Synthetic Sampling for Multi-Class Malignancy Prediction

TL;DR

This paper investigates synthetic oversampling techniques to improve minority class sensitivity in multi-label malignancy prediction, demonstrating significant gains and insights into image data preprocessing for CADx systems.

Contribution

It introduces the use of synthetic sampling methods to enhance class-specific performance in imbalanced medical image classification tasks.

Findings

Synthetic oversampling increased minority class sensitivity by up to 19.88%.

Synthetic nodules offer insights for data preprocessing and augmentation.

Overall sensitivity improved by an average of 7.22%.

Abstract

We explore several oversampling techniques for an imbalanced multi-label classification problem, a setting often encountered when developing models for Computer-Aided Diagnosis (CADx) systems. While most CADx systems aim to optimize classifiers for overall accuracy without considering the relative distribution of each class, we look into using synthetic sampling to increase per-class performance when predicting the degree of malignancy. Using low-level image features and a random forest classifier, we show that using synthetic oversampling techniques increases the sensitivity of the minority classes by an average of 7.22% points, with as much as a 19.88% point increase in sensitivity for a particular minority class. Furthermore, the analysis of low-level image feature distributions for the synthetic nodules reveals that these nodules can provide insights on how to preprocess image data for better classification performance or how to supplement the original datasets when more data acquisition is feasible.