NEW YORK (Reuters Health) – The malignancy rates of Lung-RADS category-2 and category-3 subsolid lung nodules are much higher than reported, according to an analysis of nodules from the National Lung Screening Trial (NLST).
“We were surprised by two related findings,” Dr. Mark M. Hammer from Brigham and Women’s Hospital, Harvard Medical School, in Boston, told Reuters Health by email. “The first is that about 1/3 of the subsolid nodules in the NLST data were mischaracterized – they were actually scars or benign intrapulmonary lymph nodes.”
“This artificially lowered what people thought was the malignancy risk of the subsolid nodules,” he said. “The consequence of this is that the true malignancy risk of these nodules is higher than people expected, in some cases much higher – which was the second surprising finding.”
The Lung Imaging Reporting and Data System (Lung-RADS) was developed by the American College of Radiology as a framework to follow up and manage lung nodules. Category 2 includes benign-appearing nodules with “less than 1%” chance of malignancy, and category 3 includes “probably benign” nodules with “1-2%” chance of malignancy.
Dr. Hammer’s team used 304 category-2 nodules and 67 category 3-nodules from NLST to evaluate Lung-RADS estimates of malignancy rates.
Nodules classified as category 2 at baseline were associated with a 3% risk of malignancy, which is significantly greater than the maximum of 1% listed in the Lung-RADS document, the team reports in Radiology, online September 17.
Lung-RADS category-3 lesions were associated with a 14% risk of malignancy, significantly greater than the maximum of 2% listed in the Lung-RADS document.
Among category-2 nodules, the malignancy risk in nodules smaller than 10 mm was only 1.3%, not significantly different from the upper limit of 1%, but the malignancy risk in larger nodules (6% for 10-19 mm nodules and 9% for 20-29 mm nodules) was significantly higher than the Lung-RADS risk.
Nodules classified as category 2 at the first follow-up CT were associated with a 4% risk of malignancy, and those classified as category 3 at follow-up were associated with a 7% risk of malignancy.
Using a volumetric classification system (NELSON) performed no better than the Lung-RADS linear classification scheme at discriminating between benign and malignant lesions. But the Brock University model, which includes patient and nodule characteristics, had significantly higher discriminatory ability than either Lung-RADS or NELSON.
Dr. Hammer said, “There are two important implications: (1) that the risk of cancer developing in a subsolid nodule is quite high, and (2) that risk calculators that incorporate patient demographics, such as smoking history, as well as nodule size, are potentially helpful in triaging which patients should have surgical treatment. (Many existing guidelines mostly use nodule size.)”
“It is important to understand that the cancers that arise in subsolid nodules are indolent,” he said. “In other words, they grow slowly and have a low risk of metastasizing. Thus, many patients with such cancers may actually die of other causes (e.g., smoking-related emphysema or cardiac disease) rather than of the cancers themselves.”
“This raises important questions about which patients will actually benefit from surgical treatment,” Dr. Hammer said. “We have not addressed that question in our article, but it is an important one and will be the subject of future research.”