;8:378–409.
1020
Fairweather et al.
8. Cardoso R, Coburn N, Seevaratnam R, et al.: A systematic review and meta-analysis of the utility of EUS for preoperative staging for gastric cancer. Gastric Cancer 2012;15:S19–S26. 9. Smith JW, Brennan MF, Botet JF, et al.: Preoperative endoscopic ultrasound can predict the risk of recurrence after operation for gastric carcinoma. J Clin Oncol 1993;11:2380–2385. 10. Bentrem D, Gerdes H, Tang L, et al.: Clinical correlation of endoscopic ultrasonography with pathologic stage and outcome in patients undergoing curative resection for gastric cancer. Ann Surg Oncol 2007;14:1853–1859. 11. Puli SR, Reddy J, Bechtold ML, et al.: How good is endoscopic ultrasound for TNM staging of gastric cancers? A meta-analysis and systematic review. World J Gastroenterol 2008;14:4011–4019. 12. Edge SB, Compton CC, Fritz AG, et al.: AJCC Cancer Staging Manual. New York: Springer; 2010. 13. D’Elia F, Zingarelli A, Palli D, et al.: Hydro-dynamic CT preoperative staging of gastric cancer: correlation with pathological findings. A prospective study of 107 cases. Eur Radiol 2000;10:1877–1885. 14. Habermann CR, Weiss F, Riecken R, et al.: Preoperative staging of gastric adenocarcinoma: Comparison of helical CT and endoscopic US. Radiology 2004;230:465–471. 15. D’Angelica M, Gonen M, Brennan MF, et al.: Patterns of initial recurrence in completely resected gastric adenocarcinoma. Ann Surg 2004;240:808–816. 16. Tio TL, Coene PP, Schouwink MH, et al.: Esophagogastric carcinoma: Preoperative TNM classification with endosonography. Radiology 1989;173:411–417. 17. Hwang SW, Lee DH, Lee SH, et al.: Preoperative staging of gastric cancer by endoscopic ultrasonography and multidetectorrow computed tomography. J Gastroenterol Hepatol 2010;25: 512–518.
Journal of Surgical Oncology
18. Xi WD, Zhao C, Ren GS: Endoscopic ultrasonography in preoperative staging of gastric cancer: Determination of tumor invasion depth, nodal involvement and surgical resectability. World J Gastroenterol 2003;9:254–257. 19. Lee HH, Lim CH, Park JM, et al.: Low accuracy of endoscopic ultrasonography for detailed T staging in gastric cancer. World J Surg Oncol 2012;10:190. 20. Jurgensen C, Brand J, Nothnagel M, et al.: Prognostic relevance of gastric cancer staging by endoscopic ultrasound. Surg Endosc 2013;27:1124–1129. 21. Spolverato G, Ejaz A, Kim Y, et al.: Use of endoscopic ultrasound in the preoperative staging of gastric cancer: A multi-institutional study of the US Gastric Cancer Collaborative. J Am Coll Surg 2015;220:48–56. 22. Hasegawa S, Yoshikawa T, Shirai J, et al.: A prospective validation study to diagnose serosal invasion and nodal metastases of gastric cancer by multidetector-row CT. Ann Surg Oncol 2013;20: 2016–2022. 23. Kwee RM, Kwee TC: Imaging in local staging of gastric cancer: A systematic review. J Clin Oncol 2007;25:2107–2116. 24. Kim HJ, Kim AY, Oh ST, et al.: Gastric cancer staging at multi-detector row CT gastrography: Comparison of transverse and volumetric CT scanning. Radiology 2005;236: 879–885. 25. Chen CY, Hsu JS, Wu DC, et al.: Gastric cancer: Preoperative local staging with 3D multi-detector row CT–correlation with surgical and histopathologic results. Radiology 2007;242: 472–482. 26. Kutup A, Vashist YK, roth GS, et al.: Endoscopic ultrasound staging in gastric cancer: Does it help management decisions in the era of neoadjuvant treatment? Endoscopy 2012;44: 572–576.
