AI & Big Data seem to be advancing this year with an increasing number of approvals notably in prostate, breast, lung, liver cancers, etc., curious to see how Telix's AI programs will be positioned and prioritized in this area with regards to medical imaging, genomic screening and clinical data?
Anyone has a clue and could brief us shortly as to what could be Telix's AI focus areas considering its strategic development agenda? Cheers.
An artificial intelligence startup aims to help direct cancer patients to the proper targeted therapies using only digital images of biopsied tissue—instead of relying on full molecular sequencing and analysis—and it has raised more than $21 million to kick off its efforts.
Imagene says its deep learning algorithms can pick out the biomarkers needed to guide treatments within just two minutes by visualizing the patterns put forward by tumor cells that carry different mutations.
The Tel Aviv, Israel-based company plans to help recruit participants for clinical trials, identify new biomarker targets and predict responses to treatments. As of now, its programs are only available for research use.
“Though cancer is a complex disease we might never fully understand, AI is bringing us closer to considering all relevant parameters that affect it, allowing a move towards theragnostics—a personal treatment strategy that combines therapeutics with diagnostics,” said David Agus, founding director of the University of Southern California’s Lawrence J. Ellison Institute for Transformative Medicine.
Agus co-led Imagene’s $18.5 million series A round along with Larry Ellison himself, the co-founder of Oracle, as well as Eyal Gura, co-founder of X-ray AI company Zebra Medical Vision, which was acquired last year by Nanox. Imagene also announced it had raised an additional $3 million in seed funding led by Blumberg Capital.
“Precision medicine is transforming cancer care, yet still relevant and available to less than 15% of cancer patients,” Imagene co-founder and CEO Dean Bitan said in a statement. “Our scalable technology strives to enable treatment opportunities for all cancer patients, driving precision medicine to be tailored per individual patient condition.”
Previously, AI companiessuch as Owkin have shown that by scanning thousands of tissue samples and slide images, algorithms have been able to spot new subtypes of diseases such as mesothelioma and predict which cases would fare best under different therapies. The FDA has also cleared AI pathology platforms for detecting whether a person has cancer in the first place, including Paige’s prostate cancer software.
Paige’s other programs have been used to spot breast cancer metastasis in lymph nodes, and the company joined with liver-cancer-focused Perspectum to offer its pathology software through a single platform.
RadNet scores 2 FDA clearances for AI programs in breast, prostate cancers
By Conor Hale - May 16, 2022
The diagnostic imaging provider RadNet has secured two software clearances from the FDA to help radiologists spot cases of breast and prostate cancer. The programs were developed by its artificial intelligence subsidiaries.
The first, led by its wholly owned DeepHealth unit, focuses on automatically identifying suspicious lesions in mammograms. The company’s Saige-Dx software labels each finding with a score based on its potential for harboring breast cancer, including from 2D mammograms and 3D digital breast tomosynthesis scans.
DeepHealth describes Saige-Dx as the successor to its Saige-Q AI-based workflow tool for prioritizing exams for review, which the FDA cleared last year. That program would only help manage a queue by indicating which mammograms contained at least one suspicious finding.
According to a paper published early last year in Nature Medicine, DeepHealth’s core algorithms were able to help find smaller, more subtle signs of tumors—ultimately allowing a group of expert radiologists to detect breast cancer earlier by one to two years.
“The feedback from physicians who have worked with our software tools is overwhelmingly positive, making them more accurate and efficient at interpreting mammography images," said DeepHealth co-founder and CEO Gregory Sorensen, who formerly served as CEO of Siemens Healthcare North America.
RadNet’s second AI clearance, which parses MRI scans for signs of prostate cancer, was obtained by its Quantib division.
Quantib’s software places a variety of automated tools on the radiologist’s reading station, including support for evaluating lesions using standardized scoring methods, calculating PSA density within the tissue and automatically segmenting a complex image of the prostate’s glands into different zones. It also provides a heat map of the organ highlighting areas of concern.
