Tumor classification with MALDI-MSI data of tissue microarrays: A case study.

Area of Science:

• Biomedical science
• Analytical chemistry
• Oncology

Background:

• Mass spectrometry imaging (MSI) on tissue microarrays (TMAs) enables simultaneous analysis of numerous biomolecules across many patients, facilitating biomarker discovery.
• Predicting lymph node metastasis is crucial for cancer patient management and treatment planning.

Purpose of the Study:

• To investigate the potential of matrix-assisted laser desorption/ionization (MALDI)-MSI data for predicting lymph node metastasis.
• To identify specific molecular features indicative of metastasis from MALDI-MSI data.

Main Methods:

• MALDI-MSI data were acquired from TMAs of 122 patients.
• Data underwent filtering based on spectral intensity and tumor cell percentage, followed by preprocessing.
• Univariate feature selection methods reduced data dimensionality, and selected features were used with three classifiers (decision tree, k-NN, SVM).

Main Results:

• A decision tree classifier achieved the highest accuracy, correctly classifying approximately 72% of patients.
• A single molecular peak at m/z 718.4 was identified as the primary contributor to predictive power.
• The classification approach demonstrated sensitivity when tested with artificially modified data.

Conclusions:

• MALDI-MSI data holds potential for predicting lymph node metastasis.
• The identified m/z 718.4 peak may serve as a novel biomarker for metastasis.
• The developed classification methodology is generalizable for biomarker discovery in other contexts.