Latent-CURE for Breast Cancer Diagnosis

2026-06-29

AI Summary

Current breast ultrasound diagnostic models are predominantly end-to-end black-box systems, which are highly susceptible to extreme data imbalance and often overlook rare yet critical malignant features, resulting in insufficient reliability and interpretability. This work proposes the Latent-CURE framework, which uniquely integrates latent-space reasoning with clinical BI-RADS diagnostic logic. By employing an asymmetrically weighted chain-of-thought mechanism, the model is guided to sequentially reason over morphological descriptors. Furthermore, a dual asymmetric optimization strategy coupled with dynamic margin adjustment is introduced to mitigate the scarcity of malignant samples. Evaluated on highly imbalanced real-world data, the proposed method substantially enhances the detection of rare malignant patterns while maintaining high specificity and delivering clinically coherent, interpretable diagnostic justifications.

Abstract

Multimodal Large Models have significantly advanced automated breast ultrasound diagnosis. However, most existing frameworks utilize opaque, end-to-end paradigms prioritizing global statistical correlations over structured clinical reasoning. Consequently, these models remain susceptible to shortcut learning amid extreme real-world epidemiological imbalances, often bypassing rare but decisive malignant indicators for dominant benign patterns. To address this disconnect, we propose Latent-CURE, a novel diagnostic framework driven by asymmetric weighted chain-of-thought methodology grounded in latent space reasoning. Unlike traditional approaches, our framework constructs an implicit reasoning trajectory forcing the model to sequentially infer standardized BI-RADS morphological descriptors before converging on a final diagnosis. Furthermore, to combat the extreme scarcity of critical malignant features, we couple this architecture with a dual-asymmetric optimization strategy. By dynamically adjusting margins and weights, this strategy safeguards high-specificity malignant descriptors from being overshadowed by common benign priors. Comprehensive evaluations demonstrate that our knowledge-injected approach provides transparent clinical evidence while achieving robust, accurate diagnostic performance in imbalanced medical cohorts.

Problem

• breast cancer diagnosis
• shortcut learning
• epidemiological imbalance
• multimodal large models
• clinical reasoning

Innovation

• Latent-CURE
• asymmetric weighted chain-of-thought
• latent space reasoning
• dual-asymmetric optimization
• BI-RADS descriptors