Significance of immunohistochemically demonstrated micrometastases to lymph nodes in esophageal cancer with histologically negative nodes

Significance of immunohistochemically demonstrated micrometastases to lymph nodes in esophageal cancer with histologically negative nodes Shintaro Komukai, MD, Tadashi Nishimaki, MD, Hidenobu Watanabe, MD, Yoichi Ajioka, MD, Tsutomu Suzuki, MD, and Katsuyoshi Hatakeyama, MD, Niigata, Japan

Background. We examined the prevalence, patterns, and clinical significance of nodal micrometastases in patients with esophageal cancer. Methods. Cervical, mediastinal, and abdominal lymph nodes systematically removed from 37 patients without conventional histologic evidence of lymph node metastasis from esophageal squamous cell carcinoma were immunohistochemically examined to detect cells that were stained for cytokeratins by the monoclonal antibody cocktail AE1/AE3. Postoperative care and survival were compared in cases with and without such micrometastases. Results. Nodal micrometastases were found in 14 of 37 patients (38%). Among these patients, 9, 7, and 4 had micrometastases to abdominal, mediastinal, and cervical lymph nodes, respectively. Postoperative tumor recurrence was significantly more frequent in patients with micrometastases (50%) than in those without (9%, P =.008). Overall and relapse-free survival in the former group was significantly worse than in the latter group (P = .042 and P = .002, respectively). Nodal micrometastases had an independent prognostic importance for relapse-free survival as determined by multivariate analysis. Conclusions. Metastatic tumor cells are frequently present in lymph nodes, even in patients without histologic evidence of nodal metastasis from esophageal cancer. Nodal micrometastases indicates a poorer prognosis after a curative esophagectomy procedure in histologically node-negative cases. (Surgery 2000;127:40-6.) From the First Department of Pathology and the First Department of Surgery, Niigata University School of Medicine, Niigata, Japan

ESOPHAGEAL CANCER IS an aggressive disease with a poor overall prognosis despite treatment. Even after radical esophagectomy, recurrent disease frequently develops not only at the site of the operation but also in distant organs.1 Furthermore, tumor recurrence is not rare after radical esophagectomy even when histologic examination shows no lymph node metastasis. Recently, individual or clustered tumor cells previously undetectable with a conventional histologic examination have been detected with an immunohistochemical technique in lymph nodes or bone marrow in many cases of breast cancer,2,3 colorectal cancer,4-6 gastric cancer,7,8 and lung cancer.9 However, the inciAccepted for publication September 2, 1999. Reprint requests: Shintaro Komukai, MD, First Department of Pathology, Niigata University School of Medicine, Asahimachidori 1-754, Niigata 951-8122, Japan. Copyright © 2000 by Mosby, Inc. 0039-6060/2000/$12.00 + 0

40 SURGERY

11/56/102754

dence, patterns, and clinical implications of such micrometastases to lymph nodes remain unclear in cases of esophageal cancer. Esophageal carcinomas are known to metastasize frequently to lymph nodes at a considerable distance from their primary sites, even at an early stage of tumor spread.10 Such tumor behavior necessitates extensive lymphadenectomy to accurately assess the possibility of lymph node micrometastases in cases of esophageal cancer. In this study, we sought to determine incidence, patterns, and clinical significance of micrometastases to lymph nodes from esophageal cancer in patients with no overt histologic evidence of metastasis to the lymph nodes. The operation performed was radical esophagectomy combined with extensive lymphadenectomy. To reliably evaluate the influence of the presence of nodal micrometastases upon prognosis, we analyzed patients who were followed up for at least 5 years after esophagectomy.

Surgery Volume 127, Number 1

Komukai et al 41

Fig 1. A single tumor cell (arrow) showing immunostaining for cytokeratin by an AE1/AE3 antibody cocktail (original magnification ×25).

