Home Medizin KI-identifizierte Gefäßheilung kann einen Rückfall bei UC vorhersagen

KI-identifizierte Gefäßheilung kann einen Rückfall bei UC vorhersagen

von NFI Redaktion

In Stockholm, Sweden, a novel artificial intelligence (AI)-enhanced image-enhanced endoscopy system accurately assessed vascular healing and predicted long-term clinical relapse in patients with ulcerative colitis (UC), according to data from a study of a new investigational tool.

Dr. Yasuharu Maeda, a gastroenterologist at Northern Yokohama Hospital of Showa University Center for Digestive Diseases in Yokohama, Japan, reported that 3% of patients who demonstrated vascular healing in all segments were predicted to have a clinical relapse, compared to 23.9% in patients with vascular activity (i.e. one or more segments were active).

For patients with a Mayo Endoscopic Score (MES) ≤ 1, the clinical relapse rate was 3% and 18.6% in the vascular healing and vascular active group, respectively, he said.

Endoscopic remission is a crucial „treat-to-target“ goal in patients with ulcerative colitis, and image-enhanced endoscopy is increasingly being used in routine practice to detect inflammation and predict outcomes, Maeda said.

„Image-enhanced vascular findings lead to a stronger correlation with histological activities and long-term prognosis compared to white light endoscopy assessment,“ he explained. „It also means that assessment on-site without biopsy, pathologist effort, and associated costs can be done; however, specialized training is required to achieve high accuracy of results.“

Maeda presented the data (Abstract OP16) at the 19th Congress of the European Crohn’s and Colitis Organization (ECCO).

Risk Stratification for Relapse

Maeda and colleagues developed a novel AI-based narrow-band imaging system and trained it using 8853 images from 167 patients with UC.

The AI system EndoBRAIN-UC (Cybernet System Corp, Tokyo) is currently only used and adapted for an endoscope, the Endocyto CFH290EC (Olympus EMEA, Tokyo), but for the purpose of this study, it was trained on images of five different areas.

„By combining narrow-band imaging and AI, we have developed a system that can differentiate between vascular activity and vascular healing, allowing us to predict a relapse,“ said Maeda.

In an open, prospective cohort study, they tested the system with the aim of evaluating the effectiveness of AI-identified vascular healing to stratify the risk of relapse in 100 patients showing clinical remission of UC (i.e. partial MES ≤ 1).

The patient characteristics were similar in both groups, with an average disease duration of 10 years.

In the vascular healing group (n = 33), the average age was 52 years, 20% were men, 58% had extensive colitis, and 52% had an MES score of 0.

In the vascular active group (n = 67), the average age was 56 years, 32% were men, 61% had extensive colitis, and 25% had an MES score of 0.

Using the AI system, a colonoscopy was performed to identify the mucosa of six colorectal segments of each patient as healing or active. The MES and histological assessment of these segments were also recorded. Subsequently, the patients were followed for up to 12 months and monitored for clinical relapse.

According to AI, the clinical relapse rate was higher in the vascular active group compared to the vascular healing group.

„We only evaluated the diagnostic results of the AI, but also conducted white light endoscopies and biopsies for contrast studies,“ noted Maeda.

They also investigated whether the level of experience of the endoscopist (e.g. trainee or expert) was important but found that the prediction values for clinical relapse were independent of the endoscopist’s experience.

Still in Early Stages

„AI-assisted colonoscopy is still in its early stages,“ said session co-moderator Dr. Monika Ferlitsch, Head of the Department of Internal Medicine II, Gastroenterology, and Hepatology at Evangelical Hospital in Vienna, Austria.

„We have initial results now, but I expect it will take 10 to 20 years until it is implemented into routine clinical practice,“ she said.

The best outcome for our patients is being able to predict response to therapy and recurrence rates, „and we see that this is now possible with AI. But of course, we need more clinical data to support this,“ Ferlitsch said.

Maeda and Ferlitsch did not disclose any financial disclosures.

Related Posts

Adblock Detected

Please support us by disabling your AdBlocker extension from your browsers for our website.