897 Prognostic Heterogeneity Among Patients With Bone Metastases From Well-Differentiated Digestive Neuroendocrine Tumors

Abstract

Background: There are a paucity of data pertaining to the natural history and outcomes of patients with a well-differentiated rectal carcinoid tumor. There is general consensus that a ≤10 mm, well-differentiated rectal carcinoid that does not infiltrate the muscularis propria or demonstrate nodal spread can be managed by local excision. Conversely, most favor radical resection for lesions ≥ 20mm and/or nodal metastasis. A clearer understanding of the natural history, in particular for intermediate size lesions (11-19mm), is necessary to optimize patient care. Aims: For patients with a well-differentiated rectal carcinoid who underwent endoscopic evaluation to determine the: 1.) lesion size correlating with metastasis, 2.) factors associated with disease progression, 3.) disease stage of 11-19mm lesions, and 4.) 10 year survival rates based on disease extent. Method: A histopathology database was reviewed to identify patients with a well-differentiated rectal carcinoid or neuroendocrine tumor who underwent endoscopy from 1/1/00-1/9/12. Clinical demographics, radiologic, and clinical follow-up data were analyzed. Results: Among the 87 patients diagnosed with a well-differentiated rectal carcinoid, metastatic disease (local or distant) was identified at the time of endoscopy in 2/66 (3%), 4/6 (66%) and 11/15 (73%) of patients for lesions ≤10mm, 11-19mm and ≥20mm, respectively. Metastasis was predicted with 100% sensitivity and 87% specificity using a lesion cut off size of ≥9mm, (AUC 0.948). Metastasis was subsequently discovered in 1/64 (1.6%), 1/2 (50%) and 4/4 (100%) patients with lesions measuring ≤10 mm, 11-19mm and ≥20mm, respectively. These sites included the liver (n= 4), pancreas (n=1) and liver & lung (n=1), at 34.6 (IQR 4.7-33) months from endoscopic evaluation. For patients with locally confined disease at presentation, subsequent disease progression correlated with advanced age (p=0.03), presence of initial symptoms (p=0.02), lesion size (p=0.0001) and immune-stain pattern (p=0.03). Multivariate analysis revealed one independent factor: size .20mm at the time of initial diagnosis (p,0.0001: OR 1.25, 95% CI 1.04-1.5). A lesion cut off size of .10mm, (AUC 0.957) had 83% sensitivity and 98% specificity for predicting disease progression. The ten year survival rates were 87%, 67% and 32% for patients presenting with locally confined, locally advanced and distant disease, respectively. Conclusions: Our data support the contention that small ( ≤10 mm) well differentiated lesions have a generally benign disease course. However, our findings indicate that the clinical behavior of intermediate sized tumors (11-19 mm) mimics that of large (≥20 mm) tumors and, thus, need for a more aggressive management strategy. Examination of other staging and clinical parameters may be key, particularly for patients with intermediate size lesions.