MATERIALS AND METHODS Patients. Between 1985 and 1994, a total of 194 patients with invasive thoracic esophageal cancer underwent radical esophagectomy combined with cervical, mediastinal, and abdominal lymphadenectomies at the First Department of Surgery of Niigata University. Among these patients, 59 (30%) had no metastases to lymph nodes according to conventional histologic examination. Of the 59 patients, 21 were excluded from the study for the following reasons: incomplete tumor resection (n = 3), a synchronous extraesophageal second primary cancer (n = 4), preoperative chemotherapy or radiotherapy (n = 2), a tumor histologic finding other than squamous cell carcinoma (n = 3), or either followup for less than 5 years or death from unrelated causes (n = 9). The remaining 38 patients all had curative resection, defined as an R0 resection by the Union Internationale Contre le Cancer (UICC)TNM residual tumor classification.11 All 38 patients also had a squamous cell carcinoma of the esophagus, did not receive preoperative therapy, and were followed up for at least 5 years after esophagectomy. Ages of these patients ranged from 47 to 75 years (median, 61 years), and the group included 33 men and 5 women.

Primary tumors. On the basis of the T-categories of the UICC-TNM classification,11 the 38 patients of esophageal carcinoma included 16 T1 tumors, 6 T2 tumors, and 16 T3 tumors. Of the 16 T1 tumors, 2 were intramucosal invasive carcinomas, and the others were submucosal carcinomas. Among the 38 patients, 3 had a tumor in the upper thoracic esophagus, 21 in the mid-thoracic esophagus, and 14 in the lower thoracic esophagus. Histologically, all tumors were squamous cell carcinomas, 9 were well differentiated, 24 were moderately differentiated, and 5 were poorly differentiated. Lymph nodes. Our technique of radical esophagectomy has been reported in detail elsewhere.12 The deep cervical, supraclavicular, and cervical paraesophageal lymph nodes were resected as cervical lymph nodes. Thoracic paratracheal, bifurcation, pulmonary hilar, and thoracic paraesophageal lymph nodes on both sides of the esophagus were removed as mediastinal lymph nodes. These lymph nodes were defined with the nomenclature of the Japanese Society for Esophageal Diseases.13 Perigastric and celiac lymph nodes were resected as abdominal lymph nodes. Of the 38 patients, 30 (79%) had a bilateral cervical lymphadenectomy, and the other patients underwent a left cervical lymphadenectomy.

42 Komukai et al

Surgery January 2000

Fig 2. A cluster of micrometastatic tumor cells (arrow) detected by immunostaining by an AE1/AE3 antibody cocktail (original magnification ×25).

A total of 2845 lymph nodes were dissected from the 38 patients. The average number of lymph nodes dissected per patient was 75 (range, 38-127 nodes). Histologic examination. After fixation in neutral buffered formalin for a few days, all portions of the esophagectomy specimens were serially and grossly sectioned. Each resulting tissue block was processed for routine paraffin embedding and sectioning. Sections 3 µm thick were stained routinely with hematoxylin and eosin. Likewise, all lymph nodes dissected were embedded in paraffin and processed for histologic examination. Two experienced pathologists independently examined all sections and confirmed the absence of visible metastatic deposits in all lymph nodes dissected immediately after esophagectomy. On the basis of our previous study demonstrating that 9 serial sections 3 µm thick (total thickness, 27 µm) were sufficient to detect micrometastatic foci from colorectal cancer in lymph nodes,14 we examined 5 serial sections, each 10 µm thick, from each of the paraffin blocks of the lymph nodes in the current study. The first and fifth sections were stained by hematoxylin and eosin for histologic confirmation of absence of metastatic tumor. The other 3 serial sections (30 µm in aggregate thickness) were used for the immunohistochemical examination

described below. Furthermore, to assess vascular invasion by the tumor, 2 additional 3-µm-thick sections from the tissue blocks of each primary tumor were examined with Victoria blue elastic fiber staining and CD34 antibody (Dako, Kyoto, Japan). Immunohistochemistry. The anti-cytokeratin antibody cocktail AE1/AE3 (Dako) was used at a dilution of 1:150. This mouse monoclonal antibody preparation is specific for human cytokeratins and recognizes cytokeratins in epithelial cells.15 The streptavidinbiotin (SAB) immunoperoxidase method was used. Sections to be stained with AE1/AE3 were treated with 0.1% trypsin (Sigma, St. Louis, Mo) in 0.1% calcium chloride (pH 7.8) at 37°C for 20 minutes before immunostaining. Reagents for the subsequent step— biotinylated rabbit anti-mouse immunoglobulin and SAB complex—were supplied commercially (Nichirei, Tokyo, Japan). Diaminobenzidine was used as the chromogen, and sections were counterstained with methyl green. Immunohistochemical staining with a monoclonal mouse anti-human endothelial cell CD34 antibody, diluted 1:100, was performed in the same manner. Hematoxylin was used as a counterstain. Definition of micrometastasis. Micrometastasis was defined as the presence in lymph nodes of either a single cell (Fig 1) or a cluster of 5 or fewer cells that were immunohistochemically positive for cytoker-

Komukai et al 43

Surgery Volume 127, Number 1

Fig 3. Comparison of relapse-free survival curves between patients with and without micrometastases. A significant difference (P < .01) in cumulative survival was seen between the 2 groups.

