Estimation of the Incidence of Hepatocellular Carcinoma and Cholangiocarcinoma in Songkhla, Thailand, 1989-2013, Using Multiple Imputation Method

PURPOSE: Histological specimens are not required for diagnosis of liver and bile duct (LBD) cancer, resulting in a high percentage of unknown histologies. We compared estimates of hepatocellular carcinoma (HCC) and cholangiocarcinoma (CCA) incidences by imputing these unknown histologies. MATERIALS AND METHODS: A retrospective study was conducted using data from the Songkhla Cancer Registry, southern Thailand, from 1989 to 2013. Multivariate imputation by chained equations (mice) was used in re-classification of the unknown histologies. Age-standardized rates (ASR) of HCC and CCA by sex were calculated and the trends were compared. RESULTS: Of 2,387 LBD cases, 61% had unknown histology. After imputation, the ASR of HCC in males during 1989 to 2007 increased from 4 to 10 per 100,000 and then decreased after 2007. The ASR of CCA increased from 2 to 5.5 per 100,000, and the ASR of HCC in females decreased from 1.5 in 2009 to 1.3 in 2013 and that of CCA increased from less than 1 to 1.9 per 100,000 by 2013. Results of complete case analysis showed somewhat similar, although less dramatic, trends. CONCLUSION: In Songkhla, the incidence of CCA appears to be stable after increasing for 20 years whereas the incidence of HCC is now declining. The decline in incidence of HCC among males since 2007 is probably due to implementation of the hepatitis B virus vaccine in the 1990s. The rise in incidence of CCA is a concern and highlights the need for case control studies to elucidate the risk factors.