Case Study

Google Health – Malignancy scoring. DeepHealth licensed Google Health’s AI research model for lung nodule malignancy prediction on CT

Collaboration to help improve lung cancer screening with AI


November 28, 2022

Aidence and Google Health announce an agreement to license Google Health’s AI research model for lung nodule malignancy prediction on CT imaging. Aidence will develop, validate, and bring this model to the market to support the early and accurate diagnosis of lung cancer and the reduction of unnecessary procedures in screening programmes.

Lung cancer screening

Lung cancer screening with low-dose CT has been shown to significantly reduce lung cancer mortality by as high as 24% for men and 33% for women, according to the 2020 NELSON trial. Screening initiatives, such as the UK’s Targeted Lung Health Checks, are increasingly being implemented in Europe. In the US, eligibility criteria have recently been broadened, further reflecting the benefits of lung cancer screening.

A major difficulty in lung cancer screening is establishing the nature of detected lung nodules. Most of these nodules are not cancerous. However, properly identifying and diagnosing such nodules can be time-consuming, costly, and anxiety-inducing for patients and their families. They may also be invasive, requiring follow-up CTs or surgical interventions.

Dr Raymond Osarogiagbon, Chief Scientist, Baptist Memorial Health Care Corporation and Director, Multidisciplinary Thoracic Oncology Programme, Baptist Cancer Center, Memphis, Tennessee, explained:

“One of the most exciting developments in contemporary population healthcare is the early detection of lung cancer. Unfortunately, the reality that most such nodules will be benign represents a real challenge that cries out for a technological solution. Artificial intelligence is one such solution.

The world looks forward to the rapid development and validation of software that will enhance our ability to find the many lung cancer needles in the giant haystack that is CT-detected lung nodules in today’s clinical practice.”

Deep learning

Deep learning, a subset of AI, has been shown to support the risk scoring of lung nodule malignancy. In a study published in Nature in 2019, scientists affiliated with Google Health presented a highly accurate model for malignancy classification, consistently matching the performance of experienced radiologists.

Aidence has also built a deep-learning model for this purpose. Aidence’s algorithm successfully predicted lung cancer from a single scan and was awarded in the 2017 Kaggle challenge. Its robust performance was later confirmed in a clinical study comparing its performance to that of 11 radiologists on 300 cases.

The collaboration

Aidence and Google Health intend to complete an AI application for lung nodule malignancy prediction. In this collaboration, Google Health will provide its scientific expertise. Aidence will further develop the model into a solution for clinical practice and bring it to market, complying with relevant data privacy requirements and regulatory standards. The development of this AI application is a statement of intent. No regulatory market applications have been made, and no orders for sale are being taken.

Outside of this collaboration with Google Health, Aidence has a proven track record of deploying AI in hospitals across Europe. Its application, Veye Lung Nodules, is currently running in over 80 routine practice and lung cancer screening sites.

Mark-Jan Harte, Aidence co-founder and CEO, said:

“Our mission at Aidence is to give lung cancer patients a fighting chance. This strategic partnership with Google Health allows us to accelerate and expand our efforts toward achieving it.

We are enthusiastic about working on a powerful deep learning model for lung nodule malignancy prediction based on the work of the Aidence and Google teams, as well as making sure that all the other requirements that contribute to the successful deployment of AI in clinical practice are in place, like clinical validation, certification and integration into the hospitals’ workflow.”

Akib Uddin, Product Manager at Google Health, commented:

“At Google Health, we want to be an active, catalytic force in demonstrating the real-world benefits of AI in health. We know just how important lung cancer screening is in saving lives, and we’re excited to play a role in driving impact at scale by enabling great partners like Aidence with our technology.”

About Google Health

A division of Alphabet, Inc., Google Health is a company-wide effort to help billions of people be healthier. They work toward this vision by meeting people in their everyday moments and empowering them to stay healthy, and partnering with care teams and the public health community to provide more accurate, accessible and equitable care.