Fig 4. Comparison of overall survival curves between patients with and without micrometastases. A significant difference (P < .05) was seen in cumulative survival between the 2 groups.

atin (Fig 2). We did not regard a cluster of more than 5 tumor cells as a micrometastasis because such an aggregate usually can be recognized as a metastatic deposit by conventional histologic examination. In addition, histologically overt metastatic foci newly found in the serial sections were not classified as micrometastases. Although reticular cells and plasma cells can show staining for cytokeratin, these nonneoplastic cells were discriminated from micrometastatic tumor cells by differing staining patterns as shown in our previous study.14 Follow-up data. Patients were followed regularly after discharge with routine physical and laboratory examinations at our institution or affiliated hospitals. Chest radiography, ultrasonography, or computed tomography were performed at least annually to detect possible recurrent disease. Postoperative survival information was available in all 38 patients. Tumor recurrence developed in 10 patients, and 8 died of this recurrent esophageal cancer. Local or regional recurrence occurred in 2 patients, and recurrent disease developed at distant sites in 8 patients. Twenty-eight patients survived more than 5 years with no evidence of recurrence. Statistical analysis. Differences in frequencies were detected with the chi-square test or Fisher’s exact probability test for smaller numbers. Differences between means were assessed with the Student t test. Overall and relapse-free survival rates after resection were examined with the Kaplan-Meier method. The equality of survival curves was assessed with the logrank test. To elucidate factors influencing survival, 8 factors regarding patients’ characteristics and tumor status were examined in all patients: sex, age (>61 vs ≤61 years), primary site (upper, middle vs lower), grade of differentiation (poorly differentiated vs well, moderately differentiated), T category (T2,3 vs T1), lymphatic invasion, vascular invasion, and nodal micrometastases. Cox’s proportional hazard model was used to identify whether these factors were inde-

Table I. Clinicopathologic features of patients with and without nodal micrometastases

Variable

Nodal micrometastases Absent Present (n = 23) (n = 14)

Sex Male 20 Female 3 Age* 61.7 ± 7.9 Primary site Upper 3 Middle 15 Lower 5 Grade of differentiation Well 5 Moderately 15 Poorly 3 Depth of penetration Mucosa 1 Submucosa 11 Muscularis propria 3 Adventitia 8 Lymphatic invasion Not seen 16 Seen 7 Vascular invasion Not seen 17 Seen 6

P value

12 2 59.4 ± 6.4

NS

0 6 8

NS

3 9 2

NS

1 2 3 8

NS

8 6

NS

8 6

NS

NS

*Data are expressed as mean ± standard deviation. NS, Not significant.

pendently associated with postoperative overall and relapse-free survival. This multivariate analysis was performed with the SPSS 9.0 (SPSS Japan Inc, Tokyo, Japan) software package. A P value less than .05 was considered significant. RESULTS Incidence of nodal micrometastases. Micrometa static tumor cells were shown with a immunohisto-

44 Komukai et al

Surgery January 2000

Table II. Location of lymph nodes with micrometastases Middle esophagus (n = 7) n

Lymph nodes Cervical Superficial cervical Cervical paraesophageal Deep cervical Supraclavicular Mediastinal Upper thoracic paraesophageal Thoracic paratracheal Bifurcation Middle thoracic paraesophageal Pulmonary hilar Lower thoracic paraesophageal Diaphragmatic Posterior mediastinal Abdominal Perigastric Celiac artery Splenic artery

Lower esophagus (n = 7) n

0 0 0 1

1 1 1 0

0 0 1 1 0 2 1 0

1 0 0 1 0 0 0 0

3 2 0

3 2 1

Table III. Results of Cox regression analysis for relapse-free survival and overall survival Variable Sex Age Primary site Grade of differentiation T category Lymphatic invasion Vascular invasion Micrometastasis

