Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2019 Dec 13.
Published in final edited form as: J Thorac Oncol. 2019 May 22;14(9):1662–1665. doi: 10.1016/j.jtho.2019.05.012

Insights for Management of Ground-Glass Opacities From the National Lung Screening Trial

Hilary A Robbins a,*, Hormuzd A Katki b, Li C Cheung b, Rebecca Landy b, Christine D Berg b
PMCID: PMC6909540  NIHMSID: NIHMS1059280  PMID: 31125735

Abstract

Background

In the National Lung Screening Trial (NLST), screen-detected cancers that would not have been identified by the Lung Computed Tomographic Screening Reporting and Data System (Lung-RADS) nodule management guidelines were frequently ground-glass opacities (GGOs). Lung-RADS suggests that GGOs with diameter less than 20 mm return for annual screening, and GGOs greater than or equal to 20 mm receive 6-month follow-up. We examined whether this 20-mm threshold gives consistent management of GGOs compared with solid nodules.

Methods

First, we calculated diameter-specific malignancy probabilities for GGOs and solid nodules in the NLST. Using the solid-nodule malignancy risks as benchmarks, we suggested risk-based management categories for GGOs based on their probability of malignancy. Second, we compared lung-cancer mortality between GGOs and solid nodules in the same risk-based category.

Results

Using the Lung-RADS v1.0 classifications, malignancy probability is higher for GGOs than solid nodules within the same category. A risk-based classification of GGOs would assign annual screening for GGOs 4 to 5 mm (0.4% malignancy risk); 6-month follow-up for GGOs 6 to 7 mm (1.1%), 8 to 14 mm (3.0%), and 15 to 19 mm (5.2%); and 3-month follow-up for greater than or equal to 20 mm (10.9%). This reclassification would have assigned similarly fatal cancers to 3-month follow-up (hazard ratio = 2.0 for lung-cancer death in GGOs versus solid-nodule cancers, 95% confidence interval: 0.4–8.7), but for 6-month follow-up, mortality was lower in GGO cancers (hazard ratio = 0.18, 95% confidence interval: 0.05–0.67).

Conclusions

If Lung-RADS categories for GGOs were based on malignancy probability, then 6- to 19-mm GGOs would receive 6-month follow-up and greater than or equal to 20-mm GGOs would receive 3-month follow-up. Such risk-based management for GGOs could improve the sensitivity of Lung-RADS, especially for large GGO cancers. However, small GGO cancers were less aggressive than their solid-nodule counterparts.

Keywords: Computed tomographic screening, Lung cancer screening, Malignancy probability, Nodule classification, Lung Computed Tomographic Screening Reporting and Data System, Ground-glass opacities, Risk-based screening

Introduction

The National Lung Screening Trial (NLST) showed in 2011 that three annual screens by low-dose computed tomography (CT) can reduce lung cancer mortality by 20%.1 However, this came at the cost of many false-positive results, which represented 24% of all screens and 95% of all positive screens. To reduce the false-positive rate, the 2014 Lung Computed Tomographic Screening Reporting and Data System (Lung-RADS) v1.0 nodule classification system and the National Comprehensive Cancer Network version 2.2018 guidelines require larger nodule diameters to recommend additional surveillance beyond usual screening.2,3 Specifically, for solid nodules detected at baseline, Lung-RADS raised the threshold for increased surveillance from 4 mm in longest diameter (used for all nodules in the NLST) to 6 mm in average diameter.1,2

For pure ground-glass opacities (GGOs), Lung-RADS v1.0 requires an average diameter of 20 mm or more to recommend 6-month follow-up (Lung-RADS-3) and does not recommend any GGOs for 3-month or immediate follow-up (Lung-RADS-4). It is appropriate that GGOs have a higher threshold for intervention because they are less likely to be malignant compared with a solid nodule of the same diameter.4 However, application of Lung-RADS criteria to the NLST revealed that the screen-detected cancers not identified by Lung-RADS were often GGOs with diameters less than 20 mm.5 Further, the Lung-RADS protocol was shown to be inferior to PanCan guidelines, which use a continuous calculation for malignancy probability.6 After adjusting for diameter, the PanCan model adjusts risk for GGOs downward, but only by a small amount.7 Taken together, these findings prompt the question of whether the 20-mm threshold for GGOs may be too high.

