Cancer Tissue Classification Using Supervised Machine Learning Applied to MALDI Mass Spectrometry Imaging.

Area of Science:

• Biomedical Imaging
• Computational Pathology
• Mass Spectrometry

Background:

• Matrix assisted laser desorption/ionization mass spectrometry imaging (MALDI MSI) provides spatial distribution of analytes in tissues.
• Machine learning (ML) offers potential for analyzing complex MALDI MSI data.
• Accurate cancer diagnostics are crucial for effective treatment.

Purpose of the Study:

• To explore the clinical utility of ML applied to MALDI MSI data for cancer diagnostic classification.
• To evaluate the performance of deep neural networks in discriminating tumor from normal tissue.
• To assess the prediction of lymph node metastasis using MALDI MSI data.

Main Methods:

• Tissue microarrays (TMAs) from 302 colorectal cancer (CRC) and 257 endometrial cancer (EC) patients were analyzed.
• Deep neural networks were employed for ML-based classification of MALDI MSI data.
• Balanced cross-validation was used to assess classification accuracy, sensitivity, and specificity.

Main Results:

• ML accurately discriminated colorectal tumor from normal tissue (98% overall accuracy, 98.2% sensitivity, 98.6% specificity).
• The ML approach predicted lymph node metastasis in EC primary tumors with 80% accuracy (90% sensitivity, 69% specificity).
• MALDI MSI combined with ML demonstrated high performance in cancer classification.

Conclusions:

• MALDI MSI coupled with ML shows significant potential for cancer diagnostic applications.
• This approach can complement traditional histopathological examination.
• The findings support the use of MALDI MSI for improved cancer detection and staging.