Zheng Chang

Overview:

Dr. Chang's research interests include radiation therapy treatment assessment using MR quantitative imaging, image guided radiation therapy (IGRT), fast MR imaging using parallel imaging and strategic phase encoding, and motion management for IGRT.

Positions:

Professor of Radiation Oncology

Radiation Oncology
School of Medicine

Member of the Duke Cancer Institute

Duke Cancer Institute
School of Medicine

Education:

Ph.D. 2006

University of British Columbia

Grants:

Publications:

An investigation of machine learning methods in delta-radiomics feature analysis.

PURPOSE: This study aimed to investigate the effectiveness of using delta-radiomics to predict overall survival (OS) for patients with recurrent malignant gliomas treated by concurrent stereotactic radiosurgery and bevacizumab, and to investigate the effectiveness of machine learning methods for delta-radiomics feature selection and building classification models. METHODS: The pre-treatment, one-week post-treatment, and two-month post-treatment T1 and T2 fluid-attenuated inversion recovery (FLAIR) MRI were acquired. 61 radiomic features (intensity histogram-based, morphological, and texture features) were extracted from the gross tumor volume in each image. Delta-radiomics were calculated between the pre-treatment and post-treatment features. Univariate Cox regression and 3 multivariate machine learning methods (L1-regularized logistic regression [L1-LR], random forest [RF] or neural networks [NN]) were used to select a reduced number of features, and 7 machine learning methods (L1-LR, L2-LR, RF, NN, kernel support vector machine [KSVM], linear support vector machine [LSVM], or naïve bayes [NB]) was used to build classification models for predicting OS. The performances of the total 21 model combinations built based on single-time-point radiomics (pre-treatment, one-week post-treatment, and two-month post-treatment) and delta-radiomics were evaluated by the area under the receiver operating characteristic curve (AUC). RESULTS: For a small cohort of 12 patients, delta-radiomics resulted in significantly higher AUC than pre-treatment radiomics (p-value<0.01). One-week/two-month delta-features resulted in significantly higher AUC (p-value<0.01) than the one-week/two-month post-treatment features, respectively. 18/21 model combinations were with higher AUC from one-week delta-features than two-month delta-features. With one-week delta-features, RF feature selector + KSVM classifier and RF feature selector + NN classifier showed the highest AUC of 0.889. CONCLUSIONS: The results indicated that delta-features could potentially provide better treatment assessment than single-time-point features. The treatment assessment is substantially affected by the time point for computing the delta-features and the combination of machine learning methods for feature selection and classification.
Authors
MLA Citation
Chang, Yushi, et al. “An investigation of machine learning methods in delta-radiomics feature analysis.Plos One, vol. 14, no. 12, 2019, p. e0226348. Pubmed, doi:10.1371/journal.pone.0226348.
URI
https://scholars.duke.edu/individual/pub1423241
PMID
31834910
Source
pubmed
Published In
Plos One
Volume
14
Published Date
Start Page
e0226348
DOI
10.1371/journal.pone.0226348

A Technique for Generating Volumetric Cine-Magnetic Resonance Imaging.

