Leland S. Hu
MRI biomarkers have the potential to greatly impact glioma care. However, the medical community has been slow to adopt these, in part, because we don’t yet fully understand their basic relationships with histologic correlates.
For instance, we have long assumed that pMRI measurements (ie, relative cerebral blood volume, rCBV) correspond highly with histologic quantification of microvessel number (ie, microvessel density, MVD). Although this assumption has not been rigorously tested, we have used it to explain rCBV’s correlation with tumor grade, malignant potential, and patient survival. Yet numerous tissue studies have shown MVD to be only marginally useful, at best, for predicting aggressiveness and outcome in high-grade glioma. These studies instead point to microvessel area (MVA) as a superior metric. Interestingly, MVA is often inversely related to MVD, suggesting that rCBV correlates with one but not the other.
This paper tests the basic assumptions about rCBV and its microvascular correlates. We use stereotactic coregistration to show that rCBV more strongly reports MVA, rather than MVD, which justifies rCBV’s utility as a predictive biomarker in glioma. Our hope is that this rigorous validation will increase the oncology community’s confidence in pMRI and will thereby encourage its clinical adoption at more medical centers.
Stereotactic coregistration represents an incredibly useful technique for overcoming intratumoral heterogeneity, which often confounds accurate correlation between imaging and tissue metrics. We plan to extend the techniques described in this paper to validate other glioma biomarkers in the future.