Abstract
BACKGROUND: It is difficult to predict gene mutations individually based on clinical background alone. Tumor markers may help to predict each gene mutation. Identifying tumor markers that can predict gene mutation will facilitate timely genetic testing and intervention. METHODS: We selected 134 cases of advanced or recurrent ALK-positive and 172 cases of advanced or recurrent EGFR-positive lung cancer from our clinical database. The cutoff values for the tumor markers were defined as 5.0 ng/mL or higher for carcinoembryonic antigen (CEA) and 3.5 ng/mL or higher for soluble fragment of cytokeratin 19 (CYFRA21-1) in accordance with the institutional standards. A positive CYFRA21-1:CEA ratio was defined as 0.7 or higher. RESULTS: The CEA-positivity rate was 49% for ALK-positive lung cancers and 73% for EGFR-positive lung cancers, which was significantly different (p < 0.001). The CYFRA21-1 positivity rate was significantly higher in ALK-positive lung cancer (36%) compared with EGFR-positive lung cancer (23%) (p = 0.034). The median CYFRA21-1:CEA ratio was 0.395 for the ALK group, which was significantly higher compared with 0.098 for the EGFR group (p < 0.001). These trends were similar when excluding histology other than adenocarcinoma. The median time-to-treatment failure (TTF) for initial tyrosine kinase inhibitor (TKI) therapy was 308 days for the high CYFRA21-1:CEA ratio group and 617 days for the low CYFRA21-1:CEA ratio group for ALK-positive lung cancer (p = 0.100). CONCLUSIONS: A higher proportion of patients with ALK-positive lung cancer were CYFRA21-1 positive and had higher CYFRA21-1:CEA ratios compared with EGFR-positive lung cancer patients.