Abstracts
2002 Digestive Disease Week

# 103334 Abstract ID: 103334 Specific Histologic Features Can Be Used to Predict Reliably Lymph Node Spread in T1 Colorectal Carcinoma - Beats All Biomarkers - a Novel Finding That Will Advance Medical and Surgical Management
Tarun Mullick, Preetha Ramalingam, Feza Remzi, Jason Tasch, Ian Lavery, Adrian Ormsby, Cleveland, OH

BACKGROUND: The ability to predict (+) lymph node (LN) status in early colorectal carcinoma (CRC) would significantly alter the medical and surgical management options for patients with CRC, thereby likely improving disease free survival, morbidity, and mortality. A comparison of tumor histologic features with a comprehensive group of molecular marker has not been performed to predict LN status. AIMS: To compare specific histological features with potential early biomarkers in T1 CRC to determine if there are specific features that can predict LN+ status. METHODS: From 1987 to 2001, all LN+ T1 CRC cases (n=23) and an age/sex matched control group of LN- cases (n=43) were prospectively studied. In all cases, comprehensive lymph node dissection of surgical resection specimens was performed. Exhaustive histologic assessment of tumor grade, extent of infiltration, base of tumor invasion, tumor lymphocyte response, and angiolymphatic invasion was performed on the colectomy by 2 GI pathologists blinded to the outcome of LN status. Biomarkers of telomerase, p53, COX, MIB-1, BCL-1, BCL-X, cyclin D1, MLH-1, MSH-2, and MSH-6 expression by immunohistochemistry was assessed via tissue microarray analysis by the 2 blinded GI pathologists. Independent prognostic status of the specific histologic features and biomarkers was performed with multivariate analysis. RESULTS: The histologic criteria of base of tumor invasion (p=0.05) and angiolymphatic invasion (p=0.045) correlated with LN+ status and were independent prognostic factors on multivariate analysis (p<0.05). None of the other histologic features and none of the biomarkers predicted LN+ status. Angiolymphatic invasion was 95% specific. CONCLUSION: 1) Base of tumor invasion or angiolymphatic invasion on histology of the submucosa of the colectomy predicts LN+ status without examination of the lymph nodes. This is a major advancement that will likely modify management of T1 CRC, thereby improving morbidity and reducing costs. 2) Since only the submucosa is needed to determine lymph node spread after biopsy confirmation of cancer from colonoscopy, endoscopic mucosal resection could be performed on all patients in order to stage patients lymph node status. If these factors are absent thus predicting no lymph node spread, endoscopic mucosal resection with negative tumor margins or local surgical excision may be adequate.