Correlation of CT imaging features and CT texture analysis parameters with pathologic risk stratification in gastric stromal tumors

Abstract

To analyze the correlation of the CT imaging features and CT texture analysis (CTTA) parameters with risk stratifications of the gastric stromal tumors (GSTs). Methods: CT (plain scans with portal phase contrast enhanced scans) images from 98 GSTs patients before surgery were collected. CT features of the tumors were retrospectively analyzed and pathological risk stratifications were determined. Feature parameters of CTTA at plain and portal phase were obtained by using a MaZda software. The correlations of CT imaging features and CTTA parameters with the risk stratifications were analyzed. Results: CT imaging features including tumor size, growth pattern, shape, margin, the relationship between mass and adjacent organs, enhancement pattern, necrosis or cystic change, and the abnormal tumor vessels were associated with the risk stratifications (P<0.05). For CTTA, entropy was associated with the risk stratifications (P<0.05). Conclusion: Certain CT imaging features and CTTA parameters are associated with risk classifications in GSTs.