Development and Validation of a Protein Electrophoresis Classification Algorithm: Tabular Data-Based Alternative

TL;DR

This paper introduces a machine learning method that uses numerical data from serum protein electrophoresis instead of images for accurate and interpretable analysis.

Contribution

The novel contribution is a tabular data-based machine learning approach for serum protein electrophoresis classification.

Findings

The proposed method provides a robust alternative to image-based deep learning for SPE analysis.

It leverages numerical SPE profiles directly for efficient and interpretable classification.

Abstract

Serum protein electrophoresis (SPE) is routinely interpreted through visual assessment of electropherogram images by medical laboratory scientists. We introduce an efficient tabular data–based machine learning approach that directly leverages numerical SPE profiles, offering a robust and interpretable alternative to image-based deep learning methods.

Figures