BEIJING, Aug. 25 (Xinhua) -- A liver cancer diagnostic AI model developed by a team of Chinese scientists discovered 15 cancer cases that had originally been missed during a two-month real-world prospective clinical trial, helping the patients receive timely surgical or drug treatment, China Science Daily reported on Tuesday.
The detection of early-stage liver cancer is challenging. Because the lesions are small and easily obscured by cirrhosis, fatty liver and the liver's complex internal anatomy, even experienced physicians can miss them on contrast-enhanced CT scans.
It is even more difficult when cancer from other parts of the body metastasizes to the liver. The physician's attention may be drawn to the primary lesion, causing tiny liver metastases to be overlooked.
Alibaba's DAMO Academy, in collaboration with Shengjing Hospital of China Medical University and other institutions, developed the liver cancer diagnostic AI model DAMO LiON, which can identify tiny cancerous liver lesions from CT images. This model is able to not only accurately identify primary liver cancer, but also find liver metastases that are easily missed.
Experiments showed that DAMO LiON's accuracy in identifying malignant tumors surpassed that of general radiologists. When physicians used the AI model to assist in reading scans, reading time was reduced by 27 percent, while sensitivity to malignant tumors increased by 11.5 percent, effectively reducing missed diagnoses. With AI assistance, junior physicians were able to perform at the level of senior physicians.
During the two-month real-world prospective clinical trial, DAMO LiON read contrast-enhanced CT scans from more than 10,000 patients, assisting physicians in catching 15 liver metastases that had originally been overlooked and prompting changes to these patients' treatment plans.
The liver metastases detected by DAMO LiON were generally characterized as small, faint and off-center. With an average diameter of about 1 cm, they had a low contrast against liver tissue, or were located in relatively uncommon anatomical positions.
The study was recently published in the journal Nature Medicine.
DAMO Academy has also developed AI screening models including DAMO PANDA for pancreatic cancer, DAMO GRAPE for gastric cancer and DAMO COCA for colorectal cancer. ■
