Gastrointestinal stromal tumours: Correlation of modified NIH risk stratification with diffusion-weighted MR imaging as an imaging biomarker

Purpose To evaluate the correlation of risk grade of gastrointestinal stromal tumours (GISTs) based on modified National Institutes of Health (NIH) criteria with conventional magnetic resonance (MR) imaging and diffusion-weighted (DW) imaging. Methods We included 22 patients with histopathologically proven GISTs in the stomach or small bowel who underwent pre-operative gadoxetic acid-enhanced MR imaging and DW imaging. We retrospectively assessed correlations between morphologic findings, qualitative (signal intensity, consensus from two observers) and quantitative (degree of dynamic enhancement using signal intensity of tumour/muscle ratio and apparent diffusion coefficient [ADC]) values, and the modified NIH criteria for risk stratification. Spearman partial correlation analysis was used to control for tumour size as a confounding factor. The optimal cut-off level of ADC values for intermediate or high risk GISTs was analyzed using a receiver operating characteristic analysis. Results Except tumour size and necrosis, conventional MR imaging findings, including the degree of dynamic enhancement, were not significantly different according to the modified NIH criteria (p > 0.05). Tumour ADC values were negatively correlated with the modified NIH criteria, before and after adjustment of tumour size (ρ = -0.754; p < 0.001 and ρ = -0.513; p = 0.017, respectively). The optimal cut-off value for the determination of intermediate or high-risk GISTs was 1.279 × 10^-3 mm²/s (100% sensitivity, 69.2% specificity, 81.8% accuracy). Conclusion Except tumour size and necrosis, conventional MR imaging findings did not correlate with the risk grade. However, the ADC value can be used as an imaging biomarker to assess the risk grade of GISTs, regardless of tumour size.