You are on page 1of 19

This copy is for personal use only. To order printed copies, contact reprints@rsna.

org
1975

CHEST IMAGING
Lung-RADS: Pushing the Limits1

Maria D. Martin, MD
Jeffrey P. Kanne, MD In response to the recommendation of the U.S. Preventive Services
Lynn S. Broderick, MD Task Force and the coverage decision by the Centers for Medicare
Ella A. Kazerooni, MD, MS and Medicaid Services for lung cancer screening (LCS) computed
Cristopher A. Meyer, MD tomography (CT), the American College of Radiology introduced
the Lung CT Screening Reporting and Data System (Lung-RADS)
Abbreviations: ACR = American College of in 2014 to standardize the reporting and management of screening-
Radiology, BI-RADS = Breast Imaging Report- detected lung nodules. As with many first-edition guidelines, ques-
ing and Data System, LCS = lung cancer screen-
ing, Lung-RADS = Lung CT Screening Report- tions arise when such reporting systems are used in daily practice.
ing and Data System In this article, a collection of 15 LCS-related scenarios are present-
RadioGraphics 2017; 37:00000000 ed that address situations in which the Lung-RADS guidelines are
unclear or situations that are not currently addressed in the Lung-
https://doi.org/10.1148/rg.2017170051
RADS guidelines. For these 15 scenarios, the authors of this article
Content Codes:
provide the reader with recommendations that are based on their
1
From the Department of Radiology, University collective experiences, with the hope that future versions of Lung-
of Wisconsin School of Medicine, 600 High-
land Ave, Madison, WI 53792-3252 (M.D.M., RADS will provide additional guidance, particularly as more data
J.P.K., L.S.B., C.A.M.); and Department of from widespread LCS are collected and analyzed.
Radiology, University of Michigan Health Sys-
tem, Ann Arbor, Mich (E.A.K.). Recipient of
RSNA, 2017 radiographics.rsna.org
a Certificate of Merit award for an education
exhibit at the 2016 RSNA Annual Meeting. Re-
ceived March 11, 2017; revision requested May
18 and received June 19; accepted July 28. For
this journal-based SA-CME activity, the author
J.P.K. has provided disclosures (see end of ar-
Introduction
ticle); all other authors, the editor, and the re- For lung cancer screening (LCS) computed tomography (CT), the
viewers have disclosed no relevant relationships. American College of Radiology (ACR) developed the Lung CT
Address correspondence to M.D.M. (e-mail:
mmartin3@uwhealth.org). Screening Reporting and Data System (Lung-RADS) in 2014, which
was modeled on the success of the Breast Imaging Reporting and
RSNA, 2017
Data System (BI-RADS). Lung-RADS is a tool that facilitates stan-
dardized reporting and management of abnormal findings at LCS
SA-CME LEARNING OBJECTIVES
CT (1,2). The primary goal of Lung-RADS is to minimize variation
After completing this journal-based SA-CME in the management of LCS CTdetected lung nodules so that LCS
activity, participants will be able to:
can be implemented effectively in radiology practices outside the
Describe LCS scenarios in which
Lung-RADS guidance is unclear or sce-
purview of a clinical trial. Lung-RADS is also an important com-
narios that are not currently addressed in ponent of LCS CT quality assurance and registry reporting. Data
Lung-RADS. collected from LCS programs help with practice audits and facilitate
Explain how volume doubling time can outcome monitoring and research on an institutional level and a
help guide management of a nodule that national level. Further study of the effectiveness of LCS can lead to
was depicted at imaging performed be-
the optimization of screening criteria and to ideal management of
fore baseline LCS CT.
screening-detected lung nodules and can guide the development of
Recognize that growth of a nodule can
include an increase in diameter and an future versions of Lung-RADS.
increase in density.
See www.rsna.org/education/search/RG.
1976 November-December 2017 radiographics.rsna.org

Scenario 5: Ground-glass nodule with slow


TEACHING POINTS growth rate.
For some radiologists and referring providers, confusion may Scenario 6: How to measure and classify a
arise as to when to use the Lung-RADS guidelines and when to
use the Fleischner Society guidelines for the follow-up of lung
part-solid nodule.
nodules. Table 3 summarizes the main differences between Scenario 7: Nodule that decreases in size but
these two sets of guidelines. Lung-RADS was developed spe- increases in attenuation.
cifically for use in LCS CT reporting. The recently published Scenario 8: Nodule with characteristic features
updated guidelines of the Fleischner Society for the manage- of an intrapulmonary lymph node.
ment of incidentally detected lung nodules specifically state
that the updated guidelines do not apply to LCS.
Scenario 9: Airway (endotracheal or endo-
bronchial) nodule.
Referring providers and radiologists may be confused about
whether to perform a diagnostic CT examination or a LCS Scenario 10: Incidental potentially important
CT examination when a clinical suspicion for lung cancer ex- finding other than lung cancer detected at low-
ists. One of the exclusion criteria for LCS CT (or any screening dose LCS CT.
examination) is signs or symptoms that can be attributed to Scenario 11: Reenrolling patients in the LCS
the disease for which screening is being performed. Thus, for
CT program after a stable abnormality.
patients in whom lung cancer is suspected, LCS CT should
not be performed, and patients should instead undergo a di- Scenario 12: Low-dose LCS CT of a patient
agnostic chest CT examination. with a recent respiratory infection.
Volume doubling time may be used to predict solid pulmo- Scenario 13: Categorization of a cavitary lung
nary nodule behavior as benign or malignant. Generally, the nodule or nodules.
observed volume doubling times of lung cancers range from Scenario 14: Low-dose LCS CT of a patient
20 days to 360400 days. For solid pulmonary nodules, vol- with a history of a treated lung malignancy.
ume doubling times of less than 2030 days are usually associ-
ated with benign causescommonly, inflammatory or infec-
Scenario 15: Low-dose LCS CT of a patient
tious causes. Volume doubling times greater than 400480 with a treated low-risk nonlung malignancy.
days are also usually associated with benign nodules, with
only 3%4% of nonsmall cell lung carcinomas reported to Scenario 1: New Lung-RADS Category
have volume doubling times of more than 400 days. 3 (Probably Benign) Solid Lung
Persistent subsolid nodules are more likely to be malignant Nodule in a Patient Who Is Aging out
than other nodule types, and the risk of malignancy increases
with the size of the solid component.
of the Screening Program
An 80-year-old man, who was about to turn 81
A decrease in nodule size is not always benign, particularly
with an increase in attenuation, and continued surveillance
years old in 3 months, underwent a final annual
may be appropriate. LCS CT examination, and a new 5-mm solid
nodule was detected. On the basis of Lung-
RADS, this nodule would be classified as a Lung-
RADS category 3 (probably benign) finding
As is often the case with the first iteration of (Table 1) (2). How should this case be managed?
a set of guidelines, questions and controversies For patients with a new 5-mm solid nodule,
arise once those guidelines are implemented in Lung-RADS recommends a follow-up low-dose
daily practice, affecting radiologists, referring CT examination in 6 months. This patient will be
providers, and patients. Uncertainties primar- 81 years old at that time and will no longer satisfy
ily arise from differing interpretations of the the eligibility criteria for LCS.
Lung-RADS document and from scenarios that Many organizations advocate LCS, and they
are not currently addressed or are incompletely have differing recommendations for who should
addressed in the Lung-RADS guidelines. In this be screened (Table 2) (35). Currently, the two
article, 15 clinical vignettes are used to highlight organizations that dictate reimbursement for
a variety of ambiguous scenarios, and recommen- LCS, the Centers for Medicare and Medicaid
dations are provided for managing these situa- Services and the U.S. Preventive Services Task
tions that are based on our collective experiences. Force, propose 77 years and 80 years, respec-
tively, as the upper age limit for LCS (3,4).
Fifteen Clinical Scenarios The patient in this scenario will no longer
Scenario 1: New Lung-RADS category 3 qualify for LCS because of his age. However,
(probably benign) solid lung nodule in a patient managing actionable findings in individuals who
who is aging out of the screening program. no longer qualify for nodule management under
Scenario 2: Lung mass in a patient with vague the purview of the LCS program requires guid-
symptoms. ance. Because this CT examination was per-
Scenario 3: Solid suspicious (Lung-RADS cat- formed as part of LCS, a Lung-RADs category
egory 4B) nodule with very slow growth rate. 3 should be assigned, and a 6-month follow-up
Scenario 4: Ground-glass nodule that in- diagnostic low-dose CT examination should
creases in density but remains stable in size. be recommended, because one can presume that
RG Volume 37 Number 7 Martin et al 1977

