Improving AI models for rare thyroid cancer subtype by text guided diffusion models.
Authors
Affiliations (21)
Affiliations (21)
- Department of Bioinformatics and Biostatistics, School of Life Sciences and Biotechnology, Shanghai Jiao Tong University, Shanghai, China.
- SJTU-Yale Joint Center of Biostatistics and Data Science, Technical Center for Digital Medicine, National Center for Translational Medicine, Shanghai Jiao Tong University, Shanghai, China.
- Department of Clinical Laboratory Medicine, Shanghai Tenth People's Hospital, School of Medicine, Tongji University, Shanghai, PR China.
- SJTU-Yale Joint Center of Biostatistics and Data Science, Technical Center for Digital Medicine, National Center for Translational Medicine, Shanghai Jiao Tong University, Shanghai, China. [email protected].
- Department of Critical Care Medicine, Jiuquan Hospital of Shanghai General Hospital, Shanghai Jiaotong University School of Medicine, Jiuquan, Gansu, PR China.
- Department of Ultrasound, Pudong New Area People's Hospital Affiliated to Shanghai University of Medicine and Health Sciences, Shanghai, PR China.
- Department of Automation, Tsinghua University, Beijing, PR China.
- Medical college, Nantong University, Nantong, Jiangsu, PR China.
- Shcool of Artificial Intelligence, University of Chinese Academy of sciences, Beijing, PR China.
- Xin'an League People's Hospital, Xing'an League, Inner Mongolia, PR China.
- Department of Ultrasound, Shanghai Fourth People's Hospital Affiliated to Tongji University, Shanghai, PR China.
- Department of Biomedical Sciences Laboratory, Affiliated Dongyang Hospital of Wenzhou Medical University, Dongyang, Zhejiang, PR China.
- Department of Hepatobiliary pancreatic center, Xuzhou City Central Hospital, Xuzhou, Jiangsu, China.
- Department of General Surgery, Ruijin Hospital, Shanghai Jiaotong University School of medicine, Shanghai, PR China.
- Nanjing Drum Tower Hospital, The Affiliated Hospital of Nanjing University Medical School, Nanjing, Jiangsu, PR China.
- Department of Clinical Laboratory Medicine, Shanghai Tenth People's Hospital, School of Medicine, Tongji University, Shanghai, PR China. [email protected].
- Department of General Surgery, Ruijin Hospital, Shanghai Jiaotong University School of medicine, Shanghai, PR China. [email protected].
- World Tea Organization, Cambridge, MA, USA. [email protected].
- Department of Bioinformatics and Biostatistics, School of Life Sciences and Biotechnology, Shanghai Jiao Tong University, Shanghai, China. [email protected].
- SJTU-Yale Joint Center of Biostatistics and Data Science, Technical Center for Digital Medicine, National Center for Translational Medicine, Shanghai Jiao Tong University, Shanghai, China. [email protected].
- Institute of Bioinformatics, Shanghai Academy of Experimental Medicine, Shanghai, China. [email protected].
Abstract
Artificial intelligence applications in oncology imaging often struggle with diagnosing rare tumors. We identify significant gaps in detecting uncommon thyroid cancer types with ultrasound, where scarce data leads to frequent misdiagnosis. Traditional augmentation strategies do not capture the unique disease variations, hindering model training and performance. To overcome this, we propose a text-driven generative method that fuses clinical insights with image generation, producing synthetic samples that realistically reflect rare subtypes. In rigorous evaluations, our approach achieves substantial gains in diagnostic metrics, surpasses existing methods in authenticity and diversity measures, and generalizes effectively to other private and public datasets with various rare cancers. In this work, we demonstrate that text-guided image augmentation substantially enhances model accuracy and robustness for rare tumor detection, offering a promising avenue for more reliable and widespread clinical adoption.