HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Personal Folders Download to citation manager Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Nishimaki, T. Articles by Iizuka, T. Search for Related Content PubMed PubMed Citation Articles by Nishimaki, T. Articles by Iizuka, T.

Ann Thorac Surg 1999;68:2059-2064
© 1999 The Society of Thoracic Surgeons

---

Original Articles: General Thoracic

Evaluation of the accuracy of preoperative staging in thoracic esophageal cancer

^a First Department of Surgery, School of Medicine, Niigata University, Niigata, Tokyo, Japan
^b Department of Surgery, Niigata Cancer Center Hospital, Niigata, Japan
^c Department of Surgery, School of Medicine, Keio University, Tokyo, Japan
^d Department of Surgery, Tokyo Women’s Medical College, Tokyo, Japan
^e Department of Surgery, National Cancer Center Hospital, Tokyo, Japan
^f Department of Thoracic Surgery, Aichi Cancer Center Hospital, Nagoya, Japan
^g Department of Surgery, National Shikoku Cancer Center Hospital, Matsuyama, Japan
^h First Department of Surgery, Faculty of Medicine, Kurume University, Kurume, Japan
ⁱ First Department of Surgery, Iwate Medical College, Morioka, Japan
^j Second Department of Surgery, Faculty of Medicine, Chiba University, Chiba, Japan
^k First Department of Surgery, School of Medicine, Tokyo Medical and Dental University, Tokyo, Japan
^l Department of Surgery, National Tokyo Second Hospital, Tokyo, Japan
^m Department of Surgery, School of Medicine, Tokai University, Isehara, Japan
ⁿ Department of Surgery, Kanagawa Cancer Center Hospital, Yokohama, Japan
^o First Department of Surgery, Faculty of Medicine, Kyoto University, Kyoto, Japan
^p National Oji Hospital, Tokyo, Japan

Address reprint requests to Dr Nishimaki, First Department of Surgery, Niigata University School of Medicine, Asahimachi-dori 1-757, Niigata 951, Japan

Abstract

Background. Exact clinical staging before treatment of esophageal cancer has become increasingly important in the evaluation and comparison of the results of different treatment modalities, including surgery, chemotherapy, and radiotherapy.

Methods. The accuracy of preoperative tumor staging by using an esophagography, esophagoscopy, percutaneous and endoscopic ultrasonography, and computed tomography was assessed in 224 patients with resectable esophageal cancer. The results of tumor staging by these tests were compared prospectively with the pathologic stage of the esophagectomy specimens with respect to the T and N categories defined by the International Union Against Cancer TNM classification.

Results. For the T category, the overall accuracy was 80%. For the N category, overall accuracy was 72%, with a sensitivity of 78%, a specificity of 60%, and a positive predictive value of 78%. Overall, the accuracy of stage grouping was 56%.

Conclusions. Either the T or N categories can be predicted reliably by clinical staging techniques. However, the preoperative stage grouping might not be valid in resectable, localized esophageal cancer.

The survival of patients with esophageal cancer correlates well with the stage of the disease at the time of the initiation of treatment [1]. Because the depth of primary tumor invasion and the presence or absence of lymph node and distant organ metastases are the major prognostic factors in esophageal cancer, the TNM classifications of the International Union Against Cancer (UICC) [2] and the American Joint Committee on Cancer (AJCC) [3] have been used widely to stage esophageal cancer. Recently, preoperative chemotherapy or radiotherapy have been combined with esophagectomy to further improve the long-term results of treatment. Primary chemoradiotherapy has been attempted as an alternative treatment modality, with the potential advantage of preserving esophageal function [4, 5]. Accordingly, exact clinical staging has beome increasingly important in evaluating and comparing the results of different treatment modalities for esophageal cancer, because the true extent of tumor involvement generally cannot be assessed by histologic examination after chemotherapy and radiotherapy.

Several studies assessing the accuracy of imaging techniques in staging esophageal cancer have been reported, mainly from the medical or radiologic viewpoint [6–8]. However, many of those studies did not use radical lymphadenectomy as the histologic basis for comparison, leading to a possible underestimation of the degree of lymph node metastases, because esophageal cancer frequently metastasizes to lymph nodes considerably distant from the primary site, particularly at the early stage of lymphatic spread [9].

