Radiomics Model for Differentiating Gastric Schwannomas and GISTs
Author Information
Author(s): Yang Zimei, Ma Chongfei, Ren Jialiang, Li Min, Xv Xiaosheng, Fu Xin, Yang Li
Primary Institution: Fourth Hospital of Hebei Medical University
Hypothesis
Can a radiomics model improve the differentiation between gastric schwannomas and gastrointestinal stromal tumors across risk grades?
Conclusion
The combined model based on CT imaging features and radiomics features presents a promising non-invasive approach for accurate preoperative differentiation between gastric schwannomas and gastrointestinal stromal tumors.
Supporting Evidence
- The combined model achieved AUCs of 0.989 and 0.964 in the training and validation cohorts, respectively.
- Independent predictors for differentiating GS from GIST included tumor location, cystic changes, degree of enhancement in arterial phase, and enhancement uniformity.
- The radiomics model demonstrated AUCs of 0.949 and 0.839 in the training and validation cohorts.
- Statistically significant differences were observed in tumor location and enhancement uniformity between GS and GIST.
Takeaway
This study created a new way to tell apart two types of stomach tumors using special imaging techniques, which can help doctors make better treatment decisions.
Methodology
A retrospective analysis of 26 gastric schwannomas and 82 gastrointestinal stromal tumors was conducted, using logistic regression and radiomics feature extraction to develop diagnostic models.
Potential Biases
The rarity of gastric schwannomas may affect the statistical power of the findings.
Limitations
The study is retrospective and single-center, which may limit generalizability and introduces potential selection bias.
Participant Demographics
The cohort included 26 gastric schwannomas (10 males, 16 females) and 82 gastrointestinal stromal tumors (34 males, 48 females), with an average age of around 61 years for both groups.
Statistical Information
P-Value
P<0.05
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website