Table 1: Lung-RADS Categories 3, 4A, and 4B

Category Findings Management


3: Probably benign Solid nodule 6-month low-dose CT
6 mm to <8 mm at baseline
New, 4 mm to <6 mm
Part-solid nodule
6-mm total diameter with solid component <6 mm
New, <6-mm total diameter
Ground-glass nodule
20 mm at baseline
New
4A: Suspicious Solid nodule 3-month low-dose CT;
8 mm to <15 mm at baseline PET/CT (can be used
Growing <8 mm when there is a 8-mm
solid component)
New, 6 mm to <8 mm
Part-solid nodule
6 mm with solid component 6 mm to <8 mm
New or growing <4-mm solid component
Endobronchial nodule
4B: Suspicious Solid nodule Chest CT, with or without
15 mm contrast material admin-
New or growing, and 8 mm istration; PET/CT (can
be used when there is
Part-solid nodule
a 8-mm solid compo-
Solid component 8 mm nent); tissue sampling*
New or growing 4-mm solid component
Note.Adapted and reprinted under a CC BY 4.0 international license from the ACR (2). PET/CT = dual-
modality imaging with positron emission tomography (PET) and CT.
*Depending on the probability of malignancy and comorbidities.

the patient lacks important comorbidities on the more, the Fleischner Society guidelines do not
basis of his continued enrollment in LCS. At that apply to new or enlarging lung nodules.
time, any additional follow-up can be performed, Although both sets of guidelines were devel-
if appropriate, outside the LCS program. In this oped for different scenarios and although none
scenario, management of this nodule becomes of the guidelines apply to patients aging out of
complicated at the 6-month follow-up exami- LCS, for the patient in this vignette, we suggest
nation because no specific guidelines exist for continuing to follow up the nodule outside the
managing screening-detected nodules in patients screening program with use of the time frames
who no longer qualify for LCS. suggested in Lung-RADS, because it is currently
For some radiologists and referring providers, the only practice standard that addresses new or
confusion may arise as to when to use the Lung- growing nodules, provided the patient is healthy
RADS guidelines and when to use the Fleischner enough to undergo treatment if the nodule proves
Society guidelines for the follow-up of lung to be actionable.
nodules. Table 3 (2,68) summarizes the main If the nodule is stable in this particular case,
differences between these two sets of guidelines. one could report that the stability of the nodule is
Lung-RADS was developed specifically for use in suggestive of benign etiology or behavior (on the
LCS CT reporting. The recently published up- basis of the Lung-RADS description of category
dated guidelines of the Fleischner Society for the 2, because stable category 3 nodules fall back
management of incidentally detected lung nodules into category 2). However, it would be inappro-
specifically state that the updated guidelines do priate to assign a Lung-RADS category or to rec-
not apply to LCS (6). Distinguishing incidentally ommend follow-up LCS CT, because the patient
detected lung nodules found at examinations per- is no longer eligible for LCS. There is no correct
formed for reasons other than LCS from nodules answer for when to stop monitoring this nodule,
detected in the setting of LCS is key. Further- because there are no guidelines, but it is certainly
1978 November-December 2017 radiographics.rsna.org

Table 2: Current Eligibility Criteria for LCS

Smoking History
Organization Age Criteria (y) Criteria (pack-years) Other Eligibility Criteria
ACS, ACCP, ALA, 55 to 74 30; quit smoking
ASCO, and ATS <15 y ago
Centers for Medicare 55 to 77 30; quit smoking
and Medicaid Services <15 y ago
U.S. Preventive Services 55 to 80 30; quit smoking
Task Force <15 y ago
National Comprehensive 55 to 74 30; quit smoking
Cancer Network <15 y ago
50 20 With one more risk factor (other
than secondhand smoke), includ-
ing radon or occupational expo-
sure, cancer history, history of
lung cancer in a first-degree rela-
tive, or disease history (COPD or
pulmonary fibrosis)
American Association 55 to 79 30
for Thoracic Surgery
Lung cancer survivors
starting 5 y after treat-
ment until age 79
50 20 5% risk of developing a lung cancer
in the next 5 y (COPD, environ-
mental/occupational exposure,
prior cancer/radiation therapy,
genetics, or a family history)
Sources.References 35.
Note.ACS = American Cancer Society, ACCP = American College of Chest Physicians, ALA = Ameri-
can Lung Association, ASCO = American Society of Clinical Oncology, ATS = American Thoracic Society,
COPD= chronic obstructive pulmonary disease.

Table 3: Comparison between Lung-RADS Guidelines and Fleischner Society


Guidelines for the Management of Pulmonary Nodules

Lung-RADS Guidelines Fleischner Society Guidelines


Single version published in 2014 (2) (ad- Updated version published in 2017 (6)
dresses solid and subsolid nodules) (addresses solid and subsolid nodules)
Older versions published in 2005 for solid
nodules (7) and in 2013 for subsolid
nodules (8)
Developed for the management of nodules Developed for the management of inciden-
in the setting of LCS CT tally detected nodules
Includes management of nodules that are Does not address how to manage nodules
new or growing that are new or growing
Applies to patients older than 55 years of Applies to patients older than 35 years of
age (current lower limit for LCS) and age, with no upper age limit
up to 80 years of age (upper age limit
according to the U.S. Preventive Services
Task Force)
Applies to all patients undergoing LCS CT Does not apply to immunosuppressed
patients or those with a history of ma-
lignancy
Note.Numbers in parentheses are reference citations.
RG Volume 37 Number 7 Martin et al 1979

Figure 1. Squamous cell carcinoma de-


tected at baseline LCS CT in a 63-year-old
woman. Axial unenhanced CT image of the
chest (lung window settings) shows a right
upper lobe irregular mass (arrow) extend-
ing into the mediastinum. Careful review of
the electronic health record after the LCS
CT examination disclosed that the patient
had complained of new-onset cervicalgia
(neck pain).

controversial, given the patients age. Discussing stage disease, and patients in whom limited-stage
this with the ordering provider and weighing the small cell lung carcinoma was diagnosed at an
benefits and risks may help with this decision. annual LCS CT examination were more likely to
be asymptomatic (9).
Scenario 2: Lung Mass in a Patient
with Vague Symptoms Scenario 3: Solid Suspicious (Lung-
A 63-year-old woman had a 39-mm mass de- RADS Category 4B) Nodule with Very
tected at baseline LCS CT (Fig 1). A retrospec- Slow Growth Rate
tive review of the electronic medical record At baseline LCS CT, a 74-year-old woman had
disclosed that the patient had complained of a 19-mm solid nodule, which had measured 11
new neck pain. What signs and symptoms should mm at an abdominal CT examination performed
preclude LCS CT and prompt diagnostic chest 12 years earlier (Fig 2). On the basis of Lung-
CT evaluation for lung cancer outside an LCS RADS, this screening-detected nodule should
CT program? be classified as category 4B (suspicious) because
Symptoms of lung cancer are vague and can of the increase in size (Table 1) (2). However,
overlap with other smoking-related comorbidities considering the time interval of 12 years between
such as chronic obstructive pulmonary disease, the earlier CT image (comparison image) and
cardiovascular disease, and head and neck can- the LCS CT image, how should this nodule be
cers. Referring providers and radiologists may be managed?
confused about whether to perform a diagnostic One situation not addressed in the current ver-
CT examination or an LCS CT examination sion of Lung-RADS is categorizing and manag-
when a clinical suspicion for lung cancer exists. ing a nodule that was present on examinations
One of the exclusion criteria for LCS CT (or any performed before the baseline LCS CT examina-
screening examination) is signs or symptoms that tion. One option would be to manage this nodule
can be attributed to the disease for which screen- as a nodule depicted at a follow-up LCS CT
ing is being performed. Thus, for patients in examination, but the definition of growth (at least
whom lung cancer is suspected, LCS CT should a 1.5-mm increase in diameter) complicates mat-
not be performed, and patients should instead ters because of the 12-year interval between the
undergo a diagnostic chest CT examination. two CT examinations in this patient. If we strictly
Signs and symptoms that should raise suspicion apply the Lung-RADS definition of growth, this
for lung cancer include chest pain, worsening nodule should be classified as category 4B, with
cough, hemoptysis, and unintended weight loss. a probability of malignancy of more than 15%
Symptomatic patients with lung cancer are because of a diameter increase of 8 mm, a finding
much more likely to have advanced-stage dis- potentially leading to an invasive workup. On
ease at the time of diagnosis. For example, in the the other hand, one can argue that if the 8-mm
results of one study that was based on data from increase in diameter was divided evenly over the
the International Early Lung Cancer Action Pro- 12-year time span, then the annual growth rate
gram (I-ELCAP) study, investigators found that would not meet the 1.5-mm threshold. However,
patients in whom interim small cell lung carci- with only two data points, there is no way to
noma was diagnosed because of the development determine the true growth curve.
of symptoms (cancers that were not apparent at Volume doubling time may be used to predict
LCS CT but were found before the next annual solid pulmonary nodule behavior as benign
LCS CT examination) usually had advanced- or malignant (Fig 3) (10,11). Generally, the
1980 November-December 2017 radiographics.rsna.org

Figure 2. Very slow-growing nodule in a


74-year-old woman. (a)Baseline LCS CT: Axial
unenhanced chest CT image (lung window
settings) of the right lung shows a 19-mm nod-
ule (arrow) in the posterior basal right lower
lobe. (b)Abdominal CT performed 12 years
earlier (before baseline LCS CT): Comparison
axial CT image (lung window settings) of the
right lung shows that the nodule (arrow) had
measured 11 mm.