Accuracy of EUS and CT Imaging in Preoperative Gastric Cancer Staging MARK FAIRWEATHER, MD,1 KUNAL JAJOO, MD,2 NISHA SAINANI, MD,3 MONICA M. BERTAGNOLLI, MD,1 AND JIPING WANG, MD, PhD1* 1
Department of Surgery, Division of Surgical Oncology, Brigham and Women’s Hospital, Boston, Massachusetts 2 Department of Gastroenterology, Brigham and Women’s Hospital, Boston, Massachusetts 3 Department of Radiology, Brigham and Women’s Hospital, Boston, Massachusetts
Background: Neoadjuvant therapy is recommended for locally advanced gastric cancer patients (stage IB–IIIC). The objective of this study is to evaluate the accuracy of endoscopic ultrasound (EUS) and computed tomography (CT) in identifying patients with locally advanced gastric cancer. Methods: Patients with gastric adenocarcinoma who underwent gastrectomy at our institution were reviewed. Preoperative EUS and CT staging were compared with pathologic staging to determine accuracy. Results: Between 2001 and 2013, 280 patients were identified. Forty-nine patients (18%) who underwent preoperative staging by EUS were analyzed. The accuracy of EUS in identifying individual T stage and N stage was 41.0% and 42.9%, respectively. EUS had moderate accuracy in differentiating locally advanced from early (stage 0–IA) disease (75.5%, area under curve (AUC) 0.772). For individual T stage and N stage, the accuracy of CT was 4.0% and 56%, respectively. CT had relatively poor accuracy in differentiating locally advanced from early disease (60.0%, AUC 0.649). Conclusions: EUS and CT had poor performance in identifying individual T and N stage although EUS demonstrated moderate accuracy in identifying patients with locally advanced disease. A combined staging approach, in addition to further optimization of staging modalities, is required for accurate staging for patients with gastric cancer.
J. Surg. Oncol. 2015;111:1016–1020. ß 2015 Wiley Periodicals, Inc.
KEY WORDS: endoscopic ultrasound; computed tomography; T stage; N stage; neoadjuvant
INTRODUCTION While the incidence of gastric cancer continues to decrease in the United States, approximately 21,000 new cases of gastric cancer are estimated to have been diagnosed in 2013 [1]. The 5-year survival following R0 resection remains poor (23–49%), especially for patients with locally advanced disease (14–55%) [2–4]. For patients with resectable gastric cancer, the Intergroup 0116 trial demonstrated that adjuvant chemoradiation therapy significantly improved 3-year overall survival (50% vs. 41%, P ¼ 0.005) and lowered local recurrence rates (19% vs. 29%) when compared to patients undergoing surgery alone [5]. Results from the MAGIC trial in 2006, and later from the FNCLCC/FFCD phase III trial in 2011, shifted the focus to the efficacy of neoadjuvant chemotherapy, showing improved disease-free survival and overall survival compared to surgery alone in a similar patient population [2,3]. Early results reported from the randomized, phase II COMPASS trial investigating the optimal length of neoadjuvant chemotherapy further support these results, as they found 10% of patients undergoing four cycles of neoadjuvant chemotherapy had a complete pathologic response, with limited chemotherapy-related toxicities in patients with stage IIA–IV disease [6]. As a result, the National Comprehensive Cancer Network (NCCN) recommends the consideration of neoadjuvant chemotherapy for patients with locally advanced disease (stage IB–IIIC) [7]. Previous studies investigating the accuracy of endoscopic ultrasound (EUS) and computed tomography (CT) imaging in preoperative gastric cancer staging that preceded the most recent 7th edition of the American Joint Committee on Cancer (AJCC) Cancer Staging Manual have demonstrated inconsistent results [8– 11]. In addition, there have been modifications of the T stage definitions including upstaging of subserosal involvement from T2b in the 6th edition to T3 in the newest edition and serosal
ß 2015 Wiley Periodicals, Inc.
involvement upstaging to T4 from T3 in the older edition. There are an increased number of permutations of TNM groupings with two additional stages (IIB and IIIC). These changes have resulted in increased specificity of the gastric wall layers that may lead to decreased accuracy of preoperative staging with EUS and CT imaging. The objective of this study was to evaluate the accuracy of EUS and CT imaging in identifying patients with locally advanced gastric cancer, utilizing the TNM definitions from the 7th edition AJCC Staging Manual.
PATIENTS AND METHODS Patients With Institutional Review Board approval, records of all patients with gastric adenocarcinoma undergoing preoperative assessment of TNM staging with EUS and CT imaging followed by gastrectomy with curative intent between 2001 to 2013 at the Brigham and Women’s Hospital were identified from an institutional gastric cancer database. Patients with tumors originating at the gastroesophageal junction or patients that received preoperative chemotherapy or radiation therapy were excluded.
