Google Develops AI Technology for Enhanced Ultrasound Diagnosis and Cancer Treatment

AI is not just useful for writing term papers or creating clickbait; it has the potential to save lives in the medical field. At Google's annual healthcare event, The Check Up, the company unveiled partnerships focused on enhancing ultrasound readings, medical language models, and cancer treatments—areas poised for significant benefits from AI technology.

Google aims to improve ultrasound device usage in regions lacking trained specialists. While ultrasound sensors are more accessible, expert interpretation is still required. The company’s AI models could streamline this process, aiding in the early detection of conditions such as breast cancer and assessing gestational age in pregnant women. To further this initiative, Google is collaborating with Jacaranda Health, a Kenyan nonprofit, to research AI-driven ultrasound applications for mothers and infants in government hospitals. “This partnership will help us explore how AI tools can enhance point-of-care ultrasound delivery for pregnant women in Kenya,” shared Greg Corrado and Yossi Matias from Google Health AI.

Additionally, Google is working with Chang Gung Memorial Hospital in Taiwan to investigate AI's capability to detect breast cancer via ultrasound. This method serves as a promising alternative to mammograms, which are often less accessible in under-resourced areas and may be less effective for women with denser breast tissue.

Google has also made significant strides with its medical-focused large language model (LLM), Med-PaLM 2, which recently achieved an 85 percent accuracy rate on doctor-level medical exam questions—an 18 percent improvement from its predecessor. This model not only provides accurate responses but also justifies its answers and evaluates its accuracy.

However, it's important to note that AI is not yet ready to fully replace human doctors. Google emphasizes that significant gaps still exist in answering medical questions based on criteria like scientific accuracy and reasoning. The team is committed to collaborating with researchers and the medical community to bridge these gaps and enhance healthcare delivery.

Furthermore, Google is partnering with Mayo Clinic to investigate AI’s role in streamlining radiotherapy planning for cancer treatment. This research aims to reduce the lengthy, intricate steps involved—specifically, the contouring process, which can take up to seven hours per patient. Findings from this three-year study will soon be published, along with expanded research initiatives into AI applications in radiotherapy.

Lastly, Google envisions AI playing a critical role in chest x-ray screenings for tuberculosis (TB). In collaboration with an AI-driven organization, Google aims to implement AI-enhanced TB screenings across Sub-Saharan Africa, with plans to provide 100,000 free screenings to facilitate early detection and treatment, helping to mitigate the disease's spread.