Author: Sean L. Berry, Weixing Cai, Laura I. Cervino, Maria F. Chan, Yabo Fu, Puneeth Iyengar, Hsiang-Chi Kuo, Nancy Lee, Tianfang Li, Xiang Li, Jean M. Moran 👨🔬
Affiliation: Department of Medical Physics, Memorial Sloan Kettering Cancer Center, Memorial Sloan Kettering Cancer Center 🌍
Purpose: Iodine maps derived from Dual-Energy CT (DECT) provide critical biological information for radiotherapy treatment planning, but clinical iodine maps often mistakenly include bones due to insufficient X-ray spectral separation. This study proposes a more accurate approach for quantifying iodine and calcium materials by analyzing the HU spectral curves from virtual mono-energetic images (VMIs), maximizing the separation of iodine and calcium materials to improve target contouring by accurately differentiating tumors (iodine-enhanced) from adjacent bones (calcium).
Methods: The proposed method used a multi-energy CT QA phantom to develop a voxel clustering template for varying iodine and calcium concentrations. This was performed by applying principal component analysis (PCA) to the HU spectral curves of all voxels in the VMI series. Material groups were identified in the PCA space and the iodine and calcium templates were fitted for various concentrations. During application, the patient VMI images were projected onto the same PCA space. The material concentrations were determined using the developed material decomposition template. The approach was tested on DECT images of metastatic spine patients with and without iodine contrast enhancement.
Results: The iodine and calcium voxels were effectively differentiated by the learned clustering template in the PCA space. In the testing with contrast-enhanced DECT images from metastatic spine patients, the proposed method accurately quantified the iodine-enhanced areas and the bone regions. In testing with multi-energy CT phantom images, the proposed approach successfully excluded the calcium inserts from the iodine ones in the material map unlike the clinical software.
Conclusion: The proposed approach maximizes material separation in the PCA space, reliably quantifies the iodine and calcium materials. This provides a valuable direct overlay for target and OARs contouring in radiotherapy.