Patterns of failure following combined modality therapy for esophageal cancer, 1984-1990

From 1984-1990, 143 patients with squamous cell or adenocarcinoma of the esophagus were enrolled in a Phase I/II study of neoadjuvant chemotherapy followed by concurrent chemotherapy plus radiotherapy with or without subsequent esophagectomy. Patients received one cycle of Cisplatin or Carboplatin plus Etoposide for squamous cell carcinoma, or Cisplatin or Carboplatin plus 5FU for adenocarcinoma, followed by two cycles of the same chemotherapy given concurrently with 44-46 Gy over 5 weeks. Operable patients then underwent esophagectomy. Inoperable patients and those with positive surgical margins received additional irradiation (16-18 Gy). Twelve percent of the surgical group received preoperative radiotherapy doses > or = 50 Gy. Seventy-two percent (103) had clinical Stage I-III tumors and 28% (40) were clinical Stage IV (1983 American Joint Committee on Cancer criteria). Only clinical Stage I-III patients were analyzed with respect to patterns of failure. Isolated local failure occurred in 19/103 (18%) of clinical Stage I-III patients. Both local and distant relapse occurred in 15/103 (15%), and distant metastases alone occurred in 25/103 (24%). The 3-year actuarial rates of local and distant failures were 45% and 60%, respectively. Among the clinical Stage I-III patients who underwent surgery (n = 58) versus those who did not (n = 45), the 3-year actuarial local and distant failure rates were 30% versus 60% and 45% versus 45%, respectively. Multivariate analysis was performed to identify significant predictors of local control. For all clinical Stage I-III patients, treatment with surgery (p = 0.001) and with three or more cycles of chemotherapy (p = 0.02) were significant predictors of improved local control. Patients who underwent surgery were significantly younger and had a better performance status than those who did not. The improvement in local control with surgery did not translate into better survival, likely on account of a high operative mortality rate in older patients and those receiving > or = 50 Gy preoperatively. We conclude that local control remains poor with concurrent chemotherapy + radiotherapy for esophageal cancer. The addition of surgery improved local control, but distant metastases remain a problem both in this group of patients as well as those treated without esophagectomy. Efforts to improve local control appear warranted, but it remains to be demonstrated that improved local control translates into improved survival in esophageal cancer because of a high rate of distant metastases in patients whose disease is controlled in the esophagus.