Position Paper

Interstitial lung abnormalities detected incidentally on CT:

a Position Paper from the Fleischner Society
Hiroto Hatabu*, Gary M Hunninghake, Luca Richeldi, Kevin K Brown, Athol U Wells, Martine Remy-Jardin, Johny Verschakelen,
Andrew G Nicholson, Mary B Beasley, David C Christiani, Raúl San José Estépar, Joon Beom Seo, Takeshi Johkoh, Nicola Sverzellati,
Christopher J Ryerson, R Graham Barr, Jin Mo Goo, John H M Austin, Charles A Powell, Kyung Soo Lee, Yoshikazu Inoue, David A Lynch†

Lancet Respir Med 2020; The term interstitial lung abnormalities refers to specific CT findings that are potentially compatible with interstitial
8: 726–37 lung disease in patients without clinical suspicion of the disease. Interstitial lung abnormalities are increasingly
*Chair and †co-chair of the recognised as a common feature on CT of the lung in older individuals, occurring in 4–9% of smokers and 2–7% of
Fleischner Society Writing
non-smokers. Identification of interstitial lung abnormalities will increase with implementation of lung cancer
Committee for Position Paper on
interstitial lung abnormalities screening, along with increased use of CT for other diagnostic purposes. These abnormalities are associated with
Department of Radiology radiological progression, increased mortality, and the risk of complications from medical interventions, such as
(Prof H Hatabu MD, chemotherapy and surgery. Management requires distinguishing interstitial lung abnormalities that represent
R San José Estépar PhD), and clinically significant interstitial lung disease from those that are subclinical. In particular, it is important to identify
Department of Pulmonary and
the subpleural fibrotic subtype, which is more likely to progress and to be associated with mortality. This
Critical Care Medicine
(G M Hunninghake MD), multidisciplinary Position Paper by the Fleischner Society addresses important issues regarding interstitial lung
Brigham and Women’s abnormalities, including standardisation of the definition and terminology; predisposing risk factors; clinical
Hospital, Harvard Medical outcomes; options for initial evaluation, monitoring, and management; the role of quantitative evaluation; and future
School, Boston, MA, USA;
research needs.
Unità Operativa Complessa di
Pneumologia, Università
Cattolica del Sacro Cuore, Introduction abnormalities. The definition of ILAs is purely
Fondazione Policlinico A Interstitial lung disease (ILD) comprises a diverse group radiological and is based on the incidental identification
Gemelli IRCCS, Rome, Italy
(Prof L Richeldi MD);
of lung diseases with overlapping clinical, radiological, of CT abnormality. Differentiation between ILAs and
Department of Medicine physiological, and pathological features.1 Interstitial lung clinical and subclinical ILD must be on the basis of
(Prof K K Brown MD), and abnormalities (ILAs) refer to the presence of CT scan clinical evaluation.
Department of Radiology findings that are potentially compatible with ILD in ILAs are increasingly recognised on chest CT scans.2
(Prof D A Lynch MB), National
Jewish Health, Denver, CO, USA
patients who have partial (eg, abdominal CT including Systematic evaluation of large cohorts has shown a
(Prof K K Brown); Department the lower lung zones) or complete chest CT examinations prevalence of ILAs in older individuals (>60 years) of 4–9%
of Respiratory Medicine without previous clinical suspicion of ILD. As ILAs are in smokers and 2–7% in non-smokers (table).3–9 However,
(Prof A U Wells MD) and associated with respiratory symptoms, functional impair­ their presence is not routinely recorded on radiology
Department of Histopathology
(Prof A G Nicholson DM), Royal
ment, risk of progression, and increased all-cause reports, even at academic centres.20 ILAs are likely to be
Brompton and Hospital NHS mortality,2–10 their identification has clinical implications. increasingly identified with the implementation of lung
Foundation Trust, London, UK; The term ILAs does not imply the absence of respiratory cancer screening and increased use of CT for other
National Heart and Lung signs, symptoms, or functional impairment, but when diagnostic purposes. Still, our understanding of ILAs is
Institute, Imperial College
London, London, UK
these clinically significant findings are present, ILAs are minimal, with insufficient evidence to provide definitive
(Prof A U Wells, likely to represent mild ILD rather than subclinical management recommendations. This Fleischner Society
Prof A G Nicholson); Position Paper provides multidisciplinary perspectives on
Department of Thoracic definition and terminology; predisposing risk factors;
Imaging, Hospital Calmette, Key messages
University Centre of Lille, Lille,
clinical outcomes; options for initial evaluation,
France (Prof M Remy-Jardin MD);
Background monitoring, and management; the role of quantitative
Department of Radiology, • Early interstitial lung abnormalities (ILAs) are common evaluation; and future research needs.
University Hospitals Leuven, incidental findings on CT, particularly in older individuals
Leuven, Belgium
(Prof J A Verschakelen MD);
• The presence of ILAs is an independent predictor of What definition and terminology could be used
Department of Pathology mortality to describe and characterise ILAs?
(M B Beasley MD), and • About 20% of ILAs progress over 2 years, and more than High-resolution CT is highly sensitive for detecting
Pulmonary, Critical Care and 40% progress over 5 years subclinical interstitial abnormalities in high-risk
Sleep Medicine
• Individuals with subpleural predominant fibrotic ILAs are populations, such as patients with connective tissue
(Prof C A Powell MD), Icahn
School of Medicine at Mount, most likely to progress disease (eg, systemic sclerosis) or occupational exposures
New York, NY, USA;
Management of ILAs (eg, asbestos).21–23 Systematic evaluation of large cohorts
Department of Environmental
• Identify potential risk factors for interstitial lung disease of smokers screened by CT for lung cancer, or undergoing
Health, Harvard T.H. Chan
School of Public Health, • Identify clinical or functional impairment CT as part of epidemiological evaluation of chronic
Boston, MA, USA • Establish whether there is current evidence of clinically obstructive pulmonary disease (COPD) or cardiovascular
(Prof D C Christiani MD);
significant interstitial lung disease risk factors, has shown that these incidental abnormalities
Massachusetts General
• Undertake clinical and imaging follow-up as appropriate are relatively common, particularly in older individuals
Hospital, Harvard Medical
(table).4,7,9,15 Terms applied to this finding have included

