The results confirm that ex vivo deep learning screening of drugs from patient tissues is a promising tool for identifying effective individual treatments for advanced blood cancer compared to traditional methods.
Customized deep learning algorithms and single cell analysis of over 1 billion patient cells unlock additional potential to improve patient outcomes.
VIENNA AND OXFORD, England – (BUSINESS WIRE) – Exscientia (Nasdaq: EXAI), ETH Zurich, the Medical University of Vienna and the CeMM Molecular Medicine Research Center today announce a new publication in Blood Cancer Discovery, the journal of the American Association for Cancer Research, entitled “Deep Learning morphology enhances the accuracy of medicine with image-based ex vivo drug testing,” conducted by the laboratory of Professor Berend Snyder. This post hoc analysis is based on the pioneering work of the EXALT-1 study, published in the journal Cancer Discovery, using deep learning algorithms to classify complex cell morphologies into disease ‘morphotypes’ in patient cancer tissue samples.
EXALT-1 was the first prospective study to demonstrate significantly improved outcomes for patients with advanced hematologic cancer using an AI-based precision medicine platform to provide personalized treatment recommendations by comparing physician treatment choices. In the EXALT-1 study, 40% of patients experienced an exceptional response that lasted at least three times longer than expected for their respective disease. An a posteriori analysis published today in the journal Blood Cancer Discovery shows that the combination of the technology used in EXALT-1 with new advances in deep learning that take advantage of cell-specific characteristics in high-resolution image content has revealed potential for further improvement. these results for patients.
“Following the results of the EXALT-1 study, these observations continue to confirm that our AI-based precision medicine platform is able to identify actionable clinical recommendations for blood cancer treatment, our knowledge and improvement in the platform’s clinical predictive ability to help patients,” said Gregory Wladimir, MD. philosophy, vice president of translational research at Exscientia and co-investor in platform technology. “Cell morphology, or the evaluation of cell characteristics, is fundamental to the diagnosis of cancer. Through this study, we were able to use deep learning on the platform to improve our ability to identify personalized cancer treatments, leading to better clinical outcomes for patients. At Exscientia, we are excited to expand the platform’s applications to bring personalized medicine to a wider population. »
“We believe that performing drug screening directly in the tumor tissues of cancer patients is a big step forward in understanding the complexity of tumors compared to traditional cell model systems. The fact that we can now use the power of deep learning to turn these terabytes of images into useful information is really very encouraging,” added Prof. Berend Snyder, Principal Investigator at the Institute for Systems Molecular Biology at ETH Zurich.
The impact of deep learning on the clinical predictive power of ex vivo drug screening was assessed in a retrospective analysis of 66 patients over a three-year period in a combined dataset of 1.3 billion patient cells for 136 drugs tested ex vivo in hematology studies. diagnoses including acute myeloid leukemia, T-cell lymphomas, diffuse large B-cell lymphomas, chronic lymphocytic leukemia, and multiple myeloma. Patients treated with treatment recommended by platform immunofluorescence assay or deep cell morphology showed an increased rate of exceptional clinical response, defined as progression-free survival, that was three times longer than expected for each patient’s respective disease. Subsequent analyzes confirmed that clinical predictions became more accurate when the drug’s toxicity to healthy cells in the patient’s test sample was also taken into account.
Exscientia’s precision medicine platform uses special deep learning and computer vision techniques to extract meaningful single cell data from images rich in tissue samples from each patient. This analysis provides clinically relevant information about which treatment will be most beneficial for a particular patient. Further evaluating individual patient outcomes using Exscientia’s genomic and transcriptomic capabilities could help Exscientia better understand which other patients might benefit from a similar treatment. The underlying technology was developed by Dr. Gregory Vladimir and Prof. Berend Snyder while working in the laboratory of Giulio Superti-Furg at the CeMM Research Center for Molecular Medicine in Austria.
Exscientia is an AI-powered pharmaceutical technology company dedicated to finding, designing and developing the best medicines in the fastest and most efficient way. Exscientia has developed the world’s first functional precision oncology platform for successful treatment selection and improved patient outcomes in a prospective interventional clinical trial, as well as advancing small molecules designed using artificial intelligence. Our portfolio of internal projects builds on our precision medicine platform in oncology, and our portfolio of partner projects expands our approach to other therapeutic areas. As pioneers of a new approach to drug development, we believe that the best scientific ideas can very quickly become the best medicines for patients.
Exscientia is based in Oxford (England, United Kingdom) and has offices in Vienna (Austria), Dundee (Scotland, United Kingdom), Boston (Massachusetts, United States), Miami (Florida, United States), Cambridge (England, United Kingdom) . and Osaka (Japan).
For more information visit https://www.exscientia.ai or follow us on Twitter @exscientiaAI.
Forward looking statements
This press release contains certain forward-looking statements within the meaning of the safe harbor provisions of the Private Securities Litigation Reform Act of 1995, including statements relating to Exscientia’s expectations regarding progress in the development of candidate molecules, the timing and progress of preclinical trials. research and clinical trials of Exscientia’s product candidates, and Exscientia’s expectations for its precision medicine platform and AI-based drug discovery platform. Words such as “plans,” “considers,” “expects,” “intends,” “plans,” “anticipates,” and the future tense or similar expressions are intended to refer to forward-looking statements. These forward-looking statements are subject to the uncertainties inherent in predicting future results and conditions, including the scope, progress, and expansion of Exscientia’s product development efforts; the initiation, scope and progress of clinical trials by Exscientia and its partners, and their financial implications; clinical, scientific, regulatory and technical developments; and those inherent in the process of discovering, developing and commercializing candidate products that are safe and effective for use in human therapy, as well as business development efforts around these candidate products. Exscientia undertakes no obligation to publicly update or revise any forward-looking statements as a result of new information, future events or other factors, except as required by law.
The text of a press release resulting from a translation should in no case be considered official. The only authentic version of a press release is the press release in the original language. The translation will always have to be compared with the source text, which will set a precedent.