Figure 3. Suggested interpretation for the etiology of a solid lung nodule on the
basis of the volume doubling time (VDT). Data are from Henschke et al (10) and
Truong et al (11).

observed volume doubling times of lung cancers hamartomas detected at LCS CT, seven of nine
range from 20 days to 360400 days. For solid hamartomas had volume doubling times greater
pulmonary nodules, volume doubling times of than 550 days (range, 5506000 days), one had a
less than 2030 days are usually associated with volume doubling time less than 450 days, and one
benign causescommonly, inflammatory or had a decrease in volume at follow-up (16).
infectious causes. Volume doubling times greater Another potential cause for a slowly growing
than 400480 days are also usually associated solid nodule is a carcinoid tumor. These low-
with benign nodules (10,11), with only 3%4% grade malignant neoplasms are rare, with an
of nonsmall cell lung carcinomas reported to estimated frequency of 1.57 per 100000 pa-
have volume doubling times of more than 400 tients. In the results of one retrospective analysis
days (10,12). These volume doubling times do of 28 pathologically proven carcinoid tumors,
not apply to pure ground-glass nodules, which investigators showed that when the tumors were
are discussed separately. The use of volume dou- peripheral in location, 81% (13 of 16) were
bling time as a discriminator in the context of detected incidentally (17). CT features reported
LCS has been proposed to reduce false-positive to help distinguish carcinoid tumors from be-
screening findings, unnecessary follow-up exam- nign lung nodules include lobulated margins,
inations, and additional workups (13). Currently high attenuation on unenhanced CT images,
available tools for radiologists include volume avid contrast enhancement, and proximity to
doubling time calculators and online lung the airways. In the results of the same series, 13
cancer risk assessment tools such as the McWil- (81%) of 16 peripheral carcinoids had direct
liams lung cancer risk calculator (2,14,15). airway involvement, defined as a component of
Granulomatous inflammation and mesenchy- the lesion located within or obstructing a visible
mal tumors fall into the types of solid lung nodules airway (tip of the iceberg sign) (17). Indirect
that have long volume doubling times. The pres- signs of airway involvement in peripheral carci-
ence of macroscopic fat, popcorn calcifications, or noids included peripheral hyperlucency (31%),
both in a solid nodule allows the confident diag- peripheral bronchiectasis (25%), and peripheral
nosis of a hamartoma, but these findings are often opacity or atelectasis (19%), with 10 (63%) of
lacking in smaller hamartomas (16) (Fig 4). In the 16 peripheral carcinoids having at least one in
results of one retrospective study of biopsy-proven direct finding of airway involvement (17).
RG Volume 37 Number 7 Martin et al 1981

Figure 4. Slowly growing pulmonary ham-


artoma in a 64-year-old man. (a) Baseline
LCS CT: Axial unenhanced chest CT image
(lung window settings) of the right lung
shows an irregular 11 3 7-mm solid nodule
(arrow). (b) Comparison axial CT image ob-
tained 5.5 years earlier (before baseline LCS
CT) shows that the nodule (arrow) had mea-
sured 6 3 4 mm. Comparison of this earlier
measurement with the current one indicates
a volume doubling time of approximately
939 days. A remote history of testicular can-
cer prompted biopsy of the nodule.

Figure 5.Invasive adenocarcinoma in a


66-year-old woman. (a) Axial contrast mate-
rialenhanced chest CT image (lung window set-
tings) of the right upper lobe shows a ground-
glass nodule (arrow). (b) Comparison axial CT
image obtained 3 years earlier than a shows that
the nodule had increased in attenuation centrally
with time, without an overall change in size.

The calculated volume doubling time of the of Lung-RADS may incorporate volume doubling
nodule in this vignette is 5.2 years. In this case, we time as a measure to assess solid nodules, as is be-
are confident that this nodule is not an aggressive ing suggested in Europe (13).
cancer, and we recommend assigning category
2 (benign or benign behaving) on the basis of Scenario 4: Ground-Glass
the long volume doubling time and recommend Nodule That Increases in Density
continued LCS CT in 12 months. Although it but Remains Stable in Size
may be argued that only two time points during a A 66-year-old woman had a ground-glass nodule
long time period do not guarantee that the nodule with increasing attenuation but unchanged diam-
recently accelerated growth, we advocate for a eter (Fig 5). If confronted with this situation at
conservative approach in a screening program. LCS CT, what would be the appropriate Lung-
In this case, the nodule could also be assigned to RADS category and management of this nodule
category 4B on the basis of its size and growth; the (Table 1) (2)?
recommendations can be based on footnote 9 in Subsolid nodules include a spectrum of fo-
Lung-RADS, which states that the management cal areas of increased lung attenuation. Subsolid
of category 4B nodules should be based on the nodules can be divided into (a)pure ground-glass
probability of malignancy on the basis of patient nodules, which are defined as focal areas of in-
evaluation, preference, and the risk of malignancy. creased lung attenuation through which normal
Therefore, a follow-up CT examination could be structures are visible, and (b)part-solid nodules,
performed in 3 months, and if the nodule is stable, which are defined as ground-glass nodules with
it could be assigned to category 2 and the patient solid components, the latter of which obscure un-
returned to the LCS CT program. Future versions derlying structures (8,18). Subsolid nodules have a
1982 November-December 2017 radiographics.rsna.org

Figure 6. Large ground-glass nodule in a


69-year-old man. (a) Baseline LCS CT: Axial
unenhanced chest CT image (lung window set-
tings) of the right upper lobe shows a 22-mm
irregular ground-glass nodule (arrow). (b) Axial
chest CT image obtained 6 months after baseline
LCS CT shows resolution of the nodule.

growth behavior that is different from that of solid lations, a ground-glass nodule that doubles in size
nodules. A subsolid nodule can grow not only by in 1 year, and enlarged lymph nodes.
increasing in diameter but also by increasing in
attenuation, which can manifest as the develop- Scenario 5: Ground-Glass Nodule
ment of a solid component or by enlargement of with Slow Growth Rate
only the solid component without a change in the When classifying a ground-glass nodule in
overall diameter, resulting in an overall increase Lung-RADS, all ground-glass nodules that
in mass. These features of growth have been as- either (a)measure less than 20 mm, or (b)mea-
sociated with an increased risk of malignancy. sure 20 mm or greater and are stable or slowly
Persistent subsolid nodules are more likely to be growing fall into category 2. The term slowly
malignant than other nodule types, and the risk growing is not precisely defined in Lung-RADS
of malignancy increases with the size of the solid and, as such, is open to the radiologists in-
component (6,10,11). terpretation. Until the term slowly growing is
Currently, the only definition of growth in defined in a future version of Lung-RADS, the
Lung-RADS is described in footnote 4, where following several key points about ground-glass
growth is defined as an increase in diameter of at nodules should be considered, to make appro-
least 1.5 mm. Although Lung-RADS addresses priate recommendations.
new or growing solid components as a measure of Subsolid nodules can result from infection,
growth of part-solid nodules, it does not address inflammation, a scar, hemorrhage, or a neo-
the growth of ground-glass nodules manifest- plasm (Figs 6, 7). The first step in assessing a
ing as a diffuse increase in attenuation. Future ground-glass nodule is to determine whether it
versions of Lung-RADS may clarify the concept is transient or persistent; a nodule that is due to
of growth, but until then, radiologists should rec- infection, inflammatory conditions, or hemor-
ognize other findings indicative of nodule growth rhage will resolve. If a ground-glass nodule
and view them with appropriate suspicion. Mass, persists, it can still be benign (eg, organizing
which combines the nodule density and volume, pneumonia or focal fibrosis) (11) (Fig 8). When
has been reported to be a more accurate assess- malignant, lung nodules with ground-glass at-
ment of nodule growth, when compared with tenuation are nearly always adenocarcinomas
diameter or volume measurements alone (19). (6,10). Features of pure ground-glass nodules
Nodules that diffusely increase in attenuation considered to be risk factors for malignancy
but not in size do not fall clearly into a category. include a diameter of greater than 10 mm and
Although there is not a clear solid component, the presence of cystic changes, which are also
the increase in attenuation is consistent with described as internal lucencies (6). As discussed
growth, which makes the nodule suspicious. Be- in scenario 4, development of a solid component
cause the increase in attenuation is suggestive of is also a feature suggestive of malignancy. Inves-
a developing invasive component and a change in tigators have shown that ground-glass nodules
mass, it would be appropriate to assign category smaller than 5 mm have a less than 1% risk of
4X to nodules with this behavior. Category 4X transforming into a malignant lesion (20). About
gives radiologists discretion to use their experi- 7% of ground-glass nodules 510 mm in diam-
ence and judgment when a particular situation is eter have features of invasive adenocarcinoma
not clearly described but the imaging findings are (Fig 9). Ground-glass nodules larger than 10
highly suspicious for lung cancer. In the descrip- mm have a higher risk of being adenocarcinoma
tion of category 4X, such features include spicu- in situ or minimally invasive adenocarcinoma
RG Volume 37 Number 7 Martin et al 1983

Figure 7. Ground-glass pseudonodule at base-


line LCS CT in a 56-year-old man. (a) Axialunen-
hanced chest CT image (lung window settings)
of the left lung shows an apparent ground-glass
nodule (arrow) in the inferior lingula. (b)Sagittal
reformatted CT image through the left hemitho-
rax shows that the apparent nodule (arrow) is
flat, a finding consistent with a scar or an unusual
focus of atelectasis.

