Although on a pancreatogram the main pancreatic duct looks stenotic and irregular in type 1 AIP, ductal lumens are histologically always open, and the duct epithelium is intact. The ductal inflammation consists of lymphoplasmacytic infiltration that surrounds the epithelium. It can be easily identified in the interlobular pancreatic ducts, but the intralobular ducts are also involved.

The ductal inflammation typically consists of a collar of lymphoplasmacytic infiltration with storiform fibrosis, which looks like a thickened duct wall (Fig. 4.1c). However, there are no structures that are histologically identified as a wall in the normal pancreatic ducts, and thus, they are considered to be a pathological structure caused by the inflammation itself.

The ductal inflammation can also be a thin inflammatory collar of lymphoplasmacytic infiltration just beneath the epithelium, which is surrounded by a fibrotic band. Storiform fibrosis is absent in this pattern.

Although it has not been emphasized as a significant finding in previous reports, the pancreatic lobules are inflamed in a peculiar way. The lobular inflammation is characterized by acinar cell loss and lymphoplasmacytic infiltration. Interestingly though, this is in a fashion in which the original lobular contour and size are preserved (Fig. 4.1d). This is in contrast to the lobular atrophy or loss seen in pancreatitis due to other causes. The lobular contours are sharply defined by the interlobular fibrosis, giving rise to the nodular appearance of the pancreatic parenchyma. This is especially evident when special stains for fibrosis such as Verhoeff-van Gieson stain are performed. In addition, the lobular appearance may resemble storiform fibrosis with or without true fibrosis. The pancreatic lobules are often destructive at the margin of pancreatic parenchyma, which is explained in the following section. Rarely, the pancreatic lobules are atrophic in cases of type 1 AIP with regressive changes.

The lack of neutrophilic infiltration is a feature that distinguishes type 1 from type 2 AIP. However, in about one-third of cases, neutrophils are observed in the lobules of type 1 AIP. This is most likely due to a secondary inflammation that is caused by ductal strictures. In this instance, the distinction between type 1 and type 2 AIP should be made based on the findings of other histological features in the lobules as well as the inflammation seen in the connective tissues.

Venous involvement is one of the most well known and unique histological features of type 1 AIP. Venous walls are often infiltrated and destroyed by the inflammatory process, which is histologically similar to that seen around the veins. This inflammatory process further involves venous lumens and finally obliterates them (obliterative phlebitis, Fig. 4.1e). Obliterative phlebitis is common in small veins (venules) that cannot be identified macroscopically. Although obliterated venules are difficult to identify on HE-stained slides, nodular inflammatory lesions adjacent to arterioles are the clue to identify them because the arteries and veins run parallel in the pancreas. In addition, obliterative phlebitis is common and involves relatively larger venules, which makes obliterative phlebitis easily identifiable even on HE-stained slides. According to our study, obliterative phlebitis seen in the larger-sized venules (>100 μm in diameter, as a guide) is more diagnostic for type 1 AIP, and smaller venular lesions are nonspecific [20]. Simple fibrous obstruction without inflammatory cells or only inflammatory cell infiltration without fibrosis in the veins should not be included in obliterative phlebitis in type 1 AIP, because those findings may be seen in pancreatic carcinomas and alcoholic chronic pancreatitis [20]. Obliterative phlebitis with storiform fibrosis is typical for type 1 AIP. When the venous inflammation occurs in a larger vein, such as the splenic vein or portal vein, the luminal occlusion does not occur. This is because they are too large to be occluded completely. Even so, I suggest that this finding is also unique to type 1 AIP. In this situation, it is common to observe a local well-marginated involvement of a vein with storiform fibrosis.

Arteries are also involved in type 1 AIP. This is usually limited to the border between the media and adventitia, and it is thus similar to the periaortitis/periarteritis seen in the larger arteries [21]. Because this involvement is continuous with the surrounding inflammation, it is difficult to determine whether the inflammation is present or not in the arteries themselves. Thus, the true incidence of periarteritis in type 1 AIP has been underestimated. Arteries are not completely occluded by the inflammation, probably due to an anatomical reason, that is, the thicker smooth muscle layer in the arteries.

Perinervous inflammation is common and may be prominent in type 1 AIP. This finding alone, however, is not specific for type 1 AIP because it can also be observed in alcoholic chronic pancreatitis and pancreatic carcinomas.

The inflammation in the ducts may be similar to but apparently different from that of type 1 AIP. Although the presence and absence of GEL in type 2 and type 1 AIP, respectively, is an important histological feature, there are other clues that distinguish the two subtypes. Because the duct epithelium itself is inflamed in type 2 AIP, the epithelial cells may be degenerative or may contain enlarged and vesicular nuclei which reflect a regenerative nature of the epithelium. Such epithelial changes are usually absent in type 1 AIP. Though there is typically a collar of lymphoplasmacytic infiltration surrounding the duct epithelium in type 2 AIP, it lacks storiform fibrosis. In some cases of type 2 AIP, the entire ducts appear to be entrapped by the densely packed inflammatory cells (Fig. 4.2a). This does not happen in type 1 AIP. In addition, the inflamed ducts in type 2 AIP seem tortuous and star shaped rather than the straight- or round-shaped ducts typically seen in type 1 AIP.