Background
Most patients presenting with symptoms of esophageal cancer (EC) have advanced disease. Even with resection, the cure rate is extremely low due to local recurrence and metastatic disease. Early detection and effective therapeutic intervention are essential to improve survival.
Aims
This study tested the hypothesis that the presence of EC modulates concentrations of specific plasma proteins and peptides, potentially allowing discrimination between EC and controls based on mass spectrometric analysis of the respective plasma proteomes.
Methods
Blood samples from 79 esophageal cancer patients and 40 age-matched normal subjects were processed to plasma, and protein/peptide sub-fractions were isolated using HIC8 or WAX-derivatized superparamagnetic beads. Triplicate matrix-assisted laser desorption time-of-flight mass spectra were acquired for specific plasma fractions from each subject.
Results
HIC8 and WAX-derivatized plasma eluates yielded 79 and 77 candidate features, respectively, and a Random Forest algorithm identified a subset of features whose peak intensities allowed discrimination between cancer patients and controls. Areas under the curve in receiver operating characteristic curves for HIC8 spectra were 0.88 and 0.83 for WAX spectra. The combined feature set discriminated EC from control plasma with 79 % sensitivity and 79 % specificity, with positive and negative test likelihood ratios of >14 and 0.17, respectively.
Conclusions
These data lay the foundation for the development of a clinically useful test for esophageal cancer based on statistical analysis of proteomic spectra of patient plasma samples. This approach will be validated by analysis of larger patient cohorts, development of cancer-specific classifiers, and assessment of racial origin imbalances.