Figure 8. Benign part-solid nodule in the right


lower lobe mimicking lung cancer at LCS CT in a
59-year-old man. Axial unenhanced chest CT image
(lung window settings) of the right lung shows a
part-solid irregular nodule (arrow) in the right lower
lobe. The nodule persisted at follow-up CT. After a
nondiagnostic fine-needle aspiration biopsy, lobec-
tomy was performed. The final histopathologic find-
ings from examination of the lobectomy specimen
disclosed benign lung parenchyma with scattered
noncaseating granulomas, anthracosis, emphysema,
and focal alveolar siderophages.

is best practice to compare all nodules with the


most remote examination available, to optimize
sensitivity for the detection of change.
Lung-RADS incorporates all of this informa-
(Fig 10), although approximately 20%25% of tion by simply classifying all stable or slowly
such ground-glass nodules are benign (11) . growing ground-glass nodules in category 2
The volume doubling times of ground-glass (benign appearance or behavior). Behavior is
nodules are quite variable but typically are longer what would best describe these nodules. Only
than the volume doubling times of solid nodules time will tell if these nodules are, or will ever
and part-solid nodules. The volume doubling become, clinically relevant cancers. In this era in
times of adenocarcinomas manifesting as pure which all screening tests are being scrutinized,
ground-glass nodules have been reported to be radiologists must be clear and must educate
longer than 400 days and up to 1436 days (11). referring providers that even though a ground-
Investigators have shown that the growth of most glass nodule may be malignant, it can be safely
ground-glass nodules (including adenocarcino- observed to avoid overdiagnosis. When faced
mas) is slow enough that annual surveillance with the controversial question of when to inter-
is appropriate, thereby avoiding overdiagnosis vene in a patient with a slowly changing ground-
and aggressive, potentially harmful intervention. glass nodule (Fig 11), clear communication with
Sawada et al (21) reported no adverse outcome the referring provider can avoid misinterpreting
related to a delay in therapy in patients who slow interval growth as an impetus for aggres-
underwent resection of subsolid nodules only sive intervention. Although this path may be
after growth was apparent at CT. Furthermore, more challenging because no formal definition
the most recent Fleischner Society guidelines of the term slowly growing exists, radiologists are
have been updated to reflect the indolent nature encouraged to discuss the options of surveil-
of these lesions (6). An increase in the maximum lance versus more aggressive evaluation, includ-
diameter of more than 1.72 mm is needed to ing the potential consequences of more aggres-
identify true growth of a ground-glass nodule, sive workup of these indolent lesions, with the
according to one group of investigators (22). It other members of the patient care team.
1984 November-December 2017 radiographics.rsna.org

Figure 10. LCS-detected lung adenocarci-


Figure 9. LCS-detected invasive lung noma in a 55-year-old man. Sagittal reformat-
adenocarcinoma in a 57-year-old man. ted chest CT image (lung window settings)
Axial unenhanced chest CT image (lung through the right hemithorax shows an 11-mm
window settings) of the right lung pure ground-glass nodule in the superior seg-
shows a 7-mm ground-glass nodule ment of the right lower lobe. Note the distor-
with a lucent center (arrow) in the right tion of the fissure (arrow).
lower lobe.

Two items to consider with respect to the the solid component. The inter- and intraobserver
definition of the term slowly growing in future agreement in the classification of nodules as solid
versions of Lung-RADS are the concepts of mass or subsolid is highly variable (26). The Fleischner
(attenuation) change and change during longer Society guidelines recommend evaluating part-
time intervals (eg, 23 years). Some investiga- solid lung nodules by using lung window settings
tors advocate for adjusting screening intervals with a sharp filter to better assess the extent of
to risk-defined models (13), including increas- the solid component (6). However, how does one
ing the screening interval to 2 years in selected measure a solid component when it is not discrete?
patients with subsolid nodules at LCS CT (23). We asked thoracic radiology colleagues across
In the results of another study, investigators the country how they would categorize the nod-
have shown no protective effect on mortality in ule in Figure 12, and responses included catego-
patients undergoing annual screening compared ries 3, 4A, 4B, and 4X. The disagreement was
with biennial screening (24). However, in the related to the measurement of the solid com-
findings from a study derived from the Dutch- ponent. However, all radiologists agreed that
Belgian NELSON trial, investigators showed that this nodule was suspicious for lung adenocar-
a 2.5-year screening interval was associated with cinoma. This example is another one in which
(a)higher interval cancer rates, compared with 1- the X modifier allows radiologists to use their
and 2-year screening intervals, and (b)a higher clinical judgment to decide how to best man-
proportion of patients with an advanced disease age a suspicious nodule. For this nodule, the
stage in the final screening round, compared with most conservative approach would be to repeat
previous rounds (25). This area is controversial imaging in 3 months to assess for change. If the
and will likely be studied heavily in the future as nodule is stable, Lung-RADS category 2 would
data from LCS continue to accumulate. be assigned, and yearly screening would resume.
However, given the high suspicion for primary
Scenario 6: How to Measure lung adenocarcinoma, biopsy or resection could
and Classify a Part-Solid Nodule be considered. This patient ultimately under-
A 69-year-old man had a right upper lobe nod- went resection, and the results of histopatho-
ule at baseline LCS CT (Fig 12). Which Lung- logic examination of the specimen from resec-
RADS category should be assigned? tion disclosed a primary lung adenocarcinoma
Evaluation of some lung nodules can be chal- with a mucinous component. Consider the use
lenging for radiologists, especially part-solid of the X modifier in complex subsolid nodules
nodules. Challenges include determining what is such as this one. Management options should be
the solid component, if any, and how to measure targeted to each patient.
RG Volume 37 Number 7 Martin et al 1985

Figure 11. Slowly growing ground-glass nodule in a 67-year-old man. (a) Axial unenhanced chest CT image (lung window
settings) of the left upper lobe shows a pure ground-glass nodule (arrow). (b) Axial unenhanced chest CT image obtained 2
years later than a shows minimal growth of the nodule (arrow), which now measures 7 mm. (c) Axial unenhanced chest CT
image obtained 2 years later than b shows continued growth of the nodule (arrow), which now measures 13 mm. The his-
topathologic findings from examination of the specimen obtained at wedge resection helped confirm the diagnosis of a well-
differentiated minimally invasive adenocarcinoma.

may play a role in the future as the LCS data are


analyzed with more long-term studies.
A decrease in nodule size is not always benign,
particularly with an increase in attenuation, and
continued surveillance may be appropriate. The
nodule in Figure 13 could be assigned either to
category 4B (on the basis of a part-solid nodule
with an enlarging solid component >4 mm) or to
Figure 12.Invasive mucinous category 4X, given that the appearance is highly
lung adenocarcinoma in a 69-year-
old man. Axial unenhanced chest
suspicious for malignancy, even though there is a
CT image (lung window settings) slight decrease in size.
of the right lower lobe shows a
large lobulated nodule (arrow) Scenario 8: Nodule with
with a high-attenuation rim and
a ground-glass center with some
Characteristic Features of an
small lucencies. Intrapulmonary Lymph Node
A 72-year-old man had a 4-mm perifissural
nodule with characteristics of an intrapulmonary
Scenario 7: Nodule That Decreases lymph node at baseline LCS CT (Fig 14). How
in Size but Increases in Attenuation should this nodule be classified?
Figure 13 is an example of a malignant nodule The terms perifissural nodule and intrapulmonary
that decreased in size but increased in attenu- lymph node are often used interchangeably. Many
ation. Lung-RADS currently does not address investigators have hypothesized that perifissural
nodules that decrease in size. Several investiga- nodules are normal lymph nodes located in the
tors have documented that adenocarcinomas can lung parenchyma. These solid lung nodules are
transiently decrease in size, presumably related to well circumscribed, smoothly marginated, and in
the development of a fibrous component and as- contact with or in proximity to a fissure or pleural
sociated collapse. However, this decrease in size is surface (usually within 20 mm from the pleural
usually associated with an increase in attenuation surface). Intrapulmonary lymph nodes are usually
(6,11,27). For these reasons, radiologists should triangular, oval, or polygonal; and they often have
recognize that cancer growth may not always a thin septal attachment (Figs 15, 16). On multi-
be exponential and that the overall volume of a planar reformatted images, intrapulmonary lymph
malignant nodule may decrease at some point. As nodes are often nonspherical and flat or triangular.
mentioned in other scenarios, assessment of mass They are more commonly seen in the lower lungs
1986 November-December 2017 radiographics.rsna.org

