Recently, researchers from Harvard Medical School have developed a versatile AI model similar to ChatGPT, capable of handling various diagnostic procedures for different types of cancer.
This AI model, named CHIEF, can diagnose and assess various cancers more accurately, offering powerful functionalities such as predicting patient survival rates and providing treatment suggestions. The flexibility of this AI model represents a significant leap compared to current AI systems.
Image source note: The image was generated by AI, provided by the image licensing service Midjourney
Current AI systems are typically trained to perform specific tasks, such as detecting the presence of cancer or predicting the genetic characteristics of tumors, and they often only apply to a few types of cancer. In contrast, the new model can perform a variety of tasks and has been tested on 19 types of cancer, giving it the flexibility similar to large language models like ChatGPT.
It can not only read digital slides of tumor tissue but also identify cancer cells and predict the molecular characteristics of tumors, with accuracy significantly higher than most existing AI systems. The emergence of this model could significantly improve cancer treatment strategies, helping doctors to more effectively assess patient conditions.
CHIEF was trained by analyzing 15 million unlabeled images and 60,000 full-slice images, involving cancers such as lung cancer, breast cancer, and prostate cancer.
After comprehensive testing, CHIEF outperformed the current state-of-the-art AI technology in tasks such as cancer detection, tumor origin identification, and predicting patient survival, with accuracy rates ranging from 94% to 96%.
Additionally, CHIEF can quickly identify tumor gene mutations through cell images, providing a cheaper and faster alternative for regions without access to gene sequencing. For example, CHIEF can identify 54 common gene mutations related to cancer with an accuracy of over 70%. In predicting patient survival, CHIEF also performs exceptionally well, effectively distinguishing between long-term and short-term survivors, especially outperforming other models in late-stage cancer patients.
The research team stated that they will continue to improve the functionalities of the CHIEF model in the future, including training on samples of rare diseases and non-cancerous conditions, enhancing its capabilities in molecular data identification, and even planning to predict the effectiveness and side effects of new cancer treatments.
Key Points:
🌟 CHIEF is a new AI model capable of efficiently diagnosing 19 different types of cancer, surpassing current AI technology.
🧬 The model quickly predicts gene mutations by analyzing tumor images, providing a convenient solution for patients without access to gene sequencing.
📈 CHIEF performs excellently in predicting patient survival rates, effectively distinguishing between long-term and short-term survivors, supporting clinical decision-making.