Supercomputer-Generated Models Help Better Understand Esophageal Disorders
Gastroesophageal reflux disease, more commonly known as GERD, affects about 20 percent of US citizens, according to the National Institutes of Health. If left untreated, GERD can lead to serious medical problems and sometimes esophageal cancer. Using supercomputers, imaging advances of the swallowing process of patients with GERD have been modeled Comet at the San Diego Supercomputer Center (SDSC) at UC San Diego and Bridges-2 at the Pittsburgh Supercomputing Center (PSC).
Researchers from Northwestern University, McCormick School of Engineering and Feinberg School of Medicine have teamed up and recently published these new models in Biomechanics and modeling in mechanobiology. Their work resulted in a new computer modeling system, called FluoroMech, which could help identify physio-markers for early and accurate diagnosis of esophageal pathologies.
“While nearly 60% of the adult population suffers from some form of GERD each year, there are 500,000 new cases of esophageal cancer diagnosed each year worldwide, with a projected 850,000 new cases / year by 2030, “said Neelesh Patankar, professor of mechanics. engineering at Northwestern and the lead author of the study. “In the United States, there are nearly 17,000 diagnoses per year, which is about one percent of cancer diagnoses, but less than 20 percent of these patients survive at least five years and the main curative treatment is l esophagectomy. “
Patankar said staggering statistics such as these for esophageal disorders in general inspired the team of engineers and medical researchers to create an interdisciplinary study, which resulted in FluoroMech, which complements the techniques of Common non-invasive medical imaging to quantitatively assess the mechanical health of the esophagus. .
Why it matters
The mechanical properties, such as elasticity, of the esophageal wall and its relaxation during swallowing have been shown to play an essential role in the functioning of the esophagus; thus, these qualities were considered as indicators, or physio-markers, which explain the health of the organ. In the absence of diagnostic techniques that determine these physical quantities, Northwestern’s research team developed the FluoroMech computational technique to help clinicians have more complete and accurate information about each patient’s esophagus.
Specifically, FluoroMech was designed to predict the mechanical properties of soft tubular organs such as the esophagus by analyzing medical images and fluoroscopy (real-time x-ray imaging) videos of the esophagus during a swallowing process. The new technique was also developed to allow the clinician to quantify the relaxation of the esophageal muscles during the passage of food.
“Understanding the mechanisms underlying human pathophysiology requires knowledge of both the biochemical and biomechanical functions of organs. Yet the use of biomechanics has not advanced to the same level as biochemistry or even imaging techniques such as MRI and X-rays, ”Patankar said. “This is especially true in the case of esophageal disorders, and we aim to shift the existing paradigm from studying the pathogenesis of organ diseases to a biomechanics-based approach that harnesses information about how mechanical properties of organs alter physiology. “
How Supercomputers Helped
“In our biomechanics problem, the whole system had nearly five million unknowns to solve at each time step and there were a large number of time steps to solve. Simulations of this magnitude required advanced supercomputers to get results in a reasonable amount of time, which is often five to seven days for each model, ”said Sourav Halder, a doctoral student at Northwestern and lead author of the study. “Without the availability of these computing resources, it would not have been possible to simulate such systems. “
These recent supercomputer compatible models provided the team with the tools to achieve their noble goal of creating FluoroMech. Prior to the development of his technique, it was not possible to quantify the mechanical health of the esophagus, and perhaps more importantly, to enable patient-specific predictive modeling. Halder said this was previously impossible due to the lack of automated image segmentation based on machine learning techniques and complex calculations based on the physics of esophageal function. With the help of supercomputers that are part of the National Science Foundation’s Extreme Scientific and Technical Discovery Environment (XSEDE) as well as the SDSC education and training team, Halder and his colleagues were able to develop their FluoroMech technique.
Halder commended the SDSC support team for their help in setting up proprietary software for efficient prototyping of simulation models and the ability to generate massive amounts of data for machine learning models. He said the SDSC team also offers training seminars on GPU Computing, HPC in data science and high-level programming frameworks like OpenACC. “We are extremely grateful to the SDSC education and training team for regularly organizing these workshops which are extremely accessible to all participants, regardless of their computer background,” he said.
“While FluoroMech uses fluoroscopy data to predict the properties of the esophageal wall as well as the function of the esophagus in terms of estimating active muscle wall relaxation, we recently extended our work with another device to diagnostic called Endolumenal Functional Lumen Imaging Probe (Endolumenal Functional Lumen Imaging Probe) to predict mechanically-based physio-markers such as the force of esophageal wall contraction, active relaxation and elastic properties of the wall, ”Patankar said.
Using EndoFLIP data from a large cohort of subjects and the predictions of the FluoroMech model, the team began to develop a virtual disease landscape (VDL). The VDL is a parameter space where subjects of different esophageal disorders are grouped into different regions. The locations of the groups, relative to each other, are designed to represent the similarities and differences between the ways in which boluses are transported through the esophagus. Prototypes of the VDL concept have already shown how it can provide a fundamental understanding of the underlying physics of various esophageal disorders.
“With the XSEDE allocations, we have already been able to develop improved models for gastric peristalsis, illustrate muscle activity in the stomach and simulate acid reflux,” said Halder. “More recent clusters like Scope at SDSC were built with double the amount of RAM on each node and nearly five times the number of cores compared to Comet, so this leap in computing power allows us to plan advanced models without seeing a substantial increase in the time required for computation.
Support for this research has been provided by grants from the Public Health Service (R01-DK079902 and P01-DK117824) and the National Science Foundation (NSF) (OAC 1450374 and OAC 1931372). Computing resources were provided by Northwestern University’s Quest High Performance Computing Cluster and Extreme Science and Engineering Discovery Environment (XSEDE) through allocation TG-ASC170023, which is supported by NSF (ACI-1548562 ). He also used SDSC Comet, which is supported by NSF (ACI-1548562) and the Bridges-2 system to PSC, which is supported by NSF (ACI-1928147).
The San Diego Supercomputer Center (SDSC) is a leader and pioneer in high-performance, data-intensive computing, providing cyberinfrastructure resources, services and expertise to the national research community, universities and the international community. ‘industry. Located on the UC San Diego campus, the SDSC supports hundreds of multidisciplinary programs spanning a wide variety of fields, from astrophysics and earth sciences to disease research and drug discovery. The newest SDSC supercomputer funded by the National Science Foundation, Scope, supports the “Computing Without Borders” SDSC theme with data-centric architecture, public cloud integration and state-of-the-art GPUs to incorporate experimental facilities and state-of-the-art computing.
The Pittsburgh Supercomputing Center (PSC) is a joint computer research center of Carnegie Mellon University and the University of Pittsburgh. Founded in 1986, PSC is supported by several federal agencies, the Commonwealth of Pennsylvania, and private industry and is a leading partner in XSEDE, the National Science Foundation’s cyberinfrastructure program. PSC provides academic, government and industry researchers with access to many of the most powerful systems for high performance computing, communications, and data storage available to scientists and engineers nationwide for unclassified research. PSC advances the state of the art in high performance computing, communications, and data analytics, and provides a flexible environment to solve computer science’s most important and challenging problems.