Which Lung Nodules Are Cancer? Risk Tools May Help

Nick Mulcahy

December 21, 2017

Clinicians "rarely" use available tools to quantify the cancer risk of a lung nodule and instead mostly rely on qualitative language, such as "suspicious," "very concerning," and "high pretest probability," according to a new retrospective study published online December 20 in JAMA Surgery.

At least three tools are available to make risk assessment more exact, say the authors, led by Amelia Maiga, MD, MPH, from Vanderbilt University Medical Center and Tennessee Valley Healthcare System in Nashville.

However, their analysis found that fewer than 5% of patients in two cohorts of 200-plus patients with lung conditions (known/suspected lung cancer and lung nodules) had a documented quantitative "prediction of malignancy" in their medical records.

Risk assessment that employs "imprecise" and "highly variable" qualitative, descriptive language was used for 71% of the remaining patients.

The authors called for use of a "standard scale that correlates with predicted risk" for the pulmonary nodules.

Dr Maiga told Medscape Medical News that in undertaking the study, she and her co-investigators "observed a lack of use of clinical prediction models" (which help quantify cancer risk). It's an "opportunity for training or in a decision support intervention," she said.

The study does not conclude that clinicians should abandon qualitative assessments, Dr Maiga explained, but rather that a quantitative prediction should also be included.

"Previous studies suggest that prediction models and clinician estimates are complementary — that is, have additional benefit when used together," she said.

There is a lot at stake here, say the study authors.

They explain that pulmonary growths incidentally detected by high-tech imaging of the chest are common in US medicine.

These nodules are often of "indeterminate" nature, with reported rates varying from 17% to 51% of all computed tomographic images and gross numbers of 1.5 million a year, at least.

Fortunately, only 1% to 12% of such nodules are malignant. But those numbers still represent a "significant diagnostic burden," the authors say.

The team argues out that because 88% to 99% of the nodules are benign, there needs to be a strong method for both determining and communicating (to patients and to other clinicians) the probability of cancer. In short, because not every nodule requires surgery, a precise assessment of risk is needed.

In 2007, the American College of Chest Physicians recommended that clinicians document the pretest probability of cancer for every patient with pulmonary nodules — even if only one nodule is present.

At least three tools — the Mayo Clinic Model, the Veterans Affairs pretest probability model, and the Thoracic Research Evaluation and Treatment  model — are available for that assessment.

The investigators also add some further context by highlighting the increase in lung resections for benign lung disease that has occurred in the United States since the onset of "pervasive use of video-assisted thoracoscopic surgery."

Study Details

For their study, Dr Maiga and colleagues reviewed patient records from the Tennessee Valley Healthcare Center in Nashville, which is part of the Veterans Affairs system. This study is the first to analyze the documentation habits of clinicians with regard to the probability of malignancy among patients with indeterminate pulmonary nodules, they note.

The team looked at two cohorts of patients, nearly all of whom (97%) were male. Cohort 1 included 291 veterans undergoing surgical resection of known or suspected lung cancer (from 2003 to 2015). Cohort 2 included 239 veterans undergoing evaluation of indeterminate pulmonary nodules (2003 to 2012).

Predictably, cancer prevalence was much higher in cohort 1 (88.7%) than in cohort 2 (48.9%).

The investigators comment that the relatively high percentage of nodules that turned out to be malignancies in cohort 2 probably reflects the fact that the growths were mostly high risk by virtue of having been referred to a surgery clinic.

Nevertheless, only 13 patients in cohort 1 (4.5%) had a documented quantitative prediction of malignancy before the eventual tissue diagnosis. In cohort 2, only 3 patients (1.3%) had such documentation.

Of the remaining patients, 78.1% in cohort 1 had a qualitative statement of cancer risk and 63.1% had the same in cohort 2.

The problem with the qualitative statements is that they were found to be "vague, wide ranging and not systematic" by the investigators. But the team also acknowledged that the relatively high number of such statements was, in fact, "an attempt by each clinician to document a high-risk IPN [indeterminate pulmonary nodule] that required biopsy for diagnosis."

The investigators believe that a first step to implement a program to improve evaluation of indeterminate pulmonary nodules that are suspicious for lung cancer would be to have a standardized cancer risk assessment link to evidence-based evaluation guidelines.

The study was supported in part by the Department of Veterans Affairs. The authors have disclosed no relevant financial relationships.

JAMA Surg. Published online December  20, 2017. Abstract

Follow Medscape senior journalist Nick Mulcahy on Twitter: @MulcahyNick

For more from Medscape Oncology, follow us on Twitter: @MedscapeOnc


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.