Pathologic prognostic factors in Barrett's-associated adenocarcinoma - A follow-up study of 96 patients

BACKGROUND. The objective of this study was to evaluate a variety of histol ogic features, some of which to our knowledge have never been evaluated in Barrett's-associated adenocarcinoma (BAd) (such as Crohn's-like lymphoid re action and peritumoral lymphoid response) in patients with and without preo perative neoadjuvant chemotherapy combined with radiotherapy (chemrad) to d etermine their prognostic significance in these two groups of patients. METHODS. Tumor sections from 96 patients (83 males and 13 females; mean age , 62 years) with resected BAd (61 with chemrad and 35 without chemrad) were evaluated for numerous histologic features such as pathologic stage accord ing to the American Joint Committee on Cancer TNM staging system, peritumor al lymphoid infiltrate, Crohn's-like lymphoid reaction, and degree of post chemrad residual tumor and correlated with the preoperative chemrad status and with survival (mean follow-up, 23 months). RESULTS. By univariate analysis, older patient age (P = 0.02), higher patho logic stage (P = 0.02) (including depth of invasion and lymph node status), infiltrative growth pattern (P = 0.05), perineural invasion (P = 0.05), va scular invasion (P = 0.04), and the absence of a peritumoral lymphoid infil trate (P = 0.04) were associated with shortened survival in the entire coho rt and in patients without chemrad, with the exception of infiltrative grow th pattern (P = 0.1 in the nonchemrad group only). Higher stage was the onl y feature associated with decreased survival in the chemrad group, Subcateg orization of lymph nodes according to the number involved with metastases ( fewer than four, four to seven, and greater than seven) had no further effe ct on prognosis. However, subcategorization of T1 tumors into Tla and T1b d id influence prognosis in a negative manner. Using multivariate analysis, o nly older patient age (P = 0.005) and the absence of a peritumoral lymphoid infiltrate (P = 0.05) were statistically associated with poor survival ind ependent of stage. In addition, perineural invasion (P = 0.07) showed a tre nd toward shortened survival in patients with this feature. Preoperative ch emrad had no effect on survival in this retrospective nonrandomized cohort of patients. CONCLUSIONS. This study confirms the strong prognostic usefulness of the TN M staging system in patients with resected BAd, even in those patients who received preoperative chemrad. In addition, older patient age, the absence of a peritumoral lymphoid infiltrate, and possibly perineural invasion corr elate with poor survival independent of pathologic stage in patients with t hese tumors. Cancer 1999;85: 520-8. (C) 1999 American Cancer Society.