Ying Xiao, Ph.D. Professor, Department of Radiation Oncology
Perelman Center for Advanced Medicine
3400 Civic Center Blvd TRC-2
West Philadelphia, PA 19104
Our group supports IROC RT QA Center on quality assurance (QA) of imaging and radiotherapy. Our work involves the following:
- Support establishment of appropriate QA procedures for advanced medical imaging and radiotherapy.
- Consultation to the NCTN groups in the development of protocols. Assist with hypothesis generation and trial design that can be supported by effective QA programs.
- Radiotherapy treatment planning studies.
- Qualification and credentialing policies and help ensure the delivery of appropriate protocol-specified radiotherapy and advanced imaging
- Policy and procedure maintenance and updates.
- Support on case submission and case review, pre-review and post-review data management.
- Maintaining standard structure name library.
- Providing quality assurance for the piloting personalized Radium-223 dosimetry
Our research direction can be summarized in four main categories.
♦ Knowledge Based Planning Quality Research for both IMRT and IMPT plans with Varian Rapidlan.
Knowledge based planning utilizes pre-existing clinical knowledge to generate estimated dose volume histogram based on patient-specific anatomy and prescription information. Using RapidPlan allows reduction of variability in treatment planning quality, and enhances efficiency for patient care. We build and train models with NRG clinical trial data, and use the model as a quality assurance evaluation tool to study the submitted RT plan.
♦ RT imaging analysis: Automated Contour Delineation with Deep Learning methods.
Contour accuracy is essential for the quality of Radio-Therapy. Various methods are explored in our group for contour quality assurance. Convolution neural network (CNN) based methods have demonstration high accuracy for auto-segmentation of patient structures. Our group is in the process of built CNN based auto-segmentation models for critical organs in various disease sites. Methods are being developed to use these models for contour quality assurances.
♦ Predictive modeling using machine learning methods.
The demographics, prognosis, dosimetric, and RT modality information can all be collected as features for prediction model. Features typically include gender, age, ethnicity, treatment modality, tumor clinical staging, pathological staging, surgery type, chemotherapy and treatment techniques (IMRT/CRT, PBS/Double scattering). Recently the research of Radiomics hypothesized that medical imaging provides crucial information about tumor physiology, and could be exploited to enhance cancer diagnostics. Radiomics features are also extracted in addition to the above clinical features for modeling overall survival(OS) and local recurrence(LR) preditions. Various machine learning methods such as Cox regression, random survival forest are explored. Recently 3D convolution neural network is also used to model patient dose distribution and patient toxicity.
♦ Provide quality assurance for the piloting personalized Radium-223 dosimetry.
In radiopharmaceutical therapy (RPT), a radionuclide is systemically or locally delivered with the goal of targeting and delivering radiation to cancer cells while minimizing radiation exposure to untargeted cells. Assessment of radiation doses in individual patients and their correlation with tumor and normal tissues response to radiation is essential for analyzing the outcome of clinical trials combining RPT with new chemotherapeutic agents. Thus, we aimed to help IROC to provide quality assurance for the piloting personalized Radium-223 dosimetry. As there can be inter-lesion and inter-patient heterogeneity in Radium-223 dichloride intake, dosimetry will help to measure the Radium-223 dichloride bio-distribution and absorbed dose in each bone lesion and throughout the body. Dosimetry will also measure biodistribution especially in critical organs for toxicity.
Huaizhi Geng, PhD (Research Associate)
Du Wang, MS (Research Assistant)
Sangho Lee, PhD (Research Associate)
Tingyu Wang, MS (Research Assistant)
(Nov. 2020 - current)
Tawfik Giaddui, PhD (Jan 2015-Feb 2017)
Haoyu Zhong, MS (June 2016-Mar 2019)
Mi Huang, PhD (July 2016-July 2018)
Chingyun Cheng, PhD (Oct 2016-July 2018)
Kuo Men, PhD (July 2017-August 2019)
Nishanth Sasankan (Oct 2018-Nov 2019)