[9] Four of the six had evidence of gastric varices

on imaging studies. One of these had presented with hematemesis from gastric varices and another patient had findings of portal hypertensive gastropathy on endoscopic evaluation. Takahashi et al. reported vascular involvement in 44% (11/25) of AIP patients.[10] In this issue of the Journal, these findings are extended by Ishikawa et al. who report on the prevalence and long term outcome of peripancreatic vascular involvement in 54 AIP p38 MAPK activation patients.[11] Peripancreatic vascular involvement of any form was observed using multidetector computed tomography (CT) scan in 24 (44%) patients. There was occlusion or stenosis of splenic vein in 22, superior mesenteric or portal veins in 13, and development of collaterals in 18 patients. This prevalence is higher than reported by our group,[9] but similar to that of the Mayo group.[10] The difference may be due to a broader definition for vascular involvement. CP-673451 mw The only patient characteristic increasing the risk of vascular involvement was diffuse

enlargement of the pancreas (56 versus 22%). Corticosteroid treatment was administered to 20 of the 24 patients with vascular involvement who also had symptoms (jaundice or abdominal pain). On a follow-up CT scan, improvement in vascular lesions, including resolution of portal vein thrombosis, was seen in 14 of 16 patients treated with steroids. Of the four patients with vascular involvement who did not receive treatment, interval worsening of vascular lesions occurred in two, while no change was seen in the other two patients. Among patients who did not receive steroids, new vascular lesions were not detected in any

patient who had a follow up CT scan. The pathogenesis of vascular involvement in AIP remains unclear. Similar to pancreatitis from other causes, a pre-existing hypercoagulable factor does not appear to play a role.[9, 11] One possible explanation is the extension of inflammatory changes from the pancreas to the surrounding vessels. This possibility is supported by the preferential involvement of vessels in close proximity to the inflamed pancreas. It is important to note that AIP patients Rucaparib typically do not develop pancreatic necrosis or pancreatic/peripancreatic fluid collections. The other possibility is involvement of vessels by the primary inflammatory process typical for IgG4 related systemic disorder. Although Ishikawa et al. did not find this to be present in the patient who underwent pancreatic resection,[11] Kamisawa et al. have demonstrated obliterative phlebitis in pancreatic and peripancreatic veins in resection specimens of AIP patients.[8] In addition to defining the pathogenesis and factors that determine vascular involvement, there are three other questions that should be answered by future investigations.