Caps improves MRI scans to help brain tumor diagnosis and treatment
CAPS: Clinically-Verifiable Anatomy-Prioritized Synthesis for Missing MRI Imputation in Brain Tumor Segmentation
Computer Vision and Pattern Recognition
Summary
Brain tumor diagnosis often relies on MRI scans with multiple types of images, but sometimes some images are missing. The authors propose CAPS, a method that creates realistic and reliable missing MRI images by focusing on preserving important brain anatomy features. Their method not only produces clearer images but also helps brain tumor detection tools work better, even across different patient groups. Doctors who reviewed the images found them more trustworthy for clinical use compared to earlier methods.
What this means in practice
- •For radiology departments: Create consistent multi-modal MRI scans when some image types are missing to improve brain tumor diagnosis and automated analysis accuracy.
- •For medical imaging software developers: Integrate anatomy-focused MRI synthesis to generate high-quality missing scans that improve robustness of segmentation tools across diverse patient populations.
Authors
Cecilia Diana-Albelda, Arthur Longuefosse, Mario Rubén Paz Campos, Álvaro García-Martín, Jesus Bescos
Abstract
Multi-parametric MRI is the gold standard for neuro-oncology, yet acquiring complete 4-modality protocols may be unfeasible in real-world clinical practice. When handling missing modalities, existing automated systems force a critical compromise: robust segmenters operate as uninterpretable "black boxes" that bypass image generation, while standard synthesis models often produce blurry reconstructions lacking diagnostic reliability. To address this issue, we propose CAPS, a Clinically-Verifiable Anatomy-Prioritized Synthesis framework that reconciles computational precision with clinical transparency. Our multi-modal generative architecture is driven by a hybrid objective function (L_1 + L_AFP + L_DualDice) that explicitly enforces the preservation of deep anatomical features and strict topological boundaries, rather than relying solely on pixel-wise optimization. Evaluated comprehensively on the BraTS Adult Glioma dataset, the framework achieves superior synthesis fidelity across intensity-based metrics, and demonstrates robust downstream segmentation performance (82.38% mean Dice), successfully restoring verifiable visual context. Furthermore, the framework exhibits exceptional zero-shot cross-domain generalization, maintaining high structural fidelity across completely unseen Meningioma, Pediatric, and Sub-Saharan African cohorts without domain-specific adaptation. Finally, a blinded radiological assessment confirms that our anatomy-aware approach significantly enhances both perceived image quality and clinical utility, providing clinicians with the reliable visual evidence required to trust downstream automated workflows.