A recent study published in European Radiology has found that the CT radiomics model rather than CT- and endoscopic sonography (EUS) based high-risk morphological features can potentially be applied for the preoperative risk stratification of 1–2-cm gastric gastrointestinal stromal tumors (gGISTs).

Small gastric gastrointestinal stromal tumors have a diameter of under 2 cm. Studies have reported that the malignant risk of gGISTs with 1–2-cm diameters were 10.1%, thus warranting an urgent need to identify them to develop the surgery strategy, the researchers noted. EUS is considered the most commonly used imaging technique in the identification and surveillance of small gGISTs, they added.

Read more about GIST therapies

Continue Reading

Previous studies have revealed that CT scans exhibited a promising detection rate of 100.0% for 1–2 cm diameter gGISTs as identified earlier through endoscopy or EUS, the study team noted. Hence, it can be assumed that CT scans can be applied for the risk stratification of 1–2-cm gGISTs, they added.

Recently concluded studies have exhibited exceptional accuracy of radiomics in anticipating the lethal potential of GISTs with greater sizes and could aid in clinical decision-making of treatment and surveillance, the investigators noted.

In this study, the researchers investigated if the high-risk features detected by CT and EUS and a radiomics approach based on CT images could anticipate the fatal risk of 1–2-cm gGISTs. 

The research team included 151 pathologically confirmed patients with 1–2-cm gGISTs from seven different institutions. The gGISTs were arbitrarily classified into three groups for building the radiomics model, containing 74 in the training cohort, 37 in the validation cohort, and 40 in the testing cohort. The inclusion criteria for gGISTs were patients who got complete surgical resection of lesions, patients without preoperative treatment, and patients who got contrast-enhanced abdominal CT scans less than 30 days before the surgery.

Study results showed that 20 lesions out of a total 151 were reported to be affected with pathological proven intermediate risk of malignancy while the remaining 131 were affected with very low risk of malignancy. 

“In our study, high-risk features detected by contrast-enhanced CT and EUS could not effectively predict the malignant potential of 1–2-cm gGISTs, whereas the radiomics model based on contrast-enhanced CT could identify 1–2-cm gGISTs with a higher risk of malignancy,” the authors highlighted.

Moreover, the study reported no high-risk 1–2-cm gGISTs among the enrolled patients. The results suggest that the morphological features cannot be reliable for the preoperative prediction of malignant potential of 1–2-cm gGISTs. An accurate and quantitative technique is needed for risk stratification.

“In conclusion, instead of CT- and EUS-detected morphological high-risk features, the CT radiomics model demonstrated favorable discrimination of malignant 1–2-cm gGISTs,” the authors further added.


Jia X, Wan L, Chen X, et al. Risk stratification for 1- to 2-cm gastric gastrointestinal stromal tumors: visual assessment of CT and EUS high-risk features versus CT radiomics analysis. Euro Radiol. Published November 30, 2022. doi: 10.1007/s00330-022-09228-x