Preoperative chemotherapy for clinically node-positive patients with squamous cell carcinoma of the esophagus

Introduction

The prognosis of patients with esophageal cancer is poor. Compared with other malignant gastrointestinal tumors, esophageal cancer is likely to metastasize to lymph nodes, with the incidence of lymph node metastasis in submucosal tumors being reported at 40%, which corresponds to that in advanced gastric cancers.^1,2 Many studies have shown that the nodal status such as the number of metastatic nodes and the extent of node metastasis is the strongest prognostic factor for patients who have undergone esophagectomy.^3,4 For example, patients having three or fewer metastatic nodes show substantially good prognosis of 5-year survival, whereas those having more positive nodes display dismal prognosis.^5,6 As a whole, even after curative resection by esophagectomy with extended three-field lymph node dissection, about 50% of these patients show recurrence.^3,4

Because surgery alone does not seem to offer a complete cure for node-positive patients, multimodality therapy is necessary. Theoretically, preoperative systemic chemotherapy is expected to eradicate systemic micrometastasis which may exist outside the surgical field at diagnosis, and also to reduce the extent and number of lymph node metastases. Neoadjuvant chemotherapy followed by surgery is a promising strategy for node-positive esophageal cancer.

Recent studies have shown that preoperative chemotherapy can lead to disease downstaging and increases the curative resection rate.⁷,⁸ Although randomized controlled trials comparing preoperative chemotherapy followed by surgery with surgery alone have not yet yielded definitive conclusions on the usefulness of chemotherapy,^8–10 many investigators have reported that the responders to chemotherapy show better prognosis than non-responders.^7,11

Thus far, several studies have analysed the efficacy of neoadjuvant chemotherapy for locally advanced esophageal cancers. However, few studies have focused on the node-positive tumors.

The present retrospective study evaluated the effects of chemotherapy on the nodal status such as the number of metastatic lymph nodes and examined their correlation with the prognosis.

Materials and Methods

Patient eligibility

Between October 1994 and December 2003, a total of 77 patients with esophageal cancer, who underwent preoperative chemotherapy followed by surgery at our hospital, were enrolled in this study. The patients were newly diagnosed and had no prior treatment. The eligibility for this study was as follows: the patients were 80 years old or younger and showed performance status (Eastern Cooperative Oncology Group [ECOG]) of 3 or less. Patients with any T (tumor, T1–T4) and nodal involvement including regional lymph nodes (N1) and/or distant lymph node metastasis (M1LYM) were enrolled in this study, whereas those with distant organ metastasis were excluded. The stage was assigned according to the criteria of the American Joint Committee on Cancer.¹² The T and N statuses of the disease were routinely diagnosed by computed tomography (CT), esophagogram and/or bronchoscopy. Diagnostic criteria by CT scan for clinically positive nodes is round-shaped nodes with a size larger than 10 mm in diameter. Magnetic resonance imaging (MRI) was employed in some cases to improve the accuracy of the T4 diagnosis. Bronchoscopy was carried out when tracheal invasion was suspected by the CT scan. For enrolment, all patients had to have adequate bone marrow function (white blood cell count of > 3500 cells/µL, platelet count of > 100 000 cells/µL), normal renal function (serum creatinine level of < 1.2 mg/dL or creatinine clearance of > 50 mL/dL), and normal liver function (serum transaminases < ×2 the upper limit of normal level). This study protocol was approved by the Human Ethics Review Committee of Osaka Medical Center for Cancer and Cardiovascular Diseases. Written informed consent was obtained from all patients.

Treatment regimen

The treatment regimen consisted of cisplatin, adriamycin and 5-fluorouracil (5-FU). 5-FU was administered intravenously at 750 mg/m²/day on days 1–7 in a continuous manner. Adriamycin was administered on day 1 at a dose of 30 mg/m²/day by i.v. injection and cisplatin on day 1 at 70 mg/m²/day by drip infusion for 2 h with sufficient pre- and post-hydration to prevent renal toxicity. After a 2–3-week interval, the same regimens were repeated.