There are two key factors that govern how aggressively a nodule should be followed. The first is the risk that the nodule is malignant, and the second is the risk it will cause death if it is malignant. We analyzed data from the NLST to calculate these risks for pure GGOs. First, we calculated diameter-specific malignancy probabilities for GGOs and compared them to those for solid nodules, using risks for solid nodules as benchmarks to suggest risk-based Lung-RADS categories for GGOs. Second, we investigated whether lung cancer mortality was lower for GGOs compared to solid nodules in the same risk-based category.

Methods

The U.S. NLST randomized 53,452 heavy smokers to three screens with either chest radiography or low-dose CT.1 In the CT arm, we calculated the immediate probability of malignancy for solid nodules and pure GGOs. For solid nodules, we analyzed 4887 individuals who had a solid nodule at baseline, but no part-solid nodules or GGOs. If multiple solid nodules were present, we classified diameter using the largest nodule. We similarly analyzed 1455 individuals who had a baseline or newly detected GGO, but no solid or part-solid nodules.

We classified GGOs and solid nodules into the Lung-RADS diameter categories used for baseline solid nodules (4 to 5 mm, 6 to 7 mm, 8 to 14 mm, and 15 to 19 mm), along with an additional greater than or equal to 20-mm category to reflect the current classification for GGOs.2 We calculated average diameter by averaging the longest and perpendicular diameter then rounding to the nearest integer.2,5 Within each category, we calculated immediate risk of malignancy using the NLST linked-year method.8 Using the diameter-specific risks for baseline solid nodules as benchmarks, we suggested risk-based management strategies for GGOs in each diameter category based on their probability of malignancy.9

Finally, we investigated whether such a risk-based assignment of management strategies might group together cancers from GGOs and solid nodules that do not have the same likelihood of causing death from lung cancer. Specifically, for each risk-based management category, we fit a Cox proportional hazards model to estimate the hazard ratio comparing lung cancer mortality between GGOs and solid nodules.

Results

Among 4887 individuals with solid nodules, nodules 4 to 5 mm in diameter were most common (45.9%) and nodules greater than or equal to 20 mm in diameter were least common (3.0%) (Table 1). The same pattern was present among 1455 individuals with GGOs (35.6% and 3.2%, respectively). Malignancy probability for 4- to 5-mm nodules was equally small for solid nodules (0.4%, 95% confidence interval [CI]: 0.2%–0.7%) and GGOs (0.4%, 95% CI: 0.0% −1.4%), which are both Lung-RADS-2 (continue annual screening).

Table 1.

Malignancy Probability by Average Diameter for Baseline Solid Nodules Compared to Baseline or New GGOs in the NLST

Solid Nodules at Baseline
 GGOs, Baseline or New
Average Diameter,mm n Malignanc Probability (95% CI), % Lung-RADS v1.0 Categorya n Malignancy Probability (95% CI), % Lung-RADS v1.0 Categorya Risk-Based Categoryb
4–5 2242 0.4 (0.2–0.7) 2 518 0.4 (0.0–1.4) 2 2
6–7 1442 1.0 (0.6–1.7) 3 369 1.1 (0.3–2.8) 2 3
8–14 924 7.1 (5.6–9.0) 4A 464 3.0 (1.7–5.0) 2 3
15–19 133 26.3 (19.1–34.7) 4B 58 5.2 (1.1–14.4) 2 3
≥20 146 46.6 (38.3–55.0) 4B 46 10.9 (3.6–23.6) 3 4A
Open in a new tab