Modality Male Female ≤61 >61 Lower Upper, middle Well, moderately Poorly T1 T2, T3 Absent Present Absent Present Absent Present

RR

Relapse-free survival 95% CI P value

RR

Overall survival 95% CI P value

1.21

0.18-8.38

0.85

1.62

0.17-15.28

0.68

1.02

0.22-4.89

0.97

0.79

0.12-5.07

0.81

1.51

0.32-7.06

0.60

2.05

0.34-12.31

0.43

3.26

0.39-26.99

0.27

8.65

0.91-81.91

0.06

2.12

0.31-14.67

0.45

2.21

0.24-20.56

0.49

1.80

0.31-10.48

0.52

2.09

0.31-13.89

0.45

1.22

0.24-6.17

.81

0.36

0.04-3.16

.36

6.45

1.14-36.48

.03

5.64

0.71-44.66

.10

RR, Relative risk; CI, confidence interval.

chemical technique as single cells (Fig 1) or small clusters of cells (Fig 2). Cytoplasmic staining was done with the anti-cytokeratin antibody cocktail AE1/AE3. These micrometastatic cells were found only in the marginal sinuses of the lymph nodes in 2 patients and only in the lymphoid interstitium in 7 patients. Micrometastatic tumor cells were present in both marginal sinuses and the interstitium of lymph nodes in 6 patients. All of the primary tumors resected from the 38 patients also showed diffuse staining with AE1/AE3. One previously undetected histologically overt metastatic deposit, measuring 1 mm in diameter, was

found in a perigastric lymph node resected from 1 patient after close examination of serial sections of the lymph nodes. After exclusion of this patient from data analysis, micrometastatic cells were found in 29 of 2774 lymph nodes (1.0%), and found in 14 of 37 patients (38%). The number of lymph nodes containing micrometastases of tumor cells ranged from 1 to 7 per patient (mean, 2.1; median, 1.5). Clinicopathlogic features. After the single exclusion described above, the 37 remaining patients were divided into one group of 23 without micrometastases and one group of 14 with micrometastases. No

Surgery Volume 127, Number 1

significant difference was evident between the 2 patient groups in sex, mean age, primary site, grade of cellular differentiation, depth of tumor invasion, or frequency of vessel invasion (Table I). One of 2 intramucosal and 2 of 13 submucosal carcinomas had nodal micrometastases (Table I). Location patterns of lymph nodes with micrometastases. Of the 14 patients with lymph nodes harboring micrometastases, 9 (64%) had these deposits in abdominal lymph nodes; 7 (50%) in mediastinal lymph nodes; and 4 (29%) in cervical lymph nodes. On the basis of the nomenclature of the Japanese Society for Esophageal Diseases,13 locations of lymph nodes involved by micrometastases are shown by primary site in Table II. Influence of micrometastases on long-term outcomes. Recurrent disease developed in 7 (50%) of the 14 patients with micrometastasis, while tumors recurred in 2 (8.7%) of the 23 patients without micrometastasis after an esophagectomy. Recurrence of esophageal cancer was significantly more common among patients with micrometastasis than among those without (P = .008). Of the 7 patients who had nodal micrometastases at the time of the operation and tumor recurrence after the esophagectomy, 3 had lymphogenous metastases to the cervical, mediastinal, or abdominal lymph nodes after an esophagectomy. The other patients had hematogenous metastases to the liver, lung, bone, or skeletal muscle. The 2 patients who had no nodal micrometastases at the time of the operation but who had recurrent disease after an esophagectomy had hematogenous metastases to the liver and lung. The mean number of lymph nodes involved by micrometastases per affected patient was 2.6 in the 7 patients who had tumor recurrence after an esophagectomy, while the mean number of lymph nodes containing micrometastases per patient was 1.6 in the other 7 patients who survived without recurrent disease. This difference was not statistically significant. The relapse-free survival curve of 14 patients with micrometastasis was significantly steeper than that of the 23 patients without micrometastasis (P = .002, Fig 3). Similarly, the overall survival curve of the former patients was significantly worse than that of the latter group (P = .042, Fig 4). Cox regression analysis showed that nodal micrometastases was the only independent prognostic factor for relapse-free survival. However, nodal micrometastasis had no independent prognostic importance for overall survival (Table III). DISCUSSION The prognosis for patients with esophageal cancer remains poor. The 5-year survival rate of patients free from lymph node metastasis has been