RadNet acquired Quantib this past January in a deal made alongside the separate purchase of Aidence, a company focused on AI-powered lung cancer screening. Both subsidiaries are located in the Netherlands. The Massachusetts-based DeepHealth, meanwhile, was bought by RadNet in 2020.
“With the addition of Aidence and Quantib, we will now have effective screening solutions for the three most prevalent cancers,” Radnet Chairman and CEO Howard Berger said when announcing the deals.
“While prostate MRI is a growing area of our overall MRI business, the opportunity to create a lower-cost, more accurate service offering to Medicare and private payors allows for a conversation about creating large-scale screening programs for appropriately qualified male patient populations, akin to how mammography is utilized today to detect and manage breast disease in women,” Berger added.
RadNet maintains a U.S. network of 350 outpatient imaging centers and about 9,000 employees. The financial details of the Quantib or Aidence deals were not disclosed.
In its recent first-quarter earnings report, the company posted $341.2 million in revenue from its imaging center segment, alongside losses of $41.7 million in its AI reporting segment.
“Though we project losses for the next 24 months from the investments we are making in these new technologies, we continue to be more convinced than ever that AI will have a significant impact on the growth and cost structure of our business in the coming years,” Berger said.
Paige nets European OK for AI to spot prostate cancer biomarkers
By Andrea Park - May 12, 2022
Last year, Paige became the first tech developer to score an FDA approval for an artificial intelligence program that could detect signs of prostate cancer in digital pathology slides. Just a few months later, it’s pushing the technology to new frontiers, enabling it to look for even more indicators of the disease in biopsy samples.
In contrast to the initial software, Paige Prostate Detect, which pores over a slide to identify microscopic cancer cells, the New York-based company’s new offering, Paige Prostate Biomarker Suite, instead looks for molecular signals linked to prostate cancer.
And while the biomarker suite hasn’t yet earned the same FDA authorization granted to its predecessor, Paige announced Thursday that it did secure both a CE mark and UKCA certification, allowing it to be used in the EU and U.K., respectively.
The new biomarker-spotting tool relies on the same underlying AI technology as the cell-spotting version. It analyzes a biopsy sample for androgen receptor, TP53, RB1 and PTEN biomarkers, all of which have been linked to the development and spread of prostate cancer. If any of those indicators are found, a pathologist may proceed to call for confirmatory diagnostics like immunohistochemistry or fluorescence in situ hybridization, commonly known as FISH.
With the new tool, Paige is aiming to make biomarker detection more accessible and more widely used in cancer diagnostics. It’s meant to be used to examine digital images stained with hematoxylin and eosin, the most commonly used stain in tissue analysis. It also takes a much more straightforward, image-based approach to biomarker testing, which typically requires an expensive molecular diagnostics platform and blood sample to perform.
Buoyed by the European approvals, Paige said it now plans to identify and evaluate other biomarkers linked to cancer and train the AI to recognize them, too.
“By employing Paige Prostate Biomarker Suite, clinicians can rapidly reduce laboratory turnaround time while providing a broader range of data at the point of diagnosis,” said Jill Stefanelli, Ph.D., Paige’s president and chief business officer.
“As we expand our biomarker portfolio, we also look forward to developing novel biomarkers across indications to identify patients that should receive genomic testing or could potentially respond to targeted therapies,” Stefanelli said. “This opens the door to a whole new range of biomarker applications and, in turn, new opportunities for industry collaboration.”
Prostate cancer is just one half of the current focus of Paige’s diagnostic AI. Earlier this year, for example, in between securing regulatory approvals for the cell- and biomarker-detecting tools, the company also began rolling out yet another program in its breast cancer portfolio. Still designated for research use only, the new offering can help pathologists determine whether breast cancer has spread to the lymph nodes.
Paige Announces Series C Funding Round of $100Million to Accelerate Transformation of Digital Pathology
- Casdin Capital and Johnson & Johnson Innovation – JJDC, Inc. led financing - Company has raised over $195 million to date - Funds will accelerate development of AI-based clinical applications, biomarkers and diagnostics
Paige, a global leader in AI-based digital diagnostics, today announced a Series C funding round of $100 million to accelerate the Company’s efforts to transform digital pathology by unlocking insights from each patient’s sample so pathologists, care teams and researchers can make decisions more confidently and efficiently to optimize outcomes.