Figure 13. Invasive lung adenocarcinoma in a


68-year-old man. (a)Initial LCS CT: Axial unen-
hanced chest CT image (lung window settings)
of the right lung shows a part-solid nodule (ar-
row) in the inferior right middle lobe. (b)Follow-
up axial CT image obtained 3 months later than a
shows slight contraction of the nodule, although
the solid component (arrow) has enlarged.

below the level of the carina (2830). The median


diameter of intrapulmonary lymph nodes ranges
from approximately 1 mm to 6 mm, but they can
be larger (28,30,31).
Intrapulmonary lymph nodes are extremely
common and have been detected more fre-
quently with increased routine use of thin-sec-
tion CT. For example, up to one-third of lung
nodules identified at LCS CT in one series were
classified as perifissural (28). In the results of
several LCS studies, investigators have shown Figure 14. Normal intrapulmonary lymph node in a 72-year-old
man at LCS CT. Axial (a) and sagittal (b) unenhanced reformat-
that no lung nodules with features of intrapul- ted chest CT images (lung window settings) through the left lung
monary lymph nodes ultimately proved to be a show a triangular left lower lobe nodule (arrow) abutting the
lung cancer (6,28,32). oblique fissure.
Lung-RADS addresses intrapulmonary lymph
nodes in footnote 11, recommending that they
Figure 15. Normal
should be managed as any other solid nodule on intrapulmonary lymph
the basis of the mean diameter. Because most node in a 75-year-old
intrapulmonary lymph nodes fall into Lung- man. Axial contrast-en-
RADS category 2 (benign or benign behavior), hanced chest CT image
(lung window settings)
management is to continue annual LCS as long of the right upper lobe
as the patient meets the eligibility criteria. How- shows a thin band of
ever, some radiologists do not routinely report tissue (arrow) connect-
or recommend follow-up of nodules with the ing the nodule to the
pleural surface.
classic features of intrapulmonary lymph nodes,
conflicting with the Lung-RADS requirement to
report and manage nodules on the basis of size.
We believe that if there are multireader studies
showing adequate agreement for the recognition RADS category should be assigned, and how
of intrapulmonary lymph nodes, future ver- should this case be managed?
sions of Lung-RADS may allow management Cigarette smoke promotes mucin synthesis
of intrapulmonary lymph nodes as category 1 and mucous gland metaplasia and causes stasis of
(definitely benign nodules), potentially reducing mucus with a variety of mechanisms (33). On CT
unnecessary follow-up and patient anxiety. images, these abnormalities may be apparent as
bronchial wall thickening and mucus plugging, the
Scenario 9: Airway (Endotracheal latter of which can mimic endobronchial nodules.
or Endobronchial) Nodule Lung-RADS suggests classifying endobronchial
A 74-year-old man had a new endobronchial nodules as category 4A lesions, with recommenda-
filling defect at LCS CT (Fig 17). What Lung- tions including a follow-up low-dose CT examina-
RG Volume 37 Number 7 Martin et al 1987

Figure 17. New endobronchial filling defect in a 74-year-old


man at LCS CT. (a) Baseline LCS CT: Axial unenhanced chest CT
image (soft-tissue window settings) cropped to the mediastinum
Figure 16. Normal intrapulmo-
shows a filling defect (arrow) in the bronchus intermedius. The
nary lymph node in a 60-year-old
filling defect is irregular, shows high attenuation, and is nonde-
man at LCS CT. Axial unenhanced
pendent in location. (b) Follow-up axial CT image obtained after
chest CT image (lung window set-
vigorous coughing shows that the filling defect is no longer pres-
tings) of the anterior right hemi-
ent, which indicates that it represented adherent secretions with
thorax shows a trapezoidal nodule
atypical features.
(arrow) in the right middle lobe
abutting the horizontal fissure.

Several methods can help identify airway


secretions with greater confidence. A repeat CT
examination of the area of interest with use of
a reduced dose technique after vigorous cough-
ing is one option, because coughing might help
clear secretions mimicking endobronchial lesions
(35). Although this method has proved to be
helpful to us in selected cases such as the one
in this scenario, to our knowledge, no published
Figure 18. Typical airway secre- studies exist that evaluate the utility of coughing
tions in a 72-year-old man. Axial for clearing secretions before scanning a patient.
unenhanced chest CT image (lung We suggest considering this option in patients
window settings) of the upper tra-
chea shows low-attenuation filling with questionable endobronchial lesions if this
defects containing air bubbles (ar- option can avoid a more aggressive workup such
rows) in the upper thoracic trachea. as a PET/CT examination. Maximum intensity
projections, which should be part of any LCS CT
protocol, or two-dimensional reformatted images
tion in 3 months, or PET/CT if the nodule has can improve confidence that an apparent endo-
a solid component measuring 8 mm or greater. bronchial nodule is mucus by showing a rectan-
If these guidelines are strictly followed, patients gular configuration within an airway that paral-
could end up being scanned every 3 months, lels the adjacent pulmonary artery. As with all
because mucus plugs are common in smokers and patients, mucus plugs should also prompt one to
former smokers, and new foci of plugging may be evaluate the affected airways to exclude a central
found at each subsequent CT examination. obstructing process.
To avoid unnecessary CT examinations, If the lesion remains indeterminate, an endo-
knowledge of the appearances of airway secre- bronchial nodule should be managed as a cate-
tions and mucus plugs is important. Features of gory 4A lesion according to Lung-RADS. In the
secretions in large airways (Fig 18) include a thin findings from one study of approximately 53000
and long linear configuration, a dependent or patients who underwent LCS CT, investigators
layering location, water attenuation, air bubbles showed that of the 186 subjects who underwent
within the filling defect, preservation of the adja- follow-up evaluation (CT, bronchoscopy, or
cent cartilaginous rings, a complex shape, and at- both) for endobronchial nodules, only seven
tenuation of less than 21.7 HU, all of which have had persistent endobronchial nodules, none of
a high positive predictive value for secretions. In which were malignant (36). Unless there are
contrast, features highly predictive for neoplasm features suggestive of a solid tumor such as inva-
are a round or lobulated shape, attenuation of sion of the airway wall, expansion of the airway,
21.7 HU or more, and internal fat or calcification or postobstructive atelectasis or infection, we
(34). Unfortunately, retained or adherent secre- favor a follow-up low-dose CT examination in
tions can have a variable composition and may 3 months over the performance of PET/CT or
not always demonstrate a classic appearance. bronchoscopy.
1988 November-December 2017 radiographics.rsna.org

Figure 19. Hepatocellular carcinoma incidentally detected at LCS CT in a 71-year-old man. (a) Baseline LCS CT: Axial unenhanced
chest CT image (narrow soft-tissue window settings) through the upper abdomen shows a low-attenuation mass (arrow) in the right
hepatic lobe, a finding that prompted the use of the S modifier. (b) Subsequent axial contrast-enhanced T1-weighted magnetic
resonance (MR) image helps confirms the enhancing mass (arrow).

