Concept-based Explainable Malignancy Scoring on Pulmonary Nodules in CT Images

TL;DR

This paper introduces an interpretable, concept-based model for lung nodule malignancy scoring in CT images, enhancing transparency and aligning with clinical reasoning, while maintaining competitive accuracy.

Contribution

The paper presents a novel concept-based learning framework using generalized additive models for explainable lung nodule malignancy assessment.

Findings

Model provides human-readable explanations of attributes and their contributions.

Achieves classification performance comparable to existing methods.

Demonstrates alignment with clinical patterns in diagnosis.

Abstract

To increase the transparency of modern computer-aided diagnosis (CAD) systems for assessing the malignancy of lung nodules, an interpretable model based on applying the generalized additive models and the concept-based learning is proposed. The model detects a set of clinically significant attributes in addition to the final malignancy regression score and learns the association between the lung nodule attributes and a final diagnosis decision as well as their contributions into the decision. The proposed concept-based learning framework provides human-readable explanations in terms of different concepts (numerical and categorical), their values, and their contribution to the final prediction. Numerical experiments with the LIDC-IDRI dataset demonstrate that the diagnosis results obtained using the proposed model, which explicitly explores internal relationships, are in line with similar patterns observed in clinical practice. Additionally, the proposed model shows the competitive classification and the nodule attribute scoring performance, highlighting its potential for effective decision-making in the lung nodule diagnosis.