Venous invasion in colorectal cancer: impact of an elastin stain on detection and interobserver agreement among gastrointestinal and nongastrointestinal pathologists

Abstract

Venous invasion (VI) is an independent prognostic indicator in colorectal cancer and may prompt consideration for adjuvant chemotherapy in patients with stage II tumors. Recent evidence suggests that VI is underreported in colorectal cancer and that detection may be enhanced by an elastin stain. This study aimed (1) to determine the impact of an elastin stain on VI detection and on interobserver agreement between gastrointestinal (GI) and non-GI pathologists, and (2) to identify factors associated with increased VI detection. Forty hematoxylin and eosin (H&E)-stained slides were circulated to 6 GI and 6 non-GI pathologists who independently assessed the VI status as positive, negative, or equivocal. Six weeks later, 40 corresponding Movat-stained slides were recirculated together with the original H&E slides and reassessed for VI status. Detection of VI was >2-fold higher with a Movat stain compared with an H&E stain alone (46.4% vs. 19.6%, P=0.001). GI pathologists detected VI more frequently than non-GI pathologists on both H&E (30.0% vs. 9.2%, P=0.029) and Movat (58.3% vs. 34.6%, P=0.018) stains. There was higher interobserver agreement in the case of a Movat stain, particularly for extramural VI (H&E: κ=0.23 vs. Movat: κ=0.41). A poststudy survey indicated that GI pathologists and non-GI pathologists applied similar diagnostic criteria but that GI pathologists more frequently applied "orphan arteriole" and "protruding tongue" signs as diagnostic clues to VI. This study confirms that VI is underdetected on H&E and highlights the role of elastin staining in improving VI detection and interobserver agreement. Strategies to improve VI detection are warranted.