Collaboration on Data and Computational Sciences Announces Next Round of Projects to Advance Cancer Breakthroughs

MD Anderson News Release November 11, 2021

The Oden Institute for Computational Engineering and Sciences (Oden Institute), The University of Texas MD Anderson Cancer Center (MD Anderson) and Texas Advanced Computing Center (TACC) have announced the second round of projects to be funded through their 2021-2022 cooperative research and educational program in Oncological Data and Computational Sciences.

The strategic initiative between the three institutions was designed to align mathematical modeling and advanced computing methods with MD Anderson’s oncology expertise to generate new approaches that can improve outcomes for patients with unmet needs.

Led by Karen Willcox, Ph.D., director of the Oden Institute, David Jaffray, Ph.D., chief technology and digital officer at MD Anderson, and Dan Stanzione, Ph.D., executive director at TACC, the collaborative effort continues to gain momentum with initial success of the first round of projects and its second annual retreat. Together, the institutions leverage their expertise to accelerate the development of innovative, data-driven solutions for patients, as well as to provide a solid foundation on which further cancer research breakthroughs can be made.

The initiative builds upon ongoing collaborations between the Oden Institute’s Center for Computational Oncology, led by Tom Yankeelov, Ph.D., and MD Anderson’s Department of Imaging Physics, led by John Hazle, Ph.D.

“We are excited to launch the next phase of projects and to continue our multidisciplinary approach to developing innovative solutions for complex cancer problems,” Hazle said. “With our collective resources and expertise, we can bring quantitative data and state-of-the-art computational models of disease together in a meaningful way to accelerate progress for our patients.”

The latest projects — which include new ways for identifying, characterizing and treating prostate cancer, blood-related cancer, liver cancer and skin cancer — highlight the growing opportunity to bring computational approaches deep into cancer research and care.

“This is the beginning of a strong collaboration in oncological data and computational science between the Oden Institute, MD Anderson and TACC, and we look forward to our continued cooperation in advancing digitally-enabled efforts to end cancer,” Yankeelov said.

Along with project award funding of $50,000, each collaborative team has access to 12,500 core computing hours at TACC. Ernesto Lima, Sc.D., research associate at the Oden Institute’s Center for Computational Oncology and TACC, will assist all groups in their implementation of the computational aspects of the project on the high-performance computing platforms available at UT Austin.

Read on to learn more about the next round of projects. 

Patient-specific computational models to forecast prostate cancer growth

• Aradhana Venkatesan, M.D., professor of Abdominal Imaging at MD Anderson
• Thomas J.R. Hughes, Ph.D., professor of Aerospace Engineering and Engineering Mechanics at the Oden Institute

Prostate cancer is the fifth leading cause of cancer death in the United States, affecting one in every seven men. The causes are essentially unknown. 

This study will use deterministic models of computational medicine, developed at the Oden Institute and informed by extensive clinical data obtained from MD Anderson. “We are integrating analyses of advanced multiparametric magnetic resonance imaging or MRI (mpMRI) within a computational modeling framework,” Venkatesan said. “The use of mpMRI has become integral to the diagnosis and monitoring of prostate cancer patients, both to assess tumor status and to guide clinical decision-making.”

The study will bring medical research one step closer to realizing patient-centric care. “It will provide predictions of cancerous tumor growth for individual patients, rather than statistical guidelines,” said Thomas J.R. Hughes, Ph.D., from the Oden Institute.

Establishing a single-cell spatial multi-omics reference atlas for studying human hematopoietic malignancy

• Ken Chen, Ph.D., professor of Bioinformatics and Computational Biology at MD Anderson
• Song (Stephen) Yi, Ph.D., assistant professor of Biomedical Engineering and Oncology at the Oden Institute

Every three minutes, one person in the U.S. is diagnosed with a blood cancer. The nature of such diseases, including leukemia, lymphoma and myeloma, is complex, as each tends to carry its own unique set of variables that must be considered when determining treatment. Finding more reliable ways to analyze malignancies at the single cell level would greatly assist in advancing individualized health care. In this collaboration, the research team aims to map out the molecular state of hematopoietic malignancies (the presence of tumors affecting the process through which the body manufactures blood cells) in single-cell resolution.

“Advanced technologies have made it possible to achieve even higher resolution images of cellular differences, thereby providing a better understanding of the function of an individual cell in the context of its microenvironment, in this case the body’s system for manufacturing blood cells,” Yi said. 

Chen and his team will work on the “multi-omics reference atlas,” a novel biomedical approach where the data sets of distinct omic groups — genome, proteome, transcriptome, epigenome, microbiome, etc. — are combined during analysis to provide a more comprehensive guide for studying the molecular state of hematopoietic malignancies.

“We have a novel computational methodology that allows us to integrate data produced by different single-cell modalities together to reveal novel cell populations and associated molecular signatures,” Chen said. “This is particularly exciting because these capabilities are either not yet possible or are very costly to obtain without using the proposed methodology and validation strategies.”

Characterization of thermoembolization cellular damage with computational modeling

• Erik Cressman, M.D., Ph.D., associate professor of Interventional Radiology at MD Anderson
• David Fuentes, Ph.D., associate professor of Imaging Physics at MD Anderson
• M. Nichole Rylander, Ph.D., associate professor of Mechanical Engineering the Oden Institute

Liver cancer remains an incurable disease that is fatal in the majority of cases, with more than 40,000 new cases estimated to be diagnosed in the U.S. in 2021.

This pilot project will focus on the liver with the interdisciplinary expertise in mathematical modeling, chemistry and interventional radiology at MD Anderson leveraged to investigate a mechanistic framework for guiding therapy delivery.

“Thermoembolization provides a novel conceptual endovascular approach for treating primary liver tumors. The approach we are taking is unique in that it subjects the target tumor to simultaneous hyperthermia, ischemia and chemical denaturation in a single procedure,” Fuentes said.

“My team will determine the contribution of heat, pH, and hypoxia [the absence of enough oxygen in the tissues to sustain bodily functions] in the therapeutic outcome using unique in vitro platforms we have developed here at UT Austin,” Rylander said.  

Development of advanced machine (deep) learning algorithms to rapidly detect and accurately estimate protein biomarker values in borderline melanocytic lesions

• Phyu P. Aung, M.D., Ph.D., associate professor of Pathology at MD Anderson
• Chandrajit Bajaj, Ph.D., professor of Computer Science at the Oden Institute

The goal of this unique project is to develop computational tools to help pathologists make more accurate diagnoses for managing patients who encounter borderline melanocytic lesions that could potentially develop into more serious forms of skin cancer. 

“Accurate discrimination between melanomas and benign nevi can be extremely difficult at times, even among expert dermatopathologists,” Aung said. “However, our goals are to develop and train advanced machine (deep) learning algorithms to equivariantly transform stained images to embeddings where this discrimination is disentangled, thereby enabling rapid detection and accurate estimation of the percentage of melanocytes co-expressing MART1 and Ki67 in borderline melanocytic lesions, as well as PD-L1 in tumor cells for potential treatment with immunotherapy, using multiplex histochemical studies with tumor-specific markers.”