A new tumor-associated antigen, MN, has been shown to be expressed in virtually all cervical carcinomas and the majority of cervical intraepithelial neoplasia, but not in normal cervices (S. Y. Liao et al., Am. J. Pathol., 145: 598-609, 1994). Therefore, we postulated that the exfoliative cells in cervical Papanicolaou (Pap) smears would reflect the MN immunoreactivity seen in the tissue sections, and high levels of MN expression in the exfoliative cells would indicate the presence of dysplasia in the cervix. A total of 305 cervical Pap smears, with histological confirmation, representing all categories of the Bethesda System, were immunohistologically examined. We found that high levels of MN expression in exfoliative cells were not restricted to the dysplastic cells but were observed also in the normal endocervical cells (NECs) when dysplasia was present in the tissue biopsies. Overall, the rates of positive MN immunostaining of the dysplastic cells in low- and high-grade squamous intraepithelial lesions and invasive carcinoma were 35 (65%) of 54, 44 (77%) of 57, and 12 (92%) of 13, respectively. However, diffuse MN immunoreactivity of the atypical and/or dysplastic endocervical columnar cells was seen in all cases (100%) of adenocarcinoma in situ (AIS; n = 23) and adenocarcinomas (n = 8). In the groups with cytological diagnoses of atypical squamous cells or atypical glandular cells of undetermined significance (ASCUS and AGUS, respectively), MN positivity was seen in 47% of ASCUS (22/47) and 55% of AGUS (12/22). Dysplastic tissues were identified in all MN-positive cases. In contrast, all MN-negative atypical Pap smears were confirmed histologically to be benign cervix with one exception, in which the cytological diagnosis was ASCUS and focal low-grade squamous intraepithelial lesions were found in the cervix. The study also included 89 cases with cytological diagnoses of within normal limits/benign cellular changes. Among these, 10 Pap smears expressed diffuse MN antigen in the NEC, and dysplasia (8 cases of low-grade squamous intraepithelial lesions, 2 AIS) was found in the cervices. None of MN-negative cases with "within normal limits" cytology contained dysplastic cervices. Therefore, it would seem that diffuse MN antigen expression in the NEC may be an indicator of cervical dysplasia. Thus, MN antigen might serve as an early biomarker of cervical neoplasia. The combination of detection via cytology and MN immunostaining resulted in no false negatives and also discriminated between cellular atypia due to benign reactive changes versus cellular atypia due to dysplasia in the category of ASCUS and AGUS. In particular, it was found in the AGUS group that diffuse MN immunostaining restricted to atypical columnar cells was diagnostic for AIS. These findings indicate that MN antigen expression is an important diagnostic biomarker of glandular neoplasia and a valuable adjunct to cytological diagnosis of ASCUS and AGUS.