1. ¹Department of Pathology, Tohoku University Graduate School of Medical Science, Sendai, Japan
2. ²Department of Pathology, Jikei University Medical School, Tokyo, Japan
3. ³Department of Pathology, Kawasaki Medical School Hospital, Kurashiki, Japan
4. ⁴Department of Obstetrics and Gynecology, Kawasaki Medical School Hospital, Kurashiki, Okayama, Japan
5. ⁵Department of Pathology, Tohoku University Hospital, Sendai, Japan
6. ⁶Laboratory of Anatomic Pathology, Kyoto University Hospital, Kyoto, Japan
7. ⁷Department of Obstetrics and Gynecology, Tohoku University Graduate School of Medical Science, Sendai, Japan
8. ⁸Department of Gynecology, Miyagi Prefectural Cancer Center, Natori, Japan
9. ⁹Department of Pathology, Miyagi Prefectural Cancer Center, Natori, Japan
10. ¹⁰Department of Pathology, Sendai City Hospital, Sendai, Japan

Correspondence: Dr Y Mikami, MD, Department of Pathology, Division of Histopathology, Tohoku University Graduate School of Medical Science, 2-1 Seiryo-machi, Aoba-ward, Sendai, Miyagi 980-8575, Japan. E-mail: mika@patholo2.med.tohoku.ac.jp

Received 28 July 2003; Revised 7 October 2003; Accepted 25 March 2004; Published online 14 May 2004.

Abstract

Gastrointestinal phenotype in cervical adenocarcinomas was examined by immunohistochemistry and correlated with morphologic features. Antibody panels included anti-MUC2, MUC6, CD10, chromogranin A (CGA) and HIK1083. In addition, expression of p16^INK4, a cyclin-dependent kinase inhibitor which is expressed in a variety of high-risk HPV-related conditions, was studied. A total of 94 invasive adenocarcinomas including 20 minimal deviation adenocarcinomas (MDAs) and 72 adenocarcinomas in situ (AIS) were examined. MDAs were most frequently positive for HIK1083 and/or MUC6, two representative gastric markers, with a rate of 95%, followed by intestinal-type adenocarcinomas (IAs) with a rate of 85% whereas only 27% of 56 usual endocervical-type adenocarcinomas (UEAs) were positive. MUC2, a goblet cell marker, was positive in 85% and 25% of IAs and MDAs, respectively, while in only 14% of UEAs. CD10 was positive in 15% of IAs, indicating incomplete intestinal differentiation without a brush border in most of the cases. CGA-positive cells were frequently seen in MDAs and IAs with rates of 60% and 62%, respectively. Nuclear and cytoplasmic p16^INK4 positivity was identified in 93% of UEAs, whereas 30% of MDAs were positive for p16^INK4. Results in AISs were comparable to their invasive counterparts, but morphologically usual-type AISs identified in eight cases of MDA were frequently positive for HIK1083 (75%) and MUC6 (63%), and p16^INK4. Of note was the existence of lobular endocervical glandular hyperplasia (LEGH) with atypical features including cytologic abnormalities, and/or papillary projection, which were identified in this study in pure form (n=3) or in association with MDAs (n=6), but not in cases of other types of adenocarcinomas. These observations indicate that gastrointestinal phenotype is frequently expressed in MDAs and IAs, and there seems to be a possible link between MDA, and LEGH and morphologically usual-type AIS with gastric immunophenotype in histogenesis. Frequent absence of p16^INK4 expression in MDAs suggests a possibility that high-risk HPV does not play a crucial role in development of MDAs, in contrast to the majority of endocervical adenocarcinomas. p16^INK4 immunohistochemistry appears to be a promising diagnostic tool, but pathologists should be aware of frequent negative staining in MDAs, which can be a source of erroneous diagnosis.