* Correspondence to: Jiping Wang, MD, PhD, Department of Surgery, Division of Surgical Oncology, Brigham and Women’s Hospital, 75 Francis Street, Boston, MA. Fax: 617-582-6177. E-mail: [email protected] Received 27 january 2015; Accepted 19 March 2015 DOI 10.1002/jso.23919 Published online 14 April 2015 in Wiley Online Library (wileyonlinelibrary.com).
Preoperative Gastric Cancer Staging Pathologic Staging Surgical pathology reports were reviewed and TNM staging was classified according to the 7th edition of the AJCC Cancer Staging Manual [12]. A T1 tumor invades the lamina propria, muscularis mucosa, or the submucosa and T2 tumors invade the muscularis propria. T3 tumors penetrate the subserosal connective tissue and T4 tumors invade the serosa or adjacent structures. The presence of metastatic lymph nodes were staged N1 for 1–2 lymph nodes, N2 for 3–6 lymph nodes, and N3 for 7 or more lymph nodes.
which were correctly identified. Positive predicted value was defined as the proportion of locally advanced cases identified by EUS and CT imaging which were true pathological proven stage IB–IIIC diseases. Negative predicted value was defined as proportion of early disease cases identified by EUS and CT imaging which were true pathological proven stage 0–IA disease. Receiver operating characteristic (ROC) curve analysis was used to evaluate the overall concordance between EUS/CT imaging and pathologic staging. Statistical analyses were performed using Statistical Analysis System (SAS) software (Cary, NC).
Endoscopic Ultrasound Technique
RESULTS
EUS was performed with a 5 to 10 MHz radial echoendoscope (GFUE160-AL5, Olympus America, Center Valley, PA). EUS reports were reviewed by an experienced gastroenterologist for determination of TNM stages. Node positive disease (N1þ) was recorded if a lymph node was 8 mm in the short-axis diameter.
Patient and Tumor Characteristics
CT Image Analysis Preoperative CT images were retrospectively reviewed by an expert abdominal radiologist who was blinded to the EUS and surgical pathology staging. Only those patients’ CT scans with adequate stomach distention and image quality were staged by the radiologist. TNM staging was assigned using protocols previously described in the literature [13,14]. T1 tumors are those that have focal thickening of the inner layer with a visible outer layer and a clear fat plane around the tumor. T2 tumors exhibit focal or diffuse thickening of the gastric wall with transmural involvement and smooth outer border or have only a few small linear strands of soft tissue extending into the fat plane involving less than one third of the tumor. T3 tumors are transmural with blurring of at least one third of the tumor or have wide reticular strands surrounding the outer border of the tumor. T4 tumors have obliteration of the fat plane between tumor and adjacent organs or invasion of adjacent organs. N1 nodes are >8 mm in short-axis diameter and within 3 cm of the primary lesion and nodes >3 cm away are considered N2.
Data Analysis The preoperative T and N stages determined by EUS or CT were compared with the results from surgical pathology report. Accuracy was defined as the percentage of correctly identified pathologic T and N stages by EUS and/or CT imaging. Early stage cancer was defined as stage 0–IA and locally advanced gastric cancer was defined as patients with either T2–T4 tumor or positive lymph nodes (stage IB– IIIC). Sensitivity was defined as the proportion of locally advanced (stage IB–IIIC) cases which were correctly identified by EUS and CT imaging. Specificity was defined as the proportion of early stage cases
1017
Two-hundred and eighty patients with gastric adenocarcinoma underwent gastrectomy with curative intent, of which, 88 patients (31%) underwent preoperative EUS staging. Of the 88 patients who underwent preoperative EUS, 39 patients received preoperative chemotherapy and/or radiotherapy and were excluded, with the remaining 49 patients (18%) included in the analysis. Median age was 67 years (range 31–90 years) and the majority of patients were male (n ¼ 31, 63%). The most common tumor locations were the antrum (n ¼ 15, 31%) and cardia (n ¼ 10, 20%).
