Browse By Topic and Related Items
- COSM 2012: Award Winners Tackle Sleep-Disordered Breathing, Unilateral Hearing Loss and Tumor Resections
COSM 2012: Dr. Jonas Johnson Explores the Subtleties of PET/CT for Tumors
From: ENT Today, May 2012
by Thomas R. Collins
PET combined with CT might be one of the most powerful imaging technologies available, but how effective it is in evaluating head and neck tumors and helping with treatment is not a cut-and-dried proposition, said Jonas Johnson, MD, professor and chair of otolaryngology at the University of Pittsburgh in his State of the Art Lecture.
Dr. Johnson gave the presentation at the 115th Annual Meeting of the Triological Society, held here on April 20 as part of the Combined Otolaryngology Spring Meetings.
He said that while those in his group at the University of Pittsburgh were “early adopters” of PET/CT in head and neck tumor treatment, he acknowledged that it’s a costly technology that requires understanding of the subtleties.
“It’s very sensitive and phenomenally expensive,” Dr. Johnson said. “The real problem is that while PET/CT is very reliable in terms of its negative predictive value, it is plagued by false positives.”
His address didn’t include any discussion of PET only, since by itself it gives “little or no anatomic detail.” “PET/CT is vastly better than PET alone,” he said.
He issued the caveat that neither the National Comprehensive Cancer Network (NCCN) Guidelines nor the Radiation Therapy Oncology Group (RTOG) includes any suggestions for the use of PET/CT. “If you use PET/CT, you’re out there a little in front of the rest,” he said.
A main theme of Dr. Johnson’s remarks was that the benefits of PET/CT can be blurry. The value of PET/CT varies according to the size of the tumor, he said, discussing the case of a 4-mm metastasis.
“This is a very tiny microscopic metastasis,” he said. “We all understand that you cannot identify this volume of tumor with any modality currently available, other than by accident.”
As far as determining tumor staging, Dr. Johnson said the value of PET/CT might be dubious. “PET/CT has improved sensitivity, and the findings may influence therapeutic decisions, but frankly there’s no good cost-effective data,” he said. “We don’t have any yet and we need it…. PET/CT is inadequate to identify occult metastases in either the neck or distant (metastases).” But, he said, “the best available neck node data is from histology.”
Studies have shown that PET/CT detects about half of occult neck nodes and cannot replace the accuracy of a neck dissection, Dr. Johnson said. For identifying unknown primary tumors, studies have found that PET/CT has detected as many as 87 percent of them—which seems like “a lot,” Dr. Johnson said—down to 38 percent. The best meta-analysis suggests PET will identify one-third of these primary tumors, he said. “PET/CT, our most sensitive imaging modality, will not identify microscopic disease,” he added.
The use of PET/CT for surveillance in patients who’ve received chemoradiation treatment is one of the most difficult issues the technology presents, Dr. Johnson said. Generally, he said patients should get a PET/CT sometime around eight to 12 weeks after treatment.
“We’re pretty sure that if you did it before eight weeks you will be badly confounded by false positives,” he said. “The closer that PET is to the completion of treatment, the more likely that you won’t have a false positive.” In fact, he said, even histology may be misleading if it’s performed as soon as six to eight weeks of finishing therapy.
The longer the wait before PET/CT is done, the more reliable the results will be, but that presents a problem, he said. “If you wait four months… the problem that we’ve all struggled with is delay,” he said. “I mean, we worry if we wait a long time we will lose the opportunity to cure the patient. And wouldn’t it be better to find that persistent disease sooner rather than later?”
Still, he said, he is unconvinced of the benefit of earlier imaging. “[With] these persistent cancers that have not [completely] responded to the chemoradiation, finding them early has not been demonstrated to change outcome.”
The question of how long PET/CT is needed in surveillance is not known, he said. But his group in Pittsburgh does them every three months. When it’s okay to stop doing PET/CT for surveillance is similarly undefined, he said.
“Once again, the data are not out there,” he said. “All of us have to struggle with this as we go along.”Generally, about 18 months is good guideline, he said.
Dr. Johnson said caution is important when a tumor is found that is negative for fluorodeoxyglucose (FDG) avidity after completion of chemoradiation. He used the example of a patient in whom an FDG-negative N2a neck tumor is found after completion of chemoradiation therapy. “It does not need neck dissection,” Dr. Johnson said. “In fact, neck dissection may be harmful to these patients, in terms of development of severe dysphagia in the future…. Obviously, the tumor is dead but it didn’t go away, at least not yet. And it remains a conundrum and a topic of some debate in some centers.”
At Pittsburgh, his team watches such patients and they “almost never” discover that it was the wrong course to have taken, he said.
Robert Maisel, MD, professor of otolaryngology at the University of Minnesota Medical School, where PET/CT has been used for about seven years, said Dr. Johnson’s address hit important points. In his group, he said, patients get a post-radiation PET/CT at 12 weeks to evaluate the results—that’s “the one thing we’re sure of,” he said—and then a second one is done six months later.
“We think that for small tumors, PET/CT is unnecessary,” he added. They also don’t know how long PET/CT should be performed for surveillance. “Nobody knows the answer to that,” he said.
Dr. Johnson said there is value in using PET/CT beyond mere surveillance, in finding unknown tumors, but that it’s difficult to persuade insurers of its value, largely due to the lack of data. “The radiologists now are specialized in reading it, and originally there was nobody who trained them to read it,” he said. “So the original data is not as useful as the present data.”