The Series C financing was led by Casdin Capital and Johnson & Johnson Innovation – JJDC, Inc. (JJDC), the strategic venture capital arm of Johnson & Johnson. Existing investors and other funds are also participating in the round.
“Paige is building a transformational portfolio of computational pathology products to serve clinical needs and drive precision medicine,” said Leo Grady, Ph.D., Chief Executive Officer, Paige. “This investment reaffirms the vast potential of the Paige platform for clinical and biopharmaceutical drug development applications. These funds will enable us to build additional AI-based products within and outside of oncology, deliver these products to laboratories and clinicians globally, and invest in our talent across engineering and commercial functions.”
In addition to accelerating the development of AI-based clinical applications, biomarkers and diagnostics, Paige will also invest in sales and marketing efforts to rapidly scale product adoption by hospitals and labs. Accordingly, the Company will further expand by building out its engineering and commercial teams with plans to hire about 70 new employees in 2021, doubling the size of the company.
Eli Casdin, Chief Investment Officer of Casdin Capital, adds: “Bending the mortality curve on cancer is a humbling and critical goal requiring big data, big technology and big talent. Paige combines all three: Robust AI capabilities, access to millions of digital pathology images linked to the key clinical data modalities of imaging, genomic, and clinical/EMR data, and a growing team purpose built to deliver. This is a unique opportunity to transform data into a next generation of cancer diagnostics and therapeutics, with clear application beyond cancer, and we are excited to be joined by JJDC to fuel the company to deliver on this opportunity.”
Human Cell Atlas initiative makes strides as scientists piece together 'Google Map' of human cells
The Human Cell Atlas consortium has added important pieces to the creation of the full map of human cells, offering new insights into gene activity in specific cell types and how immune cells behave in various human tissues and organs.
Through four studies simultaneously published in Science, international teams of researchers offered detailed single-cell and single-nucleus transcriptome maps of more than 1 million human cells across 500 cell types in 33 tissues and organs.
They uncovered some unexpected cell features, signaling mechanisms and genetic traits. The publicly available findings could inform the understanding of human health and disease states and aid the development of diagnostics as well as anti-tumor immunotherapies, vaccines and regenerative medicines.
In two studies, scientists from Wellcome Sanger Institute, the University of Cambridge and other collaborators looked at immune cells in the human body, one focused on the early development of the immune cells across several tissues, and the other on immune cells in adults to understand their states in different tissues.
“You can think of it as a Google Map of the human body, and it’s really that street map’s view of the individual cells and where they sit in tissues that we are aiming towards,” Sarah Teichmann, Ph.D., of the Wellcome Sanger Institute and co-corresponding author of both studies, explained during a press conference.
Beyond commonly studied immune cells in the blood, the scientists also examined the tissues that create immune cells, the tissues where immune cells migrate to, such as the skin, gut and lung, and where the immune cells mature, such as lymph node and the spleen, she said.
In one shocking finding, the team tracked blood stem and progenitor cells not only in professional immune tissues—where they’re expected—but also in embryonic gut and skin. Turns out, B-cell progenitors receive input signals for their development from immune cells in the gut in addition to the bone marrow. This finding is important for engineering immune cells outside the body, Teichmann said, because it offers new clues to what signaling should be put in.
Besides, they found that T cells in the thymus are also talking to each other rather than just from their parent thymic epithelial cells during maturation.
In the second study, Teichmann and colleagues sequenced RNA from 330,000 single immune cells from 16 tissue sites across the adult body to understand their specific functions in different tissues. Teichmann described the process as finding “the molecular GPS system for immune cells that locates them to specific organs around the body.”
The team developed a machine-learning algorithm, called CellTypist, to automate cell type identification. Using CellTypist, the researchers created a cross-tissue immune cell atlas that revealed the relationship between immune cells in different tissues.