Two weeks after completing the chemotherapy, the patients were re-evaluated for therapeutic responses of the main tumor and metastatic lymph nodes by barium study, tissue biopsy obtained by endoscopy, and chest and abdominal CT scans. The patients were scheduled for surgery approximately 4 weeks after the last day of the chemotherapy. Histological effectiveness was defined as follows: grade 3, complete disappearance of cancer cells; grade 2, more than 2/3 disappearance; grade 1b, 1/3–2/3 disappearance; grade 1a, less than 1/3 disappearance.

Statistical methods

For ordered categorical data, the Mann–Whitney U-test was used for comparisons among subgroups of patients for each clinico-pathological factor. Student's t-test and chi-square test were used to compare the number of lymph node metastases and the ratio of downstaged patients, respectively. Survival time was calculated by the Kaplan-Meier method and statistically compared among patient subgroups by the log-rank test. A two-sided P < 0.05 was considered significant.

Results

Patient and tumor characteristics

The clinico-pathological characteristics of the 77 patients are summarized in Table 1. Most of the tumors were squamous cell carcinomas. The pretherapeutic clinical disease stages were Stage III in 18 patients, Stage IVA in 19 and Stage IVB in 40. All the patients were clinically node-positive.

Clinical and histological responses to chemotherapy

All the 77 patients in this study underwent esophagectomy with lymphadenectomy after chemotherapy. Of the 77 patients, 70 received two or more cycles of the chemotherapy (2 cycles for 67 patients and 3 cycles for three). The remaining seven received only one cycle of the chemotherapy because of progressive disease in four patients and severe toxicity in three. Table 2 summarizes the histological effects of the chemotherapy and pathological examinations of the surgical specimens. The histological effects were not very good, with only one patient (1.3%) showing a grade 3 response and 10 (13.0%) showing a grade 2 response for the main tumors. Compared with the clinical stages before the chemotherapy, the postoperative pathological disease stages were down-staged. Eleven patients (14.3%) were pathologically node-negative (Table 2). Overall, 23 patients showed earlier pathological disease stages than their clinical stages (cStage III to pStage 0–IIB in six patients and cStage IV to pStage 0–III in 17).

Comparison of pathological stages by histological responses in the main tumors

We compared the degree and number of metastatic lymph nodes between the responders and non-responders. Because accurate diagnosis of lymph node metastasis before chemotherapy is difficult and therefore the assessment of the histological effects on the individual metastatic nodes is impossible, we just counted the number of histological metastatic nodes in the resected specimens and correlated them with the histological effects in the main tumors. The patients were classified into two groups, responders and non-responders, according to the histological effects of the chemotherapy on their main tumors. The responders were defined as displaying grade 1b−3 effectiveness in their main tumors, and the non-responders were defined as displaying grade 1 or less effectiveness. Table 3 shows that the pathological stages were significantly earlier in responders than in non-responders (P = 0.0001). The responders also showed significantly smaller degrees of lymph node metastasis (P = 0.0005) and fewer metastatic nodes (2.2 ± 3.1 for responders vs. 12.0 ± 20.5 for non-responders, P = 0.0482) than non-responders.

Comparison of patient prognosis by pathological factors

As shown in Fig. 1, we compared the overall survival curves according to the histological effects on the main tumors. The responders (grade 3–1b) displayed significantly better survival than non-responders (P = 0.002). Figure 2 shows that the 23 downstaged patients showed significantly better survival than non-downstaged patients (P = 0.0005). Next, we compared survival by nodal status.Figures 3 and 4 show a clear difference in the prognosis between the node-negative and node-positive patients. Only the patients who had no metastatic nodes displayed extremely good prognosis of 100% survival at 5 years. However, among the node-positive patients, the prognosis did not greatly differ by the extent of node metastasis or by the number of metastatic nodes.

Comparison of recurrence patterns after surgery of responders with those of non-responders

There were more recurrences in the non-responders than in the responders (Table 4). The most common failure pattern in the non-responders was lymphatic recurrence, with an incidence of 25.4% (15/59), while that in the responders was 16.7% (3/18). The incidence of lymphatic recurrence with or without other sites of recurrence was 47.5% (28/59) in the non-responders as compared with 16.7% in the responders.