Scenario 10: Incidental Potentially


Important Finding Other than Lung
Cancer Detected at Low-Dose LCS CT
Figures 19 and 20 show incidental findings
detected at LCS CT at our institutions. Which
incidental findings require an S modifier?
Incidental findings detected at diagnostic
imaging studies are common (37) and are one
of the criticisms of the use of low-dose CT for
LCS. In the results of the National Lung Screen-
ing Trial (NLST), the rate of incidental findings
needing additional evaluation was 10.2% at
Figure 20. Adrenal adenoma incidentally detected at LCS CT
baseline and 7.5% overall (5). In a more recent in a 60-year-old man. Axial unenhanced CT image (soft-tissue
retrospective study, investigators analyzed the window settings) of the upper abdomen shows a round, low-
data of 17309 participants of the NLST (38). attenuation left adrenal nodule (arrow) with an appearance
Of these participants, 58.7% of the patients had that is typical of a benign adenoma. The S modifier was not
used, because benign adenomas are not considered clinically
extrapulmonary findings (described as an abnor- important or potentially clinically important.
mality beyond the lungs, pleura, chest wall, and
hilar and mediastinal lymph nodes), and 19.6%
had at least one finding that was considered po- modifier for clinically significant or potentially
tentially clinically important. Of those patients in clinically significant findings other than lung
which the location of the abnormality was avail- cancer detected at LCS CT. However, the term
able (70%), these abnormalities were located in clinically significant or potentially clinically signifi-
the following organs and systems: cardiovascular cant is not defined, leaving the decision to the
system (8.5%), renal system (2.4%), hepatobili- interpreting radiologist and potentially intro-
ary system (2.1%), adrenal glands (1.2%), and ducing local or regional variation in the report-
thyroid (0.6%). The prevalence of extrapulmo- ing and management. Furthermore, although
nary malignancy was low at 0.39%, including guidelines exist outside Lung-RADS for further
malignancies of the kidney (0.26%), thyroid management of some findings, no consensus
(0.08%), and liver (0.05%) (38). exists for others. Finally, Lung-RADS does not
The joint ACRSociety of Thoracic Radiol- address reporting or management of any inci-
ogy practice parameter for the performance dental findings.
and reporting of LCS CT states that the entire We encourage radiologists to adopt published
examination should be reviewed for other poten- management guidelines such as the ACR white
tially important findings and the results reported papers on management of incidental findings at
in accordance with the practice parameter (39). CT. Relevant to LCS, the ACR has guidelines for
Lung-RADS provides radiologists with the S the management of incidentally detected thyroid
RG Volume 37 Number 7 Martin et al 1989

Table 4: Incidental Findings at Chest CT: Management Recommendations and Their Rationale
Coronary artery calcification
S modifier Although the results of studies that are based on LCS trials have shown that coronary artery
calcifications are predictive of future all-cause mortality and cardiovascular events (48), others question
whether coronary artery calcification scoring will be clinically useful (49).
Langerhans cell histiocytosis
S modifier Although uncommon, Langerhans cell histiocytosis can be seen in smokers; may lead to a
health benefit if the patient stops smokingcan progress to end-stage COPD or pulmonary hypertension if
untreated.
Any abdominal findings for which workup or management would be needed
S modifier Follow ACR white papers for further workup and management (4144).
Thyroid nodules
Follow ACR white paper guidelines (40). Only assign S modifier if the nodule warrants further workup and
has not already been evaluated (40).
Mediastinal mass
Definitely benign findings such as a duplication cyst or diverticula can be left in the body of the report to
avoid confusion and potential harm and cost from unnecessary testing.
Indeterminate findings such as a thymic solid mass S modifier. Manage on the basis of CT appearance
(size, signs of invasion, lymphadenopathy) (45).
Emphysema
Controversial; not routinely encouraged to report with an S modifier extremely common finding in smok-
ers and unlikely to change management.
Note.Numbers in parentheses are reference citations. COPD = chronic obstructive pulmonary disease.

nodules and abdominal abnormalities, including angioma, pericardial cyst). Henschke et al (45)
those in the gallbladder and biliary tract, arter- concluded that in a screening population, inci-
ies and veins, spleen, and lymph nodes (4044). dentally detected mediastinal masses could be
When no published guideline is available, local managed conservatively, including thymic masses
practice patterns should be aligned to ensure smaller than 3 cm in diameter.
consistent practice within a screening program. Adopting local or established guidelines can
Published literature with regard to each topic can minimize variation in reporting and avoid confu-
be used to develop local-regional standards until sion and nonadherence to recommendations. We
these issues are addressed. also encourage having the S modifier as a sepa-
For example, in the results of a retrospective rate line in the impression and not leaving poten-
study of 9263 baseline screening examinations tially important findings limited to the body of
from the Early Lung Cancer Action Project (EL- the report, with the description of the abnormal-
CAP) study, incidentally detected mediastinal ity and clear management suggestions outlined
masses were documented in 71 subjects (0.77%), in the impression to avoid potentially important
and only one subject (0.01%) developed a medi- findings being overlooked (46,47).
astinal mass at a subsequent screening CT (45). Table 4 (4045,48,49) summarizes common
Thymic masses were the most common, followed incidental findings at chest CT and proposes
by thyroid lesions. Of the 41 thymic masses that management options. The S modifier should be
were detected, only five were larger than 3 cm, reserved for findings that may lead to a health
of which one was a cyst, three were noninvasive benefit if a behavior is modified or those that will
thymomas, and one was a thymic carcinoma (45). lead to an adverse outcome if not further evalu-
Eighteen of the nonresected mediastinal masses ated or treated. We do not encourage the use of
were unchanged at follow-up imaging, five were the S modifier for the same findings at subsequent
larger, and two were smaller. No adverse out- examinations if the findings have been addressed.
comes were associated with 1-year surveillance
of this population. All 16 thyroid masses were Scenario 11: Reenrolling Patients
associated with goiter (45). Of the remaining 14 in the LCS CT Program after a Stable
mediastinal masses, two were esophageal can- Abnormality
cers, six were esophageal diverticula, two were A 76-year-old man had a 7-mm indeterminate
lipomas, and four were purely cystic lesions and solid nodule at a baseline LCS CT examination
presumably benign (bronchogenic cyst, lymph- (Fig 21). Because no comparison images existed,
1990 November-December 2017 radiographics.rsna.org

Figure 21.Indeterminate nodule in


a 76-year-old man at baseline LCS CT.
Axial unenhanced chest CT image (lung
window settings) shows a smoothly mar-
ginated 7-mm solid nodule (arrow) in the
right upper lobe.
Figure 22. Residua of infection in a
67-year-old man at LCS CT. Axial unen-
the CT finding was classified as Lung-RADS hanced chest CT image (lung window
category 3, and follow-up was recommended in 6 settings) of the right lung shows a nodular
focus of consolidation (arrow) in the right
months. Because the nodule did not change at the upper lobe. A recent chest radiograph
follow-up examination, the patient was returned to (not shown) demonstrated more exten-
annual screening. Should the next screening exam- sive consolidation in this region.
ination be scheduled at 12 months from the origi-
nal baseline LCS CT examination or at 12 months
from the most recent LCS CT examination? general, patients with a recent respiratory tract
Footnote 12 in Lung-RADS states, Category infection should delay LCS for approximately 3
3 and 4A nodules that are unchanged on inter- months to ensure that any residual lung inflam-
val CT should be coded as category 2, and in- mation has resolved.
dividuals returned to screening in 12 months. The Findings suggestive of infection or other in-
time point from which the 12 months is counted flammatory processes are not addressed in Lung-
is less clear. The ACR recommends that the 12 RADS. Because a recent chest radiograph was
months should be counted from the day of the available in this case and because the area of con-
LCS CT examination that prompted the follow- solidation was deemed smaller at the LCS CT
up examination, meaning that the patient should examination, this case was assigned to category 3
undergo the next round of screening 12 months (probably benign), and a 6-month follow-up CT
after the baseline LCS CT examination in this examination was recommended. The reasoning
case. Many radiologists, including some of the for this decision was that the area of consolida-
authors, believe that a change in less than a year tion persisted but was smaller.
is unlikely, so the 12 months should be counted Unnecessary follow-up imaging related to
from the most current LCS CT examination, or acute lung infection can be avoided with targeted
18 months after the baseline LCS CT examina- questioning about any recent acute respiratory
tion in this case. No data exist to support either illness at the time of LCS CT scheduling, as well
position. Future Lung-RADS iterations will as by the technologist before the LCS CT ex-
likely clarify this recommendation. amination is performed. Patients with clinical or
radiologic findings of a recent acute respiratory
Scenario 12: Low-Dose LCS CT tract infection should defer LCS CT.
of a Patient with a Recent
Respiratory Infection Scenario 13: Categorization of a
A 67-year-old man underwent baseline LCS CT Cavitary Lung Nodule or Nodules
shortly after an acute respiratory illness, which A 73-year-old man had a cavitary nodule
was not reported to the primary care provider or in the left upper lobe at baseline LCS CT
the CT technologist, and residual consolidation (Fig23). How are cavitary nodules classified in
was present at CT (Fig 22). How should the LCS Lung-RADS?
CT examination be categorized? Lung-RADS does not address categorization
As part of the shared decision-making process, and management of cavitary lung nodules. Many
patients should be asked about signs and symp- conditions can manifest as solitary or multiple
toms of a recent respiratory tract infection. In cavitary nodules, including lung cancer, metas-
RG Volume 37 Number 7 Martin et al 1991

