Deep learning proves effective in spotting liver masses in CT

The consternation of radiologists about the impact of artificial intelligence is real—but so are the benefits of machine learning. Recent research showed that deep learning with a convolutional neural network (CNN) was successful in differentiating liver masses in CT.

The retrospective study, published online Oct. 23 in Radiology, examined the diagnostic abilities of a deep learning method with a CNN. Researchers tested the CNN with 100 liver mass image sets from 2016, including 74 men and 26 women with the average age of 66 years old.

“This preliminary study, which used 55, 536 image sets (1068 image sets augmented by a factor of 52) to obtain models, indicated that classifying liver masses into five categories can be accomplished with a high degree of accuracy by using a deep learning method with a CNN on dynamic contrast-enhanced CT images,” wrote Koichiro Yasaka, MD, PhD, with the department of radiology at the University of Tokyo Hospital in Japan, and colleagues.

The study featured two stages. The first was a training stage, where researchers built the deep learning models by using CT image sets. An abdominal radiologist searched image archives for CT studies that examined liver lesions.

The test stage then examined the accuracy of the model.

“After building the models, we examined the accuracy of the trained models in distinguishing among the five liver mass categories by using test CT image sets,” wrote Koichiro et al. “These image sets were provided for the CNN.”

The deep learning method, according to the researchers, was designed to allow radiologists to more narrowly focus on tumor detection.

“Our model required the radiologists only to focus on the tumors, capture the images and provide the image to the CNN,” wrote Koichiro et al. “Our model also differs from previous studies that used conventional machine-learning methods for the differentiation of liver masses in that it does not require complex-shaped regions of interest tracing boundaries of tumors, or circular or elliptical regions of interest within tumors.”

The CNN model showed its ability to detect liver masses, but the researchers also mentioned that further improvements were necessary before such a method could be used to diagnose rare malignant masses.

""
Nicholas Leider, Managing Editor

Nicholas joined TriMed in 2016 as the managing editor of the Chicago office. After receiving his master’s from Roosevelt University, he worked in various writing/editing roles for magazines ranging in topic from billiards to metallurgy. Currently on Chicago’s north side, Nicholas keeps busy by running, reading and talking to his two cats.

Trimed Popup
Trimed Popup