EUS and Surgical Pathologic Accuracy Determination of EUS staging accuracy for individual surgical pathology T stage is shown in Table I. The overall diagnostic accuracy of EUS for individual T staging was 41.0%. When differentiating between early and locally advanced disease, the accuracy improved to 75.5% (Table II). Ten patients (25.6%) with locally advanced disease were understaged (sensitivity¼74.4%) and 2 (20%) patients with early disease were overstaged by EUS (specificity¼80.0%). The positive and negative predictive values for locally advanced disease were 93.5% and 44.4%, respectively. The overall accuracy for the presence or absence of nodal disease for EUS was 42.9% (Table II). Twenty-five of 27 (93%) patients with node positive disease were understaged (sensitivity ¼ 7.4%) and 3 patients with N0 disease were overstaged as having node positive disease (specificity ¼ 86.4%). The positive and negative predictive values for nodal metastases were 40.0% and 43.2%, respectively. ROC analysis indicated that EUS had moderate accuracy in differentiating early versus locally advanced disease (area under curve (AUC) 0.772) (Fig. 1).
CT Imaging and Surgical Pathology Accuracy Twenty-five of the 49 patients who underwent EUS had good quality preoperative CT imaging which allowed T and N staging to be
TABLE I. Accuracy of Preoperative Endoscopic Ultrasound (EUS) (20/49 ¼ 41.0%) and Computed Tomography (CT) Imaging (1/25 ¼ 4.0%) in Predicting Individual Surgical Pathologic T Staging Pathologic staging EUS Stage (n ¼ 49) T0 (n ¼ 4) T1 (n ¼ 14) T2 (n ¼ 16) T3 (n ¼ 9) T4 (n ¼ 6) Total
Pathologic staging
T0
T1
T2
T3
T4
CT Stage (n ¼ 25)
T0
T1
T2
T3
T4
1 0 0 0 0 1
3 7 2 0 1 13
0 1 4 0 0 5
0 4 6 4 1 15
0 2 4 5 4 15
T0 (n ¼ 11) T1 (n ¼ 3 T2 (n ¼ 9) T3 (n ¼ 2) T4 (n ¼ 0) Total
0 0 1 0 0 1
5 0 2 0 0 7
1 0 0 0 0 1
4 0 5 1 0 10
1 3 1 1 0 6
*The concordance between EUS and pathology stages: sum of the diagonal (shaded)/total ¼ 20/49 ¼ 41.0%; The concordance between CT and pathology stages: 1/ 25 ¼ 4%
Journal of Surgical Oncology
1018
Fairweather et al.
TABLE II. Accuracy of Preoperative Endoscopic Ultrasound (EUS) in Predicting Early and Locally Advanced Disease (37/49 ¼ 77.5) and the Presence or Absence of Nodal Disease (21/49 ¼ 42.9%)
TABLE III. Accuracy of Preoperative Computed Tomography (CT) Imaging in Predicting Early (0–IA) and Locally Advanced (IB–IIIC) Disease (15/ 25 ¼ 60.0%) and the Presence or Absence of Nodal Disease (14/25 ¼ 56.0%)
Pathologic staging EUS Stage 0–IA IB–IIIC Total
N0 N1þ Total
0-IA
IB-IIIC
Pathologic Staging Total
8 10 2 29 10 39 Specificity ¼ 80.0% Sensitivity ¼ 74.4% N0
N1þ
19 3 22 Specificity ¼ 86.4%
25 2 27 Sensitivity ¼ 7.4%
18 31 49
CT Stage NPV ¼ 44.4% PPV ¼ 93.5%
0-IA IB–IIIC Total
Total 44 5 49
NPV¼43.2% PPV¼40.0%
NPV, negative predictive value; PPV, positive predictive value.
N0 N1þ Total
0–IA
IB–IIIC
Total
4 3 7 Specificity ¼ 57.1%
7 11 18 Sensitivity ¼ 61.1%
11 14 25
N0
N1þ
Total
9 3 12 Specificity ¼ 75.0%
8 5 13 Sensitivity ¼ 38.5%
17 8 25
NPV ¼ 36.4% PPV ¼ 78.6%
NPV ¼ 53.0% PPV ¼ 62.5%
NPV, negative predictive value; PPV, positive predictive value.