726 www.thelancet.com/respiratory Vol 8 July 2020

 Position Paper

School, Boston, MA, USA

Population-based cohorts Smoking and lung cancer screening cohorts (Prof D C Christiani);
MESA11,12,13,14 Nagano, FHS6,8,9 AGES- ECLIPSE9 NLST7,16 COPDGene4,9,17 MILD18 DLCST19 Department of Radiology, Asan
Japan*15 Reykjavik9 Medical Center, University of
Ulsan College of Medicine,
Study characteristics Seoul, South Korea
Total number of chest 3137 3061 2633 5320 1670 884 9292 692 1990 (Prof J B Seo MD); Department
CT scans evaluated of Radiology, Kansai Rosai
Prevalence of ILAs 310 (10%) 80 (3%) 177 (7%) 377 (7%) 157 (9%) 86 (10%) 708 (8%) 28 (4%) 332 (17%) Hospital, Hyogo, Japan
(Prof T Johkoh MD); Department
Mean age of those with ILAs 75 62 70 78 64 62 64 60 60
of Medicine and Surgery,
(years)
University of Parma, Parma,
Radiological progression Italy (N Sverzellati MD);
Overall progression, NA 46%, 43%, 63%, NA 20%, NA 20%, NA Department of Medicine,
follow-up time 4 years 6 years 5 years 2 years 2 years University of British Columbia
and Centre for Heart Lung
Mortality
Innovations, St Paul’s Hospital,
Relative risk of death, NA NA 2·7 1·3 1·4 NA 1·8 NA 2·0 Vancouver, BC, Canada
(hazard ratio [95% CI]) (1·1–6·5) (1·2–1·4) (1·1–2·0) (1·1–2·8) (1·4–2·7) (C J Ryerson MD); Department
ILAs=interstitial lung abnormalities. NA=not available. *Patients participating in a health screening programme from Nagano prefecture, Japan. of Medicine and Department of
Epidemiology
Table: Interstitial lung abnormalities across study populations (Prof R Graham Barr MD), and
Department of Radiology
(Prof J H M Austin, MD),
Columbia University Medical
interstitial lung changes at an early phase,15 early ILD,4 published series. Other findings not considered as ILAs Center, New York, NY, USA;
Department of Radiology,
ILD,18 subclinical ILD,23 and preclinical ILD.24 Quantitative are shown in panel 1 and figure 2.
Seoul National University
abnormalities, such as an abnormally high proportion of Ensuring that specific descriptors of CT findings are College of Medicine, Seoul,
high-attenuation areas of the lung, have also been provided in radiology reports is essential, as different South Korea (Prof J M Goo MD);
identified in cohort studies and are thought to suggest imaging findings have very different implications. Department of Radiology,
Samsung Medical Center,
subclinical parenchymal lung disease.11 Relevant descriptors include craniocaudal and axial
Sungkyunkwan University
ILAs are not synonymous with subclinical ILD because a distribution and individual features, such as ground-glass School of Medicine, Seoul,
subset of patients with ILAs has symptoms and lung or reticular abnormalities, traction bronchiectasis, South Korea (Prof K S Lee MD);
function impairment without suspected ILD. A further architectural distortion, honeycombing, and non-emphy­s­ and Clinical Research Center,
National Hospital Organization
subset of patients with ILAs is at risk of progression to clin­ ematous cysts (panel 1). Among these findings, the
Kinki-Chuo Chest Medical
ically significant disease. Abnormalities identified during following subcategories are of prognostic significance: Center, Osaka, Japan
screening for ILD in high-risk groups (eg, those with first, ground-glass opacity and reticular opacities without a (Prof Y Inoue MD)
rheumatoid arthritis, systemic sclerosis, or familial ILD) predominant subpleural localisation; second, ground- Correspondence to:
are not considered as ILAs because they are not incidental; glass opacity and reticular opacities with a predominant Prof Hiroto Hatabu, Department
of Radiology, Center for
these might be referred to as preclinical ILD and their subpleural localisation without evidence of fibrosis;
Pulmonary Functional Imaging,
management is beyond the scope of this Position Paper. and finally, traction bronchiectasis, architectural distor­ Brigham and Women’s Hospital,
ILAs have been described as non-dependent abnor­ tion, and honeycombing, providing evidence of lung Harvard Medical School, Boston,
malities affecting more than 5% of any lung zone (upper, fibrosis.7,29,30,33 Non-emphysematous cysts, defined as MA 02215, USA
hhatabu@partners.org
middle, and lower lung zones are demarcated by the levels lucencies with irregular, well-defined walls, are often seen
of the inferior aortic arch and right inferior pulmonary in cigarette smokers34 with or without other features of
vein). In initial descriptions, ILAs included ground-glass ILAs. These cysts can be distinguished from emphy­sema
or reticular abnormalities, diffuse centri­lobular nodularity, by the presence of a well-defined wall and from
traction bronchiectasis, honeycomb­ing, and non-emphy­ honeycombing by their irregular shape, varying size, and
sematous cysts (figure 1).4 In the definition proposed in the absence of subpleural predominance.35 On histo­logical
this Position Paper, centrilobular nodul­arity, which is the analysis, non-emphysematous cysts usually correlate with
typical presentation of smoking-related respiratory bron­ airspace enlargement and fibrosis or smoking-related
chiolitis,27,28 is not included as this feature is common, interstitial fibrosis,34,35 and might have prognostic
typically non-progressive, and not associated with fibrosis significance,7,29,30 although they are not usually associated
(panel 1).7,29,30 Although the 5% threshold is arbitrary and with imaging evidence of fibrosis.
imprecise, it is retained to exclude minimal opacities and ILAs with a non-subpleural distribution are usually non-
to conform to previous published literature. Focal or progressive29 and not associated with increased mortality.
unilateral patchy ground-glass opacity seldom represents Subpleural ILAs have potentially greater clinical
an ILD, and is classified as equivocal. Dependent significance and are further subcategorised according to
abnormalities are regarded as equivocal unless persistent the presence or absence of fibrosis (figure 1, panel 1).29
in the prone position. Pleuropulmonary fibroelastosis, Fibrotic ILAs are associated with a higher rate of
sometimes an incidental finding on CT, is a clearly defined progression and death on 5-year follow-up.7,29 If fibrosis is
entity,31,32 which has not been included within ILAs in present, the pattern can be further classified according to