Figure 24.Pulmonary blastomycosis in a


44-year-old woman. Axial unenhanced chest CT
image (lung window settings) of the right lung
shows an irregular cavitary nodule (arrow) in the
right lower lobe. The dominant nodule measured
20 mm. The presence of satellite airway-centric
nodules makes infection a more likely cause. Al-
though this case was not found in the setting
of LCS, if similar findings were encountered at
LCS, the Lung-RADS category should be based
on the size of the largest nodule. The Smodifier
should be used, specifically stating that infection
is suspected and that further evaluation is recom-
mended to establish the diagnosis.
Figure 23. Squamous cell car-
cinoma of the lung in a 73-year-
old asymptomatic man. Axial un-
enhanced chest CT image (lung
window settings) of the left lung evaluation is usually needed to establish an ac-
shows an 18-mm lobulated cavi- curate diagnosis (Fig 24) (53,54).
tary nodule (arrow) in the left up- We suggest classifying and managing a solitary
per lobe. It would be appropriate
to classify this nodule as category
cavitary lung nodule at LCS CT as if it were a
4B or category 4X, given the size, solid nodule and appending the X modifier when
thick walls, and irregular inner mar- the suspicion for malignancy is high or the S mod-
gins that make the cavitary nodule ifier if infection is suspected so that the appropri-
highly suspicious for malignancy.
ate workup can ensue. If multiple cavitary nodules
are present, the S modifier would be the correct
designation unless diffuse lung adenocarcinoma is
tasis, infection, granulomatosis with polyangiitis, suspected, in which case the X modifier should be
and pulmonary Langerhans cell histiocytosis. appended. If the S modifier is used, management
Diffuse lung adenocarcinoma can manifest as and follow-up of these nodules should be per-
multiple lung nodules, which can be cavitary. formed outside of the screening program, either
Discussion of the differential diagnosis of cavi- with chest radiography or diagnostic CT.
tary nodules is beyond the scope of this review,
but it is important to remember that encounter- Scenario 14: Low-Dose LCS CT of a
ing patients with cavitary metastasis, vasculitis, Patient with a History of a Treated
or pulmonary Langerhans cell histiocytosis in Lung Malignancy
the setting of LCS would be unlikely. Because A 67-year-old woman with a 35pack-year his-
current or recent infectious symptoms are a tory of smoking had undergone lobectomy for
contraindication for LCS, encountering infection stage IA lung adenocarcinoma 7 years earlier. She
manifesting as cavitary nodules would also be was considered to be free of disease after 5 years
unusual, although it is worth remembering that of surveillance and currently wishes to undergo
some patients with infections such as nontuber- LCS. Is this patient eligible?
culous mycobacterial infection or endemic fungal The peak incidence of recurrent lung cancer
infection can be asymptomatic, with single or after definitive therapy is 23 years after treatment,
multiple cavitary pulmonary nodules (50). and the annual risk of developing a new lung
Up to 22% of nonsmall cell lung carcinomas cancer is 3% (55). The American Association for
demonstrate cavitation (51). In order of fre- Thoracic Surgery advocates annual screening for
quency, the most common histologic types that patients cured of lung cancer after 4 years of post-
show cavitation are squamous cell carcinoma, therapeutic surveillance as long as they are able
adenocarcinoma, and large cell carcinoma (52). and willing to undergo curative resection (55). The
Usually the cavity wall is thicker than 4 mm; and U.S. Preventive Services Task Force and the Cen-
the thicker the wall, the higher the likelihood of ters for Medicare and Medicaid Services do not
malignancy. Irregular inner surfaces and mural address screening patients with previously treated
nodules are more commonly associated with ma- and presumptively cured lung cancer. However,
lignant lesions. Because of the overlap between Lung-RADS provides the C modifier for these
malignant cavities and benign cavities, further patients, should they undergo LCS CT. Given the
1992 November-December 2017 radiographics.rsna.org

current requirements for adherence to eligibil- 3. Centers for Medicare & Medicaid Services. Decision memo for
screening for lung cancer with low dose computed tomography
ity criteria, patients with previously treated lung (LDCT) (CAG-00439N). Centers for Medicare & Medicaid
cancers must still meet the standard LCS eligibil- Services website. https://www.cms.gov/medicare-coverage-
ity criteria. In this case, we would recommend that database/details/nca-decision-memo.aspx?NCAId=274.
Published February 5, 2015. Accessed March 9, 2017.
the patient enroll in the LCS program. 4. Moyer VA; U.S. Preventive Services Task Force. Screen-
ing for lung cancer: U.S. Preventive Services Task Force
Scenario 15: Low-Dose LCS CT of recommendation statement. Ann Intern Med 2014;
160(5):330338.
a Patient with a Treated Low-Risk 5. Fintelmann FJ, Bernheim A, Digumarthy SR, et al. The
Nonlung Malignancy 10 pillars of lung cancer screening: rationale and logis-
Should a 62-year-old woman who was treated tics of a lung cancer screening program. RadioGraphics
2015;35(7):18931908.
for stage I breast cancer 18 months ago who is 6. MacMahon H, Naidich DP, Goo JM, et al. Guidelines for
eligible for LCS undergo LCS CT? management of incidental pulmonary nodules detected on
Current recommendations do not address CT images: from the Fleischner Society 2017. Radiology
2017;284(1):228243.
patients who have been treated for nonlung 7. MacMahon H, Austin JH, Gamsu G, et al. Guidelines for
malignancies, including those with a low risk of management of small pulmonary nodules detected on CT
recurrence, such as early-stage breast cancer, scans: a statement from the Fleischner Society. Radiology
2005;237(2):395400.
localized prostate cancer, or low-grade lymphoma. 8. Naidich DP, Bankier AA, MacMahon H, et al. Recommen-
The ACR suggests inquiring about previously dations for the management of subsolid pulmonary nodules
diagnosed and treated malignancies as part of the detected at CT: a statement from the Fleischner Society.
Radiology 2013;266(1):304317.
shared decision-making process. We believe that 9. Austin JH, Yip R, DSouza BM, Yankelevitz DF, Henschke
LCS should be offered to eligible patients with a CI; International Early Lung Cancer Action Program Inves-
treated low-risk malignancy, given the reported tigators. Small-cell carcinoma of the lung detected by CT
screening: stage distribution and curability. Lung Cancer
mortality benefits of LCS. One proposed thresh- 2012;76(3):339343.
old to determine whether an eligible patient 10. Henschke CI, Yankelevitz DF, Yip R, et al. Lung cancers
should undergo LCS CT is if the patient is more diagnosed at annual CT screening: volume doubling times.
Radiology 2012;263(2):578583.
likely than not to be alive in the next 5 years. 11. Truong MT, Ko JP, Rossi SE, et al. Update in the evalu-
ation of the solitary pulmonary nodule. RadioGraphics
Conclusion 2014;34(6):16581679.
12. Detterbeck FC, Gibson CJ. Turning gray: the natural history
Lung-RADS was modeled after the ACRs suc- of lung cancer over time. J Thorac Oncol 2008;3(7):781792.
cessful BI-RADS, with the intention to stan- 13. Kauczor HU, Bonomo L, Gaga M, et al. ESR/ERS white paper
dardize reporting and management of screening- on lung cancer screening. Eur Radiol 2015;25(9):25192531.
14. Brock University. Lung cancer risk calculators. Brock Uni-
detected lung nodules and to reduce the risks versity website. https://brocku.ca/lung-cancer-risk-calculator.
of overdiagnosis or unnecessary intervention in Published 2010. Accessed March 8, 2017.
screened patients. Because Lung-RADS is in its 15. Chestx-ray.com. Doubling time calculator. www.chestx-
ray.com website. http://www.chestx-ray.com/index.php
first iteration at the time of this writing, man- /calculators/doubling-time. Accessed March 8, 2017.
agement in several scenarios remains unclear. 16. Huang Y, Xu Dm, Jirapatnakul A, et al. CT- and com-
We expect that future versions of Lung-RADS puter-based features of small hamartomas. Clin Imaging
2011;35(2):116122.
will take into consideration these scenarios and 17. Meisinger QC, Klein JS, Butnor KJ, Gentchos G, Leavitt
provide guidance to radiologists. In the mean- BJ. CT features of peripheral pulmonary carcinoid tumors.
time, we offer our recommendations on the AJR Am J Roentgenol 2011;197(5):10731080.
18. Hansell DM, Bankier AA, MacMahon H, McLoud TC,
basis of the current literature and our collective Mller NL, Remy J. Fleischner Society: glossary of terms
experience. for thoracic imaging. Radiology 2008;246(3):697722.
19. de Hoop B, Gietema H, van de Vorst S, Murphy K, van
Disclosures of Conflicts of Interest.J.P.K. Activities related to Klaveren RJ, Prokop M. Pulmonary ground-glass nodules:
the present article: disclosed no relevant relationships. Activities increase in mass as an early indicator of growth. Radiology
not related to the present article: personal fees from HealthMyne; 2010;255(1):199206.
personal fees from Parexel Informatics. Other activities: dis- 20. Kakinuma R, Muramatsu Y, Kusumoto M, et al. Solitary
closed no relevant relationships. pure ground-glass nodules 5 mm or smaller: frequency of
growth. Radiology 2015;276(3):873882.
21. Sawada S, Komori E, Nogami N, Segawa Y, Shinkai T,
References Yamashita M. Evaluation of lesions corresponding to ground-
1. American College of Radiology. First edition of ACR Lung- glass opacities that were resected after computed tomography
RADS now available. American College of Radiology follow-up examination. Lung Cancer 2009;65(2):176179.
website. https://www.acr.org/About-Us/Media-Center/ 22. Kakinuma R, Ashizawa K, Kuriyama K, et al. Measurement of
Press-Releases/2014-Press-Releases/First-Edition-of-ACR- focal ground-glass opacity diameters on CT images: interob-
LungRADS-Now-Available. Published May 1, 2014. Ac- server agreement in regard to identifying increases in the size
cessed March 9, 2017. of ground-glass opacities. Acad Radiol 2012;19(4):389394.
2. American College of Radiology. Lung-RADS version 1.0 23. Heuvelmans MA, Oudkerk M. Management of subsolid
assessment categories release date: April 28, 2014. Ameri- pulmonary nodules in CT lung cancer screening. J Thorac
can College of Radiology website. https://www.acr.org/~/ Dis 2015;7(7):11031106.
media/ACR/Documents/PDF/QualitySafety/Resources/ 24. Pastorino U, Rossi M, Rosato V, et al. Annual or biennial CT
LungRADS/AssessmentCategories.pdf. Published April 28, screening versus observation in heavy smokers: 5-year results
2014. Accessed March 9, 2017. of the MILD trial. Eur J Cancer Prev 2012;21(3):308315.
RG Volume 37 Number 7 Martin et al 1993