The purpose of the present study was to assess prospectively the accuracy of preoperative clinical staging of tumor involvement in patients with localized esophageal cancer who had radical esophagectomy.

Material and methods

Patients
The present study was conducted from the beginning of September 1993 until the end of August 1994. It was a prospective, multi-institutional study performed at 16 institutions belonging to the Japan Esophageal Oncology Group (a subgroup of the Japan Clinical Oncology Group) to assess the accuracy of preoperative staging of resectable esophageal cancer. Radical esophagectomy was planned for the 224 patients with invasive, localized esophageal cancer who were included in this study. All 224 patients were considered to be acceptable candidates for radical esophagectomy, with routine functional assessments of the vital organs. The age of these patients ranged from 35 to 83 years (mean, 62 years). There were 198 men and 26 women. The 16 institutions were requested to submit a preoperative report describing the results of clinical staging, based on the UICC-TNM classification [2], and postoperatively to submit an additional report describing the results of the pathologic staging. At the end of the study, the preoperative staging was compared with the pathologic stage of the esophagectomy specimens. Preoperative staging was done within 3 weeks of esophagectomy.

Tumor characteristics
Of the 224 tumors, 29 were located in the upper, 129 in the middle, and 66 in the lower thoracic esophagus. The tumors ranged from 0.5 cm to 14 cm, with a mean length of 5.4 cm. Histologically, 216 tumors were squamous cell carcinoma. One was adenocarcinoma, two were undifferentiated carcinoma, and five were carcinomas of other histologic types.

Esophagectomy
Transthoracic esophagectomy was done in 220 patients, of whom 85 simultaneously had systematic mediastinal and abdominal lymphadenectomies, ie, two-field lymphadenectomy. The remaining 135 patients had extended radical lymphadenectomy involving bilateral cervical, mediastinal, and abdominal lymph nodes, ie, three-field lymphadenectomy. Transhiatal esophagectomy was done in 4 patients with early esophageal cancer.

In the 85 patients who had two-field lymphadenectomy, the bilateral thoracic paratracheal, bifurcation, bilateral pulmonary hilar, diaphragmatic, posterior mediastinal, and thoracic paraesophageal lymph nodes were removed as the mediastinal lymph nodes. The perigastric lymph nodes including the right cardiac, left cardiac, lesser curvature, greater curvature, and left gastric artery lymph nodes, common hepatic artery, celiac artery, and splenic artery lymph nodes were removed as the abdominal lymph nodes. The nodes were defined according to the nomenclature of the Japanese Society for Esophageal Diseases [10]. In addition to these lymph nodes, the bilateral deep cervical, supraclavicular, and cervical paraesophageal lymph nodes were removed as the cervical lymph nodes in the 135 patients who had three-field lymphadenectomy.

Of the 29 patients with upper thoracic esophageal tumor, 25 and 4 had three- and two-field lymphadenectomy, respectively. Of the 129 patients with midthoracic esophageal tumors, 127 had transthoracic esophagectomy: 75 had three-field lymphadenectomy and 52 had two-field lymphadenectomy. Of the 66 patients with lower thoracic esophageal tumor, 64 had transthoracic esophagectomy: 35 had three-field lymphadenectomy and 29 had two-field lymphadenectomy.

Assignment of T categories
The depth of the primary tumor, in terms of the T categories defined by the UICC-TNM classification [2], was assessed preoperatively in all 224 patients. An esophagography was done in 222, esophagoscopy in 224, endoscopic ultrasonography (EUS) in 176, and computed tomography (CT) in 218 patients. Endoscopic ultrasonography was done with an echoendoscope of either 7.5 or 12 MHz frequency.

