Author: Cem Dede, Clifton David Fuller, Renjie He, Laia Humbert Vidan, Stephen Y. Lai, Amy Moreno, Mohamed Naser, Kareem Abdul Wahid, Natalie A West 👨🔬
Affiliation: Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center; The University of Texas MD Anderson Cancer Center UTHealth Houston Graduate School of Biomedical Sciences, Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, The University of Texas MD Anderson Cancer Center, Department of Radiation Oncology, Department of Head and Neck Surgery, The University of Texas MD Anderson Cancer Center 🌍
Purpose: Head and neck cancers (HNC) treated with radiation therapy can result in radiation-induced bone injury. In CT, changes in radiodensity correlate to changes in bone integrity. Most studies use the mandible as one structure, neglecting anatomical differences across sub-volumes. We propose black bone (BB), a low-flip angle MRI sequence that can visualize bone comparable to CT. We aim to determine if BB is a reliable surrogate to assess relative intensity differences across mandible sub-volumes compared to CT. If successful, this method in correlating changes in intensity and tissue density on BB may help detect radiation damage.
Methods: 12 HNC patients were scanned on CT and 1.5T MR simulation scanners. 7 mandible sub-volumes were segmented using a deep learning-based autosegmentation model on CT. BB was registered to CT via mutual information-based rigid registration. Intensities were calculated for each sub-volume on CT and BB and the modalities were normalized by computing the ratio of sub-volumes. A Wilcoxon signed-rank test assessed for significant differences.
Results: CT and BB had basal/alveolar ratios ranging from [1.02,1.35] and [0.77,1.32], respectively, across all 12 patients. A Wilcoxon signed-rank of the ratios showed no significant difference (p=0.13), indicating comparability. BB intensities are inverted compared to CT, thus BB alveolar/basal ratios of [0.76,1.30] were computed. Comparing this ratio with the CT alveolar/basal, Wilcoxon signed-rank indicated a significant difference (p=0.001).
Conclusion: BB shows promise as a non-ionizing surrogate to CT in mandibular sub-volume intensity ratios. This exploratory study sets a pipeline for a testable hypothesis using a cross-sectional dataset at one timepoint across imaging modalities. It is known that MRI intensities are inconsistent across scan parameters and timepoints, resulting in intra- and inter-patient variability, which needs to be investigated further. Future work will consist of optimizing the autosegmentation and registration processes on BB to minimize error in intensity measurements.