NEW YORK (GenomeWeb) – A chart review study by researchers associated with Integrated Diagnostics indicates treatment of patients with pulmonary nodules could be significantly improved were physicians to more consistently apply clinical standards when evaluating these features.
On their face, the findings suggest the potential usefulness of Indi's Xpresys Lung cancer test as an objective measure of nodule status that can aid physician decision making. However, they also call into question the extent to which doctors will tailor their treatment decisions based on results of the Xpresys test given the apparent low compliance with existing methods of nodule evaluation.
Launched in 2013, Xpresys is intended to aid doctors in identifying lung nodules detected via CT scans as likely benign. The test uses multiple-reaction monitoring mass spec to quantify the levels of 11 proteins in patient blood samples.
The rationale for Xpresys is that, by identifying nodules that are likely benign, physicians can more comfortably put patients on a regimen of watchful waiting. A non-invasive test like Xpresys is desirable due to the expense and significant risks of the invasive procedures currently used for determining lung cancer status, like fine needle aspiration and thoracotomies.
The recent chart review study, published last week in the journal Chest, bolstered the case for Xpresys by demonstrating significant overtreatment in the management of pulmonary nodules.
In the study, which Indi funded, the researchers looked at the management of 377 patients aged 40 to 89 in 18 community pulmonary practices around the country who had been referred for the indeterminate pulmonary nodules Xpresys is intended to evaluate.
Of the 377 patients with nodules, 25 percent were cancerous. Forty six percent of the 377 patients were put on a regimen of surveillance, 33 percent underwent biopsies, and 20 percent underwent surgery. Of those who had surgery, 35 percent had benign nodules.
Using pre-test probability for malignancy calculations (pCA) that combine data like patient age, sex, smoking history, and nodule size to determine the likelihood that a nodule is cancerous, 10 percent of patients were classified as low risk, 80 percent were classified as moderate risk, and 10 percent were classified as high risk.
The Chest study found that pCA was effective in ruling out patients as likely benign, with 100 percent of nodules evaluated as having a <5 percent pCA turning out to be benign. However, surgical resections were performed at similar rates across all patient risk groups — 17 percent for patients with low-risk pCA scores, 21 percent for patients with moderate risk scores, and 17 percent for patients with high risk scores.
"I think the paper shows that indeterminate nodules are very difficult to diagnose, and that the way they are diagnosed is not standard across clinics or physicians," Paul Kearney, Indi's co-founder, president, and CSO, told GenomeWeb. "There appears to be a high use of risky invasive procedures, and a lot of this is a reflection on the fact that there are very few objective tools out there for doctors to use."
However, the fact that patients with low-risk pCA scores underwent invasive procedures at roughly equivalent rates to those with moderate and high risk scores suggests that physicians are not necessarily basing their decisions on the objective measures that they, in fact, do have. This raises the question of the extent to which the Xpresys test will influence physician decisions.
In a study published in January in the Journal of Thoracic Oncology, Indi researchers found that the Xpresys test adds predictive value independent of the clinical variables used in generating pCA scores and that the combination of Xpresys and pCA offered a significant improvement in performance over the pCA alone.
Kearney said that Indi knows "anecdotally" that the Xpresys test does influence doctor's treatment decisions, but he acknowledged that a formal study is required to determine exactly the extent to which that is the case. The company is currently "setting up a study to capture that kind of information," he said.
"It's a complex scenario because the patient is involved, and so the doctor provides recommendations that may or may not agree with the [treatment] guidelines,"Kearney said. "You may have a patient who is very anxious and wants to proceed with aggressive treatment, and then there are definitely patients who are on the other end of the spectrum who do not want to move to aggressive procedures."
Indeed, in an interview earlier this year with GenomeWeb, Indi CEO Al Luderer alluded to challenges the company had encountered in introducing physicians to the test.
"I have to admit that providing a test that gives a probabilistic value is still kind of a new thing for docs," Luderer said. "It has been interesting to give them a mathematical number that says, 'you're most likely benign.'"
These comments and the findings of the Chest paper also reflect observations made at the time of the test's launch by some physicians who indicated that doctors might be uncomfortable relying on a test that left even a small amount of doubt as to whether a lesion was benign.
For instance, Howard West, a thoracic oncologist at Seattle's Swedish Cancer Institute, suggested that the test's low positive predictive value could, in fact, lead to higher numbers of follow-up scans and invasive procedures, rather than fewer.
Indi has been quite clear that the Xpresys test is intended only to rule out the possibility of cancer, with indeterminate or positive results not to be taken as indicating that a nodule is cancerous. "The problem is," West said, "that's not the way the world works."
"If you have somebody who has a nodule and the test comes back as anything other than negative, I think it is going to be human nature that the patient or doctors are going to potentially be far more likely to escalate the work-up and now do biopsies when they otherwise wouldn't have, or do scans more frequently than they would have," he said.