🤖 AI Summary
This study addresses the limited generalizability of deep learning models in multi-center brain metastasis segmentation, which arises from variations in imaging devices, acquisition protocols, and patient populations. To mitigate this issue, the authors propose a VAE-MMD preprocessing pipeline that integrates a variational autoencoder with a maximum mean discrepancy (MMD) loss, enhanced by skip connections and self-attention mechanisms. This approach effectively aligns feature distributions across centers without requiring target-domain labels and is combined with nnU-Net to achieve high-precision segmentation. The method substantially improves generalization performance: domain classifier accuracy drops to 0.50, average F1 score increases by 11.1%, symmetric Dice (sDice) improves by 7.93%, Hausdorff distance at the 95th percentile (HD95) decreases by 65.5%, and reconstruction peak signal-to-noise ratio (PSNR) exceeds 36 dB.
📝 Abstract
Background: Deep learning has demonstrated significant potential for automated brain metastases (BM) segmentation; however, models trained at a singular institution often exhibit suboptimal performance at various sites due to disparities in scanner hardware, imaging protocols, and patient demographics. The goal of this work is to create a domain adaptation framework that will allow for BM segmentation to be used across multiple institutions.
Methods: We propose a VAE-MMD preprocessing pipeline that combines variational autoencoders (VAE) with maximum mean discrepancy (MMD) loss, incorporating skip connections and self-attention mechanisms alongside nnU-Net segmentation. The method was tested on 740 patients from four public databases: Stanford, UCSF, UCLM, and PKG, evaluated by domain classifier's accuracy, sensitivity, precision, F1/F2 scores, surface Dice (sDice), and 95th percentile Hausdorff distance (HD95).
Results: VAE-MMD reduced domain classifier accuracy from 0.91 to 0.50, indicating successful feature alignment across institutions. Reconstructed volumes attained a PSNR greater than 36 dB, maintaining anatomical accuracy. The combined method raised the mean F1 by 11.1% (0.700 to 0.778), the mean sDice by 7.93% (0.7121 to 0.7686), and reduced the mean HD95 by 65.5% (11.33 to 3.91 mm) across all four centers compared to the baseline nnU-Net.
Conclusions: VAE-MMD effectively diminishes cross-institutional data heterogeneity and enhances BM segmentation generalization across volumetric, detection, and boundary-level metrics without necessitating target-domain labels, thereby overcoming a significant obstacle to the clinical implementation of AI-assisted segmentation.