To study the association of qat chewing with the occurrence of oral cancer, the frequency of oral cancer among whole body cancers and the patients' histories of tobacco consumption and qat chewing were examined in Yemen where qat chewing has been most popular. All primary malignant tumors listed in the surgical pathology files at Al-Thawra Hospital, University of Sana'a, in the year 2004 were analyzed, and the patients' histories of tobacco consumption and qat chewing were examined. A total of 649 cases of primary malignant tumors (348, 53.6% males and 301, 46.4% females) were extracted. Oral cancer was the most frequent body cancer in both males (17.2%) and females (19.6%). Squamous cell carcinoma (SCC) was the most frequent oral cancer (84%), and the tongue (42%), gingiva (23%) and buccal mucosa (20%) were the most common sites. Among the 119 patients with oral cancer, information on chewing habits and smoking was obtained in 92 patients (77.3%). There were 70 tobacco chewers (76.1%), 55 qat chewers (59.8%), and 22 smokers (23.9%). Simultaneous chewing of tobacco and qat was found in 48 cases (52.2%). The present survey has disclosed for the first time that oral SCC is the most frequent cancer in this study area in Yemen, and that the high relative frequency of oral SCC may be related to the habits of chewing tobacco and qat.

Although the histopathogenetic process of keratin pearls is still poorly understood, acceleration of keratinization in squamous cell carcinoma (SCC) cells may represent one possible therapeutic avenue. Based on our histopathological observations, we have hypothesized that SCC cells are keratinized by phagocytosis of extravasated erythrocytes. To confirm this hypothesis, we firstly examined immature keratin pearls in oral carcinoma in situ (CIS) and mature ones in SCC by immunohistochemistry. Concentric dyskeratotic cells in CIS keratin pearls became positive for keratin (K) 10, K17, heme oxygenase-1 (HO-1), or protease activated receptor-2 (PAR-2), a candidate regulator for hemophagocytosis. When ZK-1 cells, an SCC cell system, were incubated with human peripheral blood erythrocytes, or with crude and purified hemoglobins (Hbs), their erythro-hemophagocytotic activities were confirmed by immunofluorescence. Immunofluorescence signals for K10, K17, and HO-1 were enhanced due to hemophagocytosis in time-dependent manners. mRNA expression levels for the three molecules were most enhanced by purified Hb, followed by crude Hb and erythrocytes. K17/K10 mRNA expression levels were more elevated when PAR-2 was activated in ZK-1 cells. The results indicated that immature and mature keratin pearls in CIS and SCC were generated by oxidative stresses derived from erythro-hemophagocytosis, which might mediate HO-1 expression and be regulated by PAR-2. Thus, hemorrhage from the rupture of blood vessels can be one of the triggers for keratin pearl formation in oral CIS and SCC.

It can be difficult to make a certain diagnosis in case of an oral borderline malignant lesion on hematoxylin-eosin-stained sections only. Furthermore, assessment of surgical margins of borderline lesions is difficult with the naked eye. We set out to determine the topographical distribution of capillary blood vessels within the epithelial zone and to assess its use as an aid for histopathological diagnosis and a framework for clinical assessment of lesional margins using optical techniques, such as narrow-band imaging (NBI) endoscopy. Capillary blood vessels entrapped in the epithelial compartment, which we have designated as intraepithelially entrapped blood vessels (IEBVs), were examined for their frequency, location, and shape in normal mucosa, dysplasia, and carcinoma in-situ (CIS) of the tongue using immunohistochemistry for CD31 and type IV collagen. When counted per unit length of epithelial surface, IEBVs increased in number significantly in CIS (5.6 ± 2.8), which was two times more than in normal (1.9 ± 1.6) and dysplastic (2.4 ± 1.5) epithelia. In addition, IEBVs in CIS had compressed shapes with occasional obstruction or collapse with hemorrhage and were arranged perpendicular to and extending up to the epithelial surface. These characteristic IEBVs in CIS were considered to be generated by complex expansion of rete ridges due to carcinoma cell proliferation within the limited epithelial space determined by the basement membrane. The recognition of IEBVs was helpful in the differential diagnosis of oral CIS, and the present data provide a valuable frame of reference for detecting oral CIS areas using such NBI-based optical devices.