Role of Image Acquisition and Patient Phenotype Variations in Automatic Segmentation Model Generalization

TL;DR

This study investigates how variations in image acquisition methods and patient phenotypes affect the ability of automated medical image segmentation models to generalize across different datasets, emphasizing the importance of diverse training data.

Contribution

It demonstrates that training on a diverse dataset improves the out-of-domain performance of segmentation models, providing a data-centric approach for better clinical applicability.

Findings

Models trained on diverse data perform comparably to in-domain models on similar data.

Broader training datasets enhance model generalization to new image acquisition types.

Data diversity is key to improving the robustness of medical image segmentation models.

Abstract

Purpose: This study evaluated the out-of-domain performance and generalization capabilities of automated medical image segmentation models, with a particular focus on adaptation to new image acquisitions and disease type. Materials: Datasets from both non-contrast and contrast-enhanced abdominal CT scans of healthy patients and those with polycystic kidney disease (PKD) were used. A total of 400 images (100 non-contrast controls, 100 contrast controls, 100 non-contrast PKD, 100 contrast PKD) were utilized for training/validation of models to segment kidneys, livers, and spleens, and the final models were then tested on 100 non-contrast CT images of patients affected by PKD. Performance was evaluated using Dice, Jaccard, TPR, and Precision. Results: Models trained on a diverse range of data showed no worse performance than models trained exclusively on in-domain data when tested on in-domain data. For instance, the Dice similarity of the model trained on 25% from each dataset was found to be non-inferior to the model trained purely on in-domain data. Conclusions: The results indicate that broader training examples significantly enhances model generalization and out-of-domain performance, thereby improving automated segmentation tools' applicability in clinical settings. The study's findings provide a roadmap for future research to adopt a data-centric approach in medical image AI model development.