News | Computer-Aided Detection Software | October 06, 2015

New Hybrid CAD Scheme May Improve CT Lung Nodule Detection

Study suggests computer-aided detection could reduce number of missed nodules and false positives

hybrid CAD scheme, computer-aided detection, lung nodule detection, Medical Physics, low-dose CT scans

Image courtesy of Vital Images

October 6, 2015 — A novel computer-aided detection (CAD) scheme based on a hybrid multi-classifier method could enhance early detection of lung nodules on computed tomography (CT) images and help in the earlier diagnosis of lung cancer, suggests new research. The research was published in the September issue of Medical Physics, the monthly journal of the American Association of Physicists in Medicine (AAPM).

Every year 160,000 people die of lung cancer, making it the number one cancer killer. Early detection and treatment of lung cancer is key to saving lives. The National Lung Screening Trial (NLST) — which was launched in 2002 and enrolled 53,454 high-risk subjects at the time of reporting — showed a 15 to 20 percent lower risk of dying from lung cancer when screened using low-dose CT scans compared to the chest X-rays. However, reading the large number of CT scans can be a burden for a radiology practice. The CAD method with high sensitivity and specificity could help reduce both the number of nodules that could be missed but may be lung cancer, as well as the number of false positives that are unnecessarily followed up and/or biopsied.

“A big challenge of CAD for lung cancer screening is the wide diversity of lung nodule morphology seen on CT,” said Lin Lu, Ph.D., first author of the article and postdoc research fellow at Columbia University Medical Center, New York. “Our strategy is to categorize and partition nodule candidates into different groups based on their location, size and shape, so that each group possesses more homogenous morphological characteristics of nodules and can be assigned a more appropriate algorithm to improve the performance of CAD.”

In the study, researchers evaluated 631 lung nodules on 294 CT scans from the National Cancer Institute’s Lung Image Database Consortium (LIDC) database, which were acquired to facilitate a test-bed for CAD algorithms. The image data sets were randomly divided into two independent subsets: a training set of 196 scans and a test set of 98 scans. The hybrid method achieved a sensitivity of 87 percent with 2.61 false positives per scan on the training dataset and a sensitivity of 85.2 percent and 3.13 false positives per scan on the testing dataset. The hybrid method was compared to and outperformed six recent CAD schemes.

Co-authors of the study in addition to Lu are Yongqiang Tan, Ph.D., Lawrence H. Schwartz, M.D., and Binsheng Zhao, D.Sc. (corresponding author).

For more information:

Related Content

Digital Mammography Increases Breast Cancer Detection
News | Mammography | January 16, 2019
The shift from film to digital mammography increased the detection of breast cancer by 14 percent overall in the United...
Artificial Intelligence Used in Clinical Practice to Measure Breast Density
News | Artificial Intelligence | January 15, 2019
An artificial intelligence (AI) algorithm measures breast density at the level of an experienced mammographer,...
Machine Learning Uncovers New Insights Into Human Brain Through fMRI
News | Neuro Imaging | January 11, 2019
An interdisciplinary research team led by scientists from the National University of Singapore (NUS) has successfully...
Mobile App Data Collection Shows Promise for Population Health Surveys
News | Population Health | January 10, 2019
Mobile app data collection can bring access to more potential clinical study participants, reduce clinical study...
3-D Reconstruction of Ichthyosaurus Skull

A 3-D reconstruction of the ichthyosaurus skull from a computed tomography (CT) scan. Image courtesy of Nigel Larkin, taken at Royal Veterinary College, London.

News | Computed Tomography (CT) | January 09, 2019
A nearly meter-long skull of a giant fossil marine ichthyosaur found in a farmer's field more than 60 years ago has...
SCCT Releases New Guideline for CT Use During TAVR
News | Computed Tomography (CT) | January 08, 2019
The Society of Cardiovascular Computed Tomography (SCCT) has released a new expert consensus document for computed...
Hypertension With Progressive Cerebral Small Vessel Disease Increases Cognitive Impairment Risk
News | Magnetic Resonance Imaging (MRI) | January 08, 2019
Patients with high blood pressure and progression of periventricular white matter hyperintensities showed signs of...
Artificial Intelligence Pinpoints Nine Different Abnormalities in Head Scans

A brain scan (left) showing an intraparenchymal hemorrhage in left frontal region and a scan (right) of a subarachnoid hemorrhage in the left parietal region. Both conditions were accurately detected by the tool. Image courtesy of Nature Medicine.

News | Artificial Intelligence | January 07, 2019
The rise in the use of computed tomography (CT) scans in U.S. emergency rooms has been a well-documented trend1 in...
Electronic Brachytherapy Effective in Long-Term Study of 1,000 Early-Stage Breast Cancers
News | Brachytherapy Systems, Women's Healthcare | January 07, 2019
Breast cancer recurrence rates of patients treated with intraoperative radiation therapy (IORT) using the Xoft Axxent...
Brachytherapy Alone Superior Treatment for Intermediate-Risk Prostate Cancer
News | Brachytherapy Systems | January 04, 2019
Patient-reported outcomes (PROs) indicated a significantly different clinician and patient-reported late toxicity...