Background: To test the hypothesis that intraventricular ADC values can be used to determine the presence of neoplastic leptomeningeal disease (LMD).
Materials and methods: ADC values were measured at multiple sites in the ventricular system in 32 patients with cytologically-proven LMD and 40 control subjects. Multiple linear regression analysis was used to determine the mean difference of ADCs between the LMD and control groups after adjusting for ventricle size and tumor type. Receiver operating characteristics (ROC) analysis was performed and optimal ADC value cut-off point for predicting the presence of LMD. ADC was compared to T1 enhancement and FLAIR signal hyperintensity for determining the presence of LMD.
Results: After adjusting for ventricular volume and tumor type, the mid body of lateral ventricles showed no significant difference in ventricular volume and a significant difference in ADC values between the control and LMD groups (p > 0.05). In the mid-body of the right lateral ventricle the AUC was 0.69 (95% CI 0.57-0.81) with an optimal ADC cut off point of 3.22 × 10- 9 m2/s (sensitivity, specificity; 0.72, 0.68). In the mid-body of left lateral ventricle the AUC was 0.7 (95% CI 0.58-0.82) with an optimal cut-off point of 3.23 × 10- 9 m2/s (0.81, 0.62). Using an average value of HU measurements in the lateral ventricles the AUC was 0.73 (95% CI 0.61-0.84) with an optimal cut off point was 3.11 × 10- 9 m2/s (0.78, 0.65). Compared to the T1 post-contrast series, ADC was predictive of the presence of LMD in the mid-body of the left lateral ventricle (p = 0.036).
Conclusion: Complex interactions affect ADC measurements in patients with LMD. ADC values in the lateral ventricles may provide non-invasive clues to the presence of LMD.
Keywords: Apparent diffusion coefficient; Cerebrospinal fluid; Magnetic resonance imaging; Ventricles.