DISCUSSION
adequately evaluated. The overall accuracy of CT imaging in determination of individual T stage was 4.0% (Table I). In determining early and locally advanced disease, the accuracy improved to 60% (Table III). Seven of 18 (36.8%) patients with locally advanced disease were understaged (sensitivity ¼ 61.1%) and 3 patients with early disease were overstaged by CT imaging (specificity ¼ 57.1). The positive predictive value for locally advanced disease was 78.6% and the negative predictive value was 36.4%. The accuracy of CT imaging in identifying the presence or absence of nodal disease was 56.0% (Table III). Eight of 13 (61.5%) patients with surgical pathology confirmed node positive disease were understaged (sensitivity ¼ 38.5%) and 3 patients with N0 disease were overstaged as having node-positive disease (specificity ¼ 75.0%). The positive predictive value for node positive disease was 62.5% and the negative predictive value was 53%. ROC analysis indicated that CT imaging had relatively poor performance in differentiating early versus locally advanced disease (AUC 0.649) (Fig. 2).
Locally advanced gastric cancer is associated with a high recurrence rate and poor prognosis [15]. Thus, it is important to identify high risk patients preoperatively that may benefit from neoadjuvant therapy. EUS has evolved into a valuable tool in preoperative gastric cancer staging; however, the accuracy and reliability of this modality has been variable. Early studies reported EUS staging accuracies of 57–86% for individual T staging and 50–71.2% for N staging [9,10,16–19]. EUS has been shown to have variable prognostic capability. Smith et al. found that at a median of 13 months, only 1 of 13 patients (7.7%) with EUS stage T1 or T2 had recurred and 17 of 30 (57%) with EUS stage T3 or T4 had recurred (P ¼ 0.03), thus concluding that EUS staging could potentially predict the risk of recurrence and identify patients who may benefit from neoadjuvant therapy [9]. With regards to predicting survival, contrasting data have been reported. Jurgensen et al. reported preoperative EUS T staging and pathologic T staging equally predicted survival while more recently, Spolverato et al., concluded
Fig. 1. Receiver operating characteristic (ROC) curve for endoscopic ultrasound (EUS) in differentiating early versus locally advanced disease.
Fig. 2. Receiver operating characteristic (ROC) curve for computed tomography (CT) imaging in differentiating early versus locally advanced disease.
Journal of Surgical Oncology
Preoperative Gastric Cancer Staging that survival in patients with advanced preoperative EUS T staging was significantly overestimated when compared to surgical pathologic staging [20,21]. Multi-detector row CT imaging is an additional modality for determining preoperative TNM staging. In Asia, this is a common staging modality and radiologic protocols have been established to aid in analysis and optimization of radiographic images [13,14]. These protocols include having patients drink 400–800 cc of tap water to distend the stomach for improved visualization of the stomach wall layers. Accuracy rates in recent studies using multi-detector row CT for T and N staging have been reported to be 71.4–88.9% and 64–78%, respectively [22–25]. In the United States, preoperative CT imaging is primarily used to identify the presence of M1 disease and is not routinely used to categorize T and N stages. In this study, we explored the accuracy of both EUS and CT imaging in the preoperative gastric cancer staging utilizing the most recent AJCC 7th edition guidelines. The results demonstrate a lower overall accuracy of EUS than reported in the literature for individual T stage (47.1% vs. 57–86%) and N stage (42.9% vs. 50–71.2%). One of the largest studies to date included 223 patients from Memorial SloanKettering Cancer Center reported an accuracy of 57% for individual EUS T staging and 50% for identification of nodal disease [10]. Recent data from the US Gastric Cancer Collaborative that included 149 patients from 7 academic institutions demonstrated a comparably poor accuracy with our results for T staging (46.2%), but a more favorable accuracy for N staging (66.7%). The overall accuracy of CT scan is unacceptably low for individual T stage (4%); however, the accuracy for N stage (56%) is comparable with the Asian experience (71.4– 88.9% for T stage and 64–78% for N stage, respectively). Due to the innate operator variability in performing EUS, it could be postulated that high volume geographic areas may report improved accuracy rates; however, this has not been shown to be the case. A recent study from Asia that included 277 patients reported an overall accuracy of 74.7% for T staging and 62.8% for N staging [17]. An additional study from Asia, including 309 patients, reported an overall accuracy of 70.2% for T staging and 71.2% for N staging [19]. In looking at the effect of annual volume, a meta-analysis by Cardoso et al. found no association between EUS performance in T and N staging with regards to EUS annual volume (P ¼ 0.836, 0.99, respectively) [8]. Retrospective studies such as this are becoming increasingly difficult to conduct when patients undergoing neoadjuvant therapy are excluded to avoid overstaging. Our patient numbers in this study are certainly limited