Komukai et al 45

reported to be approximately 70%, even after extended radical esophagectomy.12 Recurrence of the disease in distant organs is not rare even in these patients. These outcomes suggest that single tumor cells or clusters that cannot be detected with clinical or histologic techniques already have disseminated before esophageal resection. Our data support this idea because such micrometastases were found in lymph nodes of approximately 40% of patients without histologically detectable nodal metastasis. Incidences of nodal micrometastasis similar to ours have been reported by Izbicki et al16 and Natsugoe et al.17 Esophageal cancer is known to metastasize frequently to lymph nodes, even at the early stage of local tumor spread.10 Furthermore, extensive lymphatic spread not only to regional lymph nodes but also to cervical lymph nodes or distant abdominal lymph nodes has been found after the introduction of extended radical esophagectomy to treat esophageal cancer.18,19 The aggressive nature of this disease is reconfirmed by the present study. Although the number of lymph nodes involved by micrometastases was only 1 or 2 in more than half of our affected patients, an extensive distribution of these micrometastases among cervical, mediastinal, and abdominal lymph nodes was evident overall. Moreover, nodal micrometastases occurred in 1 patient with mucosal carcinoma infiltrating the lamina muscularis mucosae and in 2 patients with submucosal carcinoma. Whether micrometastasis has clinical significance has been controversial in the study of many tumor types. Some authors have reported a lack of significant impact of micrometastasis on prognosis,2,4,7,8 but other investigators have described adverse effects on postoperative survival curves.3,5 Glickman et al20 did not demonstrate a correlation between occult lymph node metastasis and poor survival rates for patients with esophageal cancer. In contrast, Izbicki et al16 and Natsugoe et al17 have found micrometastasis to lymph nodes to be a significant indicator of a poorer prognosis after esophagectomy in patients with esophageal cancer. Our results are consistent with the findings of the latter authors. In the present study, nodal micrometastasis was the only independent risk factor of postoperative recurrence of esophageal cancer in patients without histologically overt nodal metastasis. Half of these patients had tumor recurrences after esophagectomy. Furthermore, a significant adverse influence of nodal micrometastasis on the survival curves could be shown in our patients. The great majority of patients with postoperative tumor

46 Komukai et al

recurrence died of hematogenous metastasis to distant organs. These results indicate that patients without histologically overt lymph node metastasis but with immunohistochemically detectable nodal micrometastasis are at an increased risk for recurrent esophageal cancer, and they therefore should receive postoperative chemotherapy. Recently, genetic examination of K-ras or p53 mutations in archival tissue samples has been reported to be more sensitive in detecting occult lymph node metastasis than conventional histologic examination.21,22 However, our previous study has shown immunohistochemical staining with anti-cytokeratin antibodies to be more simple, rapid, sensitive, specific, and accurate for this purpose than such genetic analyses in cases of adenocarcinoma of the colon and rectum.14 More studies are needed to determine whether immunohistochemical staining is superior to those genetic analyses in cases of esophageal squamous cell carcinoma. Such immunohistochemical examination is relatively inexpensive. CONCLUSIONS We recommend routine immunohistochemical examination for lymph node micrometastases in esophageal cancer patients with histologically negative nodes. Immunohistochemistry frequently discloses micrometastatic tumor cells in cervical, mediastinal, and abdominal lymph nodes in patients without histologically demonstrable lymph node metastasis from esophageal cancer. The presence of nodal micrometastasis is an independent risk factor of postoperative recurrent disease in these patients. REFERENCES 1. Müeller JM, Erasmi H, Stelzner M, Zieren U, Pichlmaier H. Surgical therapy of oesophageal carcinoma. Br J Surg 1990;77:845-57. 2. McGuckin MA, Cummings MC, Walsh MD, Hohn BG, Bennett IC, Wright RG. Occult axillary node metastases in breast cancer: their detection and prognostic significance. Br J Cancer 1996;73:88-95. 3. Galea MH, Athanassiou E, Bell J, Robertson JF, Elston CW, Blamey RW, et al. Occult regional lymph node metastases from breast carcinoma: immunohistological detection with antibodies CAM5.2 and NCRC-11. J Pathol 1991;165:221-7. 4. Greenson JK, Isenhart CE, Rice R, Mojzisik C, Houchens D, Martin Jr EW. Identification of occult micrometastases in pericolic lymph nodes of Duke’s B colorectal cancer patients using monoclonal antibodies against cytokeratin and CC49. Cancer 1994;73:563-9. 5. Jeffers MD, O’Dowd GM, Mulcahy H, Stagg M, O’Donoghue DP, Toner M. The prognostic significance of immunohistochemically detected lymph node micrometastases in colorectal carcinoma. J Pathol 1994;172:183-7.

