AI isn’t just good for writing term papers or clickbait financial explanations; could help save lives in the medical field. At Google’s annual The Check Up healthcare event, he announced partnerships related to AI for ultrasound readings, medical language models and cancer treatments, areas where the technology could one day serve as a force for good.
Google sees AI as crucial for reading ultrasound devices in regions without enough trained specialists. Although sensors are more accessible than ever, they require experts to perform examinations and interpret images. The company’s artificial intelligence models could help simplify that process by identifying data such as early detection of breast cancer and gestational age in pregnant women. To help make that a reality, the search giant is partnering with Jacaranda Health, a Kenyan nonprofit, to research AI-based ultrasound treatments for mothers and babies in government hospitals. “Through this partnership, we will conduct exploratory research to understand the current approach to ultrasound delivery in Kenya and explore how new AI tools can support point-of-care ultrasound for pregnant women,” said Greg Corrado. , Google Health’s head of AI and vice president of engineering. Yossi Matias in a blog post today.
The company is also working with Chang Gung Memorial Hospital in Taiwan to investigate how AI can detect breast cancer through ultrasound as an alternative to mammograms, which have limited availability in low-resource regions. Mammograms may also be less effective in populations with higher breast density.
Evelyn Hockstein/Reuters
In other areas, Google says its medical-focused large language model (LLM) has improved significantly. Med-PaLM 2, the company’s next-generation healthcare LLM, recently scored an 85 percent on physician-level medical exam questions, an 18 percent improvement over the previous version’s score. “This model not only accurately answered open-ended and multiple-choice questions, but also provided rationale and tested their own answers,” Corrado and Matías said.
Don’t expect a ChatGPT-like bot to replace your doctor any time soon, though, as Google cautions that the technology isn’t quite ready for real-world work settings yet. For example, an evaluation on criteria such as scientific feasibility, precision, medical consensus, reasoning, bias, and harm found “significant gaps” in answering medical questions. Corrado and Matias noted, “We look forward to working with researchers and the global medical community to close these gaps and understand how this technology can help improve healthcare delivery.”
Google has also partnered with the Mayo Clinic to explore the role of AI in radiation therapy planning for cancer treatment. Research is focused on reducing the tedious and time-consuming steps of the radiation therapy process, in particular, “contouring.” This technique requires doctors to draw lines on CT scans to separate cancerous areas from nearby healthy tissue that radiation could damage, a process that can take up to seven hours for a patient. The company says it will soon publish the research from the three-year study as it formalizes an agreement with Mayo Clinic to explore more radiation therapy-based research, AI models and commercial uses.
Finally, Google believes that AI helps with the detection of tuberculosis through chest X-rays. The company is partnering with an AI-based organization to make AI-powered TB screening tests widely available in sub-Saharan Africa. Its partners have pledged to donate 100,000 free screenings to help detect TB early and provide early treatment to slow its spread.