Analyses showed some immune cells like macrophages share common signatures across tissues, whereas others like memory T cells have “different flavors” depending on which tissues they reside in, Teichmann said.
“Knowing the code of which molecules direct and maintain T cells to specific tissues is important for engineering and targeting cells for cancer therapies,” Teichmann said. With the findings, future researchers could find inspiration to either enhance or suppress specific immune cells to fight disease or help design vaccines, she added.
In the third study led by the Broad Institute of MIT and Harvard, scientists used single-nucleus RNA sequencing to generate a cross-tissue atlas of the nuclei profiles from nearly 210,000 cells.
One of the main applications of a full human cell atlas is to identify the cell types where disease genes act. Identifying the precise cells where disease arises would allow for the development of more precise diagnostic and new treatment, Aviv Regev, Ph.D., of the Broad Institute and currently head of Roche’s Genentech Research and Early Development, said during the press briefing.
For their study, Regev and colleagues compared the cells across tissues. They found that macrophages take on a particular “pair of flavors” in every tissue they looked at, but the flavors arise in different ratios to support different functions.
The researchers also observed a similar phenomenon for fibroblasts, which are connective tissues, as the cells acquire different properties at different locations. In the lungs, fibroblast cells express genes that sense mechanic tensions to help the lung contract, the team found.
Using machine learning algorithms, the team also associated the cells in the atlas with 6,000 single-gene diseases and 2,000 complex genetic diseases involving multiple genes. They identified cell types and genetic signaling that could be implicated in disease, opening new venues for further research into these conditions.
In one example, the team found non-muscle cell types that are implicated in muscular dystrophy. Even though a problematic mutation exists in genes that are not expressed by muscle cells, it’s affecting other cells in muscle tissues that are critical for muscle function to cause muscular dystrophy, Regev explained.
In another finding that emphasizes the potential of the human cell atlas in assisting precision medicine, Regev’s team showed that genes that are associated with the risk of atrial fibrillation are also used by cells in the skeletal muscle of the esophagus and the prostate.
“Now we can try and devise ways to target more specifically to the cells where we want to have an impact but not to target other cells that are also using these genes in a body,” she said.
The fourth study was conducted by the Tabula Spaiens Consortium, a team of more than 160 experts led by scientists at the Chan Zuckerberg Biohub. Led by Stephen Quake, Ph.D., at Stanford University, the team mapped gene expression in nearly 500,000 cells from 24 human tissues and organs with a focus on collecting samples from the same donors, which allowed researchers to screen out certain background variations such as age.
In one unexpected finding, researchers found that sets of housekeeping genes—which have been thought to handle basic functions in much the same way in every cell—likely have many more roles across the body than was previously thought.
The team also showed that the CD47 protein, which is implicated both in cancer and in the buildup of dangerous plaques on artery walls, may differ widely among cells. Once again, the finding could guide the development of more effect drugs with fewer side effects, the researchers argue.
Based on the tissue samples, Quake and colleagues also took the opportunity to look at gut microbiome profiles. After sequencing the bacterial genes and working out their spatial relationships, the team discovered complex structures to the microbiome species as they move through the intestines, Quake said.
Overall, “[t]hese pan-tissue human cell atlases form important reference datasets for understanding and predicting the side effects and safety issues of new medicines,” two researchers from China's Peking University wrote in an accompanying editorial.
In addition, “understanding both shared and tissue-specific features of tumorigenesis is key to the development of histology-agnostic and cancer type-specific therapeutics,” they added.
Mount Sinai pairs with LabCorp and Philips to build out its AI-driven pathology center
LabCorp, Philips and the Mount Sinai Health System are coming together to develop an artificial intelligence-driven pathology center of excellence, initially aimed at cancer diagnosis.
The center will work to integrate digital pathology into clinical practices across Mount Sinai’s hospitals, starting with interpretations of prostate cancer and other genitourinary tumors, as well as cancers of the head and neck.
