Proteomics-driven noninvasive screening of circulating serum protein panels for the early diagnosis of hepatocellular carcinoma

Abstract

Abstract Early diagnosis of hepatocellular carcinoma (HCC) lacks highly sensitive and specific protein biomarkers. Proteomics-driven discovery of tumor biomarkers is an important direction for omics study. Here, we described a staged mass spectrometry (MS)-based discovery-verification-validation proteomics workflow to explore serum proteomic biomarkers for HCC early diagnosis in 662 individuals (373 HCC patients and 289 non-HCC patients). Our workflow reproducibly quantified 451serum proteins using a data independent acquisition mass spectrometry (DIA-MS) strategy from discovery cohort, and proteins with significantly altered abundance in HCC were validated as candidates in an independent validation cohort using targeted proteomics based on parallel reaction monitoring (PRM). Machine learning models determined as P4 serum protein-panels (two serum proteomics biomarkers: HABP2, CD163 and two clinical used serum biomarkers: AFP, PIVKA-II) could clearly distinguish HCC patients from LC patients in an independent validation cohort (AUC 0.979, sensitivity 0.925, specificity 0.915), outperforming existing clinical prediction strategies (p < 0.05). Moreover, the P4 panels showed high sensitivity in AFP negative (0.857) HCC patients and PIVKA-II negative HCC patients (0.813). Most importantly, the P4 panels were validated to be perfectly accurate in predicting the conversion of LC to HCC (accuracy: 100.0%) with predicting HCC at a median of 12.6 months prior to imaging in a prospective external validation cohort, which was superior to existing clinical prediction strategies. These results suggested that proteomics-driven serum biomarker discovery provided a valuable reference for the liquid biopsy, and had great potential to improve early diagnosis of HCC.