by this fact. Additionally, despite previous studies that have advocated the use of EUS, this has not been strongly adopted as only a small portion of patients with gastric cancer undergo preoperative EUS staging. In a study by Bentrem et al., only 40% of patients who underwent a gastrectomy underwent preoperative EUS staging [10]. The US Gastric Cancer Collaborative reported rates of EUS utilization of 13.3–45.2% for the 7 participating academic institutions included in their study [21]. At our institution, 31% of patients who underwent gastric resection were staged preoperatively by EUS. Our reported accuracies may be influenced by the recent changes made to the AJCC 7th edition staging system. T categories have been harmonized with those of the esophagus and small and large intestine, whereby T2 tumors invade the muscularis propria, T3 tumors invade the subserosal connective tissue, and T4 tumors invade the serosa or adjacent structures [12]. Our reported accuracy rate of 41.0% for individual T staging may reflect the increased difficulty of delineating the margin between the muscularis propria and the subserosal connective tissue endoscopically. Previously, using the AJCC 6th edition, this differentiation occurred between the muscularis propria or subserosa and the more easily identified serosa. Additionally, T4 tumors are likely to have a lower accuracy rate with the new staging system as they were previously defined as involvement of an adjacent organ, which is intuitively easier to identify, while the 7th edition includes Journal of Surgical Oncology
1019
tumor involving the serosa. Lastly, tumors arising at or within 5 cm of the gastroesophageal junction are staged according to the TNM system for esophageal cancer and were thus not included in this study [12]. While our overall accuracy rate of 41.0% is lower than reported in previous studies, it is important to note that the accuracy of the individual T stage is less important than distinguishing between early and locally advanced disease. By over or understaging T or N stages, patients are potentially being mis-managed by either receiving or not receiving neoadjuvant chemoradiotherapy. We report a low accuracy rate of 42.9% for identifying the presence of nodal disease with EUS. It is concerning that of the 27 patients with node positive disease identified by surgical pathology, 25 patients were reported as not having nodal disease. Potentially contributing to the low accuracy of detecting nodepositive disease, in this study as well as historically, is the presence of metastatic disease in normal-size lymph nodes. Nevertheless, EUS had moderate accuracy in differentiating between early and advanced cancer stages (75.5%). We did find that for T3/T4 tumors, EUS was more likely to understage tumors as T0–T2 (53.3%). This is in contrast to previously reported results in which EUS more commonly resulted in overstaging [19,21,26]. It should be noted, however, that this may be confounded by inclusion of patients who received neoadjuvant therapies. In summary, our results suggest that EUS and CT imaging are not accurate assessment tools for preoperative staging of individual T and N staging in patients with gastric adenocarcinoma. A combination of modalities is likely to improve accuracy in addition to distinguishing between early and locally advanced disease. Optimization of CT imaging protocols is required for improved accuracy of T and N staging. Future investigation whereby patients can be restaged following neoadjuvant therapy to increase the number of patients able to be included in studies such as this is required. Until further data is available, interpretation of EUS and CT imaging should be done with caution with the knowledge that these modalities tend to understage gastric tumors but have reliable accuracies in distinguishing early and locally advanced disease.
ACKNOWLEDGMENT The authors would like to thank Pebbles Ashley and Jacob Barron for their assistance in data collection. No further acknowledgments of assistance or funding.
REFERENCES 1. Cancer Facts & Figures 2013.http://www.cancer.org/research/ cancerfactsstatistics/cancerfactsfigures2013/index Assessed October 17, 2013. 2. Cunningham D,Allum WH, Stenning SP, et al.: Perioperative chemotherapy versus surgery alone for resectable gastroesophageal cancer. N Engl J Med 2006;355:11–20. 3. Ychou M, Boige V, Pignon JP, et al.: Perioperative chemotherapy compared with surgery alone for resectable gastroesophageal adenocarcinoma: An FNCLCC and FFCD multicenter phase III trial. J Clin Oncol 2011;29:1715–1721. 4. Karpeh MS, Leon L, Klimstra D, et al.: Lymph node staging in gastric cancer: Is location more important than number? An analysis of 1,038 patients. Ann Surg 2000;232:362–371. 5. Macdonald JS, Smalley SR, Benedetti J, et al.: Chemoradiotherapy after surgery compared with surgery alone for adenocarcinoma of the stomach or gastroesophageal junction. N Engl J Med 2001;345:725–730. 6. Yoshikawa T, Tanabe K, Nishikawa K, et al.: Induction of a pathological complete response by four courses of neoadjuvant chemotherapy for gastric cancer: Early results of the randomized phase II COMPASS trial. Ann Surg Oncol 2014;21:213–219. 7. Ajani JA, Barthel JS, Bekaii-Saab T, et al.: Gastric cancer. J Natl Compr Canc Netw 2010;8:378–409.