From there, the project is slated to expand in scope, to tap Mount Sinai pathologists for consultations for nationwide cases through LabCorp’s Dianon Pathology specialty laboratory, the companies said in a joint statement.
Over the next several months, the center of excellence will deploy Philips’ IntelliSite Pathology Solution for primary diagnosis and consultations at the eight Mount Sinai hospitals across the New York metropolitan area, as well as in select ambulatory care centers. The system’s academic department already processes more than 80 million diagnostic tests a year, the companies said, making it one of the largest of its kind.
“Digital pathology gives us the unprecedented opportunity to expand our services to the community at large and engage members of our department, considered key opinion leaders in their field, to provide expert diagnostic opinions in complex cases,” said Carlos Cordon-Cardo, chairman of Mount Sinai’s pathology department, and a professor of oncological sciences, pathology, and genetics and genomic sciences at the Icahn School of Medicine.
“This, in addition to our new predictive AI-based tests, introduces the potential for optimization of treatment efficacy and provides the opportunity for improved clinical outcomes,” Cordon-Cardo said.
Mount Sinai had previously participated in the evaluation of Philips’ digital pathology system as it obtained market clearance in the U.S., which helped provide a foundation for the development of Mount Sinai’s AI diagnostics, the companies said.
The Amsterdam-based company’s IntelliSite platform helps pathologists review digital images of surgical pathology slides through an automated management system, including a scanner, high-resolution display and workflow software tools.
The collaboration follows on from Mount Sinai’s plans to establish a broader center for AI and healthcare development at its Icahn School of Medicine.
The Hamilton and Amabel James Center for Artificial Intelligence and Human Health aims to combine AI with data science and genomics at a standalone site in Manhattan, slated to open in late 2021. It plans to launch with about 40 principal investigators, plus 250 graduate students, postdoctoral fellows, computer scientists and support staff.
The center will focus on applying big data and machine learning techniques to genomic data, including single-cell epigenomics and pharmacogenomics, alongside data from patient health records and wearable devices.
In addition, researchers will work to bring AI to diagnostic imaging technologies, including X-ray, MRI, CT and PET scans, as well as molecular imaging.
“We see a huge potential in using algorithms to automate the image interpretation and to acquire images much more quickly at high resolution—so that we can better detect disease and make it less burdensome for the patient,” said Zahi Fayad, director of Mount Sinai’s Translational and Molecular Imaging Institute, and vice chair for research for the Department of Radiology.
“In addition to AI, we envision advance capabilities in two important areas: computer vision and augmented reality, and next-generation medical technology enabling development of new medical devices, sensors and robotics,” Fayad said.
Icahn School of Medicine at Mount Sinai to Establish World Class Center for Artificial Intelligence - Hamilton and Amabel James Center for Artificial Intelligence and Human Health
First center in New York to seamlessly integrate artificial intelligence, data science and genomic screening to advance clinical practice and patient outcomes.
The Icahn School of Medicine at Mount Sinai today announced the launch of a new center dedicated to advancing the delivery of health care through research, development, and implementation of innovative artificial intelligence tools and technologies. The Hamilton and Amabel James Center for Artificial Intelligence and Human Health in Manhattan will combine artificial intelligence with data science and genomics in a standalone site. The building will enable researchers to enhance their understanding, diagnosis, and treatment of human diseases—including the most debilitating—and promote improved health and well-being.
Made possible by a generous donation from Hamilton Evans ‘Tony’ and Amabel James, the interdisciplinary center is projected to open in late 2021. Mr. James is the Executive Vice Chairman of Blackstone, a New York-based investment firm. The new Center will open with approximately 40 principal investigators, and 250 graduate students, postdoctoral fellows, computer scientists, and support staff.
“Mount Sinai has consistently been at the forefront of advancing health care across medical disciplines and this initiative represents our next step forward in building on that legacy,” said Kenneth L. Davis, MD, President and Chief Executive Officer of the Mount Sinai Health System. “We see the potential of artificial intelligence to radically transform the care that patients receive, and we want to shape and lead this effort. We are grateful to Mr. and Mrs. James for their generous gift, which will create a hub where our talented researchers can collaborate in unprecedented ways and bring forward ideas and innovative technologies that achieve better outcomes for our patients.”
