Research

Medical Imaging

Organ Segmentation and Treatment Target Delineation

Image Reconstruction and Restoration

 


 

Organ Segmentation and Treatment Target Delineation

Accurate delineation of tumors and sensitive structures is important for many medical applications. One example is treatment planning in cancer radiotherapy. Although many auto-segmentation algorithms have been developed and implemented in clinical practice, none of them are satisfactory and manual contouring is often required. Such failure is mainly because conventional segmentation methods are purely based on local information in the images. Especially for treatment target delineation we often rely on information beyond the images which is very challenging, if not impossible, for conventional methods. We have been investigating deep learning based strategies to solve this problem.

The figure below shows an example of delineating the post-prostatectomy clinical target volume using deep learning for prostate cancer radiotherapy, where in addition to CT images, we need to use information like surgery reports, surgical pathology, pre-operative MRI knowledge of tumor location and organ invasion, etc.

 

Delineation of the post-prostatectomy clinical target volume using deep learning for prostate cancer radiotherapy

 

Publications

  1. Kazemifar, S., Balagopal, A., Nguyen, D., McGuire, S., Hannan, R., Jiang, S., & Owrangi, A. (2018). Segmentation of the prostate and organs at risk in male pelvic CT images using deep learning. Biomedical Physics and Engineering Express. (journal)(arXiv)
  2. Balagopal, A., Kazemifar, S., Nguyen, D., Lin, M. H., Hannan, R., Owrangi, A., & Jiang, S. (2018). Fully automated organ segmentation in male pelvic CT images. Physics in Medicine & Biology, 63(24), 245015. (journal)(arXiv)
  3. Li, S., Xu, P., Li, B., Chen, L., Zhou, Z., Hao, H., ... & Wang, J. (2018). Predicting Lung Nodule Malignancies by Combining Deep Convolutional Neural Network and Handcrafted Features. arXiv preprint arXiv:1809.02333. (arXiv)
  4. Zhou, Z., Li, S., Qin, G., Folkert, M., Jiang, S., & Wang, J. (2018). Automatic multi-objective based feature selection for classification. arXiv preprint arXiv:1807.03236. (arXiv)
  5. Liu, Y., Stojadinovic, S., Hrycushko, B., Wardak, Z., Lau, S., Lu, W., ... & Nedzi, L. (2017). A deep convolutional neural network-based automatic delineation strategy for multiple brain metastases stereotactic radiosurgery. PloS one, 12(10), e0185844. (journal)
  6. Rozario, T., Long, T., Chen, M., Lu, W., & Jiang, S. (2017). Towards automated patient data cleaning using deep learning: A feasibility study on the standardization of organ labeling. arXiv preprint arXiv:1801.00096. (arXiv)
  7. Liu, Y., Stojadinovic, S., Hrycushko, B., Wardak, Z., Lu, W., Yan, Y., ... & Gu, X. (2016). Automatic metastatic brain tumor segmentation for stereotactic radiosurgery applications. Physics in medicine and biology, 61(24), 8440. (journal)

Back to top 

 


 

Image Reconstruction and Restoration

By introduction of machine learning and data-driven methods, a paradigm shift has happened in the field of imaging in general and accordingly these new methods have a high translational impact in the area of medical imaging.

These impacts are not limited to only image analysis and pattern recognition techniques but also they are used widely in the field of image reconstruction. Recently, multiple groups worldwide, with encouraging results and increasing interest, are actively exploring deep learning techniques for image reconstruction and other inverse problems.

Currently, our team is working on developing new machine learning based methods for implementing MRI-only planning into the clinic. Challenges include producing robust MRI-only patient models and synthetic CT scans with accurate geometry and electron densities.

However, our novel machine learning based method allows us to simultaneously achieve electron density map for dose calculation and automatic segmentation of the target and OAR.

The successful completion of this research will provide essential tools to establish effective MRI-based RT planning in routine clinical practice, which will improve normal and target tissue delineation and localization for more accurate radiation delivery. 

 

Publications

  1. Zhong, Y., Vinogradskiy, Y., Chen, L., Myziuk, N., Castillo, R., Castillo, E., & Wang, J. (2019). Deriving ventilation imaging from 4 DCT by deep convolutional neural network. Medical physics. (journal)(arXiv)
  2. Shen, C., Gonzalez, Y., Chen, L., Jiang, S. B., & Jia, X. (2018). Intelligent parameter tuning in optimization-based iterative CT reconstruction via deep reinforcement learning. IEEE transactions on medical imaging, 37(6), 1430-1439. (journal)(arXiv)
  3. Iqbal, Z., Nguyen, D., & Jiang, S. (2018). Super-Resolution 1H Magnetic Resonance Spectroscopic Imaging utilizing Deep Learning. arXiv preprint arXiv:1802.07909. (arXiv)
  4. Iqbal, Z., Nguyen, D., Thomas, M. A., & Jiang, S. (2018). Acceleration and Quantitation of Localized Correlated Spectroscopy using Deep Learning: A Pilot Simulation Study. arXiv preprint arXiv:1806.11068. (arXiv)
  5. Liang, X., Chen, L., Nguyen, D., Zhou, Z., Gu, X., Yang, M., ... & Jiang, S. (2018). Generating Synthesized Computed Tomography (CT) from Cone-Beam Computed Tomography (CBCT) using CycleGAN for Adaptive Radiation Therapy. arXiv preprint arXiv:1810.13350. (arXiv)
  6. Ma, G., Shen, C., & Jia, X. (2018). Low dose CT reconstruction assisted by an image manifold prior. arXiv preprint arXiv:1810.12255. (arXiv)

Back to top