Endoscopic ultrasonography assisted by artificial intelligence is highly accurate in diagnosing gastrointestinal stromal tumors (GISTs), according to a meta-analysis.

The results showed that the pooled sensitivity of artificial intelligence-assisted endoscopic ultrasonography in differentiating GIST from other submucosal tumors was 0.92. Its specificity was 0.82, and the positive and negative likelihood ratios were 4.55 and 0.12, respectively. The summary diagnostic odds ratio was 64.7, and the area under the curve was 0.95.

It could therefore be “a valuable complementary method for the differentiation of [submucosal tumors] in the future,” the authors wrote. The results were recently published in the journal Surgical Endoscopy.

Continue Reading

Endoscopic ultrasonography can help diagnose GIST. However, subjective interpretation limits its accuracy.

Read more about the diagnosis of GIST

In the present study, a team of researchers led by Jing Zhu, of the First Affiliated Hospital of Chongqing Medical University, in Chongqing, China, conducted a meta-analysis of the literature to assess how useful artificial intelligence can be in diagnosing GIST using endoscopic ultrasonography images.

The researchers included 7 studies in their analysis. These studies included 2431 patients and contained more than 36,000 images. Of these, 2431 patients, 480 were used for the final testing. 

The researchers concluded that artificial intelligence-assisted endoscopic ultrasonography is a “promising noninvasive tool to diagnose GISTs” and can help distinguish it from other submucosal tumors with high accuracy.

More research is necessary to assess the diagnostic accuracy of the approach as well as to predict the malignant potential of GISTs, they added.

GIST is a rare type of submucosal tumor that develops from the interstitial cells of Cajal. The symptoms of the disease differ depending on the size and location of the tumor. They may include anemia, hematemesis, abdominal pain, and bloody bowel movements. In many cases, the disease remains asymptomatic until the tumor grows to around 6 cm.


Zhang B, Zhu F, Li P, Zhu J. Artificial intelligence-assisted endoscopic ultrasound in the diagnosis of gastrointestinal stromal tumors: a meta-analysis. Surg Endosc. Published online September 13, 2022. doi:10.1007/s00464-022-09597-w