“Artificial Intelligence and machine learning are spurring innovation across many different fields, but perhaps most significantly in health care,” Mr. James said. “Mount Sinai has proven itself a pioneer in data mining to improve patient diagnosis and treatment, and I am pleased to support its mission and accelerate the development of cutting-edge therapies and technologies that have the potential to change lives around the world.”
Dennis S. Charney, Anne and Joel Ehrenkranz Dean of the Icahn School of Medicine, and President for Academic Affairs for the Mount Sinai Health System, said, “We are looking at a future where artificial intelligence has the capacity to completely disrupt health care and Mount Sinai is going to be at the forefront of that revolution, driving the conversation, engaging stakeholders worldwide in developing solutions, and making this bold future a reality.”
Mount Sinai clinicians and investigators have been early adopters of artificial intelligence and currently use the technology in many different initiatives in precision medicine. AI technology is helping to characterize tissue samples of patients with prostate cancer, for example, and is being used to assist Mount Sinai doctors in identifying and prioritizing patients at risk for developing diseases and hazards such as falling.
The Hamilton and Amabel James Center for Artificial Intelligence and Human Health will focus on three key areas:
Center for Genomic Health— The new Center for Genomic Health—to be housed in the new Center for Artificial Intelligence and Human Health building—is accelerating the integration of genomics into clinical care throughout the Mount Sinai Health System. “Our goal is to use artificial intelligence to translate vast knowledge from deep databases of genomic information to improve the lives of every patient at Mount Sinai,” said Eimear Kenny, PhD, Founding Director of the Center for Genomic Health and Associate Professor of Medicine, (General Internal Medicine), and Genetics and Genomic Sciences. “The new building will bring together a generation of scientists and physicians who are trained in big data and artificial intelligence—tools that can that can enable the development of precise genomic tests and increasingly sophisticated ways to integrate genomic information into routine patient care,” said Noura Abul-Husn, MD, PhD, Clinical Director of the Center for Genomic Health and Senior Faculty of Medicine, (General Internal Medicine), and Genetics and Genomic Sciences at Icahn School of Medicine at Mount Sinai.
Integrative Omics and Multi-Scale Disease Modeling— Artificial intelligence and machine learning approaches developed at the Icahn Institute have been extensively used for identification of novel pathways, drug targets, and therapies for complex human diseases such as cancer, Alzheimer’s, schizophrenia, obesity, diabetes, inflammatory bowel disease, and cardiovascular disease. Researchers will combine insights in genomics—including state-of-the-art single-cell genomic data—with ‘omics,’ such as epigenomics, pharmacogenomics, and exposomics, and integrate this information with patient health records and data originating from wearable devices in order to model the molecular, cellular, and circuit networks that facilitate disease progression. “Novel data-driven predictions will be tightly integrated with high-throughput experiments to validate the therapeutic potential of each prediction,” said Adam Margolin, PhD, Professor and Chair of the Department of Genetics and Genomic Sciences and Senior Associate Dean of Precision Medicine at Mount Sinai. “Clinical experts in key disease areas will work side-by-side with data scientists to translate the most promising therapies to benefit patients. We have the potential to transform the way care givers deliver cost-effective, high quality health care to their patients, far beyond providing simple diagnoses. Mount Sinai wants to be on the frontlines of discovery.”
Precision Imaging—Researchers will use artificial intelligence to enhance the diagnostic power of imaging technologies—X-ray, MRI, CT, and PET—and molecular imaging, and accelerate the development of therapies. “We see a huge potential in using algorithms to automate the image interpretation and to acquire images much more quickly at high resolution – so that we can better detect disease and make it less burdensome for the patient,” said Zahi Fayad, PhD, Director of the Translational and Molecular Imaging Institute, and Vice Chair for Research for the Department of Radiology, at Mount Sinai. Dr. Fayad plans to broaden the scope of the Translational and Molecular Imaging Institute by recruiting more engineers and scientists who will create new methods to aid in the diagnosis and early detection of disease, treatment protocol development, drug development, and personalized medicine. Dr. Fayad added, “In addition to AI, we envision advance capabilities in two important areas: computer vision and augmented reality, and next generation medical technology enabling development of new medical devices, sensors and robotics.”