On esophagography and esophagoscopy, tumors with a superficial appearance were diagnosed as T1 (tumors invading the lamina propria or submucosa). Tumors with a protruded or shallow ulcerated appearance were diagnosed as T2 (tumors invading the muscularis propria). Tumors with deep ulceration or a circular stricture were diagnosed as T3 (tumors invading the adventitia), and tumors with fistula formation or esophageal axis deviation were diagnosed as T4 (tumors invading adjacent structures). The normal esophageal wall was usually depicted as five layers by EUS. The outermost layer invaded by tumor was defined as the depth of tumor invasion in the EUS studies, as previously described [7, 8, 11]. On CT examinations, thickening of the esophageal wall was diagnosed as T3, and direct involvement of adjacent organs as T4. In addition, individual results of different diagnostic modalities were integrated into the final diagnosis, which was described as a consensus evaluation in the preoperative report.

Assignment of N categories
The 4 patients who had a transhiatal esophagectomy were excluded from the evaluation of nodal status, because the extent of the lymphadenectomy, particularly of the mediastinal lymph nodes, was considered to be too small to accurately assess the nodal status. Consequently, the nodal status, ie, the N categories defined by the UICC-TNM classification [2], was assessed in the remaining 220 patients, of whom ultrasonographic examination, including percutaneous and endoscopic ultrasonography was done in 201 and CT was done in 217. Lymph nodes located in the lower neck, mediastinum, and upper abdomen were assessed by these imaging techniques. Of the 201 patients who had ultrasonographic examinations, 172 had EUS. Of these patients, percutaneous cervical and abdominal ultrasonography was also done in 168 and 167, respectively. Tumors were located in the upper, mid, and lower thoracic esophagus in 17, 102, and 53, respectively. Three-field and two-field lymphadenectomy was done in 104 and 68, respectively. Ten-millimeter-thick contiguous sections were used in the CT examinations.

Lymph nodes more than 5-mm in short-axis diameter with a clear, hypoechoic, round shape, and lymph nodes more than 10 mm in diameter were defined as node metastases on ultrasonographic and CT examinations, respectively [7, 8, 12]. The consensus evaluation, which integrated the individual results of the different imaging techniques, was also reported preoperatively in the assessment of N categories.

The sites of lymph node metastasis were also determined preoperatively. In the present study, lymph nodes were classified into the following seven groups by anatomic location, to facilitate the localization of lymph node metastases by the imaging techniques: cervical paraesophageal, deep cervical, and supraclavicular lymph nodes were grouped as the cervical nodes, which were further divided into the right and left cervical nodes; the upper thoracic paraesophageal, and thoracic paratracheal lymph nodes as the upper mediastinal nodes; the bifurcation, middle thoracic paraesophageal, and pulmonary hilar lymph nodes as the middle mediastinal nodes; the lower thoracic paraesophageal, diaphragmatic, and posterior mediastinal lymph nodes as the lower mediastinal nodes; the right cardiac, left cardiac, lesser curvature, greater curvature, and left gastric artery nodes as the perigastric nodes; and the common hepatic artery, celiac artery, and splenic artery lymph nodes as the distant abdominal nodes.

Furthermore, to investigate the spatial relationships between the extent of lymph node metastases assessed clinically and that verified pathologically, lymph node metastases were classified into the following five patterns: concordance, indicating the cases in which all of the metastatic sites were predicted correctly at clinical staging; overestimation, indicating the cases in which true extent of nodal metastases was clinically overestimated; underestimation, indicating the cases in which the true extent of lymph node metastases was clinically underestimated; mixed type, indicating the cases in which clinical metastatic sites overlapped with pathologic metastatic sites; and discordance, indicating the cases in which clinical metastatic sites were completely different from pathologic metastatic sites (Fig 1).

View larger version (32K): [in this window] [in a new window]   Fig 1. Patterns of the spatial relationships between the extent of lymph node metastases which were clinically predicted and pathologically verified.

Results

T categories
Of the 224 tumors assessed histologically, 75 were T1, 23 were T2, 108 were T3, and 18 were T4 (Table 1). T1 and T3 tumors were predicted correctly in more than 80% of the patients by any of the diagnostic modalities, whereas T2 and T4 tumors were often erroneously predicted. Esophagography, endoscopy, and EUS each correctly predicted T1 tumors in more than 80% of the patients. In contrast, none of the T1 tumors were diagnosed correctly by CT, mainly because most of these tumors were not depicted by CT. T2 tumors were predicted poorly by esophagography and endoscopy; EUS had the best accuracy (61%) for predicting T2 tumors. The main reason for the incorrect prediction of T2 tumors was overestimation of their depth of invasion. Computed tomography could not correctly predict T2 tumors. The accuracy of each diagnostic technique for predicting T4 tumors was less than 30%, because the present study included only resectable T4 tumors; therefore, the degree of adjacent organ involvement was less extensive in all cases. Overall, the accuracy of the consensus evaluation for predicting the T category was 79.9%.

