CONTEXT:Vascular invasion is an important diagnostic and prognostic feature of hepatocellular carcinoma (HCC) in cirrhosis. Intravascular free-floating tumor clusters (IvCs) of HCC are found histologically in the vicinity of HCC. Thrombus formation is not seen morphologically in association with these IvCs, which are usually covered by endothelium.OBJECTIVE:Our hypothesis is that these IvCs are the result of a nondestructive form of vascular invasion by HCC, and we tried to define this aspect of microvascular invasion more accurately.DESIGN:Tissue sections were stained with hematoxylin-eosin, and consecutive sections were stained for fibrin (Martius scarlet blue, fibrinogen), platelets (factor XIIIa), smooth muscle actin, and endothelium (CD34). We studied cirrhotic livers removed at transplantation between 1997 and 1999. Of the livers studied, 35 of 81 consecutive cirrhotic livers contained HCC, and 17 showed microscopic vascular invasion. Five of these 17 cases showed IvCs and were subjected to the study.MAIN OUTCOME MEASURE:Presence or absence of thrombus formation in association with IvC.RESULTS:Usually, IvCs were covered by endothelium, and no associated thrombus formation was seen. In 1 case of HCC, thrombus formation was seen focally in association with disruption of the endothelial coating.CONCLUSIONS:We propose that the endothelial-lined trabecular structure of HCC everts, frondlike, via vascular structures within the tumor capsule into peritumoral vascular lumens without destruction of the endothelial coating. This may protect these HCC tumor projections from thrombus formation but may also act as a barrier to tumor extravasation, and this may be exploited from a therapeutic point of view.

I have read with interest the paper by Rene Buesa in your December edition and felt that some comment was required. Without wishing to accuse the author of re-inventing the wheel, I should like to point out that the problem of staining paraffin sections with Azure-Eosin mixtures was solved in 1932 by R.D. Lillie and J.G. Pasternack (I), and yet these workers are not quoted or referenced in the new paper. Lillie and Pasternack determined that

AIMS:To examine the vascular changes occurring in three archival cases of acute multiple sclerosis, and to provide immunohistochemical evidence of early endothelial cell activation and vascular occlusion in this condition.METHODS:Central nervous system tissues from three cases of acute active multiple sclerosis and six non-inflammatory controls were stained using the following methods: haematoxylin and eosin, Luxol fast blue, cresyl violet, Bielschowsky's silver, and reticulin. Tissues were also immunostained with specific antibodies against collagen type IV, factor XIIIa, class II antigens, glial fibrillary acidic protein, and fibrinogen.RESULTS:Early vascular endothelial cell activation which may progress to vasculitis and vascular occlusion including class II antigen expression and fibrin deposition were identified. The vascular changes were seen prior to cerebral parenchymal reaction and demyelination, and were not seen in control cerebral tissues.CONCLUSION:It is proposed that vascular endothelial cell activation may be an early and pivotal event in the evolution of multiple sclerosis, and that demyelination may have an ischaemic basis in this condition. The vascular endothelium may contain an early element in the evolution of multiple sclerosis.