By bringing a cross-section of researchers together in one dedicated space, the new Hamilton and Amabel James Center for Artificial Intelligence and Human Health is expected to foster ideas that will significantly advance treatments.
To date, Mount Sinai has made progress with AI in many areas:
- In 2016, researchers led by Joel Dudley, PhD, Director of the Institute for Next Generation Healthcare and the Co-Director of the newly formed Hasso Plattner Institute for Digital Health at Mount Sinai, used an advanced AI algorithm to improve prediction of diseases by analyzing de-identified data from patients across the Mount Sinai Health System. In the study published in Scientific Reports, Dr. Dudley and his colleagues found that their algorithm significantly outperformed evaluations based on raw data from electronic health records (EHR) and had impressive results in predicting severe diabetes, schizophrenia, and various cancers. “The findings indicated that using artificial intelligence and applied learning with EHRs can offer improved clinical predictions and help augment decision making by patients and their health care providers,” said Dr. Dudley, who is also Professor of Genetics and Genomic Sciences and Executive Vice President for Precision Health for the Mount Sinai Health System. “Wearable and other digital technologies can also enhance data from EHRs and lead to predictive and preventive health solutions.” The Hasso Plattner Institute, co-led by Erwin P. Bottinger, MD, Professor of Digital Health-Personalized Medicine, Hasso Plattner Institute, University of Potsdam, Germany, will develop digital technologies and wearable devices with Dr. Dudley.
- Mount Sinai pathologists currently use artificial intelligence to characterize tissue samples in patients with certain diseases, including prostate and breast cancer, to more accurately predict the course of the disease; as well as to recognize and quantify accumulation of abnormal proteins, such as those in Alzheimer’s disease. “Our goal is to provide a precise mathematical approach to classifying and treating disease, which assists our clinicians with information for effective patient care and health management,” said Carlos Cordon-Cardo, MD, PhD, Director of the Center for Computational and Systems Pathology and Precise Diagnostics” (Precise Dx); Chair of the Department of Pathology; and Professor of Pathology, Genetics and Genomic Sciences, and Oncological Sciences at the Icahn School of Medicine at Mount Sinai. “By refining diagnoses, we can save patients from unnecessary treatments and improve outcomes.”
- Mount Sinai’s AI-powered system assists clinicians in identifying and prioritizing patients at risk for conditions such as cardiopulmonary deterioration, malnutrition, and falls. Developed by the Clinical Data Science team, and assembled by David Reich, MD, President and Chief Operating Officer of The Mount Sinai Hospital, the system supports clinical decision making for thousands of patients each day. As the algorithms continue to “learn” from real-time data, their accuracy and performance improves.
- Judy Cho, MD, Director of the Charles Bronfman Institute for Personalized Medicine, which runs the BioMe Biobank that houses more than 47,000 DNA and blood serum specimens from Mount Sinai’s diverse population of patients, said her team is using AI as a predictive tool. “By combining traditional clinical measures, genetics, and new blood biomarkers with AI methods, we can much more efficiently predict and treat patients earlier to try to prevent kidney failure and the need for dialysis,” Dr. Cho said. BioMe is linked to Mount Sinai’s EHRs, and enables scientists to rapidly and efficiently conduct genetic, epidemiologic, molecular, and genomic studies on large collections of research specimens linked with medical information.
