Lung Tumor Study Reveals Variability of CT Scans - July 12, 2011 • Volume 8 / Number 14

Lung Tumor Study Reveals Variability of CT Scans
July 12, 2011 • Volume 8 / Number 14

Doctors and researchers use imaging tools such as computed tomography (CT) to monitor the size of a patient’s tumor and to assess responses to new therapies during clinical trials. To evaluate the performance of CT for these purposes, researchers compared measurements of tumor size on scans of the same patient taken just minutes apart. On many repeat scans, the study found, radiologists identified changes in tumor size that, if real, would be considered clinically important.

In the report of their findings, published online July 5 in the Journal of Clinical Oncology (JCO), the researchers noted that decreases in tumor size of less than 10 percent may not be distinguishable from changes caused by inherent variability in the scanning process. Dr. Gregory Riely of Memorial Sloan-Kettering Cancer Center (MSKCC) led the study.

Radiologists have known that many nonbiological factors can create apparent changes in tumor size on repeat CT scans. This study, for the first time, has quantified the effect.

The researchers scanned 30 patients with non-small cell lung cancer twice within 15 minutes using the same CT machine. Three radiologists then read the images without knowing the interval between the scans. They reported changes of more than 2 mm in tumor diameter on many of the repeat scans.

Clinical trials have increasingly reported small changes in tumor size as evidence of drug activity, the study authors noted. They point out, however, that changes of this magnitude could be solely the result of variability in imaging. As such, these kinds of changes should not by themselves be a marker of efficacy in clinical trials, the authors said.

This unprecedented study of the performance characteristics of CT may have “important implications for the future of drug development,” noted Dr. Michael Maitland of the University of Chicago and his colleagues in an accompanying editorial.

full-text ►
http://www.cancer.gov/ncicancerbulletin/071211/page3