In Nature Reviews Clinical Oncology, Colin Jacobs and Bram van Ginneken discussed the recent paper by Google AI in Nature Medicine on deep learning models for automatic detection of lung cancer from computed tomography.
The authors of the Nature Medicine paper report a level of performance similar to, or better than, that of radiologists. Jacobs and van Ginneken explain that these claims are currently too strong. The developed models are promising but need further validation and could only be implemented if screening guidelines were adjusted to accept recommendations from black- box proprietary AI systems.
Google's lung cancer AI: a promising tool that needs further validation.
Jacobs C, van Ginneken B.
Colin Jacobs and Bram van Ginneken are members of theme Rare cancers.
Related news items
Real-time dialogue with a dreaming person is possible25 February 2021
Dreams take us to what feels like a different reality. They also happen while we're fast asleep. So, you might not expect that a person in the midst of a vivid dream would be able to perceive questions and provide answers to them. But a new international study shows that, in fact, they can.read more
Vulnerable Nijmegen citizens less likely to visit GP physically due to corona23 February 2021
The COVID-19 pandemic in 2020 in Nijmegen and the surrounding area led to a substantial decrease in GP consultations for patients with chronic physical health problems.read more
SATB1 - Three classes of mutations and their unique rare diseases18 February 2021
Recent advances in DNA sequencing technologies have made it possible to uncover the causes of multiple rare diseases. A new collaborative study describes how three classes of mutation within the same gene result in different neurodevelopmental disorders.read more
New Physical, Mental, and Cognitive Problems 1-year Post-ICU18 February 2021
Half of the Intensive Care Unit (ICU) survivors suffer from new physical, mental and/or cognitive problems one year after ICU admission. This is evident from the large-scale MONITOR-IC study led by the Radboudumc.read more
Trained immunity: a tool for reducing susceptibility to and the severity of SARS-CoV-2 infection17 February 2021
In a review in Cell Mihai Netea, Frank van de Veerdonk, Reinout van Crevel and Jorge Dominguez Andres propose that induction of trained immunity by whole-microorganism vaccines may represent an important tool for reducing susceptibility to and severity of SARS-CoV-2.read more