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

Will AI Improve Tumor Delineation Accuracy for Radiation Therapy?

Authors
MLA Citation
Chang, Zheng. “Will AI Improve Tumor Delineation Accuracy for Radiation Therapy?Radiology, vol. 291, no. 3, June 2019, pp. 687–88. Pubmed, doi:10.1148/radiol.2019190385.
URI
https://scholars.duke.edu/individual/pub1379609
PMID
30917296
Source
pubmed
Published In
Radiology
Volume
291
Published Date
Start Page
687
End Page
688
DOI
10.1148/radiol.2019190385

Real Time Dosimetry for Gynecologic Brachytherapy: Initial Results of a Prospective Clinical Trial (vol 93, pg S203, 2015)

Authors
Chino, JP; Belley, MD; Chang, Z; Langloss, B; Yoshizumi, TT; Therien, MJ; Craciunescu, OI
MLA Citation
Chino, J. P., et al. “Real Time Dosimetry for Gynecologic Brachytherapy: Initial Results of a Prospective Clinical Trial (vol 93, pg S203, 2015).” International Journal of Radiation Oncology Biology Physics, vol. 95, no. 2, ELSEVIER SCIENCE INC, June 2016, pp. 858–59.
URI
https://scholars.duke.edu/individual/pub1149272
Source
wos
Published In
International Journal of Radiation Oncology, Biology, Physics
Volume
95
Published Date
Start Page
858
End Page
859

Susceptibility-based positive contrast MRI of brachytherapy seeds.

PURPOSE: To provide visualization of the brachytherapy seeds and differentiation with natural structures in MRI by taking advantage of their high magnetic susceptibility to generate positive-contrast images. METHODS: The method is based on mapping the susceptibility using an equivalent short-TE sequence and a kernel deconvolution algorithm with a regularized L1 minimization. An appealing aspect of the method is that signals from the surrounding areas where signal to noise ratio (SNR) is sufficiently high are used to derive the susceptibility of the seeds, even though the SNR in the immediate vicinity of the seeds can be extremely low due to rapid signal decay. RESULTS: The method is tested using computer simulations and experimental data. Comparing to conventional methods, the proposed method improves seed definition by a factor of >70% in the experiments. It produces the enhanced contrast at the exact seed location, whereas methods based on susceptibility gradient mapping produce highlighted regions surrounding the seeds. The proposed method is capable to perform the function for a wide range of resolutions and SNRs. CONCLUSION: The results show that the proposed method provides positive contrast for the seeds and correctly differentiates them from other structures that appear similar to the seeds on conventional magnitude images.
Authors
Dong, Y; Chang, Z; Xie, G; Whitehead, G; Ji, JX
MLA Citation
Dong, Ying, et al. “Susceptibility-based positive contrast MRI of brachytherapy seeds.Magn Reson Med, vol. 74, no. 3, Sept. 2015, pp. 716–26. Pubmed, doi:10.1002/mrm.25453.
URI
https://scholars.duke.edu/individual/pub1083946
PMID
25251865
Source
pubmed
Published In
Magn Reson Med
Volume
74
Published Date
Start Page
716
End Page
726
DOI
10.1002/mrm.25453

MO-G-18C-04: Improved Synthetic 4D-MRI Using Linear Polynomial Fitting Model.

PURPOSE: To reduce deformable image registration error by fitting the displacement vector field (DVF) to smooth the motion trajectory of each pixel in synthetic 4D-MRI. METHODS: Five patients with cancers in the liver were enrolled in this study. For a 4D MR image data set, the DVF matrices relative to a specific reference phase were calculated using an in-house deformable image registration based on b-spline. The displacement trajectory of each voxel throughout the respiratory cycle was constituted by concatenating the corresponding displacement values from all DVF matrices. A linear polynomial fitting model was then used to fit the DVFs in three spatial and the temporal dimension, respectively. By warpping the source MR images using the remodeled DVFs, we synthesized MR images at selected phases. Tumor motion trajectories were derived from source 4DMRI, original synthetic images and improved synthetic images. These were analyzed in the superior-inferior (SI), anterior-posterior (AP), and mediallateral (ML) directions, respectively. Correlation coefficients (CC) and differences in motion amplitude (D) were calculated for comparison. RESULTS: For all patients, tumor motion trajectories were strongly correlated between source 4D-MRI images and improved synthetic 4D-MRI (mean CC 0.98±0.01). Differences in motion amplitude were small (mean D 0.46±0.14 mm) in all directions. Correlation between source 4D-MRI and original synthetic 4D-MRI was slightly less strong (mean CC 0.97±0.01) and motion amplitude differences were slightly larger (0.55±0.19 mm). CONCLUSION: The feasibility of synthesizing T2w 4D-MRI using remodeled DVFs has been investigated in this study. Preliminary results in oncologic patients demonstrated the potential of reducing inaccuracies in original synthetic 4DMRI caused by registration errors using the linear polynomial fitting model without much loss of respiratory motion information. NIH (1R21CA165384-01A1), Golfers Against Cancer (GAC) Foundation, The China Scholarship Council (CSC).
Authors
MLA Citation
Yang, J., et al. “MO-G-18C-04: Improved Synthetic 4D-MRI Using Linear Polynomial Fitting Model.Med Phys, vol. 41, no. 6, June 2014, p. 440. Pubmed, doi:10.1118/1.4889215.
URI
https://scholars.duke.edu/individual/pub1163893
PMID
28037732
Source
pubmed
Published In
Medical Physics
Volume
41
Published Date
Start Page
440
DOI
10.1118/1.4889215