Serum Tumor Markers and Cervical Intraepithelial Neoplasia

Abstract

Objective: To investigate the relationship between the serum levels of the tumor markers squamous cell carcinoma antigen (SCC), carcinoma antigen (CA125) and carcinoembryonic antigen (CEA) and cytologic, colposcopic and histologic findings in women with cervical intraepithelial neoplasia (CIN). We also assessed whether a coefficient based on the three markers, namely the quotient (SCC × CEA)/CA125, could detect women with histologically proven high-grade lesions missed by cytology or colposcopy. Materials and Methods: Ninety-eight women attending our colposcopy clinic underwent cytologic and colposcopic examination and measurement of the serum levels of the three tumor markers. Seventy women underwent cervical biopsy and/ or endocervical curettage. The cytologic, colposcopic and histologic findings were classified according to severity. Analysis of variance, univariate analysis and multiple regression models were used for the statistical evaluation. Results: The cytologic results showed a statistically significant positive correlation with SCC values (p = 0.023), a negative relation with CA125 values (p = 0.050) and no correlation with CEA values. The colposcopic findings showed no correlation with SCC values, but the values of CA125 and CEA showed significant differences among the groups (p = 0.034 and p = 0.009, respectively). The analysis of tumor marker levels according to histologic findings showed no significant differences for CEA while SCC and CA125 showed the same significant correlations with the presence and grade of CIN (p = 0.005 and p 0.040, respectively). The coefficient SCC × CEA/CA125 showed a highly significant direct correlation with the cytologic (p = 0.0002), colposcopic (p = 0.0018) and histologic findings (p = 0.0005). Univariate analysis showed a positive correlation between cytology/ histology and SCC (p<0.001), a negative correlation among all morphologic parameters and CA125 (p < 0.001), and no correlation between them and CEA. These findings remained statistically significant in multiple regression models with cytologic, colposcopic or histologic findings as independently correlated variables. In contrast to the individual tumor marker values the coefficient SCC × CEA/CA125 remained in multiple regression analysis as a statistically significant variable for predicting the underlying histologic diagnosis (p <0.05). Conclusion: Serum levels of SCC, CA125 and CEA combined as a quotient would probably allow, without invasive procedures such as biopsies or endocervical curettages, the detection of masked high-grade intraepithelial lesions and invasive cancers in patients with equivocal cytologic and/or colposcopic findings and thus reduce the percentage of premalignant or malignant lesions which elude diagnosis.