www.thelancet.com/respiratory Vol 8 July 2020 727

 Position Paper

precursor to idiopathic pulmonary fibrosis (IPF) or other

A B C
progressive fibrotic ILDs.

What CT protocol should be used to evaluate

and follow up patients with ILAs?
When ILAs are detected, a dedicated chest CT examination
could help to confirm and characterise the abnormality,
especially if dependent atelectasis was present, if the initial
Figure 1: Subcategories of interstitial lung abnormalities scan of the lungs was incomplete (eg, an abdominal CT),
(A) Non-subpleural and non-fibrotic. CT shows widespread ground-glass abnormality with central predominance
(circled). (B) Subpleural non-fibrotic. CT shows predominantly subpleural ground-glass and linear abnormality or if the scan was done without thin sections, with an ultra-
without evidence of fibrosis (arrows). (C) Subpleural fibrotic. Traction bronchiectasis and architectural distortion low dose technique, or using first-generation, hybrid-type,
are indicated by the ovals in the lingula and left lower lobe. This pattern would correspond to a probable usual iterative reconstruction methods, all of which might
interstitial pneumonia pattern.25,26 obscure fine lung details. On the dedicated CT examination
(if indicated), thin sections (<1·5 mm) with moderate
edge-enhancing reconstruction are helpful. Prone views
Panel 1: Definitions and subcategories of interstitial lung abnormalities are particularly important to distinguish dependent
What are interstitial lung abnormalities (ILAs)? atelectasis and true interstitial abnormality, whereas
• Incidental identification of non-dependent abnormalities, including ground-glass or expiratory imaging could potentially identify lobular air
reticular abnormalities, lung distortion, traction bronchiectasis, honeycombing, and trapping as a clue to hypersensitivity pneumonitis.
non-emphysematous cysts Potential recommended imaging protocols are outlined in
• Involving at least 5% of a lung zone (upper, middle, and lower lung zones are 2018 guidelines.25 Subsequent CTs to evaluate for pro­
demarcated by the levels of the inferior aortic arch and right inferior pulmonary vein) gression should use similar scanning protocols.
• In individuals in whom interstitial lung disease is not suspected
Pathological correlation of ILA
What are not ILAs? ILA is a radiological term with few published studies on
Imaging findings restricted to: pathological correlates. Pulmonary resection specimens
• Dependent lung atelectasis for lung cancer in current or former cigarette smokers
• Focal paraspinal fibrosis in close contact with thoracic spine osteophytes (figure 2A) have a high frequency of background inter­stitial fibrosis.
• Smoking-related centrilobular nodularity in the absence of other findings (figure 2B) Katzenstein and colleagues36 reported a 60% prevalence of
• Mild focal or unilateral abnormality (figure 2C) clinically occult fibrosis occupying more than 25% of the
• Interstitial oedema (eg, in heart failure) resected lobe. Most of these cases were viewed as smoking-
• Findings of aspiration (patchy ground-glass, tree in bud; figure 2C) related interstitial fibrosis, but usual interstitial pneu­
Preclinical and clinical identification: monia, pulmonary Langerhans’ cell histiocytosis, and
• Preclinical interstitial abnormalities identified during screening of high-risk individuals asbestosis were also found. Similarly, in a larger study by
(eg, those with rheumatoid arthritis, scleroderma, occupational exposure, familial Kawabata and colleagues,37 most cases of fibrosis were
interstitial lung disease) defined as airspace enlarge­ment and fibrosis or respiratory
• Findings in patients with known clinical interstitial lung disease bronchiolitis, the remainder being usual interstitial
pneumonia. ILAs were not specifically identified on CT in
Subcategories of ILAs
these studies, but on the basis of the association between
• Non-subpleural: ILAs without predominant subpleural localisation (figure 1A)
smoking and interstitial fibrosis it seems likely that many
• Subpleural non-fibrotic: ILAs with a predominant subpleural localisation and without
ILAs in smokers represent subclinical smoking-related
evidence of fibrosis* (figure 1B)
fibrosis or macrophage accumulation.38–46 In 2018, Miller
• Subpleural fibrotic: ILAs with a predominant subpleural localisation and with evidence
and colleagues47 evaluated histological correlates of ILAs
of pulmonary fibrosis* (figure 1C)
in 424 lung nodule resections. Of 26 patients with ILAs,
* Fibrosis is characterised by the presence of architectural distortion with traction bronchiectasis or honeycombing (or both). histology showed fibrosis in 19 (73%), with usual
interstitial pneumonia in two (8%) individuals. Of note,
207 (52%) of 398 patients with no ILAs or an indeterminate
the 2018 Fleischner and American Thoracic Society, status also showed histological fibrosis, suggesting that
European Respiratory Society, Japanese Respiratory fibrosis can be below the resolution of imaging. Apart
Society, and Latin American Thoracic Society criteria as from usual interstitial pneumonia, the histological fibrosis
typical, probable, or indeterminate for usual interstitial seemed predominantly smoking related. In a similar
pneu­monia.25,26,29 About 2% of patients in the AGES- study, Hung and colleagues48 found fibrotic ILD in 10% of
Reykjavik cohort had a probable or definite usual 406 specimens from 397 patients, consisting of smoking-
interstitial pneu­monia pattern, were more likely to have related interstitial fibrosis in 7%, usual interstitial
subpleural ILA progression, and had worse survival pneumonia in 1%, non-specific interstitial pneumonia in
compared with individuals without these patterns.29 It 1%, and undefined in 1%. ILAs were present in 10% of
seems likely that fibrotic ILAs might be an important cases with smoking-related interstitial fibrosis and in