25. Yousaf-Khan U, van der Aalst C, de Jong PA, et al. Final 42. Sebastian S, Araujo C, Neitlich JD, Berland LL. Managing
screening round of the NELSON lung cancer screening incidental findings on abdominal and pelvic CT and MRI.
trial: the effect of a 2.5-year screening interval. Thorax IV. White paper of the ACR Incidental Findings Commit-
2017;72(1):4856. tee II on gallbladder and biliary findings. J Am Coll Radiol
26. van Riel SJ, Snchez CI, Bankier AA, et al. Observer vari- 2013;10(12):953956.
ability for classification of pulmonary nodules on low-dose 43. Khosa F, Krinsky G, Macari M, Yucel EK, Berland LL.
CT images and its effect on nodule management. Radiology Managing incidental findings on abdominal and pelvic CT
2015;277(3):863871. and MRI. II. White paper of the ACR Incidental Find-
27. Lindell RM, Hartman TE, Swensen SJ, Jett JR, Midthun ings Committee II on vascular findings. J Am Coll Radiol
DE, Mandrekar JN. 5-year lung cancer screening experi- 2013;10(10):789794.
ence: growth curves of 18 lung cancers compared to histo- 44. Heller MT, Harisinghani M, Neitlich JD, Yeghiayan P, Ber-
logic type, CT attenuation, stage, survival, and size. Chest land LL. Managing incidental findings on abdominal and
2009;136(6):15861595. pelvic CT and MRI. III. White paper of the ACR Incidental
28. Ahn MI, Gleeson TG, Chan IH, et al. Perifissural nod- Findings Committee II on splenic and nodal findings. J Am
ules seen at CT screening for lung cancer. Radiology Coll Radiol 2013;10(11):833839.
2010;254(3):949956. 45. Henschke CI, Lee IJ, Wu N, et al. CT screening for lung
29. Sykes AM, Swensen SJ, Tazelaar HD, Jung SH. Computed cancer: prevalence and incidence of mediastinal masses.
tomography of benign intrapulmonary lymph nodes: retro- Radiology 2006;239(2):586590.
spective comparison with sarcoma metastases. Mayo Clin 46. Mehta HJ, Mohammed TL, Jantz MA. The American
Proc 2002;77(4):329333. College of Radiology Lung Imaging Reporting and Data
30. Shaham D, Vazquez M, Bogot NR, Henschke CI, Yankelevitz System: potential drawbacks and need for revision. Chest
DF. CT features of intrapulmonary lymph nodes confirmed 2017;151(3):539543.
by cytology. Clin Imaging 2010;34(3):185190. 47. Ravenel JG, Tanner NT, Silvestri GA. Viewing all the trees
31. Takenaka M, Uramoto H, Shimokawa H, et al. Discriminative in the forest: the importance of reporting abnormal find-
features of thin-slice computed tomography for peripheral ings on CT scan when screening for lung cancer. Chest
intrapulmonary lymph nodes. Asian J Surg 2013;36(2):6973. 2017;151(3):525526.
32. McWilliams A, Tammemagi MC, Mayo JR, et al. Probability 48. Jacobs PC, Gondrie MJ, van der Graaf Y, et al. Coronary
of cancer in pulmonary nodules detected on first screening artery calcium can predict all-cause mortality and cardiovas-
CT. N Engl J Med 2013;369(10):910919. cular events on low-dose CT screening for lung cancer. AJR
33. Williams OW, Sharafkhaneh A, Kim V, Dickey BF, Evans Am J Roentgenol 2012;198(3):505511.
CM. Airway mucus: from production to secretion. Am J 49. Bernheim A, Auffermann WF, Stillman AE. The dubious
Respir Cell Mol Biol 2006;34(5):527536. value of coronary calcium scoring on lung cancer screening
34. Hong SR, Lee YJ, Hong YJ, et al. Differentiation between CT. J Am Coll Radiol 2017;14(3):343344.
mucus secretion and endoluminal tumors in the airway: 50. Hahm CR, Park HY, Jeon K, et al. Solitary pulmonary nod-
analysis and comparison of CT findings. AJR Am J Roentgenol ules caused by Mycobacterium tuberculosis and Mycobacterium
2014;202(5):982988. avium complex. Lung 2010;188(1):2531.
35. Bankier AA, ODonnell CR, Boiselle PM. Quality initiatives: 51. Onn A, Choe DH, Herbst RS, et al. Tumor cavitation in
respiratory instructions for CT examinations of the lungsa stage I nonsmall cell lung cancer: epidermal growth factor
hands-on guide. RadioGraphics 2008;28(4):919931. receptor expression and prediction of poor outcome. Radiol-
36. Kim HJ, Kim DK, Kim YW, et al. Outcome of incidentally ogy 2005;237(1):342347.
detected airway nodules. Eur Respir J 2016;47(5):15101517. 52. Gill RR, Matsusoka S, Hatabu H. Cavities in the lung in
37. Brown SD. Professional norms regarding how radi- oncology patients: imaging overview and differential diagno-
ologists handle incidental findings. J Am Coll Radiol ses. Appl Radiol 2010. http://appliedradiology.com/articles/
2013;10(4):253257. cavities-in-the-lung-in-oncology-patients-imaging-overview-
38. Nguyen XV, Davies L, Eastwood JD, Hoang JK. Extrapul- and-differential-diagnoses. Published June 9, 2010. Accessed
monary findings and malignancies in participants screened March 9, 2017.
with chest CT in the National Lung Screening Trial. J Am 53. Honda O, Tsubamoto M, Inoue A, et al. Pulmonary
Coll Radiol 2017;14(3):324330. cavitary nodules on computed tomography: differentiation
39. Kazerooni EA, Austin JH, Black WC, et al. ACR-STR of malignancy and benignancy. J Comput Assist Tomogr
practice parameter for the performance and reporting of lung 2007;31(6):943949.
cancer screening thoracic computed tomography (CT): 2014 54. Woodring JH, Fried AM, Chuang VP. Solitary cavities of
(resolution 4). J Thorac Imaging 2014;29(5):310316. the lung: diagnostic implications of cavity wall thickness.
40. Hoang JK, Langer JE, Middleton WD, et al. Managing AJR Am J Roentgenol 1980;135(6):12691271.
incidental thyroid nodules detected on imaging: white paper 55. Jaklitsch MT, Jacobson FL, Austin JH, et al. The American
of the ACR Incidental Thyroid Findings Committee. J Am Association for Thoracic Surgery guidelines for lung cancer
Coll Radiol 2015;12(2):143150. screening using low-dose computed tomography scans for
41. Berland LL, Silverman SG, Gore RM, et al. Managing lung cancer survivors and other high-risk groups. J Thorac
incidental findings on abdominal CT: white paper of the Cardiovasc Surg 2012;144(1):3338.
ACR Incidental Findings Committee. J Am Coll Radiol
2010;7(10):754773.

TM
This journal-based SA-CME activity has been approved for AMA PRA Category 1 Credit . See www.rsna.org/education/search/RG.