Abstract Esophageal varices (EV) are a critical complication of portal hypertension. While endoscopy is the gold standard for diagnosis, its invasive nature limits frequent screening. Non-contrast CT (NCCT) offers a safer alternative but suffers from low vascular contrast. This paper introduces MOON++ (Multi-Organ-COhesion Network++), a multimodal framework that leverages multi-organ features (esophagus, liver, spleen) and clinical priors to accurately grade EV from NCCT scans. MOON++ achieves an AUC of 0.921 for moderate-to-severe EV detection, significantly outperforming single-organ models and experienced radiologists.
Introduction
Esophageal varices—enlarged veins in the esophagus—are a life-threatening complication of portal hypertension, often caused by liver cirrhosis. If left undiagnosed, they can rupture, leading to massive internal bleeding with a mortality rate of up to 20%.
The current "gold standard" for diagnosis is invasive endoscopy, which is uncomfortable and carries risks. Contrast-enhanced CT (CE-CT) is a non-invasive alternative but requires injecting contrast agents that can be harmful to patients with kidney or liver issues. Non-contrast CT (NCCT) is safer and more common, but because the contrast between blood vessels and surrounding tissue is so low, it is notoriously difficult for even expert radiologists to grade EV severity accurately. The authors of MOON++ tackle this "low-contrast" hurdle by looking beyond the esophagus, using multi-organ analysis to provide the necessary context for diagnosis.
Methods
The MOON++ framework is built on several sophisticated components designed to mimic a radiologist’s holistic diagnostic process:
- Multi-Organ Backbone: The system uses a UniFormer architecture to extract features from three regions: the esophagus, liver, and spleen.
- Organ Representation Interaction (ORI): This novel module allows the model to exchange information between organs. For example, features from a swollen spleen can "inform" the analysis of the esophagus, helping the model focus on subtle signs of vascular dilation.
- Clinical Prior Embedding: The authors calculate clinical metrics like organ volumes and the Liver-to-Spleen Volume Ratio (LSVR). They then use a medical CLIP-based encoder (M3D) to transform these numerical/categorical values into a high-dimensional feature space, effectively "prompting" the vision model with clinical domain knowledge.
- Ordinal Regression & DCCA: To maintain the logical progression of disease (Grade 0 to Grade 3), they use an ordinal regression loss. They also employ Deep Canonical Correlation Analysis (DCCA) to maximize the correlation between the features extracted from different organs.
Experimental Setup
The study utilized a substantial dataset of 1,631 patients with endoscopically confirmed EV grades. This included an internal dataset of 1,474 patients from Shengjing Hospital of China Medical University and an external validation set of 157 patients from the Second Affiliated Hospital of Baotou Medical College. All cases were graded (G0–G3) based on invasive endoscopy, the clinical gold standard.
Results and Discussion
MOON++ showed a significant leap in performance over baseline methods. For severe EV detection (G3 vs. <G3), it achieved an AUC of 0.894, compared to 0.803 for single-organ methods. For moderate-to-severe detection (≥G2 vs. <G2), it achieved an AUC of 0.921, significantly outperforming the 0.793 of single-organ approaches.
In a reader study, MOON++ achieved a multi-class accuracy of 65.3%, whereas two experienced radiologists scored 48.4% and 52.2%. The model was particularly superior at identifying high-risk G3 cases, which are critical for surgical intervention. Key insights from the study include:
- Context is King: The paper demonstrates that the esophagus alone is insufficient for NCCT-based grading. Adding liver and spleen data significantly boosts performance because their morphological changes (like splenomegaly or liver contraction) are much easier to detect on NCCT.
- Categorical > Numerical: Ablation studies showed that describing organ volumes in categorical terms (e.g., "Very Low," "Average," "High") performed better than using raw numerical values, likely because it helps the model capture clinically meaningful thresholds.
Conclusion
MOON++ demonstrates the power of multi-organ analysis and clinical prior integration for challenging medical imaging tasks. While the model relies on accurate organ segmentation, which can be difficult in cases of severe ascites, it represents a major step toward non-invasive, accurate EV grading. Future work may incorporate blood test markers to further refine the clinical priors and improve diagnostic reliability.
Summary updated