This article is more than 5 years old. Information may no longer be current.
Lung-RADS may reduce false-positives in lung cancer screening
The use of Lung Imaging Reporting and Data System criteria during interpretation of low-dose CT scan results may yield lower false-positive results than observed in the National Lung Screening Trial, according to results of a retrospective study.
However, the Lung Imaging Reporting and Data System (Lung-RADS) criteria — developed by the American College of Radiology — also were associated with reduced sensitivity.
“The findings suggest that the implementation of Lung-RADS can substantially reduce the false-positive rate in CT screening for lung cancer,” William C. Black, MD, study author and professor of radiology at the Dartmouth-Hitchcock Medical Center, told HemOnc Today. “However, the findings also suggest that the diagnosis of some lung cancers may be delayed with the implementation of Lung-RADS. Whether this delay will substantively reduce the effectiveness of CT screening is unknown and will have to be further studied.”
William C. Black
Black and colleagues retrospectively applied Lung-RADS criteria to findings of the National Lung Screening Trial (NLST).
The NLST included 26,455 individuals aged 55 to 74 years who were at high risk for lung cancer who received at least one low-dose CT screen. These participants received a total of 48,671 subsequent screens.
Lung-RADS criteria classify findings in one of four categories: category 1 (negative), category 2 (benign appearance), category 3 (probably benign) or category 4 (suspicious). All category 3 and category 4 findings are considered positive.
Lung-RADS criteria deem solid nodules at least 6 mm in size identified during a baseline screen as positive, whereas the NLST used a threshold of 4 mm. During subsequent screenings, Lung-RADS criteria deem nodules 4 mm in size or those that have grown by at least 1.5 mm as positive.
Baseline screenings in NLST yielded 26,090 results absent of cancer. Lung-RADS classified most of those samples as category 1 (56.2%) or category 2 (31%). However, 6.4% of baseline screening results absent of cancer were classified as category 3, and 5.4% were classified as category 4.
Subsequent screenings in NLST yielded 48,197 results absent of cancer. Lung-RADS criteria classified 52.2% of those as category 1 and 42.5% as category 2. A small proportion of results were classified as category 3 (1.2%) or category 4 (3.6%).
Using these data, researchers determined the false-positive rate on baseline screens was significantly lower with Lung-RADS criteria than in NLST (12.8% vs. 26.6%; P ˂ .001). The false-positive rate on subsequent screens was 5.3% (95% CI, 5.1-5.5) for Lung-RADS vs. 21.8% (95% CI, 21.4-22.2) in the NLST (P ˂ .001).
However, sensitivity with Lung-RADS was lower at baseline (84.9% vs. 93.5%; P ˂ .001) and on subsequent screenings (78.6% vs. 93.8%; P ˂ .001) than the NLST.
“The effect of delaying diagnosis of these Lung-RADS missed cancer cases is unknown, but it cannot be assumed that most are indolent and would not affect lung cancer mortality rates,” Black and colleagues wrote. “Although we have considered NLST screen-detected cancer cases that are negative on Lung-RADS as contributing to decreased sensitivity of Lung-RADS, it is unknown what proportion would have presented clinically within the next year as true-interval cancer cases or would have been screen detected with Lung-RADS on the next round of screening. Still, even delayed screen-detected diagnosis could adversely affect survival.” – by Alexandra Todak
For more information:
William C. Black, MD, can be reached at Dartmouth-Hitchcock Medical Center, One Medical Center Drive, Lebanon, NH 03756; email: william.c.black@dartmouth.edu.
Disclosure: The researchers report no relevant financial disclosures.
Perspective
Back to Top
The value of the NLST has been established: For every 1,000 patients screened, approximately three deaths would be averted with low-dose CT. One concern about adopting this paradigm is the presence of false-positive screening tests or lung nodules that prompt additional CT or PET scans — as well as invasive staging procedures — that can add significant cost as well some risk for complications for the patients. In the NLST, approximately one-third of screened patients had a false-positive result either at the first or subsequent scan.
The American College of Radiology has created a reporting system — the Lung-RADS score — that narrows the definition of a positive lung screen. A positive screen is not necessarily one that identifies cancer, but rather one that prompts additional tests. In the NLST, for example, a nodule that measured 4 mm or larger was considered a positive screen on the baseline scan. When using the Lung-RADS criteria, additional testing is only prompted for a solid nodule that measures 6 mm or larger, or for a ground-glass nodule that measures more than 2 cm.
In the current study, the authors retrospectively evaluated the scans obtained during the NLST by using the Lung-RADS criteria. They determined that the use of Lung-RADS significantly reduced the false-positive screening rate, eliminating 50% of unnecessary CT scans and almost 25% of unnecessary invasive procedures.
The sensitivity of Lung-RADS — or the ability to successfully detect cancer — remains high, although there were a small number of lung cancers detected in the NLST that would have been missed.
There have been a variety of screening guidelines proposed, with differences in the included patient populations as well as in various criteria for evaluating the scans obtained. The adoption of the Lung-RADS criteria will make low-dose CT lung cancer screening significantly more cost-effective, will reduce unnecessary CT and PET scans, and will reduce the unnecessary use of invasive procedures to evaluate patients who ultimately do not have cancer.
Perspective
Back to Top
Screening for lung cancer utilizing low-dose CT has only recently been recognized by governing panels and insurers, with the U.S. Preventive Services Task Force recommending screening (grade B) in December 2013, and CMS issuing final guidance on Feb. 5 that there was sufficient evidence for screening in a high-risk population.
The path was slow and filled with controversy dating back to trials that showed no mortality benefit from screening with chest radiographs and/or sputum sampling, to early studies by groups such as the Early Lung Cancer Action Program (ELCAP), which demonstrated benefit of low-dose CT vs. chest radiography for both increased and earlier detection (ie, a higher proportion of stage I diagnoses) of lung cancers.
The National Lung Screening Trial (NLST) further reported a 20% reduction in lung cancer-related deaths associated with low-dose CT screening in a defined high-risk population. One of the largest concerns of screening and of the NLST is its high false-positive rates (26.6% at baseline screening), which may lead to unnecessary testing, potential harms and increased health care costs.
The American College of Radiology — based on a multitude of studies demonstrating a need for refinement of the definition of a positive test — standardized reporting with its Lung-RADS classification, similar to the Breast Imaging Reporting and Data System used in mammography. In addition to increasing the nodule size threshold of a positive test, it was hoped that refinements in growth assessments and nodule density would decrease the false-positive rate without missing life-threatening cancers.
Although the study by Pinsky and colleagues was retrospective, a possible reduction of false-positives at baseline screening from 26.6% to 12.8% with the Lung-RADS algorithm underscores the importance of criteria refinement and signifies a turning point in low-dose CT screening research. Efforts are now focused on fine-tuning this preventive measure. A lesson learned from screening mammography is the importance of maintaining high-quality standards and strict guidelines for analysis and follow-up. Continued experience should allow improvements in imaging characterization through efforts in volumetric and densitometry measurements. Additional investigations in lung cancer risk modeling also may allow for incorporation of other possible risk factors outside of the current criteria, which consist only of age and smoking history.
References:
Aberle DR, et al. N Engl J Med. 2011;doi:10.1056/NEJMoa1102873.
Henschke CI, et al. N Engl J Med. 2006;355:1763-1771.