Visual versus region-of-interest based diffusion evaluation and their diagnostic impact in adult-type diffuse gliomas

Area of Science:

• Neuroimaging
• Oncology
• Radiology

Background:

• Adult gliomas require accurate assessment of isocitrate dehydrogenase (IDH) mutation status for prognosis and treatment.
• Diffusion-weighted imaging (DWI) is a key MRI technique for evaluating brain tumor characteristics.

Purpose of the Study:

• To compare the reproducibility of visual versus region-of-interest (ROI)-based diffusion assessment in adult gliomas.
• To evaluate the capacity of these methods to predict IDH mutation status.

Main Methods:

• Preoperative MRI scans of 303 adult gliomas underwent visual and ROI-based diffusion assessment by three raters, repeated after one month.
• ROI-based assessment utilized apparent diffusion coefficient (ADC) values (absolute/aADC and normalized/nADC).
• Logistic regression analyzed IDH prediction in 99 non-necrotic, non-hemorrhagic cases.

Main Results:

• ROI-based assessment showed superior inter- and intra-rater agreement (ICC=0.56) compared to visual assessment (W/kappa=0.34).
• ADC thresholds correlated significantly with visual assessments (P < .001).
• IDH classification accuracy was comparable (visual: 69%, ROI: 73%/70%), but visual assessment had high specificity (99%) and low sensitivity (14%), while ROI improved sensitivity (57%/61%) with reduced specificity (81%/75%).

Conclusions:

• ROI-based diffusion assessment, guided by visual input, offers superior reproducibility over visual assessment alone.
• Visual assessment provides high specificity but low sensitivity for IDH prediction.
• ROI-based assessment enhances sensitivity for IDH prediction at the expense of specificity.