It seems to me that if one has an axe to grind, and would like to do so in a peer-reviewed journal, then the easiest way would be to press into service a meta-analysis, like the one just published in JAMA (February 21, 2001) on four decades of study of the use of FDG-PET for diagnosis of pulmonary nodules and mass lesions.
The good news is that the researchers from the VA Palo Alto Health Care System in California concluded from the analysis that FDG-PET is an accurate test for identifying malignant pulmonary lesions at least 1 cm in diameter, though they noted that most studies operated at a test threshold favoring sensitivity rather than specificity, which minimized the number of false-negative results. They placed current practice of PET at a point on a ROC curve at which sensitivity and specificity are approximately 96.8% and 77.8%, respectively.
The not-so-good news is that the authors were highly critical of the quality of these studies, stating that not one satisfied all of the researchers’ admittedly rigorous study quality criteria, but adding that the studies adequately described their technical methods and followed most of the guidelines published by the Society of Nuclear Medicine.
Had the researchers chosen melanoma, or lymphoma, or esophageal cancer, it is impossible to predict whether the published studies would have fared any better or any worse. But by focusing on pulmonary nodules and mass lesions, the researchers gave their study profound economic resonance due to the fact that early detection of lung cancer has not been proven conclusively to have verifiable impact on long-term outcomes for patients. And in the absence of clear medical value, cost becomes more important. A PET scan of the lungs could be perceived as cost-effective only in the event that it yielded information that eliminated futile surgeries. The cost of a PET diagnosis is about $1,912, considerably more than the $560 reimbursed for a CT-guided needle biopsy (assuming there are no complications).
“PET scans and technology assessment: d?j? vu?” ask Ethan Balk, MD, and Joseph Lau, MD, with the division of clinical care research at New England Medical Center in their commentary, published in the same issue. They begin their editorial by acknowledging that few diagnostic technologies have had as profound an effect on clinical medicine as have MRI and CT, and then they cut to the chase: the role these technologies have played in rising health care costs. The authors proceed to state their concern that MRI and CT have been adopted before being adequately assessed. Ironically, the two supporting references for this statement include an article dating from 1985, and another published in the New England Journal of Medicine in 1993 called “Magnetic resonance coronary angiography: assessing an emerging technology.” In last month’s issue we quoted Geoffrey D. Rubin, MD, chief of cardiovascular imaging at Stanford University School of Medicine, as saying that, where CTA and MRA have been adopted, diagnostic contrast angiography has been virtually replaced, reducing patient discomfort and health care costs. Now medicine needs proof.
The editorialists raise a real issue with their concern that PET will help inflate the total health care bill. But is technology the issue, or is it overutilization? Our quasi-free-market health care system has sidestepped the issue of developing clinical care pathways. The questions we are not asking include: Are emergency physicians relying unnecessarily on imaging to make diagnoses? Are primary care physicians caving in to patients who demand MRI for a pain in the back or a swollen knee? What about the chain-smoking 70-year-old woman who is hospitalized because she is dying of emphysema? Her physician has ordered diagnostic imaging tests for suspected cancer, a diagnosis that would not be acted upon, yet would ensure that an exhausted daughter can have the assistance of hospice care if and when her mother comes home to die. Is that test necessary ?
The solution to high health care costs is not to quash an important, 30-year-old technology that provides new information that cannot be obtained by existing technology. HCFA’s conservative decision to reimburse for PET for broadened indications for five types of cancers, myocardial viability following inconclusive SPECT, and presurgical evaluation of refractory seizures means that insured Americans, not just those affluent enough to pay out of pocket for a state-of-the-art diagnosis, will have access to tests for the functional information that FDG-PET supplies.
The challenge, then, is to devise those efficacious diagnostic pathways and then convince the medical community to act upon them-not to figure out how many studies can dance on the head of a pin.