728 www.thelancet.com/respiratory Vol 8 July 2020

 Position Paper

100% with usual interstitial pneumonia. Similar to Miller A B C

and colleagues,47 Hung and colleagues48 found fibrotic
changes in 51% of specimens with no radiological ILAs,
although there were no cases of usual interstitial
pneumonia in this category. A small number of cases with
ILAs had granulomatous disease, non-specific interstitial
pneumonia, undefined fibrosis, aspiration, or pulmonary
Langerhans’ cell histiocytosis.48 Overall, these studies
suggest that although usual interstitial pneu­ monia is
sometimes present in patients with ILAs, a larger
proportion of ILAs represent smoking-related changes.
Figure 2: Imaging abnormalities that do not represent interstitial lung abnormalities
However, there is potential bias because the few studies (A) Focal paraspinal fibrosis. Coronal CT shows a linear fibrotic band in the medial right lower lobe, closely related
published focus on findings in smokers or consist of lung to osteophytes (arrow). (B) Centrilobular nodularity in a heavy smoker. CT shows numerous poorly defined
cancer resections. Further study is needed to determine ground-glass centrilobular nodules without other findings of interstitial abnormality (circled). (C) Unilateral mild
focal abnormality. CT shows patchy ground-glass abnormality in the left lower lobe that is thought to be due to
the frequency with which incidental histological fibrotic
aspiration (circled).
changes progress to clinically significant disease. To
facilitate this research, reporting guidelines recommend
that pathologists should record and categorise the familial interstitial pneumonia has been increased copies
presence in resection specimens of non-neoplastic lung of the minor allele of a common variant in the promoter of
parenchymal changes, such as emphysema, respiratory the MUC5B gene (rs35705950).5,57–60 Similarly, the assoc­
bronchiolitis, and interstitial fibrosis (with identification iations between the MUC5B promoter genotype, ILAs,6,17,61
of a discernible pattern if possible).49,50 and ILA progression27,62 have been consistently replicated.
For example, in the Framingham Heart Study, COPDGene
What are the risk factors for ILAs? study, and AGES-Reykjavik cohorts,6,8,29,30,63 each copy of the
Advanced age, a common feature of patients with IPF,49 is minor allele of the MUC5B promoter poly­morphism is
strongly associated with ILAs in almost all studies in associated with between a 1·5 and 2·7 times increase in
which it has been assessed.4,6,7,9,12,15,18,19,51,52 For example, in the risk of presenting with ILAs, particularly in those with
smokers with and without COPD, each 10-year increase in subpleural ILA (eg, the MUC5B promoter genotype is
age was associated with about a 2·2 times increase in the more strongly associated with subpleural abnormalities
odds of detecting ILAs.3 Male sex has also been identified than with centrilobular nodules).5,29 Although additional
as a demographic risk factor in some studies of IPF53 and overlapping findings of genetic association have been
has been associated with ILAs in some,9,16 but not in shown in comparisons of genome-wide association
all,4,6,7,9,12,15,18,19,51,52 cohorts. For example, in smokers with and analyses of IPF5,57–60,64 and ILAs,59 novel genetic association
without COPD, each male patient had about a 1·7 times with ILAs suggests that disorders other than IPF are also
increase in the odds of having ILAs compared with female likely to be present among some research participants
patients.3 with these imaging abnormalities.
Tobacco smoke exposure, commonly cited as an
environmental risk factor for IPF as well as for other What are the clinical outcomes of ILAs?
forms of ILD,53,54 is associated with ILAs in nearly all ILAs have been associated with adverse clinical outcomes
populations in which it has been evaluated.4,6,7,9,12,15,18,51,52 in numerous populations.7–9,12,14–19,29,65–68 These include
ILAs are associated with both the activity (eg, current general population cohorts8,9,12,15,29,67,68 and populations of
smoking status) and the quantity (eg, pack years) of smokers enriched for the presence of COPD or under­
tobacco smoke exposure.4,6,7,9,12,15,18 For example, current going lung cancer screening.7,9,16–19
smokers had about a 1·8 times increase in the odds of
having ILAs compared with former smokers.3 Progression of ILAs
In a general population-based cohort, analyses of Estimates of the rate of imaging progression of ILAs range
participants in MESA showed that self-reported from 20% over 2 years in the National Lung Screening
occupational exposures to vapours, gases, dusts, and Trial7 to 48% over 5 years in the AGES-Reykjavik study
fumes were associated with an increased prevalence of (figure 3).29 Thus, although not all cases of ILAs progress,
high-attenuation areas and ILAs.13 In the MESA-Lung progression is more likely to be detected when followed
study,55 increased exposure to nitrogen oxides, a marker up over longer time periods. Additionally, patients with
of exposure to traffic-related air pollution, was also ILAs without clear pulmonary fibrosis might subsequently
associated with ILAs. Similarly, in the Framingham develop traction bronchiect­ asis, honeycombing, or
Heart Study,56 elemental carbon exposure (another patterns consist­­ent with usual interstitial pneumonia.8,29
common metric of traffic-related air pollution) was However, the proportion of such cases that evolve to usual
associated with ILAs and ILA progression. interstitial pneumonia on long-term follow-up remains
The most consistent genetic risk for both IPF and unclear.

