Ampullary carcinoma is often of mixed or hybrid histologic type: an analysis of reproducibility and clinical relevance of classification as pancreatobiliary versus intestinal in 232 cases

Creative Commons License

Reid M. D., Balci S., Ohike N., Xue Y., Kim G. E., Tajiri T., ...More

MODERN PATHOLOGY, vol.29, no.12, pp.1575-1585, 2016 (SCI-Expanded) identifier identifier identifier

  • Publication Type: Article / Article
  • Volume: 29 Issue: 12
  • Publication Date: 2016
  • Doi Number: 10.1038/modpathol.2016.124
  • Journal Name: MODERN PATHOLOGY
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Page Numbers: pp.1575-1585
  • Hacettepe University Affiliated: Yes


Histologic classification of ampullary carcinomas as intestinal versus pancreatobiliary is rapidly becoming a part of management algorithms, with immunohistochemical classification schemes also being devised using this classification scheme as their basis. However, data on the reproducibility and prognostic relevance of this classification system are limited. In this study, five observers independently evaluated 232 resected ampullary carcinomas with invasive component >3 mm. Overall interobserver agreement was 'fair' (kappa 0.39; P< 0.001) with complete agreement in 23%. Using agreement by 3/5 observers as 'consensus' 40% of cases were classified as 'mixed' pancreatobiliary and intestinal. When observers were asked to provide a final diagnosis based on the predominant pattern in cases initially classified as mixed, there was 'moderate' agreement (kappa 0.44; P< 0.0001) with 5/5 agreeing in 35%. Cases classified as pancreatobiliary by consensus (including those with pure-pancreatobiliary or mixed-predominantly pancreatobiliary features) had shorter overall (median 41 months) and 5-year survival (38%) than those classified as pure-intestinal/mixed-predominantly intestinal (80 months and 57%, respectively; P= 0.026); however, on multivariate analysis this was not independent of established prognostic parameters. Interestingly, when compared with 476 cases of pancreatic ductal adenocarcinornas, the pancreatobiliary-type ampullary carcinomas had better survival (16 versus 41 months, P< 0.001), even when matched by size and node status. In conclusion, presumably because of the various cell types comprising the region, ampullary carcinomas frequently show mixed phenotypes and intratumoral heterogeneity, which should be considered when devising management protocols. Caution is especially warranted when applying this histologic classification to biopsies and tissue microarrays. While ampullary carcinomas with more pancreatobiliary morphology have a worse prognosis than intestinal ones this does not appear to be an independent prognostic factor. However, pancreatobiliary-type ampullary carcinomas have a much better prognosis than their pancreatic counterparts.