Quantitative validation of MRI mapping of cerebral venous oxygenation with direct blood sampling: A graded-O study in piglets.

PURPOSE

To validate two neonatal cerebral venous oxygenation (Y ) MRI techniques, T relaxation under phase contrast (TRUPC) and accelerated TRUPC (aTRUPC) MRI, with oxygenation measured with direct blood sampling.

METHODS

In vivo experiments were performed on seven healthy newborn piglets. For each piglet, a catheter was placed in the superior sagittal sinus to obtain venous blood samples for blood gas oximetry measurement as a gold standard. During the MRI experiment, three to five venous oxygenation levels were achieved in each piglet by varying inhaled O content and breathing rate. Under each condition, Y values of the superior sagittal sinus measured by TRUPC, aTRUPC, and blood gas oximetry were obtained. The Y quantification in TRUPC and aTRUPC used a standard bovine blood calibration model. The aTRUPC scan was repeated twice to assess its reproducibility. Agreements among TRUPC Y , aTRUPC Y , and blood gas oximetry were evaluated by intraclass correlation coefficient (ICC) and paired Student's t-test.

RESULTS

The mean hematocrit was 23.6 ± 6.5% among the piglets. Across all measurements, Y values were 51.9 ± 21.3%, 54.1 ± 18.8%, and 53.7 ± 19.2% for blood gas oximetry, TRUPC and aTRUPC, respectively, showing no significant difference between any two methods (P > .3). There were good correlations between TRUPC and blood gas Y (ICC = 0.801; P < .0001), between aTRUPC and blood gas Y (ICC = 0.809; P < .0001), and between aTRUPC and TRUPC Y (ICC = 0.887; P < .0001). The coefficient of variation of aTRUPC Y was 8.1 ± 9.9%.

CONCLUSION

The values of Y measured by TRUPC and aTRUPC were in good agreement with blood gas oximetry. These findings suggest that TRUPC and aTRUPC can provide accurate quantifications of Y in major cerebral veins.