Identifying the Primary Tumor Site and Distinguishing False-positives in Patients With Elevated Serum Carcinoembryonic Antigen

Abstract

Background/aim: Carcinoembryonic antigen (CEA) is a tumor marker that is frequently evaluated clinically for gastrointestinal and lung cancers. CEA is a glycoprotein antigen, and only the "value" measured by enzyme-linked immunosorbent assay is provided in the clinical setting. At present, no method has been established to indicate whether the value is a false-positive elevation or whether there is a primary cancer site. To obtain clues on how to identify the originating site in patients with cancer with high CEA levels and to identify CEA false-positives in healthy individuals, we conducted an exploratory study.

Patients and methods: A pilot study was performed using the multivariate analysis method and principal component analysis-discriminant analysis on proteomic results obtained using liquid chromatography-mass spectrometry (LC/MS) in two patients with lung cancer, one patient with gastric cancer, and one healthy control individual.

Results: No differences in specific proteins associated with high CEA levels were detected between lung and gastric cancers using LC/MS. Therefore, we performed statistical analysis using principal component analysis-discriminant analysis to determine whether there were differences in the protein signal patterns obtained using LC/MS. The results showed that the plots obtained for each patient and the healthy control were located in different quadrants of the four-quadrant matrix scatter plot.

Conclusion: Our results suggest the possibility of visually differentiating the primary tumor site in patients with elevated CEA levels. This method may also help recognize false-positive CEA results.

Keywords: CEA; Tumor marker; carcinoembryonic antigen; differential diagnosis.

Figure 1

Results of liquid chromatography/mass spectrometry-total ion current chromatogram in three patients (two with lung adenocarcinomas and one with gastric cancer) and a healthy control individual.

Figure 2

Results of principal component analysis (PCA)-discriminant analysis (DA) of serum samples. (A) Results of PCA score plots of the serum samples from two patients with lung adenocarcinoma, one patient with gastric cancer (circled in red), and a healthy control individual. (B) PCA-DA loading plots of the serum samples showing that the samples from the two patients with lung cancer, the patient with gastric cancer, and the healthy control individual were visually separated into distinct quadrants.