Improving AI models for rare thyroid cancer subtype by text guided diffusion models
Fang Dai, Siqiong Yao, Min Wang, Yicheng Zhu, Xiangjun Qiu, Peng Sun, Cheng Qiu, Jisheng Yin, Guangtai Shen, Jingjing Sun, Maofeng Wang, Yun Wang, Zheyu Yang, Jianfeng Sang, Xiaolei Wang, Fenyong Sun, Wei Cai, Xingcai Zhang, Hui Lü
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. Artificial intelligence (AI) is becoming increasingly relevant to assist with oncology imaging, but diagnosing rare tumours remains challenging. Here, the authors develop an AI approach to detect rare thyroid cancer subtypes by integrating clinical knowledge with image generation based on ultrasound imaging data from large patient cohorts.