GE Healthcare and SOPHiA Genetics: Exciting Times for Radiomics

At GE, we understand that practicing precision medicine requires that an immense amount of health data be synthesized in new ways and shared within more comprehensive networks to create novel, actionable insights. At the forefront of data-driven medicine is the field of radiomics, the science of converting digital medical images into mineable, high-dimensional data.[1] Radiomics maximizes the usefulness of information that’s already collected during routine examinations, using automated data extraction from medical images to analyze details far too miniscule for the naked eye to see, and providing in-depth assessments and modeling of image features that correlate to various different diseases.[2]

SOPHiA Genetics, a Swiss biotech company, uses the rich potential of radiomics to unearth the abundant data within standard medical imaging. SOPHiA DDM? is a novel, AI-powered cloud based software that uses patented machine-learning algorithms to provide genomic insights and treatment recommendations.[3] That’s why GE Healthcare has chosen to collaborate with SOPHiA Genetics, using their platform with our extensive medical imaging solutions, as well as in tandem with our Edison-platform enabled data aggregation, in an effort to break down data silos and create more personalized, accurate patient care pathways that are data informed.

SOPHiA’s AI is able to give clinicians a stunning amount of detail about their patients, including tumor characteristics and indications about changes in growth after treatment. SOPHiA’s algorithms can even generate predictive models. One of the most exciting prospects of radiomics is the potential for MR/PET/CT and digital biopsy pathology data to be integrated, creating Al-driven radiogenomic solutions that could eventually remove the need for some surgical biopsies.[4]

Our collaboration with SOPHiA Genetics offers multiple other benefits in the field of radiogenomics. It’s a highly efficient solution that helps to ameliorate the workload of often overburdened radiologists and clinicians. Making greater use of medical imaging that’s already a basic step in the oncology workflow, it automates the time-consuming task of working with antiquated computer programs to analyze images with far greater comprehension, allowing medical professionals to focus on more cognitively difficult tasks.

Radiomics can already improve the standard of care, providing clinicians with multiple dimensions of crucial data in order to chart out the best treatment path for patients. Tumors can be tracked in real time, and that data can then be shared; each unique case has the potential to benefit massively from connection with other patients and clinical trials around the world. The scalability of SOPHiA DDM? makes the platform ideal for use in breaking down data silos. For AI to evolve and maintain a high standard of accuracy, it’s also essential that algorithms are consistently trained on reliable datasets.

For a dataset to be statistically relevant, the general rule is that it should have ten times the amount of samples than the parameters you’re modeling.[5] Data sharing is a crucial step on the path to better accuracy. However, as radiomics technology has become more advanced and readily available, that’s increased the amount of poorly curated datasets. When those datasets are used to program automated tools, it results in ineffectual models.[6] A particular benefit of SOPHiA GENETICS is the high quality of the data, as the efficacy of what is analyzed and shared on the platform is consistently subject to examination.

What solutions will GE’s collaboration with SOPHiA Genetics undoubtedly unearth? The cloud-based aspect of?SOPHiA DDM? is particularly exciting, as it massively increases the scalability of radiology platforms. In 2020, more than 1,500 peer-reviewed radiomics papers were published. But almost none of that research had an impact on the clinical world because of the aforementioned difficulty in sharing radiomics data.[7]

Better data sharing isn’t just the future of healthcare, it’s crucial to practicing precision medicine now. It might also be the key to advancing the field of radiomics, creating more detailed maps of tumor habitats, enabling the sharing of similarities and differences between lesions and how they are likely to progress and respond to different treatments. Combining the reach of GE’s vast clinical network with the SOPHiA DDM? platform could also lead to an advancement in the standardization of biomarkers, image acquisition, and data extraction. Stay up-to-date on the collaboration between GE Healthcare and SOPHiA Genetics to find out where this exciting joint effort leads

[1] Gillies RJ, Kinahan PE, Hricak H. Radiomics: Images Are More than Pictures, They Are Data. Radiology 2016;278:563-77. 10.1148/radiol.2015151169

[2]https://www.gehealthcare.com/article/integrating-cancer-care-with-powerful-insights-data-integration-and-collaboration

[3] https://www.sophiagenetics.com/technology/sophia-ddm-for-radiomics/

[4]https://www.gehealthcare.com/article/integrating-cancer-care-with-powerful-insights-data-integration-and-collaboration

[5] Peeken JC, Bernhofer M, Wiestler B, et al. Radiomics in radiooncology—challenging the medical physicist. Phys Med. 2018;48:27–36. doi: 10.1016/j.ejmp.2018.03.012

[6] Hosny A, Parmar C, Quackenbush J, Schwartz LH, Aerts HJWL. Artificial intelligence in radiology. Nat Rev Cancer. (2018) 18:500–10. 10.1038/s41568-018-0016-5

[7] 1 Pinto dos Santos, D., et al. (2021) 'A decade of radiomics research: Are images really data or just patterns in the noise?' European Radiology, 31, pp1-4.

?