www.thelancet.com/respiratory Vol 8 July 2020 729

 Position Paper

and AGES-Reykjavik cohorts, this increase in mortality

A B
was most strongly associated with imaging progression
of ILAs.8,29 In the AGES-Reykjavik cohort, specific
imaging patterns indicative of pulmonary fibrosis were
associated with earlier mortality.29 In addition to increased
all-cause mortality, ILAs were associated with increased
respiratory mortality in the AGES-Reykjavik cohort.9
Figure 3: Progression from subpleural non-fibrotic to subpleural fibrotic
interstitial lung abnormality
From a Brigham and Women’s Hospital cohort of
High-resolution CT examination obtained in an asymptomatic, 61-year-old patients with systemic inflammatory response syndrome
ex-smoker referred for suspicion of radiographic abnormality. (A) Prone high- or sepsis, ILAs were associated with increased rates of
resolution CT section at the lung bases shows subpleural lung abnormality, acute respiratory distress syndrome and increased in-
primarily ground-glass opacity (arrows). (B) Prone high-resolution CT section
obtained 7 years later shows increased severity and extent of abnormality, with
hospital mortality.65 ILAs are also associated with
new traction bronchiectasis and honeycomb cysts in the anterior and posterior increased mortality in patients with COPD and lung
left lower lobe, indicating the interim development of lung fibrosis (arrows). cancer,9,19,66 and in individuals who undergo transcatheter
This patient was still asymptomatic at the time of this follow-up CT aortic valve replacement.69 Increased quantitative metrics
examination.
of ILAs (based on an increased number of high-
attenuation areas of the lung12 and local histogram-based
Specific imaging features and patterns can identify methods designed to identify ILAs17) are also associated
ILAs that are most likely to progress over a 5-year with increased mortality.
interval.29 For example, in a study by Putman and Although ILAs have been associated with increased
colleagues,29 patients with subpleural reticular changes, mortality from pulmonary fibrosis,9 it is important to
lower lobe predominant changes, or traction bronchi­ recognise that the contribution of ILAs to these elevated
ectasis had more than a six times increase in their odds mortality rates far exceeds the expected rate of progression
of imaging progression than those with ILAs without to clinically detectable ILD. It is also important to note
these features, even after adjusting for important that respiratory-related deaths, which were more common
covariates (eg, age and smoking history). In that study, all among those with ILAs in the AGES-Reykjavik cohort,
cases of honeycombing progressed over 5 years. were reported in less than 15% of those with ILAs.9 This
Conversely, the presence of centrilobular nodules was observation suggests that although some of the
associated with ILAs that were unlikely to progress.29 association between ILAs and death could be due to
ILA progression and lung function decline were pulmonary fibrosis, those patients with ILAs could
explored in a Framingham Heart Study cohort.8 possibly be at an excess risk of death because of
Progression (including both the development of new accelerated physiological ageing or other causes of death
ILAs and the progression of existing ILAs) occurred in that are not directly related to pulmonary disease.
6% of the population over approximately 6 years. Patients
with imaging progression in the Framingham Heart Lung cancer mortality and treatment toxicity
Study had an accelerated decline in forced vital capacity Several studies have shown an association between
(FVC) compared with those patients without ILAs or pretreatment ILAs and cancer-associated mortality,
those with ILAs that did not progress. However, the including patients with early stage cancer treated with
annual decline in FVC in patients with ILA progression surgical resection,70,71 as well as patients with advanced
in the Framingham Heart Study (about 60 mL per year, stage 4 disease treated with systemic therapy.66,72 The cause
with an excess annual decline of about 30 mL per year of increased mortality is not clear, but other studies suggest
compared with those without ILAs) is substantially less that lung injury risk associated with ILAs and cancer
than the annual decline in FVC generally noted among therapies might be important. Specifically, lung irradiation
patients with IPF (approximately 200 mL per year). and systemic treatment with chemotherapy and targeted
Whether the excess FVC decline associated with ILA tyrosine kinase inhibitors, immunotherapy checkpoint
progression on imaging represents a small subgroup inhibitors, and antibody–drug conjugates are associated
with major FVC decline (averaging to a small FVC with an increased risk of pneumonitis in patients with pre-
decline across all progressing patients) or a larger existing ILAs (figure 4). Pre-existing ILAs increase the risk
subgroup with subclinical disease that tends to be less of extensive radiation pneumonitis in patients with early
pronounced than clinically apparent IPF is not clear. stage lung cancer treated with stereotactic body radio­
therapy73 and in patients with small-cell lung cancer treated
Mortality with 50–60 Gy of thoracic radiotherapy.74
One of the most consistent findings with regard to ILAs Immune checkpoint inhibitors have emerged as
is the association with increased mortality, both in standard first-line therapy for patients with advanced non-
general population samples and among populations of small-cell lung cancer and for other malignancies. Overall,
smokers enriched for COPD or undergoing lung cancer the rate of immunotherapy-associated pneumonitis is
screening (table).8,9,16,19,29 In the Framingham Heart Study approximately 5%, and this toxicity is manageable when