The extent of esophageal cancer has been staged by using various staging modalities, including endoscopy, esophagography, percutaneous and endoscopic ultrasonography, CT, and magnetic resonance imaging. Many studies have assessed the accuracy of these modalities for staging esophageal cancer [6]. The extent and quality of the esophagectomy is the most essential component affecting the sampling accuracy in such studies. However, radical esophagectomy and systematic regional lymphadenectomy were not done in many of the previous studies. In the present study, the sampling accuracy in the assessment of nodal status was maximal because all patients involved in the nodal status assessment had a thorough mediastinal and abdominal lymphadenectomy with or without bilateral cervical lymphadenectomy.

The superiority of EUS for staging both the T and N categories of the UICC-TNM classification [2] was confirmed in the present study. Tio and colleagues [7] reported that EUS correctly staged the depth of tumor invasion in 89% and the regional lymph node metastasis in 80% of patients. Likewise, Botet and associates [8] reported excellent accuracy rates of 92% and 88% for EUS assessment of the depth of the primary tumor and the regional lymph node status, respectively. More recently, Reed and associates [13] reported that EUS was more accurate than CT in preoperative staging of celiac axis lymph nodal status [13].

Although EUS is the most reliable staging modality for locoregional esophageal cancer, conventional esophagography and esophagoscopy still remain useful diagnostic modalities for identifying T1 esophageal tumors, because high accuracy rates (79% and 81%, respectively) were demonstrated by these methods in the present study. In accordance with previous reports, the accuracy for predicting T2 esophageal tumors was poor in the present study because the depth of the T2 tumors was overestimated in many cases. Because the present study included only resectable, localized esophageal tumors, the accuracy rate for predicting T4 tumors was extremely low (less than 30%), in contrast with the high accuracy rate for staging T3 tumors. The low accuracy for diagnosing T4 tumors might be due to the less extensive involvement of adjacent organs in these tumors, a result of the patient selection in the present study. Excellent accuracy rates of more than 90% have been reported for predicting T4 tumors by EUS in other studies [7, 8].

High accuracy rates for predicting lymph node status defined by the UICC-TNM classification [2] have been repeatedly reported and were confirmed in the present study. The high accuracy (72%), sensitivity (78%), and positive predictive value (78%) indicate that N1 disease can be reliably predicted by clinical staging modalities before esophageal resection. However, the localization of metastatic lymph node sites appears to be less accurate in localized esophageal cancer, because the sensitivity of the imaging techniques to detect lymph node metastases was poor when the anatomic sites of the lymph nodes were divided into seven regions. Therefore, the clinical staging of nodal status was reliable only when defined by the UICC-TNM classification [2].

Although the overall accuracy of clinical staging was 80% for the T category and 72% for the N category, the accuracy of stage grouping was 56% in the present study. This low rate of the accuracy of stage grouping is probably because the stage grouping consists of the combination of the T and N categories, both of which cannot be predicted perfectly by the current clinical staging tools. Therefore, preoperative stage grouping might not provide the physician with reliable information in cases of resectable, localized esophageal cancer. Recently, Rice and colleagues [14] reported that depth of tumor invasion (T category) was the most useful clinical predictor of regional lymph node metastasis (N category), and the probability of regional lymph node metastasis could be predicted from an assessment of depth of tumor invasion.

In conclusion, either the local tumor extent (the T category) or the lymph nodal status (the N category) can be predicted reliably by clinical staging techniques. However, the preoperative stage grouping consisting of the combination of the T and N categories is likely invalid in resectable, localized esophageal cancer.

Acknowledgments

This study was supported by a Grant-in-Aid for Cancer Research from the Ministry of Health and Welfare of Japan.

