Google Cloud

New Way Now: How DeepHealth shifts cancer detection to curative using Google Cloud

*Featured in this video:* Sham Sokka, COO and CTO, DeepHealth *Executive summary:* Radiology imaging volumes are growing 10-15% annually while a 15% radiologist shortage looms by 2030. DeepHealth, the digital health subsidiary of the largest imaging services provider in the US, built DeepHealthOS on Google Cloud to address this challenge, unifying imaging, informatics, and clinical AI applications within a single connected environment that accelerates clinical insight, collaboration, and innovation across the continuum of care. Using Google Cloud's Healthcare API, Vertex AI, BigQuery, Looker, and MedGemma, DeepHealth delivers measurable and meaningful impact across large-scale screening programs and diagnostic workflows. For example,driving a 21% increase in breast cancer detection rate, enabling the majority of lung cancers to be found at earlier stages, and providing automated thyroid cancer detection and reporting. *Challenge:* Demand for radiology services far exceeds available workforce capacity in the US and globally. Over the next five years, demand is expected to exceed supply by 15% in the US and up to 50% in some regions. Shortages of technologists, radiologists, and operational staff are driving up costs and limiting access. This is important, as clinical variation in radiology interpretation can be as high as 30%, meaning two radiologists agree only two out of three times, which can affect diagnostic accuracy and patient outcomes. In addition, ​​across the imaging journey — from patient engagement and scheduling to imaging, interpretation, and follow-up — systems are fundamentally disconnected, forcing clinicians and staff to move data across different tools. These are areas where AI has already shown the potential for significant impact. DeepHealth, a digital health company that serves RadNet's 400+ imaging centers and over 2,700 customer contracts across North America, Europe, and Asia Pacific, looked for a partner to accelerate its mission of empowering breakthroughs in care through imaging. *Solution:* DeepHealth built DeepHealthOS on Google Cloud, choosing the platform for two primary reasons: healthcare-forward expertise through the Healthcare API and the robust Google AI stack which includes research prowess in medical applications. The company uses Gemini Enterprise Agent Platform to run agentic tooling that powers radiology workflows, BigQuery to create and curate data cohorts for model training, Looker for analytics, and MedGemma for report summarization and medical language processing. The platform brings imaging data together, automates tedious tasks, and standardizes radiology outputs to reduce variability. This all supports DeepHealth’s vision of advancing automated diagnostics where images are fed into the system and standardized draft reports are automatically generated. *Result:* DeepHealth's clinical AI applications have delivered a 21% increase in breast cancer detection rates, stage-shifted lung cancer detection with the majority found at a more curative stage, and reduced scan slot time by up to 30% within existing thyroid ultrasound workflows. By standardizing radiology outputs with AI consistency, DeepHealth aims to reduce the interpretation burden and making it easier and faster for radiologists to detect disease. *Highlights and key takeaways from our interview with Sham Sokka, COO and CTO at DeepHealth:* → “With the use of clinical AI, we've improved cancer detection rates by 21% in breast cancer. ” → “DeepHealth is focused on imaging platforms to bring the data together, to make it easier for the operators to do their jobs, and then we empower them with AI to make their everyday lives easier.” → “We evaluated multiple language models, and MedGemma is the strongest for medical awareness and medical use cases.” Google Cloud products used: BigQuery, Gemini Enterprise Agent Platform, Healthcare API, Looker, MedGemma Learn more: → https://deephealth.com/