730 www.thelancet.com/respiratory Vol 8 July 2020

 Position Paper

recognised early and treated appropriately in accordance

A B
with the National Cancer Institute’s Common Terminology
Criteria for Adverse Events grade at presentation.75,76
Evidence indicates that pneumonitis risk is increased by
ILAs. Nakanishi and colleagues77 examined pretreatment
chest CT scans for ILAs in 83 patients treated with the anti
PD-1 antibodies nivolumab or pembrolizumab. The
incidence of immunotherapy-associated ILD was high at
17% (n=14). Multivariate analysis showed that pre-existing Figure 4: Interstitial lung abnormalities after surgery and chemotherapy for
ILAs were associated with a six times increase in the risk lung cancer
of drug-associated ILD, with a predominant ground-glass High-resolution CT examination in a 79-year-old man with lung cancer.
(A) A preoperative CT showed mild subpleural interstitial abnormality without
pattern of pneumonitis. Given the life-threatening nature evident fibrosis (subpleural non-fibrotic). The patient developed shortness of
of malignancies treated with immune checkpoint breath following a fourth cycle of pemetrexed. (B) Subsequent CT showed
inhibitors, the benefits of the therapy, and the undefined bilateral peribronchovascular ground-glass abnormality compatible with drug-
risks associated with ILAs, clinicians should discuss the related pneumonitis, as well as a postoperative effusion (arrows).
possible increased risk of pneumonitis with patients who
have ILAs. Additionally, clinicians should consider active
monitoring for symptoms, physiological alterations, and ILA identified (see definition in panel 1)
radiological progression of drug-associated pneumonitis
(figure 5). Consider high-resolution CT if initial scan
is incomplete or equivocal
How should ILAs be evaluated and monitored?
To date, only minimal evidence exists to support a specific
Is there evidence of clinically significant disease?
management plan for ILAs. The following proposal is • Respiratory symptoms or physical exam
based on the available published literature and the findings possibly attributable to ILD Yes
Potentially clinically significant ILD
• Extensive disease on CT*
consensus clinical opinion of the authors. The first goal is • Decrements in pulmonary function or gas
to separate those patients with current clinically exchange possibly attributable to ILD
significant disease from individuals who might be at risk No
of developing disease. This distinction could be Evaluation by pulmonologist, ideally
ILA with ILD experience, with access to
established by a series of questions that incorporate a multidisciplinary discussion
general approach to ILD (figure 5). In all patients, a
standard evaluation of potential explanations for the Are there radiological or clinical features (panel 2)
presence of ILAs should occur, including factors such as that indicate increased risk of progression Clinically significant ILD
(eg, usual interstitial pneumonia or probable
cigarette smoking or other inhaled exposures, drug usual interstitial pneumonia pattern)?
toxicity, systemic disease (eg, occult connective tissue Clinically driven management
disorders), or recurrent aspiration of oroesophageal Yes No
contents. Individuals with respiratory symptoms or signs,
Active monitoring Expectant management
clinically relevant pulmonary physiological or gas transfer • Recommend risk factor reduction • Recommend risk factor reduction
abnormalities, or extensive CT abnormality (disease • Reassess clinically and repeat pulmonary • Reassess if symptoms or other evidence
involving three or more of the six lung zones consisting function tests in 3–12 months of progression emerge

of the right and left upper, middle, and lower lung zones)
should be referred for pulmonary evaluation, ideally with Repeat CT at 12–24 months, or sooner
access to multidisciplinary discussion. Management of if there is clinical or physiological
progression
patients with a diagnosed ILD should follow standard
guidelines.
Once ILD is excluded, ILAs can be separated into those Figure 5: Proposed schema for management of interstitial lung abnormalities detected on chest CT
Action items for the radiologist are in blue, action items for the treating physician or pulmonologist are in
at higher risk of progression to ILD and those at lower green, and action items for a pulmonologist, ideally with ILD experience, are in orange. ILA=interstitital lung
risk. Risk factors for progression include cigarette abnormality. ILD=interstitial lung disease. *Non-trivial abnormalities present in three or more lung zones (above
smoking, other inhalational exposures, medications, the bottom of the aortic arch, between the aortic arch and top of the inferior pulmonary vein, and below the
physiological or gas exchange findings not felt to reach inferior pulmonary vein).
the threshold of clinical significance, and specific
radiological features such as evidence of fibrosis and impairment. For example, non-subpleural ILAs seldom
subpleural, basal predominant distribution (panel 2). progress, and individuals with only these findings can be
Follow-up of patients with ILAs can be based on the followed up expectantly. In individuals with one or more
presence of risk factors for progression. Individuals risk factors, systematic follow-up should be considered.
without risk factors should be advised to return for The appropriate timing of repeated clinical evaluation
evaluation if they develop symptoms of respiratory (including a focused history and chest exam, chest

