Researchers explore power of PET/CT for lung cancer diagnosis

December 2, 2005

Quantitative data and new radiotracers developed for PET/CT boost detection of lung cancer while reducing false positives, according to research presented at the RSNA meeting.

Quantitative data and new radiotracers developed for PET/CT boost detection of lung cancer while reducing false positives, according to research presented at the RSNA meeting.

Some of the studies contradict long-held beliefs about how best to use PET/CT for this application. For example, while previous research has supported the view that CT's tissue density attenuation and FDG-PET's standardized uptake value (SUV) have no business together, researchers at the RSNA offered a radically different view.

Dr. Kiyoshi Nishikawa and colleagues from the Miyazaki and Tsukuba universities in Japan assessed 93 patients with PET/CT. The investigators measured average Hounsfield units (HU) and SUV in the region of interest around pulmonary abnormalities.

The team found the combined anatomic/functional quantitative assessments were clinically useful to detect malignancies, even if these lesions are missed or misinterpreted by the PET component.

In another study, Dr. Irfan Farukhi from the University of Texas Southwestern Medical Center in Dallas presented results of his research based on 53 randomized patients. The team had noted previously the difficulty in differentiating benign granulomatose disease from neoplasms using hybrid imaging. In the current study, Farukhi and colleagues combined HU density values from 3D volumetric CT with conventional SUV to improve characterization. They found the technique increased PET/CT sensitivity and specificity to differentiate benign from malignant pulmonary lesions.

Another research team, headed by Drs. Qiang Li and Yongkang Nie from the University of Chicago, developed a computer-aided diagnostic system to improve PET interpretation, based on a database of PET/CT images.

The CAD system uses a selective enhancement filter that boosts tumors while suppressing normal anatomic structures. The system then applies a discrimination algorithm to weed out false positives. Using this scheme, investigators achieved a high detection rate for true malignancies and lowered PET's false-positive rate.

A final study by Jamey Weichert, Ph.D., and colleagues from the University of Wisconsin-Madison unveiled preliminary results with the new SPECT ligand iodine131-

labeled NM404 on six patients with advanced non-small cell lung cancer. They found the patients tolerated the NM404 well. They also observed similar tumor uptake and retention properties in human NSCLC as previously observed in mice undergoing imaging with this agent. Studies are underway to extend the utility of NM404 to PET.

"While it is not logistically feasible to expect any one agent to trace every molecular process, knowing of one, such as NM404, that may have a broad applicability is certainly intriguing," said Dr. Richard L. Wahl, director of nuclear medicine at Johns Hopkins. "Although we have years of work left ahead, PET/CT seems to be working well for lung cancer imaging worldwide," he said.