We analyzed data from the NLST CT arm that mapped the NLST attenuation categories of soft tissue, mixed, and ground-glass to solid, part-solid, and groundglass, respectively.5 Among individuals with a solid nodule, the largest solid nodule is used for classification, and individuals with a GGO or part-solid nodule are excluded. Among individuals with a GGO, the largest GGO is used for classification, and individuals with a solid or part-solid nodule are excluded. For solid nodules, only those at baseline are considered, whereas for GGOs both baseline and new nodules are considered, to reflect current Lung-RADS classifications.2 Average diameter was calculated as the mean of the longest and perpendicular diameter5 then rounded to the nearest integer and categorized.2

a

Lung-RADS categories imply the following management: 2: continue annual screening with low-dose CT in 12 months; 3: 6-month follow-up with low-dose CT; 4A: 3-month follow-up with low-dose CT; PET/CT may be used if ≥8-mm solid component is present; and 4B: immediate chest CTwith or without contrast, PET/CT, and/or tissue sampling.

b

The Lung-RADS category that would be assigned to GGOs if they were managed according to the same risk of malignancy as are solid nodules. For example, solid nodules of 8 to 14 mm have 7.1% malignancy risk and are managed as Lung-RADS-4A. Because GGOs ≥20 mm have no less risk (10.9%), they would also be managed as Lung-RADS-4A.

NLST, National Lung Screening Trial; Lung-RADS, Lung Computed Tomographic Screening Reporting and Data System; CT, computed tomography; PET, positron-emission tomography.

In contrast, malignancy probabilities did not correspond with Lung-RADS categories for larger nodules (Table 1). For example, for 6- to 7-mm nodules, malignancy probability was similar between GGOs (1.1%, 95% CI: 0.3%–2.8%) and solid nodules (1.0%, 95% CI: 0.6%–1.7%). However, 6- to 7-mm GGOs are recommended usual annual screening (Lung-RADS-2) whereas 6- to 7-mm solid nodules receive 6-month follow-up (Lung-RADS-3). Similar inconsistencies were present for larger GGOs.

If management for GGOs were assigned based on the benchmark malignancy probabilities from solid nodules, then GGOs 6 to 7 mm would be classified as Lung-RADS-3 because their malignancy probability (1.1%) is similar to that for solid nodules in Lung-RADS-3 (1.0%) (Table 1, risk-based categories). Using similar logic, 8- to 14-mm GGOs would be assigned Lung-RADS-3, as 3.0% exceeds 1.0%. GGOs of 15 to 19 mm could be assigned to Lung-RADS-3 or 4A, as 5.2% exceeds 1.0% and approaches 7.1%; we chose to consider these as Lung-RADS-3. Finally, GGOs greater than or equal to 20 mm would be assigned Lung-RADS-4A, as 10.9% exceeds 7.1%. No GGOs had sufficiently high malignancy probability (solid-nodule benchmark = 26.3%) to be assigned to Lung-RADS-4B (immediate follow-up).

Importantly, even when GGOs and solid nodules have similar malignancy probability, less aggressive follow-up may be warranted if lung cancers arising from GGOs are less fatal. Therefore, we compared lung cancer–specific mortality between lung cancer cases in GGOs versus solid nodules within each risk-based and current Lung-RADS category. For risk-based Lung-RADS-2 (continue usual screening), there were insufficient data (10 cases) to calculate a hazard ratio (HR) comparison. For risk-based Lung-RADS-3 (6- to 7-mm solid nodules and 6- to 19-mm GGOs, 6-month follow-up) mortality was substantially lower for GGO compared with solid-nodule cancers (HR = 0.18, 95% CI: 0.05–0.67, n = 36 cases). In contrast, for risk-based Lung-RADS-4A (8- to 14-mm solid nodules and greater than or equal to 20-mm GGOs, 3-month follow-up), mortality was nonsignificantly higher for GGOs versus solid nodules (HR = 2.0, 95% CI: 0.4–8.7, n = 71 cases). Notably, diameter did not predict survival after diagnosis of GGO cancers (p= 0.33); therefore, even in current Lung-RADS-2 (6- to 7-mm solid nodules and 4- to 19-mm GGOs, continue usual screening), mortality was lower among the GGO versus solid-nodule cancers (HR = 0.22, 95% CI: 0.05–0.998).