www.thelancet.com/respiratory Vol 8 July 2020 731

 Position Paper

Panel 2: Risk factors for progression of interstitial lung A B

abnormalities
Clinical risk factors
• Cigarette smoking
• Other inhalational exposures
• Medications (eg, chemotherapy, immune checkpoint
inhibitors)
• Radiation therapy
• Thoracic surgery
• Physiological or gas exchange findings at lower limits of
normal
Radiological risk factors
• Non-fibrotic interstitial lung abnormalities (ILAs) with
basal and peripheral predominance Figure 6: Computer-based classification of interstitial lung abnormalities
• Fibrotic ILAs with basal and peripheral predominance but with the histogram approach
without honeycombing (ILAs with probable usual (A) CT images with subpleural non-fibrotic interstitial lung abnormalities and
emphysema in a participant in the COPDGene study. (B) CT image overlays of
interstitial pneumonia pattern) computer-based classification of interstitial lung abnormalities using artificial
• Fibrotic ILAs with basal and peripheral predominance and intelligence, showing objective quantification of different injury patterns.
honeycombing (ILAs with usual interstitial pneumonia Regions of interstitial lung abnormality are shown in blue. Normal parenchyma
pattern) (red), emphysema (green), and paraseptal emphysema (yellow) are also
subtyped.88

imaging, pulmonary physiology, and gas exchange) is ventilation. Medications that are known to cause ILD
unknown. In the absence of prospective data, clinical should be avoided if possible.
experience suggests that a first follow-up at 3–12 months
to look for symptomatic or physiological progression is What is the role of quantitative evaluation?
probably appropriate in most patients at increased risk. Methods for quantitative evaluation of ILAs include
Individuals with ILAs are likely to benefit from additional assessment of the proportion of high-attenuation areas,
clinical follow-up beyond 1 year, but the optimal local histogram evaluation, and deep learning-based
frequency and duration of follow-up is unknown. textural evaluation. Automatic quantification of CT density
Similarly, the optimal interval for follow-up CT scans is of the lungs has been used to identify the proportion of
unknown, but might include a follow-up scan at lung voxels with high-attenuation areas, typically between
12–24 months or sooner in patients who develop –600 and –250 Hounsfield units (the normal CT
symptoms or impaired pulmonary function. Progression attenuation of the lung is about –750 Hounsfield units).12,13
can be defined by the develop­ment of clinically signif­ High-attenuation areas are associated with elevated serum
icant respiratory symptoms and signs (eg, the new concentrations of inflammatory biomarkers, reduced FVC
presence of exercise limitation or characteristic crackles and exercise capacity, and higher mortality (including
on auscultation, or both), the development of abnormal higher mortality from ILD).12,55 The presence of high-
pulmonary physio­logy or gas exchange (or a clinically attenuation areas is also associated with a higher
significant decline in normal values), or an increase in prevalence of ILAs at follow-up CT.78 Despite these clear
the extent of CT abnormal­ ities, particularly with the epidemiological associations, high-attenuation areas
development of specific fibrotic features. Optimal appear to be neither sensitive nor specific for subsequent
management of progressive ILAs is unknown, so this appearance of ILAs.79 For this reason, the use of the term
patient group might be an appropriate population for a subclinical ILD as a synonym for increased high-
prospective treatment trial. attenuation areas is not recommended. The clinical
In patients with ILAs undergoing surgery or other significance of high-attenuation areas remains unclear,
therapy, caution should be exercised because they appear and assessment in individual patients is limited by the
to be at increased risk of rapid disease acceleration or an multiple technical and patient-related reasons for an
acute exacerbation. The clinician should regard ILAs as elevated proportion of high-attenuation areas, including
an important comorbidity that should be considered in scanner variation, inadequate inspiration, obesity, and
planning treatment and subsequent monitoring. Because pulmonary atelectasis.24
positive pressure ventilation might be a risk factor for Quantitative imaging provides an objective recognition
developing acute respiratory distress in patients with of regional disease pattern of the lung that can increase the
ILAs, a low-volume, low-pressure ventilatory approach diagnostic reliability and severity assessment of ILD.
should be considered for those needing mechanical Computer-based CT approaches to identify interstitial

