Objective: The 1997 staging system for esophageal carcinoma subdivides distant metastatic disease (M1) into M1a (nonregional lymph node metastases) and M1b (other metastases). This study evaluates the relevance of this classification.
Methods: One hundred forty patients were identified with M1 disease, 36 (26%) M1a and 104 (74%) M1b. The histologic type was adenocarcinoma in 118 (84%), squamous cell in 18 (13%), and adenosquamous in 4 (3%), with a similar distribution for M1a and M1b (P =.3). Forty-five underwent surgery, 28 (78%) with M1a disease and 17 (16%) with M1b disease (P <.001). Chemotherapy and/or radiation therapy was given to 33 (73%) surgical patients and 63 (66%) nonsurgical patients (P =.4), 28 (78%) with M1a disease and 68 (66%) with M1b disease (P =.17).
Results: Median and 5-year survivals were 11 months and 6% in patients with M1a disease and 5 months and 2% in those with M1b disease (P =.001). Surgery provided no advantage in M1b (P =.6) or M1a disease (P =.2). Multivariable analysis demonstrated that patients with M1b disease had 1.8 times the mortality risk of those with M1a disease (CI 1.2-2.7, P =.004), and patients without chemotherapy and/or radiotherapy had 2.2 times the mortality risk of those with chemotherapy and/or radiotherapy (CI 1.5-3.2, P <.001). Despite the prevalence of surgery in patients with M1a disease, the analysis suggests that M1a and use of chemotherapy and/or radiotherapy, rather than surgery, account for the small, clinically unimportant differences in survival.
Conclusions: We conclude that (1) although there are statistically significant survival differences between M1a and M1b disease, these differences are not clinically important; (2) chemotherapy and/or radiotherapy is associated with a modest survival benefit; and (3) surgery offers no survival advantage.