Compared With Elastin Stains, h-Caldesmon and Desmin Offer Superior Detection of Vessel Invasion in Gastric, Pancreatic, and Colorectal Adenocarcinomas

Background. The presence of vessel invasion is considered indicative of a poor prognosis in many malignant tumors. We aimed to compare the sensitivity of elastin stains (van Gieson's and orcein methods) with 2 smooth muscle markers (h-caldesmon and desmin) in gastric, pancreatic, and colorectal adenocarcinoma specimens. Materials and Methods. We used 27 (29.3%) gastric, 35 (38.0%) pancreatic, and 30 (32.6%) colorectal resection specimens. We applied a provisional classification of vessel invasion patterns: type A, a focus with a nearby artery unaccompanied by a vein; type T, a focus at the invasive front without an unaccompanied artery; and type X, foci that only appeared by any of the 4 stains used. Results. There were 369 foci. The smooth muscle markers were more sensitive than the elastin stains, and h-caldesmon more sensitive than desmin, in all types. Among the 139 type A foci, 33 (23.7%) were positive by desmin and h-caldesmon, whereas the elastin stains were not (P = .001). h-Caldesmon was the only positive marker in 11 (7.9%; P = .011). Among the 78 type T foci, 21 (26.9%) were positive by desmin and h-caldesmon, when both elastin stains were negative (P = .000). In 16 (20.5%) foci, h-caldesmon was the only positive marker (P = .002). Among 152 type X foci, 91 (59.9%) were positive by all markers, 26 (17.1%) by both desmin and h-caldesmon, and 9 (5.9%) by only the 2 elastin stains (P = .001). Conclusion. We recommend these stains for suspect foci in gastric, pancreatic, and colorectal adenocarcinoma specimens. They might highlight both predictable and unpredictable foci.