Surgery January 2000

6. Cutait R, Alves VAF, Lopes LC, Cutait DE, Borges JLA, Singer J, et al. Restaging of colorectal cancer based on the identification of lymph node micrometastases through immunoperoxidase staining of CEA and cytokeratins. Dis Colon Rectum 1991;34:917-20. 7. Ishida K, Katsuyama T, Sugiyama A, Kawasaki S. Immunohistochemical evaluation of lymph node micrometastases from gastric carcinomas. Cancer 1997;79:1069-76. 8. Maehara Y, Oshiro T, Endo K, Baba H, Oda S, Ichiyoshi Y, et al. Clinical significance of occult micrometastases in lymph nodes from patients with early gastric cancer who died of recurrence. Surgery 1996;119:397-402. 9. Pantel K, Izbicki J, Passlick B, Angstwurm M, Haussinger K. Frequency and prognostic significance of isolated tumour cells in bone marrow of patients with non-small-cell lung cancer without overt metastases. Lancet 1996;347:649-53. 10. Nishimaki T, Suzuki T, Kanda T, Obinata I, Komukai S, Hatakeyama K. Extended radical esophagectomy for superficially invasive carcinoma of the esophagus. Surgery 1999;125:142-7. 11. Sobin LH, Wittekind C. TNM classification of malignant tumors. 5th ed. Berlin: Springer, Union Internationale Contre le Cancer (UICC); 1997. 12. Nishimaki T, Suzuki T, Suzuki S, Kuwabara S, Hatakeyama K. Outcomes of extended radical esophagectomy for thoracic esophageal cancer. J Am Coll Surg 1998;186:306-12. 13. Japanese Society for Esophageal Diseases. Guidelines for clinical and pathologic studies on carcinoma of the esophagus. Jpn J Surg 1976;6:69-78. 14. Sasaki M, Watanabe H, Jass JR, Ajioka Y, Kobayashi M, Hatakeyama K. Immunoperoxidase staining for cytokeratins 8 and 18 is very sensitive for detection of occult node metastasis of colorectal cancer: a comparison with genetic analysis of K-ras. Histopathology 1998;32:199-208. 15. Woodcock MJ, Eichner R, Nelson WG. Immunolocalization of keratin polypeptides in human epidermis using monoclonal antibodies. J Cell Biol 1982;95:580-8. 16. Izbicki JR, Hosch SB, Pichlmeier U, Rehders A, Busch C, Niendorf A, et al. Prognostic value of immunohistochemically identifiable tumor cells in lymph nodes of patients with completely resected esophageal cancer. N Engl J Med 1997;337:1188-94. 17. Natsugoe S, Mueller J, Stein HJ, Feith M, Höfler H, Siewert JR. Micrometastases and tumor cell microinvolvement of lymph nodes from esophageal squamous cell carcinoma. Cancer 1998;83:858-66. 18. 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. 19. Akiyama H, Turumaru M, Udagawa H, Kajiyama Y. Radical lymph node dissection for cancer of the thoracic esophagus. Ann Surg 1994;220:364-73. 20. Glickman JN, Torres C, Wang HH, Turner JR, Shahsafaei A, Richards WG, et al. The prognostic significance of lymph node micrometastases in patients with esophageal carcinoma. Cancer 1999;85:769-78. 21. Hayashi N, Ito I, Yanagisawa A, Kato Y, Nakamori S, Imaoka S, et al. Genetic diagnosis of lymph-node metastasis in colorectal cancer. Lancet 1995;345:1257-9. 22. Luketich JD, Kassis ES, Shriver SP, Nguyen NT, Schauer PR, Weigel TL, et al. Detection of micrometastases in histologically negative lymph nodes in esophageal cancer. Ann Thorac Surg 1998;66:1715-8.