References

1. Japanese Committee for Registration of Esophageal Cancer, Japanese Society for Esophageal Diseases. Statistics of esophageal cancer registry. In: Watanabe S, Tominaga S, Kakizoe T, eds. Cancer treatment and survival: site-specific registries in Japan (Gann monograph on cancer research; No. 43). Tokyo: Japan Scientific Society Press, 1995:39–46.
2. In: Hermanek P., Sobin L.H., eds. UICC TNM Classification of malignant tumours, 4th ed. Berlin: Springer-Verlag, 1987.
3. In: Beahrs O.H., Henson D.E., Hutter R.V.P., Kennedy B.J., eds. Manual for staging of cancer, 4th ed. Philadelphia: JB Lippincott, 1992.
4. Ilson D.H., Kelsen D.P. Combined modality therapy in the treatment of esophageal cancer. Semin Oncol 1994;21:493-507.[Medline]
5. Daly J.M., Karnell L.H., Menck H.R. National cancer data base. Report on esophageal carcinoma. Cancer 1996;78:1820-1828.[Medline]
6. Hoelscher A.H., Dittler H.J., Siewert J.R. Staging of squamous esophageal cancer. World J Surg 1994;18:312-320.[Medline]
7. Tio T.L., Coene P.P.L.O., Schouwink M.H., Tytgat G.N.J. Esophagogastric carcinoma. Radiology 1989;173:411-417.[Abstract/Free Full Text]
8. Botet J.F., Lightdale C.J., Zauber A.G., Gerdes H., Urmacher C., Brennan M.F. Preoperative staging of esophageal cancer. Radiology 1991;181:419-425.[Abstract/Free Full Text]
9. Nishimaki T., Tanaka O., Suzuki T., Aizawa K., Hatakeyama K., Muto T. Patterns of lymphatic spread in thoracic esophageal cancer. Cancer 1994;74:4-11.[Medline]
10. Japanese Society for Esophageal Diseases. Guidelines for the clinical and pathologic studies on carcinoma of the esophagus. Jpn J Surg 1976;6:69-78.[Medline]
11. Aibe T., Fuji T., Okita K., Takemoto T. A fundamental study of normal layer structure of the gastrointestinal wall visualized by endoscopic ultrasonography. Scand J Gastroenterol 1986;21:6-15.[Medline]
12. Chandawarkar R.Y., Kakegawa T., Fujita H., Yamana H., Toh Y., Fujitoh H. Endosonography for preoperative staging of specific nodal groups associated with esophageal cancer. World J Surg 1996;20:700-702.[Medline]
13. Reed C.E., Mishra G., Sahai A.V., Hoffman B.J., Hawes R.H. Esophageal cancer staging. Ann Thorac Surg 1999;67:319-322.[Abstract/Free Full Text]
14. Rice T.W., Zuccaro G., Jr, Adelstein D.J., Rybicki L.A., Blackstone E.H., Goldblum J.R. Esophageal carcinoma. Ann Thorac Surg 1998;65:787-792.[Abstract/Free Full Text]

This article has been cited by other articles:

				C. J. Lightdale and K. G. Kulkarni Role of Endoscopic Ultrasonography in the Staging and Follow-Up of Esophageal Cancer J. Clin. Oncol., July 10, 2005; 23(20): 4483 - 4489. [Abstract] [Full Text] [PDF]

				Y. Nishimura, H. Osugi, K. Inoue, N. Takada, M. Takamura, and H. Kinosita Bronchoscopic Ultrasonography in the Diagnosis of Tracheobronchial Invasion of Esophageal Cancer J. Ultrasound Med., January 1, 2002; 21(1): 49 - 58. [Abstract] [Full Text] [PDF]

				H. Igaki, H. Kato, Y. Tachimori, H. Daiko, M. Fukaya, S. Yajima, and Y. Nakanishi Clinicopathologic characteristics and survival of patients with clinical Stage I squamous cell carcinomas of the thoracic esophagus treated with three-field lymph node dissection Eur. J. Cardiothorac. Surg., December 1, 2001; 20(6): 1089 - 1094. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Personal Folders Download to citation manager Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Nishimaki, T. Articles by Iizuka, T. Search for Related Content PubMed PubMed Citation Articles by Nishimaki, T. Articles by Iizuka, T.