Background and objective: MR based radiomics can potentially response to treatment in intracranial tuberculoma, but very scarce literature is available in this regard. The purpose of this study was to determine whether MR based radiomic features can be used to predict response to antituberculosis (AT) treatment. Methods: Data of patients with intracranial tuberculomas who underwent MR imaging and AT treatment at our institution during the last 10 years was analyzed. In each case follow-up imaging performed at 6 months post initiation of treatment was reviewed to establish response to treatment. The textural analysis was performed by two consultant neuroradiologists, using open-source software (Lifex) with FLAIR coronal image after contrast administration from pretreatment MRI study radiomic analysis.
Results: Twenty-four patients with mean age 33.8 years were included in the study. Sixteen patients were in the treatment responsive group while eight patients were in the treatment resistant group. Thirty-eight radiomic parameters were extracted for each patient. There was a significant difference in three out of 38 parameters (histogram skewness, GLCM correlation and NGLDM Coarseness) in patients amongst the two groups. Logistic regression model was developed using these parameters which accurately predicted 83.3% of the cases according to the response to the AT treatment (χ2=11.517, p=0.003). ROC curve analysis was performed using histogram skewness which showed acceptable discrimination (p=0.037 and 95% CI =0.577-0.954) for predicting the response to treatment.
Conclusion: MR textural parameters (histogram skewness, GLCM correlation and NGLDM Coarseness) may be used as imaging biomarkers to predict response to treatment in patients with intracranial tuberculoma.
Awais, Muhammad; Khan, Shahmeer; Wasay, Mohammad; Azeemuddin, Muhammad; Shoukat, Ayesha; and Khan, Hafsa
"Mr Textural Features (Radiomics) For Predicting Response to Treatment in Patients with Intracranial Tuberculoma: A Retrospective Cross-Sectional Study,"
Pakistan Journal of Neurological Sciences (PJNS): Vol. 17:
3, Article 9.
Available at: https://ecommons.aku.edu/pjns/vol17/iss3/9