Utilization of a hybrid finite-element based registration method to quantify heterogeneous tumor response for adaptive treatment for lung cancer patients. (21st March 2018)

• Record Type: Journal Article
• Authors: Sharifi, Hoda; Zhang, Hong; Bagher-Ebadian, Hassan; Lu, Wei; Ajlouni, Munther I; Jin, Jian-Yue; Kong, Feng-Ming (Spring); Chetty, Indrin J; Zhong, Hualiang
• Is Part Of: Physics in medicine & biology. Volume 63:Number 6(2018:Mar.)
• Publication Date: 2018-03-21
• Subjects: treatment response -- PET imaging -- tumor regression -- deformable image registration -- adaptive treatment planning; Biophysics -- Periodicals; Medical physics -- Periodicals; 610.153
• DOI: 10.1088/1361-6560/aab235
• Languages: English
• ISSNs: 0031-9155

Abstract

Abstract: Tumor response to radiation treatment (RT) can be evaluated from changes in metabolic activity between two positron emission tomography (PET) images. Activity changes at individual voxels in pre-treatment PET images (PET1), however, cannot be derived until their associated PET-CT (CT1) images are appropriately registered to during-treatment PET-CT (CT2) images. This study aimed to investigate the feasibility of using deformable image registration (DIR) techniques to quantify radiation-induced metabolic changes on PET images. Five patients with non-small-cell lung cancer (NSCLC) treated with adaptive radiotherapy were considered. PET-CTs were acquired two weeks before RT and 18 fractions after the start of RT. DIR was performed from CT1 to CT2 using B-Spline and diffeomorphic Demons algorithms. The resultant displacements in the tumor region were then corrected using a hybrid finite element method (FEM). Bitmap masks generated from gross tumor volumes (GTVs) in PET1 were deformed using the four different displacement vector fields (DVFs). The conservation of total lesion glycolysis (TLG) in GTVs was used as a criterion to evaluate the quality of these registrations. The deformed masks were united to form a large mask which was then partitioned into multiple layers from center to border. The averages of SUV changes over all the layers were 1.0 ± 1.3, 1.0 ± 1.2, 0.8 ± 1.3, 1.1 ± 1.5 for the B-Spline, B-Spline + FEM, Demons and Demons + FEM algorithms, respectively. TLG changes before and after mapping using B-Spline, Demons, hybrid-B-Spline, and hybrid-Demons registrations were 20.2%, 28.3%, 8.7%, and 2.2% on average, respectively. Compared to image intensity-based DIR algorithms, the hybrid FEM modeling technique is better in preserving TLG and could be useful for evaluation of tumor response for patients with regressing tumors.