Discussion

Our data highlight important considerations for managing pure GGOs detected during lung CT screening. In the NLST, Lung-RADS v1.0 guidelines do not assign similar management for nodules with similar malignancy probability, as cancer risk is higher for GGOs compared with solid nodules within the same Lung-RADS category. If management strategies for GGOs were reassigned based on malignancy probability, then GGOs of 6 to 19 mm would receive 6-month follow-up and GGOs greater than or equal to 20 mm would receive 3-month follow-up. However, for 6- to 19-mm GGOs, this risk-based reassignment may group more aggressive solid-nodule cancers together with less aggressive GGO cancers.

There are multiple ways that nodule management guidelines could consider accounting for our findings. One approach would be to reassign GGOs based on our risk-based categories; that is, 6-month follow-up for 6- to 19-mm GGOs and 3-month follow-up for greater than or equal to 20-mm GGOs. For GGOs greater than or equal to 20 mm, lethality was similar to the 8- to 14-mm solid nodules in Lung-RADS-4A. Together with the similar probability of malignancy, this supports more aggressive management for GGOs greater than or equal to 20 mm. In contrast, the risk-based reassignment we suggest would require caution for GGOs of 6 to 19 mm, because when malignant they may be less aggressive than their solid-nodule counterparts.

Alternatively, reassignment of GGOs could attempt to account for survival differences by downgrading smaller GGOs (6 to 7 mm) from our risk-based Lung-RADS-3 back to Lung-RADS-2. However, because diameter did not predict survival among malignant GGOs, this approach would still not result in similar survival among GGOs and solid nodules in the same management category.

Other studies support that GGOs may have overlooked malignancy potential, except for those less than or equal to 5 mm.10,11 With regard to subsequent cancer survival, one study concluded that malignant GGOs had longer time to diagnosis than other cancers, along with lower mortality.12 Stable (non-growing) GGOs, which we did not specifically analyze, may be more likely than stable solid nodules to eventually grow and be found malignant.13

Because of the lower likelihood of diagnostic follow-up for GGOs compared to solid nodules (i.e., ascertainment bias), we note that our malignancy probabilities for GGOs may be underestimates. In the NLST, among 4- to 5-mm nodules, GGOs were less likely than solid nodules to receive follow-up other than usual screening (77% versus 87%, p< 0.001). The same was true for 6- to 7-mm (77% versus 89%, p< 0.001) and 8- to 14-mm nodules (84% versus 93%, p< 0.001), although not larger nodules (15 to 19 mm: 90% versus 89%, p = 0.99; greater than or equal to 20 mm: 85% versus 93%, p = 0.15).

NLST diagnostic follow-up was not recorded in a nodule-specific way, so we cannot confirm that cancer diagnosis in an individual with a GGO was due to that particular GGO. Therefore, to prevent confounding, we restricted to GGOs occurring without a solid/part-solid nodule, and similarly for solid nodules. This restriction may limit generalizability to individuals with multiple types of nodules. Because Lung-RADS currently classifies nodules by average diameter, we did not examine the impact of volumetric assessment, which may more accurately capture nodule size and growth.14,15 Finally, our analysis makes an implicit assumption that Lung-RADS v1.0 recommendations for baseline solid nodules represent optimal management. In the future, if Lung-RADS classifications for baseline solid nodules change, then our diameter-specific assessment of GGOs would need to be repeated.

Lung-RADS is a critical step to reduce false-positive findings in screening, but at the cost of reduced lung cancer detection. Our results suggest that larger (greater than or equal to 20 mm) ground-glass opacities should be followed more closely than Lung-RADS recommends. For smaller GGOs, more research is needed to understand the joint context of malignancy probability and lethality. Refinement of management for GGOs could improve the ability of Lung-RADS guidelines to reduce lung cancer deaths.