PURPOSE: The purpose of this study was to develop a techique to generate on-board volumetric cine-magnetic resonance imaging (VC-MRI) using patient prior images, motion modeling, and on-board 2-dimensional cine MRI. METHODS AND MATERIALS: One phase of a 4-dimensional MRI acquired during patient simulation is used as patient prior images. Three major respiratory deformation patterns of the patient are extracted from 4-dimensional MRI based on principal-component analysis. The on-board VC-MRI at any instant is considered as a deformation of the prior MRI. The deformation field is represented as a linear combination of the 3 major deformation patterns. The coefficients of the deformation patterns are solved by the data fidelity constraint using the acquired on-board single 2-dimensional cine MRI. The method was evaluated using both digital extended-cardiac torso (XCAT) simulation of lung cancer patients and MRI data from 4 real liver cancer patients. The accuracy of the estimated VC-MRI was quantitatively evaluated using volume-percent-difference (VPD), center-of-mass-shift (COMS), and target tracking errors. Effects of acquisition orientation, region-of-interest (ROI) selection, patient breathing pattern change, and noise on the estimation accuracy were also evaluated. RESULTS: Image subtraction of ground-truth with estimated on-board VC-MRI shows fewer differences than image subtraction of ground-truth with prior image. Agreement between normalized profiles in the estimated and ground-truth VC-MRI was achieved with less than 6% error for both XCAT and patient data. Among all XCAT scenarios, the VPD between ground-truth and estimated lesion volumes was, on average, 8.43 ± 1.52% and the COMS was, on average, 0.93 ± 0.58 mm across all time steps for estimation based on the ROI region in the sagittal cine images. Matching to ROI in the sagittal view achieved better accuracy when there was substantial breathing pattern change. The technique was robust against noise levels up to SNR = 20. For patient data, average tracking errors were less than 2 mm in all directions for all patients. CONCLUSIONS: Preliminary studies demonstrated the feasibility of generating real-time VC-MRI for on-board localization of moving targets in radiation therapy.
Authors
Harris, W; Ren, L; Cai, J; Zhang, Y; Chang, Z; Yin, F-F
MLA Citation
Harris, Wendy, et al. “A Technique for Generating Volumetric Cine-Magnetic Resonance Imaging.Int J Radiat Oncol Biol Phys, vol. 95, no. 2, June 2016, pp. 844–53. Pubmed, doi:10.1016/j.ijrobp.2016.02.011.
URI
https://scholars.duke.edu/individual/pub1130881
PMID
27131085
Source
pubmed
Published In
Int J Radiat Oncol Biol Phys
Volume
95
Published Date
Start Page
844
End Page
853
DOI
10.1016/j.ijrobp.2016.02.011

A phase 1 trial of preoperative partial breast radiation therapy: Patient selection, target delineation, and dose delivery.

PURPOSE: Diffusion of accelerated partial breast irradiation into clinical practice is limited by the need for specialized equipment and training. The accessible external beam technique yields unacceptable complication rates, likely from large postoperative target volumes. We designed a phase 1 trial evaluating preoperative radiation therapy to the intact tumor using widely available technology. METHODS AND MATERIALS: Patients received 15, 18, or 21 Gy in a single fraction to the breast tumor plus margin. Magnetic resonance imaging (MRI) was used in conjunction with standard computed tomography (CT)-based planning to identify contrast enhancing tumor. Skin markers and an intratumor biopsy marker were used for verification during treatment. RESULTS: MRI imaging was critical for target delineation because not all breast tumors were reliably identified on CT scan. Breast shape differences were consistently seen between CT and MRI but did not impede image registration or tumor identification. Target volumes were markedly smaller than historical postoperative volumes, and normal tissue constraints were easily met. A biopsy marker within the breast proved sufficient for setup localization. CONCLUSIONS: This single fraction linear accelerator-based partial breast irradiation approach can be easily incorporated at most treatment centers. In vivo targeting may improve accuracy and can reduce the dose to normal tissues.
Authors
Blitzblau, RC; Arya, R; Yoo, S; Baker, JA; Chang, Z; Palta, M; Duffy, E; Horton, JK
MLA Citation
Blitzblau, Rachel C., et al. “A phase 1 trial of preoperative partial breast radiation therapy: Patient selection, target delineation, and dose delivery.Pract Radiat Oncol, vol. 5, no. 5, Sept. 2015, pp. e513–20. Pubmed, doi:10.1016/j.prro.2015.02.002.
URI
https://scholars.duke.edu/individual/pub1062048
PMID
25834942
Source
pubmed
Published In
Pract Radiat Oncol
Volume
5
Published Date
Start Page
e513
End Page
e520
DOI
10.1016/j.prro.2015.02.002

Review of treatment assessment using DCE-MRI in breast cancer radiation therapy.

