Decision region analysis for generalizability of artificial intelligence models: estimating model generalizability in the case of cross-reactivity and population shift

Alexis Burgon; Berkman Sahiner; Nicholas Petrick; Gene Pennello; Kenny H. Cha; Ravi K. Samala

doi:10.1117/1.JMI.11.1.014501

25 January 2024 Decision region analysis for generalizability of artificial intelligence models: estimating model generalizability in the case of cross-reactivity and population shift

Alexis Burgon, Berkman Sahiner, Nicholas Petrick, Gene Pennello, Kenny H. Cha, Ravi K. Samala

Author Affiliations +

Journal of Medical Imaging, Vol. 11, Issue 1, 014501 (January 2024). https://doi.org/10.1117/1.JMI.11.1.014501

Abstract

Purpose

Understanding an artificial intelligence (AI) model’s ability to generalize to its target population is critical to ensuring the safe and effective usage of AI in medical devices. A traditional generalizability assessment relies on the availability of large, diverse datasets, which are difficult to obtain in many medical imaging applications. We present an approach for enhanced generalizability assessment by examining the decision space beyond the available testing data distribution.

Approach

Vicinal distributions of virtual samples are generated by interpolating between triplets of test images. The generated virtual samples leverage the characteristics already in the test set, increasing the sample diversity while remaining close to the AI model’s data manifold. We demonstrate the generalizability assessment approach on the non-clinical tasks of classifying patient sex, race, COVID status, and age group from chest x-rays.

Results

Decision region composition analysis for generalizability indicated that a disproportionately large portion of the decision space belonged to a single “preferred” class for each task, despite comparable performance on the evaluation dataset. Evaluation using cross-reactivity and population shift strategies indicated a tendency to overpredict samples as belonging to the preferred class (e.g., COVID negative) for patients whose subgroup was not represented in the model development data.

Conclusions

An analysis of an AI model’s decision space has the potential to provide insight into model generalizability. Our approach uses the analysis of composition of the decision space to obtain an improved assessment of model generalizability in the case of limited test data.

Published by SPIE

Citation Download Citation

Alexis Burgon, Berkman Sahiner, Nicholas Petrick, Gene Pennello, Kenny H. Cha, and Ravi K. Samala "Decision region analysis for generalizability of artificial intelligence models: estimating model generalizability in the case of cross-reactivity and population shift," Journal of Medical Imaging 11(1), 014501 (25 January 2024). https://doi.org/10.1117/1.JMI.11.1.014501

Received: 30 June 2023; Accepted: 28 December 2023; Published: 25 January 2024

ACCESS THE FULL ARTICLE

JOURNAL ARTICLE
18 PAGES

DOWNLOAD PAPER SAVE TO MY LIBRARY

GET CITATION

RIGHTS & PERMISSIONS

Get copyright permission Get copyright permission on Copyright Marketplace

KEYWORDS

Data modeling

Statistical modeling

Artificial intelligence

Education and training

Performance modeling

Chest imaging

Interpolation

Purpose

Approach

Results

Conclusions

Show All Keywords

Keywords/Phrases

Search In:

Publication Years