Core Principles
X-ray: Uses high-energy electromagnetic waves (ionizing radiation) to penetrate tissues. Differential absorption by tissues of varying density and atomic number generates grayscale contrast images. Essential for skeletal imaging, pulmonary pathologies (e.g., fractures, pneumonia), and dentistry.
Ultrasound: Generates real-time dynamic images via reflection of high-frequency mechanical sound waves (non-ionizing radiation) at tissue interfaces. Doppler techniques assess blood flow. Dominates obstetrics, abdominal organ evaluation, cardiovascular diagnostics, and superficial organ imaging.
Convergence Breakthroughs
X-ray-induced Acoustic Computed Tomography (XACT):
X-ray absorption induces transient thermal expansion in tissues, emitting ultrasound waves. Detection of these waves enables high-contrast 3D reconstruction. 3D images generated from a single X-ray projection reduce radiation exposure, with transformative potential in deep-tissue imaging and real-time radiotherapy verification.
Artificial Intelligence (AI) Innovations:
X-ray:
Multimodal large models (e.g., XrayGLM) automate chest radiograph interpretation, generating structured reports and enabling multi-turn doctor-patient dialogue.
AI detects fatty liver disease from routine chest X-rays (AUC 0.82-0.83), enabling low-cost screening.
Ultrasound:
Integration of fuzzy knowledge with multimodal large language models (e.g., enhanced LLaVA-OV frameworks) improves accuracy and cross-device robustness for breast/thyroid/liver ultrasound reports.
Fragment-based bilingual training resolves dataset standardization issues, producing precise Chinese/English reports.
Future Trajectory
Multimodal imaging (X-ray/US/CT/MRI/PET) combined with genomics enables precision medicine.
Ultrasound elastography quantifies tissue biomechanical properties.
Portable device advancements expand diagnostic access in resource-limited regions.