A Clinically Applicable 24-Protein Model for Classifying Risk Subgroups in Pancreatic Ductal Adenocarcinomas using Multiple Reaction Monitoring-Mass Spectrometry
Open Access
- 16 February 2021
- journal article
- research article
- Published by American Association for Cancer Research (AACR) in Clinical Cancer Research
- Vol. 27 (12), 3370-3382
- https://doi.org/10.1158/1078-0432.ccr-20-3513
Abstract
Pancreatic ductal adenocarcinoma (PDAC) subtypes have been identified using various methodologies. However, it is a challenge to develop classification system applicable to routine clinical evaluation. We aimed to identify risk subgroups based on molecular features and develop a classification model that was more suited for clinical applications. We collected whole dissected specimens from 225 patients who underwent surgery at Seoul National University Hospital [Seoul, Republic of Korea (South)], between October 2009 and February 2018. Target proteins with potential relevance to tumor progression or prognosis were quantified with robust quality controls. We used hierarchical clustering analysis to identify risk subgroups. A random forest classification model was developed to predict the identified risk subgroups, and the model was validated using transcriptomic datasets from external cohorts (N = 700), with survival analysis. We identified 24 protein features that could classify the four risk subgroups associated with patient outcomes: stable, exocrine-like; activated, and extracellular matrix (ECM) remodeling. The “stable” risk subgroup was characterized by proteins that were associated with differentiation and tumor suppressors. “Exocrine-like” tumors highly expressed pancreatic enzymes. Two high-risk subgroups, “activated” and “ECM remodeling,” were enriched in terms such as cell cycle, angiogenesis, immunocompetence, tumor invasion metastasis, and metabolic reprogramming. The classification model that included these features made prognoses with relative accuracy and precision in multiple cohorts. We proposed PDAC risk subgroups and developed a classification model that may potentially be useful for routine clinical implementations, at the individual level. This clinical system may improve the accuracy of risk prediction and treatment guidelines. See related commentary by Thakur and Singh, p. 3272Keywords
Other Versions
Funding Information
- the Collaborative Genome Program for Fostering New Post-Genome Industry (2019000499)
- the Industrial Strategic technology development program (20000134)
- the Korea Health Industry Development Institute (HL19C0020, HI19C1132)
This publication has 50 references indexed in Scilit:
- FOLFIRINOX in Locally Advanced Pancreatic Cancer: The Massachusetts General Hospital Cancer Center ExperienceThe Oncologist, 2013
- Pancreatic ductal cells in development, regeneration, and neoplasiaJCI Insight, 2011
- Tissue Biomarkers for Prognosis in Pancreatic Ductal Adenocarcinoma: A Systematic Review and Meta-analysisClinical Cancer Research, 2011
- FOLFIRINOX versus Gemcitabine for Metastatic Pancreatic CancerNew England Journal of Medicine, 2011
- Hallmarks of Cancer: The Next GenerationCell, 2011
- Automated Detection of Inaccurate and Imprecise Transitions in Peptide Quantification by Multiple Reaction Monitoring Mass SpectrometryClinical Chemistry, 2010
- Supervised Risk Predictor of Breast Cancer Based on Intrinsic SubtypesJournal of Clinical Oncology, 2009
- Systematic and integrative analysis of large gene lists using DAVID bioinformatics resourcesNature Protocols, 2008
- Molecular prognostic markers in pancreatic cancer: A systematic reviewEuropean Journal Of Cancer, 2005
- Gene expression profiling predicts clinical outcome of breast cancerNature, 2002