A study recently published in Cancers showed that artificial intelligence (AI) is a may be useful in helping doctors better identify patients at risk of endometrial cancer (EC).
Researchers conducted a review on available studies on the topic of AI in medicine, focusing specifically on how AI can spur innovations in EC care. By combing through academic search engines, the research team was able to identify 32 studies that were ultimately included in their final analysis on the subject.
Histopathological analysis, which is the analyzing of a sample of affected tissue under a microscope, remains the gold standard for diagnosing endometrial cancer. Researchers found that AI has been used to streamline this process by training it to recognize suspicious images and teaching it to differentiate between normal tissue and cancer cells. Specifically, HIENet, a novel computer-aided diagnosis system, has a diagnostic accuracy of 84.5%.
Read more about EC testing and diagnosis
Multi-omics is another way in which doctors diagnose EC. In simple terms, multi-omics is the integration of data from various “omics”, such as genomics, metabolomics and proteomics. When combined, these data help doctors build a profile about what is likely going on in a patient.
Because this process involves the analysis of complex data, AI is poised to be helpful, and AI models such as the HECTOR excelled in predicting the likely long-term outcomes of patients, such as 10-year recurrence-free survival.
Medical imaging is another area that has been particularly revolutionized by AI, especially after it is trained to detect patterns across various imaging modalities. AI can often detect subtle features that may be missed by doctors. In the case of endometrial cancer, ultrasound is a form of medical imaging typically used for diagnosis. An AI model used to assess myometrial invasion depth achieved an accuracy of 84.78%, demonstrating AI’s vast potential in aiding clinical processes in this disease area.
“The integration of artificial intelligence (AI) into diagnostic workflows represents a significant advancement in the detection and management of endometrial cancer,” the researchers wrote. “Current evidence supports the role of AI in enabling earlier diagnosis and improving risk stratification and more personalized treatment strategies, ultimately aiming to improve patient outcomes in [endometrial cancer] care.”
