Related Experiment Videos
CLNAD-Net: A Multi-Task Learning Framework for Cervical Lymph Node Computer-Aided Diagnosis Network
Weihua He1, Fet Ouyang2, Guangming Yang2
1School of Electronics and Communication Engineering, Sun Yat-sen University, Shenzhen, 518107, Guangdong, China.
Journal of Imaging Informatics in Medicine
|August 5, 2026
Summary
A new AI tool, CLNAD-Net, accurately diagnoses cervical lymph node metastasis in papillary thyroid cancer patients using ultrasound images. This advanced network improves diagnostic accuracy and efficiency, aiding clinical decisions.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Oncology
Background:
- Accurate diagnosis of cervical lymph node metastasis (CLNM) is crucial for papillary thyroid cancer (PTC) treatment.
- Current ultrasound-based diagnosis relies heavily on physician expertise, facing challenges with image quality and variability.
Purpose of the Study:
- To develop an automated diagnostic system for CLNM in PTC using deep learning.
- To improve the accuracy and consistency of CLNM diagnosis compared to human interpretation.
Main Methods:
- Proposed CLNAD-Net, a multi-task learning network with a dual-branch encoder (CNNs and Transformers).
- Integrated a feature alignment module (FAM) for local and global feature fusion.
- Incorporated a multi-modal auxiliary classification module (MACM) using ultrasound images and clinical data.
Main Results:
- CLNAD-Net achieved 84.49% accuracy in CLNM classification, outperforming existing methods and physicians.
- Segmentation and classification tasks showed superior performance.
- Clinical data integration improved classification accuracy by 1.78%.
Conclusions:
- CLNAD-Net demonstrates significant potential for accurate and efficient CLNM diagnosis in PTC.
- The proposed framework offers a promising approach for automated auxiliary diagnostic systems in medical imaging.