Background: Determining the differential diagnosis between typical (TCs) and atypical carcinoid tumors (ACs) is imperative, as the distinction between TCs and ACs is currently based on histologic criteria that are not always correlated with the unfavorable clinical outcomes.
Patients and methods: We conducted a retrospective study of patients who were diagnosed with carcinoid tumors between 1990 and 2005 at M. D. Anderson Cancer Center. We reviewed the slides for the following pathologic features: infiltrative growth; pleural, blood, or lymphatic vessel invasion; tumor stroma; presence of active fibroblastic proliferation; chromatin pattern; presence of nucleolus; and nuclear pleomorphism. We also evaluated the necrotic patterns. Finally, we evaluated three methods for calculating the number of mitoses: randomly selected, the most mitotically active in 10 high-power fields (HPFs), or overall mean mitotic count.
Results: Our cohort consisted of 80 patients (68 with TCs and 12 with ACs). Older age (P=0.002), pathologic stage III or IV disease (P=0.04), active fibroblastic proliferation (P=0.041), and comedo-like necrosis (P=0.001) were significantly associated with tumor recurrence or patient's death. Among the three mitotic counting methods, the overall mean number of mitoses was significantly correlated with recurrence-free survival (P<0.0001). Our criteria for distinguishing AC from TC included the presence of comedo-like necrosis and/or an overall mean number of mitoses ≥0.2/HPF.
Conclusions: Using an overall mean number in counting mitoses and detecting comedo-like necrosis is important for classifying lung carcinoid tumors.
Copyright © 2010. Published by Elsevier Ireland Ltd.