Discussion

The purpose of this study was to examine the effects of neoadjuvant chemotherapy on the node-positive patients. Of the 77 patients who were diagnosed to be clinically node-positive before chemotherapy, 11 patients were found to be node-negative by pathological examination of surgical specimens. Because the accuracy of preoperative staging by conventional diagnosis is limited,^13–15 the possibility that the pathologically node-negative patients had been node-negative from the beginning and did not become so due to the chemotherapy is doubtful. However, because systemic chemotherapy is theoretically considered to be effective against both the main tumors and metastatic nodes, the finding of a significantly higher pathological node negative rate in the responders (38.9%vs. 6.8%) strongly suggests that the chemotherapy led to downstaging in the nodal status.

In this study, downstaged patients, whose postoperative pathologic stages were earlier than pretherapeutic clinical stages, showed significantly better survival than non-downstaged patients. Among the factors related to downstaging, the ‘n’ factor seems to be important. The prognosis of patients dramatically depended on their pathological nodal status. The pathologically node-negative patients showed extremely good survival of 100% at 5 years, whereas the prognosis of pathologically node-positive patients was dismal and became worse as the number of metastatic nodes increased. Although such a correlation had been reported by many investigators for patients treated with surgery alone,^5,6 similar results were seen for patients in a neoadjuvant setting. This may suggest that only responders can obtain survival benefit by chemotherapy, with non-responders being disadvantaged by the treatment. The prognosis of node-negative patients was better than pathologically node-negative patients who underwent surgery without preoperative chemotherapy at our hospital during the same periods (data not shown). In addition, non-responders showed poorer survival than surgery-alone patients at the same pathological stages at our hospital (data not shown). Cytotoxicities due to chemotherapy such as bone marrow suppression and gastrointestinal toxicities may deteriorate the host defense system against cancer cells, which leads to poor prognosis.^16,17

Because the sensitivity for preoperative diagnosis of lymph node metastasis is very low by conventional imaging such as computed tomography, endoscopic ultrasonography or positron emission tomography (PET),^13–15 assessment of the pathological effects on individual metastatic nodes is difficult. In this study, we counted the number of metastatic nodes by pathological examination and correlated them with the histological effects on the main tumors. A highly significant correlation was found between the effects on the main tumors and the number of metastatic nodes or the extent of node metastasis. The average number of metastatic nodes in responders was only 2.2, and more than 70% of them had three or less metastatic nodes, which is considered to be controllable by surgery.^5,6 In contrast, non-responders had an average number of 12 metastatic nodes. These results suggest that the chemotherapy may be equally effective in main tumors and metastatic nodes. In other words, if the main tumors are sensitive, then the metastatic nodes should be sensitive, too. The patients having grade 1b or higher effect at the primary site show good indication for subsequent radical operation.

More accurate staging would limit the use of chemotherapy to appropriate groups and thus exclude those who do not require the chemotherapy. In addition, it would help better analysis of the response of metastatic nodes to the chemotherapy. Fine needle aspiration cytology (FNA) under endoscopic ultrasonography (EUS) is considered to be more accurate for lymph node staging, but they are powerless for tumors with severe stenosis or for distant lymph node metastasis. Staging thoracoscopy/laparoscopy is also a candidate, but is highly invasive. It has been difficult to accurately assess the pathological effects before surgery. Recently, however, several reports have shown PET scans to be a non-invasive and effective tool for diagnosing the histological effects of chemotherapy and chemoradiotherapy.^18–20 In the future, a lymph node staging system including all the above modalities should be developed for more accurate assessment. Accurate assessment would allow determination of the indication for surgery after chemotherapy.

In conclusion, preoperative chemotherapy may reduce the number of positive nodes and thus improve the prognosis in responders, whereas it may worsen the prognosis of non-responders. Methods for selecting the responders before treatment should be developed. Because this is a relatively small study number, a large-scale multicenter study is necessary in the future.

References