CT texture analysis helps characterize lung lesions

Tuesday, December 2 | 11:10 a.m.-11:20 a.m. | SSG07-05 | Room S402AB
Researchers from Massachusetts General Hospital have found that texture analysis of CT images can help differentiate malignant and benign lung lesions.

Lung biopsies are associated with complications that can be serious and potentially life-threatening in some patients. It would be beneficial to identify benign lung nodules ahead of time to avoid unnecessary lung biopsies and potentially reduce the morbidity associated with biopsy complications, according to presenter Dr. Roberto Lo Gullo.

"As CT scans are performed to detect lung nodules and lung biopsies are performed under CT guidance, it makes sense to do further analysis of CT images with texture analysis," he said.

In a study involving 62 patients who had received CT-guided biopsy and had a definitive diagnosis from pathology, the researchers retrospectively used commercially available software (TexRad) to perform tissue analysis of eight features on prebiopsy CT images. Based on the tissue analysis, it was possible to differentiate benign and malignant lung nodules, Lo Gullo said.

"We can also differentiate between different histological types of lung cancers such squamous and aggressive adenocarcinomas," he told AuntMinnie.com. "The distinction between benign and a few low-grade tumors, however, is difficult to make."

Page 1 of 155
Next Page