1020
Fairweather et al.
8. Cardoso R, Coburn N, Seevaratnam R, et al.: A systematic review and meta-analysis of the utility of EUS for preoperative staging for gastric cancer. Gastric Cancer 2012;15:S19–S26. 9. Smith JW, Brennan MF, Botet JF, et al.: Preoperative endoscopic ultrasound can predict the risk of recurrence after operation for gastric carcinoma. J Clin Oncol 1993;11:2380–2385. 10. Bentrem D, Gerdes H, Tang L, et al.: Clinical correlation of endoscopic ultrasonography with pathologic stage and outcome in patients undergoing curative resection for gastric cancer. Ann Surg Oncol 2007;14:1853–1859. 11. Puli SR, Reddy J, Bechtold ML, et al.: How good is endoscopic ultrasound for TNM staging of gastric cancers? A meta-analysis and systematic review. World J Gastroenterol 2008;14:4011–4019. 12. Edge SB, Compton CC, Fritz AG, et al.: AJCC Cancer Staging Manual. New York: Springer; 2010. 13. D’Elia F, Zingarelli A, Palli D, et al.: Hydro-dynamic CT preoperative staging of gastric cancer: correlation with pathological findings. A prospective study of 107 cases. Eur Radiol 2000;10:1877–1885. 14. Habermann CR, Weiss F, Riecken R, et al.: Preoperative staging of gastric adenocarcinoma: Comparison of helical CT and endoscopic US. Radiology 2004;230:465–471. 15. D’Angelica M, Gonen M, Brennan MF, et al.: Patterns of initial recurrence in completely resected gastric adenocarcinoma. Ann Surg 2004;240:808–816. 16. Tio TL, Coene PP, Schouwink MH, et al.: Esophagogastric carcinoma: Preoperative TNM classification with endosonography. Radiology 1989;173:411–417. 17. Hwang SW, Lee DH, Lee SH, et al.: Preoperative staging of gastric cancer by endoscopic ultrasonography and multidetectorrow computed tomography. J Gastroenterol Hepatol 2010;25: 512–518.
Journal of Surgical Oncology
18. Xi WD, Zhao C, Ren GS: Endoscopic ultrasonography in preoperative staging of gastric cancer: Determination of tumor invasion depth, nodal involvement and surgical resectability. World J Gastroenterol 2003;9:254–257. 19. Lee HH, Lim CH, Park JM, et al.: Low accuracy of endoscopic ultrasonography for detailed T staging in gastric cancer. World J Surg Oncol 2012;10:190. 20. Jurgensen C, Brand J, Nothnagel M, et al.: Prognostic relevance of gastric cancer staging by endoscopic ultrasound. Surg Endosc 2013;27:1124–1129. 21. Spolverato G, Ejaz A, Kim Y, et al.: Use of endoscopic ultrasound in the preoperative staging of gastric cancer: A multi-institutional study of the US Gastric Cancer Collaborative. J Am Coll Surg 2015;220:48–56. 22. Hasegawa S, Yoshikawa T, Shirai J, et al.: A prospective validation study to diagnose serosal invasion and nodal metastases of gastric cancer by multidetector-row CT. Ann Surg Oncol 2013;20: 2016–2022. 23. Kwee RM, Kwee TC: Imaging in local staging of gastric cancer: A systematic review. J Clin Oncol 2007;25:2107–2116. 24. Kim HJ, Kim AY, Oh ST, et al.: Gastric cancer staging at multi-detector row CT gastrography: Comparison of transverse and volumetric CT scanning. Radiology 2005;236: 879–885. 25. Chen CY, Hsu JS, Wu DC, et al.: Gastric cancer: Preoperative local staging with 3D multi-detector row CT–correlation with surgical and histopathologic results. Radiology 2007;242: 472–482. 26. Kutup A, Vashist YK, roth GS, et al.: Endoscopic ultrasound staging in gastric cancer: Does it help management decisions in the era of neoadjuvant treatment? Endoscopy 2012;44: 572–576.
