AI-Powered Deep Visual Proteomics Reveals Critical Molecular Transitions in Pancreatic Cancer Precursors

Area of Science:

• Oncology
• Proteomics
• Computational Pathology

Background:

• Pancreatic ductal adenocarcinoma (PDAC) progression involves precursor stages, but the underlying protein dynamics are not well understood.
• Identifying early molecular events is crucial for improving detection and treatment strategies for PDAC.

Purpose of the Study:

• To comprehensively profile protein expression across early precursor lesions of PDAC using an integrated multi-omics approach.
• To uncover the molecular programs associated with the transition from normal pancreatic ducts to invasive carcinoma.

Main Methods:

• Deep Visual Proteomics, combining computational pathology, laser microdissection, and mass spectrometry.
• Analysis of protein profiles from normal ducts, acinar-to-ductal metaplasia (ADM), low- and high-grade pancreatic intraepithelial neoplasia (PanIN), and invasive PDAC.
• Quantification of over 9,000 proteins from approximately 100 cells per region.

Main Results:

• A molecular field effect was observed in histologically normal ducts adjacent to tumors.
• Proteomic profiles of low-grade PanINs varied depending on the cancer context.
• Four distinct stage-associated molecular programs were identified, including early stress adaptation, immune engagement, metabolic reprogramming, and mitochondrial remodeling.
• KRAS hotspot mutant peptides were detected in precursor lesions from cancer-free individuals.

Conclusions:

• Molecular reprogramming occurs before histological transformation in PDAC development.
• Early detection of PDAC may be possible by identifying these pre-invasive molecular signatures.
• The identified protein programs provide insights into the early pathogenesis of pancreatic cancer.