A groundbreaking AI model may significantly improve cancer diagnosis and evaluation, outperforming current deep learning methods, according to a recent study.
Researchers have developed an advanced artificial intelligence (AI) model named the Clinical Histopathology Imaging Evaluation Foundation (CHIEF), which demonstrates up to 36% greater efficacy in identifying various cancers, determining tumor origins, and predicting patient outcomes compared to existing deep learning technologies.
The team, led by experts from Harvard Medical School, aimed to create a model with broader applicability across different diagnostic scenarios. Many existing AI models for cancer are designed for specific tasks, limiting their versatility.
“Unlike traditional models, our AI tool offers clinicians precise, real-time second opinions on cancer diagnoses by evaluating a wide range of cancer types and variations,” said Kun-Hsing Yu, assistant professor of biomedical informatics at Harvard Medical School and senior author of the study, in a statement to Euronews Health.
How Does CHIEF Work?
The CHIEF model was trained using over 15 million pathology images, enhancing its accuracy in diagnosing cancers with unusual features. Researchers refined the model with more than 60,000 high-resolution tissue slide images, focusing on genetic and clinical prediction tasks.
The model was tested on over 19,400 images from 24 hospitals and patient cohorts worldwide, with findings published in the journal Nature. CHIEF operates by analyzing digital slides of tumor tissues, predicting their molecular profile based on image features, and identifying characteristics that may influence a patient’s treatment response.
With nearly 94% accuracy in detecting cancer across 11 types and performance reaching up to 99.43% in specific applications like identifying colon cancer cells or predicting genetic mutations, CHIEF represents a significant leap forward in AI-assisted oncology.
Ajit Goenka, a professor of radiology at Mayo Clinic, noted that while CHIEF shows great promise, its robustness across diverse clinical environments needs thorough validation. Goenka, who was not involved in the study, highlighted the importance of validating the model in real-world settings to ensure it performs reliably across various patient demographics and clinical conditions.
Next Steps for the AI Model
Before CHIEF can be implemented in clinical practice, it must undergo regulatory approval. The research team is initiating a prospective clinical study to validate the model in practical settings and is also working on expanding its capabilities to detect rare cancers.
Goenka emphasized that extensive validation is essential to confirm the model’s practical reliability and effectiveness in everyday clinical use. “This validation is crucial to ensure the model is not only theoretically superior but also dependable in routine medical practice,” he said.
Related articles:
Emerging Treatment For Her2-Positive Extramammary Paget Disease
What Is the Best Diet for a Man in His 30s?