They look through hundreds of 2D images of a single CT scan where cancer is minuscule and can be easily missed.
Google's lung cancer prediction model was built and trained on TensorFlow and comprises two frameworks: a full CT volume model to generate lung cancer malignancy predictions (viewed in 3D volume), and a malignant lesion detection model to identify subtle malignant tissues in lung nodules.
The model can also factor in information from previous scans to track the growth rate of suspicious lung nodules - which can be useful in predicting lung cancer risk, she adds.
While different radiologists may have their own interpretations of an individual scan, the Google study shows computers were as good as, or better than, doctors at detecting small lung cancer on CT scans. "AI in 3D can be much more sensitive in its ability to detect early lung cancer than the human eye looking at 2-D images".
Google's "promising" findings, published this week in the journal Nature Medicine, are only the beginning. Specifically, the new study shows that artificial intelligence being developed by the company can be used to help expand on the number of lung cancer screenings worldwide and save lives by getting the fight started earlier. In one trial, during which the participants were only allowed to examine one CT image per patient, the AI identified 5.5% more cases of cancer than the experts with 11% fewer false positives.
"Where prior computed tomography imaging was available, the model performance was on-par with the same radiologists", the study concludes.
A study by Google and several medical centres in the USA shows artificial intelligence (AI) can be used to detect lung cancer and provide doctors with more accurate readings of computerised tomography (CT) scans.
Etemadi leads his research team while also in anesthesiology residency training at Northwestern as part of a unique residency research track. And this novel system identifies both a region of interest and whether the region has a high likelihood of lung cancer. "Not only can we better diagnose someone with cancer, we can also say if someone doesn't have cancer, potentially saving them from an invasive, costly, and risky lung biopsy".
"By showing that deep learning can increase specificity without sacrificing sensitivity, we hope to spur more research and conversation around the role AI can play in tipping the cost-benefit scale for cancer screening", said Tse and Google technical lead Shravya Shetty in a blog post.
Tse and colleagues applied a form of AI called deep learning to 42,290 LDCT scans, which they accessed from the Northwestern Electronic Data Warehouse and other data sources belonging to the Northwestern Medicine hospitals in Chicago, IL.