Mount Sinai To Deploy New Supercomputer, Integrating Genomic And Clinical Data
Mount Sinai Health System said today its researchers will be able to explore more complex scientific questions more quickly once the New York City health system builds its second “Big Omics Data Engine” (BODE 2) supercomputer, to be funded through a $2 million grant awarded by the U.S. Department of Health and Human Services. Through BODE 2, Mount Sinai said, its researchers will be able to use machine learning approaches to mine deep databases of genomic and clinical information, with precision medicine in mind. “With BODE 2, we are renewing our commitment to push the boundaries of scientific research, tackle questions that we did not previously have the computational power to take on, and achieve breakthroughs that transform clinical care worldwide,” said Dennis Charney, MD, the Anne and Joel Ehrenkranz Dean of the Icahn School of Medicine at Mount Sinai and president of academic affairs for the Mount Sinai Health System. “This new supercomputer will enable us to mine deep databases of genomic and clinical information using machine-learning approaches to propel the personalized medicine of today into better medicine tomorrow,” said Eimear Kenny, PhD, associate professor of medicine, general internal medicine, and genetics and genomic Sciences, at the Icahn School of Medicine at Mount Sinai and principal investigator of the TOPMed Program. “Computing capability of this size and speed is not available widely, and Mount Sinai’s investment in building this infrastructure will translate into more robust genetics and population analysis, gene expression, machine learning, and structural and chemical biology investigations, and result in new insights and advances in a wide range of diseases including Alzheimer’s, autism, influenza, prostate cancer, schizophrenia, and substance use disorders,” said Patricia Kovatch, BS, associate professor of genetics and genomic sciences, pharmacological sciences at the Icahn School of Medicine at Mount Sinai.
28M Core Compute Hours per Year
The new BODE 2 supercomputer is a Lenovo ThinkSystem SR360 consisting of 3,840 Intel Cascade Lake cores, with 15 terabytes of memory, 14 petabytes of raw storage, and 11 petabytes of usable storage. It will produce approximately 28 million core compute hours per year at a frequency of 2.6 GHz and it will have a peak speed of 220 teraflops per second—approximately double that of BODE.
“Computing capability of this size and speed is not available widely, and Mount Sinai’s investment in building this infrastructure will translate into more robust genetics and population analysis, gene expression, machine learning, and structural and chemical biology investigations, and result in new insights and advances in a wide range of diseases including Alzheimer’s, autism, influenza, prostate cancer, schizophrenia, and substance use disorders,” Patricia Kovatch, Senior Associate Dean for Scientific Computing and Data Science at the Icahn School of Medicine at Mount Sinai, said in a statement.
Kovatch is also a member of the Icahn Institute for Data Science and Genomic Technology, and Associate Professor of Genetics and Genomic Sciences, and Pharmacological Sciences. Before joining Mount Sinai, she directed the National Science Foundation’s supercomputer center at the University of Tennessee, Knoxville, located at the U.S. Department of Energy’s Oak Ridge National Laboratory—where she deployed the world’s third fastest supercomputer in 2009, a $75 million Cray XT3.
BODE 2 is set to launch at year’s end. When it does, it will supersede BODE, an earlier-generation Cray supercomputer that consists of 2,484 Intel Haswell cores and 5 petabytes of storage. BODE was built in 2015 using a $2 million NIH grant as a second expansion of Mount Sinai’s first supercomputer, named MINERVA and deployed in May 2012. BODE nodes were fully integrated into the Minerva computer complex and were limited only to users with NIH-funded Genomics-based research projects.
According to Mount Sinai, BODE was used by 61 basic and translational researchers at Mount Sinai representing more than $100 million in NIH funding, along with their collaborators at 75 external institutions. BODE enabled scientific findings that appeared in more than 167 publications, including Nature and Science, with a total of 2,427 citations in three years.
“Based on our experiences with BODE, BODE 2 is designed to provide our researchers and clinicians, and their external partners in Mount Sinai-led national research projects, with the necessary infrastructure to achieve faster results for greater scientific throughput, increased fidelity in their simulations and analysis, and seamless migration of research applications to the software environment for enhanced scientific productivity,” Kovatch added.
TLX Price at posting:
$4.12 Sentiment: Buy Disclosure: Held
(20min delay)