- AI
- Molecular Imaging
- CT
- X-Ray
- Ultrasound
- MRI
- Facility Management
- Mammography
False positives hamper lung CAD sensitivity
Computer-aided detection systems have proved beneficial for detecting lung nodules, but how do they fare in helping radiologists identify lung cancer? Not too well, according to a study presented at the European Congress of Radiology by a team of researchers from various institutions in South Korea.
Computer-aided detection systems have proved beneficial for detecting lung nodules, but how do they fare in helping radiologists identify lung cancer? Not too well, according to a study presented at the European Congress of Radiology by a team of researchers from various institutions in South Korea.
Led by Dr. Jin Mo Goo, investigators evaluated 150 chest CT exams that included 23 lung cancers less than 20 mm in size as well as normal cases.
Five chest radiologists and five radiology residents independently recorded the locus of each nodule candidate and assigned confidence scores to each based on likelihood of nodule and malignancy without CAD. They then repeated analysis with CAD.
Lung nodule detection significantly increased with CAD for all radiologists and subgroups of chest radiologists and radiology residents. CAD itself detected 18 of 23 lung cancers. Four lung cancers missed by three radiology residents on initial reading were additionally detected with CAD.
However, because the number of false-positive detections for lung cancer increased with the use of CAD, the overall performance of lung cancer detection was not significantly different with and without CAD for all radiologists and subgroups, according to Goo.
Computed Tomography Study Assesses Model for Predicting Recurrence of Non-Small Cell Lung Cancer
January 31st 2025A predictive model for non-small cell lung cancer (NSCLC) recurrence, based on clinical parameters and CT findings, demonstrated an 85.2 percent AUC and 83.3 percent sensitivity rate, according to external validation testing in a new study.