Acknowledgments

This study was supported in part by the Intramural Research Program of the U.S. National Cancer Institute, National Institutes of Health. Dr. Robbins was supported by a National Cancer Institute Ruth L. Kirschstein individual predoctoral fellowship (F31CA210660) and by the INTEGRAL program (National Cancer Institute U19 CA203654). The authors thank the participants from the National Lung Screening Trial.

Footnotes

Disclosure: Dr. Berg has received personal fees from Grail, Inc., and Medial EarlySign, LLC. The remaining authors declare no conflict of interest.

References

  • 1.National Lung Screening Trial Research Team, Aberle DR, Adams AM, et al. Reduced lung-cancer mortality with low-dose computed tomographic screening. N Engl J Med 2011;365:395–409. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.American College of Radiology. Lung CT Screening Reporting and Data System (Lung-RADS). https://www.acr.org/Quality-Safety/Resources/LungRADS. Accessed May 18, 2017.
  • 3.Wood DE, Kazerooni EA, Baum SL, et al. Lung cancer screening, version 3.2018, NCCN clinical practice guidelines in oncology. J Natl Compr Canc Netw 2018;16(4):214–441. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Martin MD, Kanne JP, Broderick LS, Kazerooni EA, Meyer CA. Lung-RADS: pushing the limits. Radiographics. 37: 1975–1993. [DOI] [PubMed] [Google Scholar]
  • 5.Pinsky PF, Gierada DS, Black W, et al. Performance of Lung-RADS in the National Lung Screening Trial: a retrospective assessment. Ann Intern Med 2015;162:485–491. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.van Riel SJ, Ciompi F, Jacobs C, et al. Malignancy risk estimation of screen-detected nodules at baseline CT: comparison of the PanCan model, Lung-RADS and NCCN guidelines. Eur Radiol 2017;27:4019–4029. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.McWilliams A, Tammemagi MC, Mayo JR, et al. Probability of cancer in pulmonary nodules detected on first screening CT. N Engl J Med 2013;369, 910–909. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Kovalchik SA, Tammemagi M, Berg CD, et al. Targeting of low-dose CT screening according to the risk of lung-cancer death. N Engl J Med 2013;369:245–254. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Katki HA, Schiffman M, Castle PE, et al. Benchmarking CIN3þ risk as the basis for incorporating HPV and Pap cotesting into cervical screening and management guidelines. J Low Genit Tract Dis 2013;17(5 suppl 1):S28–S35. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Ichinose J, Kohno T, Fujimori S, Harano T, Suzuki S, Fujii T. Invasiveness and malignant potential of pulmonary lesions presenting as pure ground-glass opacities. Ann Thorac Cardiovasc Surg 2014;20:347–352. [DOI] [PubMed] [Google Scholar]
  • 11.Kakinuma R, Muramatsu Y, Kusumoto M, et al. Solitary pure ground-glass nodules 5 mm or smaller: Frequency of growth. Radiology. 2015;276:873–882. [DOI] [PubMed] [Google Scholar]
  • 12.Sonavane SK, Pinsky P, Watts J, et al. The relationship of cancer characteristics and patient outcome with time to lung cancer diagnosis after an abnormal screening CT. Eur Radiol 2017;27:5113–5118. [DOI] [PubMed] [Google Scholar]
  • 13.Shin KE, Lee KS, Yi CA, Chung MJ, Shin M-H, Choi Y-H. Subcentimeter lung nodules stable for 2 years at LDCT: long-term follow-up using volumetry. Respirology. 2014;19:921–928. [DOI] [PubMed] [Google Scholar]
  • 14.Heuvelmans MA, Walter JE, Vliegenthart R, et al. Disagreement of diameter and volume measurements for pulmonary nodule size estimation in CT lung cancer screening. Thorax 2018;73:779–781. [DOI] [PubMed] [Google Scholar]
  • 15.Han D, Heuvelmans MA, Vliegenthart R, et al. Influence of lung nodule margin on volume- and diameter-based reader variability in CT lung cancer screening. Br J Radiol 2018;91:20170405. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES