RESUMO
The guidelines recently recognized the intra-ampullary papillary tubular neoplasm (IAPN) as a distinct tumor entity. However, the data on IAPN and its distinction from other ampullary tumors remain limited. A detailed clinicopathologic analysis of 72 previously unpublished IAPNs was performed. The patients were: male/female=1.8; mean age=67 years (range: 42 to 86 y); mean size=2.3 cm. Gross-microscopic correlation was crucial. From the duodenal perspective, the ampulla was typically raised symmetrically, with a patulous orifice, and was otherwise covered by stretched normal duodenal mucosa. However, in 6 cases, the protrusion of the intra-ampullary tumor to the duodenal surface gave the impression of an "ampullary-duodenal tumor," with the accurate diagnosis of IAPN established only by microscopic correlation illustrating the abrupt ending of the lesion at the edge of the ampulla. Microscopically, the preinvasive component often revealed mixed phenotypes (44.4% predominantly nonintestinal). The invasion was common (94%), typically small (mean=1.2 cm), primarily pancreatobiliary-type (75%), and showed aggressive features (lymphovascular invasion in 66%, perineural invasion in 41%, high budding in 30%). In 6 cases, the preinvasive component was pure intestinal, but the invasive component was pancreatobiliary. LN metastasis was identified in 42% (32% in those with ≤1 cm invasion). The prognosis was significantly better than ampullary-ductal carcinomas (median: 69 vs. 41 months; 3-year: 68% vs. 55%; and 5-year: 51% vs. 35%, P =0.047). In conclusion, unlike ampullary-duodenal carcinomas, IAPNs are often (44.4%) predominantly nonintestinal and commonly (94%) invasive, displaying aggressive features and LN metastasis even when minimally invasive, all of which render them less amenable to ampullectomy. However, their prognosis is still better than that of the "ampullary-ductal" carcinomas, with which IAPNs are currently grouped in CAP protocols (while IAPNs are kindreds of intraductal tumors of the pancreatobiliary tract, the latter represents the ampullary counterpart of pancreatic adenocarcinoma/cholangiocarcinoma).
RESUMO
PURPOSE: To investigate the performance of computed tomography (CT) in the local staging of colon cancer in different segments, with emphasis on parameters that have been found to be significant for rectal cancer, namely, extramural venous invasion (EMVI) and tumor deposits (TDs). METHODS: CT and pathology data from 137 patients were independently reviewed by radiology and pathology teams. The performance of CT in categorizing a given patient into good, versus poor prognostic groups was assessed for each segment, as well as the presence of lymph nodes (LNs), TDs and EMVIs. Discordant cases were re-evaluated to determine potential sources of error. Elastic stain was applied for EMVI discordance. RESULTS: The T staging accuracy was 80.2%. For T stage stratification, CT performed slightly better in the left colon, and the lowest accuracy was in the transverse colon. Under-staging was more common (in 12.4%), and most of the mis-staged cases were in sigmoid colon. According to the first comprehensive correlative analysis, the sensitivity, specificity, and accuracy of CT for detecting TDs were found to be 57.9%, 92.4%, 87.6%, respectively. These figures were 44.7%, 72.7%, and 63.5% for LN, and 58.5%, 82.1% and 73% for EMVI. The detection rate was better for multifocal EMVI. The detection rate was also comparable (although substantially underestimated) for LNs, with the half of the LNs missed by CT being < 5 mm. Four patients that were classified as TD by CT, disclosed to be LNs by pathology. Correlative analysis led to refinement of the pathology criteria, with subsequent modifications of the initial reports in 13 (9.5%) patients. CONCLUSION: Overall, CT performed well in the evaluation of colon cancer, as did TD and EMVI. It is advisable to include these parameters in CT-based staging. Radiologists should be aware of the pitfalls that occur more commonly in different segments.
