Deep learning helps radiologists target missed lung cancer diagnoses on chest x-rays

Chest radiography is the first-line tool for identifying lung cancer, but it has a spotty record, with sensitivity rates as low as 44%. Wanting to close this care gap, South Korean researchers have developed a new deep learning tool that has reduced the number of overlooked cases without creating a huge uptick in unnecessary follow-up CT imaging.

Scientists detailed their new “automatic detection algorithm” in a new study published Tuesday in Radiology. They noted that using the system helped to “significantly” improve physician performance and warrants further investigation, wrote lead author Sowon Jang, from the Department of Radiology at Seoul National University Bundang Hospital, and colleagues.

“Using a deep learning-based automatic detection algorithm may help observers reduce the number of overlooked lung cancers on chest radiographs, without a proportional increase in the number of follow-up chest CT examinations,” the team concluded.

Jang et al. reached their conclusions by pinpointing 117 patients diagnosed with lung cancer between 2010-2014 who also had previous lung imaging with visible signs of the disease, prior to the prognosis. They also targeted a control group of 234 healthy patients with normal chest radiographs. Nine observers—including six chest radiologists and three residents—then reviewed each x-ray with and without the help of deep learning, locating cancers and determining which required further investigation.

The average area under the alternative free-response receiver operating characteristic curve rose “significantly” with the help of deep learning from 0.67 up to 0.76, Jang and colleagues reported. And with DL automatic detection, they recorded a higher sensitivity rate than without its help (53% versus 40%), while also recommending chest CT more frequently (62% versus 47%). Meanwhile, in the healthy control group, the number of unnecessary chest CT orders due to false-positive findings was similar with and without the artificial intelligence assist.

Read more of the analysis in RSNA’s flagship publication here.

Marty Stempniak

Marty Stempniak has covered healthcare since 2012, with his byline appearing in the American Hospital Association's member magazine, Modern Healthcare and McKnight's. Prior to that, he wrote about village government and local business for his hometown newspaper in Oak Park, Illinois. He won a Peter Lisagor and Gold EXCEL awards in 2017 for his coverage of the opioid epidemic. 

Trimed Popup
Trimed Popup