Development of a Class Prediction Model to Discriminate Pancreatic Ductal Adenocarcinoma from Pancreatic Neuroendocrine Tumor by MALDI Mass Spectrometry Imaging

Proteomics Clin Appl. 2019 Jan;13(1):e1800046. doi: 10.1002/prca.201800046. Epub 2018 Dec 19.

Abstract

Purpose: To define proteomic differences between pancreatic ductal adenocarcinoma (pDAC) and pancreatic neuroendocrine tumor (pNET) by matrix-assisted laser desorption/ionization mass spectrometry imaging (MALDI MSI).

Experimental design: Ninety-three pDAC and 126 pNET individual tissues are assembled in tissue microarrays and analyzed by MALDI MSI. The cohort is separated in a training (52 pDAC and 83 pNET) and validation set (41 pDAC and 43 pNET). Subsequently, a linear discriminant analysis (LDA) model based on 46 peptide ions is performed on the training set and evaluated on the validation cohort. Additionally, two liver metastases and a whole slide of pDAC are analyzed by the same LDA algorithm.

Results: Classification of pDAC and pNET by the LDA model is correct in 95% (39/41) and 100% (43/43) of patients in the validation cohort, respectively. The two liver metastases and the whole slide of pDAC are also correctly classified in agreement with the histopathological diagnosis.

Conclusion and clinical relevance: In the present study, a large dataset of pDAC and pNET by MALDI MSI is investigated, a class prediction model that allowed separation of both entities with high accuracy is developed, and differential peptide peaks with potential diagnostic, prognostic, and predictive values are highlighted.

Keywords: formalin-fixed paraffin embedded (FFPE); matrix-assisted laser desorption/ionization (MALDI) mass spectrometry imaging (MSI); neuroendocrine-tumor; pancreas; tumor typing.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Carcinoma, Pancreatic Ductal / diagnostic imaging
  • Carcinoma, Pancreatic Ductal / metabolism*
  • Carcinoma, Pancreatic Ductal / pathology
  • Discriminant Analysis
  • Humans
  • Models, Statistical*
  • Molecular Imaging*
  • Neuroendocrine Tumors / diagnostic imaging
  • Neuroendocrine Tumors / metabolism*
  • Neuroendocrine Tumors / pathology
  • Pancreatic Neoplasms / diagnostic imaging
  • Pancreatic Neoplasms / metabolism*
  • Pancreatic Neoplasms / pathology
  • Paraffin Embedding
  • Prognosis
  • Proteomics / methods*
  • Spectrometry, Mass, Matrix-Assisted Laser Desorption-Ionization*