As a noninvasive functional imaging technique, dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) is being used in oncology to measure properties of tumor microvascular structure and permeability. Studies have shown that parameters derived from certain pharmacokinetic models can be used as imaging biomarkers for tumor treatment response. The use of DCE-MRI for quantitative and objective assessment of radiation therapy has been explored in a variety of methods and tumor types. However, due to the complexity in imaging technology and divergent outcomes from different pharmacokinetic approaches, the method of using DCE-MRI in treatment assessment has yet to be standardized, especially for breast cancer. This article reviews the basic principles of breast DCE-MRI and recent studies using DCE-MRI in treatment assessment. Technical and clinical considerations are emphasized with specific attention to assessment of radiation treatment response.
Authors
Wang, C-H; Yin, F-F; Horton, J; Chang, Z
MLA Citation
Wang, Chun-Hao, et al. “Review of treatment assessment using DCE-MRI in breast cancer radiation therapy.World J Methodol, vol. 4, no. 2, June 2014, pp. 46–58. Pubmed, doi:10.5662/wjm.v4.i2.46.
URI
https://scholars.duke.edu/individual/pub1048364
PMID
25332905
Source
pubmed
Published In
World Journal of Methodology
Volume
4
Published Date
Start Page
46
End Page
58
DOI
10.5662/wjm.v4.i2.46

TH-C-141-07: T2-Weighted 4D-MRI with Combined Phase and Amplitude Sorting.

PURPOSE: T2-weighted MR provides excellent delineation of malignant liver lesions due to its superior tumor-to-tissue contrast. This study aims to develop a novel T2-weighted retrospective 4D-MRI technique for imaging organ/tumor respiratory motion with improved soft-tissue contrast. METHOD AND MATERIALS: Determine the number of repeated scans (NR) required obtaining sufficient phase information for each slice is the critical component in developing this technique and needs substantial testing with many samples. To do that, computer simulations using RPM respiratory signals of 29 cancer patients were performed to derive the relationships between NR and the following factors: number of slice to scanned (Ns), number of respiratory phases of the 4D-MRI (Np), and starting phase at image acquisition (P0). Assuming T2-w HASTE/SSFSE MR sequence to be used to acquire raw images for 4D-MRI, frame rate of 2 frames/s was used in the simulation. To validate our technique, 4D-MRI acquisition and reconstruction were simulated on a 4D digital human phantom using parameters derived from the above studies. Retrospective sorting of 4D-MRI was achieved using a novel phase and amplitude hybrid sorting algorithm by effectively utilizing redundant images. RESULT: Percentage of complete acquisition of all required phases (Cp) increased as NR increased in an inverse-exponential (Cp=100*[1-exp(-0.11*NR)],when Ns=50,Np=10) fashion. NR to achieve 95% completion (Cp=95%) of all required phases, defined as the NR needed for 4D-MRI, is linearly proportional to Np (Nr∼2.86*Np, r=1.0) but independent of NS and P0. Simulated 4D-MRI on the digital phantom showed clear pattern of respiratory motion. Tumor motion trajectories measured on 4D-MRI were comparable to the average input signal, with a mean relative difference in motion amplitude of 16%, presumably due to breathing irregularity. CONCLUSIONS: A novel T2-weighted 4D-MRI technique based on HASTE/SSFSE sequence have been developed and validated. Future evaluation on patients is desired. This work is partly supported by funding from NIH (1R21CA165384-01A1) and a research grant from the Golfers Against Cancer (GAC) Foundation.
Authors
Liu, Y; Chang, Z; Czito, B; Bashir, M; Yin, F; Cai, J
MLA Citation
Liu, Y., et al. “TH-C-141-07: T2-Weighted 4D-MRI with Combined Phase and Amplitude Sorting.Med Phys, vol. 40, no. 6Part32, June 2013, p. 540. Pubmed, doi:10.1118/1.4815775.
URI
https://scholars.duke.edu/individual/pub1164326
PMID
28517889
Source
pubmed
Published In
Med Phys
Volume
40
Published Date
Start Page
540
DOI
10.1118/1.4815775