732 www.thelancet.com/respiratory Vol 8 July 2020

 Position Paper

A B Panel 3: Key uncertainties with interstitial lung

abnormalities
• Reproducibility of radiological criteria
• Validity and efficacy of follow-up regimen
• Prevalence in the lung cancer screening population
• Prevalence in younger cohorts
C D • Risk factors for progression
• Natural history of non-fibrotic interstitial lung
abnormalities (ILAs)
• Optimal extent thresholds for predicting significant
physiological progression, development of clinically
significant disease, and mortality by visual and
Figure 7: Quantification of progression for interstitial lung abnormalities quantitative CT evaluation
with data-driven texture analysis • Importance of incidentally identified histological evidence
(A) A baseline CT scan shows subpleural non-fibrotic interstitial lung
of interstitial abnormality
abnormalities with fibrotic changes. (B) CT 5 years later shows clear progression.
(C) Baseline data-driven textural analysis shows overall extent of fibrosis as 1·5% • Quantitative techniques: predictive value for adverse
(red). (D) Data-driven textural analysis of follow-up scan at 5 years shows that outcome, technical variability, and inter-patient variability
the extent of fibrosis increased to 4·6% (red). • Role of biomarkers in predicting progression
• Strategy for cohort enrichment in clinical trials for
subtypes are based on density histogram analysis, texture- patients with ILAs that are likely to progress
based analysis, and deep learning approaches.80–87 In • Risk factors and preventive strategies for complications of
general, these approaches are sensitive enough to detect cancer treatment surgery, chemotherapy,
ILD in patients at high risk and provide a more reproducible immunotherapy, and radiotherapy in patients with ILAs
assessment than visual CT scoring.79 For ILAs, individuals
with a lower percentage of normal lung by local histogram
measurements had a higher prevalence of clinical impair­ practice, including in lung cancer screening (panel 3).
ment, poorer quality of life, higher risk of death, and The effect of the proposed management plan on inter­
association with the common variant in the promoter of mediate and long-term outcomes must be evaluated
the MUC5B gene (figure 6).17 Using a different local (figure 5). To under­ stand the prevalence and natural
histogram-based system in 217 individuals undergoing course of ILAs in the lung cancer screening population,
resection for lung cancer, the fibrosis score correlated ILAs should be considered as a specific subcategory
with the presence of ILAs and was an independent under the significant other findings modifier in the
predictor of decreased disease-free survival.71 A study in LungRADS scoring system,95 as used in the USA; the
family members of individuals with familial pulmonary Korean Society of Radiology has already implemented a
fibrosis showed that data-driven texture analysis could similar change.96,97 In addition to clarifying criteria for
detect early interstitial changes with 84% sensitivity and visual evaluation of ILAs, quantitative CT methods for
86% specificity (figure 7).89 However, the role of quantitative evaluation of disease extent and determination of
CT as a screening tool for ILAs requires further validation. progress will need to be developed and validated. An
Potential sources of variation in quantitative imaging important element will be the determination of optimal
of ILAs include dependence on training data, variation thresholds on visual and quantitative CT that define
in inspired lung volumes, sensitivity to image noise significant disease by predicting significant physio­logical
from CT acquisition dose, vendor differences and re­ progression, the development of clinically signifi­ cant
construction method, and variation in segmenting the disease, and mortality.
subpleural fibrotic lung from the chest wall. Annotated Since antifibrotic treatment slows the rate of physio­
datasets are needed to provide a reference benchmark to logical progression in patients with IPF98,99 and in other
establish the robustness of each approach. Considering forms of progressive lung fibrosis,100,101 it is possible that
that the characteristics of ILAs are subtle and varied, the early treatment in a high-risk population with ILAs
stability of assessment by computer-based analysis might reduce the rate of progression. However, there is
should be tested, improved, and applied in further a clear need for further epidemiological, biomarker,
studies. New advances in artificial intelligence and deep and machine learning studies in existing and novel
learning might overcome some of these limitations.86,87,90–94 cohorts to identify groups at higher risk of progression
and to understand the trajectory of progression of
Outline of future research needs and priorities ILAs to clinically significant pulmonary fibrosis.
The preliminary radiological criteria for ILAs presented The availability of this information would potentially
in this Position Paper require robust evaluation to support the design of future clinical trials in higher-risk
determine their reproducibility and application to clinical individuals.

www.thelancet.com/respiratory Vol 8 July 2020 733

 Position Paper

Panel 4: Recommendations for the evaluation and Search strategy and selection criteria
reporting of interstitial lung abnormalities
A medical librarian searched in Medline, Embase, Cochrane
CT protocol Central Registry of Controlled Trials, and the Health
• Thin sections (<1·5 mm) are essential Technology Assessment database to identify publications
• Prone and expiratory scans might be necessary to confirm related to interstitial lung abnormalities. We included studies
and characterise interstitial lung abnormalities (ILAs) from database inception through to Feb 13, 2019, and
restricted to English language. Details of the search strategy
CT description
See Online for appendix are provided in the appendix (pp 1–3). Key search terms were
• Axial and craniocaudal distribution
“interstitial$”, “lung”, “abnormal$”, and “subclinical or
• CT findings: including ground-glass abnormality, reticular
pre-clinical or preclinical”. The literature search resulted in
abnormality, traction bronchiectasis, honeycombing, and
700 references, of which 616 were excluded (duplicates
cysts
[n=11] and 605 references with little relevance to the key
• CT category: non-subpleural ILA, subpleural non-fibrotic
questions based on screening of the reference title [n=455] or
ILA, or subpleural fibrotic ILA
the reference abstract [n=150]), yielding 84 manuscripts that
Clinical evaluation underwent review for inclusion. Review of the text found that
• Distinguish ILAs from clinically significant interstitial lung 60 of these manuscripts were not relevant to the key
disease (figure 5) questions, resulting in 24 references that were analysed for
• Identify risk factors for progression (panel 2) the final Position Paper. Additional references were added by
• Follow-up evaluation (figure 5) members of the writing group.

Pathology evaluation
• On lung cancer resections, assess background lung from of non-fibrotic ILAs, of smoking-related interstitial
cancer resections and document histological patterns fibrosis, and of pleuroparenchymal fibro­elastosis. There is
diagnostic of suspicion for interstitial lung disease also a need for better understanding of risk factors for the
• Review such cases in a multidisciplinary team setting to development of ILAs. Finally, the recognition of ILAs
determine whether ILAs or clinically significant interstitial offers exciting opportunities for identifying pathogenetic
lung disease is present abnormalities in early pulmon­ary fibrosis and early usual
interstitial pneumonia; sequential evaluation of bio­
The importance of ILAs as a risk factor for complications markers in individuals with ILAs might help to identify
in the treatment of lung cancer requires further evaluation, biological abnormalities that predispose to subsequent
for example, in clinical trials of checkpoint inhibitors development of IPF.
with a focus on prevention and management of acute
pneumonitis. This analysis could be done retrospectively Conclusions
from existing CT datasets in clinical trials and could ILAs are important because they are associated with
inform a prospective comparison of specific approaches to mortality as well as increased risk of complications from
preventing progression in the context of thoracic surgery, surgery, chemotherapy, and radiotherapy. Separating
radiotherapy, and chemotherapy. The role of ILAs as a clinically significant ILD from ILAs is essential. The
predictor of acute interstitial pneumonia also merits morphology and distribution of ILAs should be clearly
further evaluation, given that unsuspected usual interstitial described and the descriptive categories of non-sub­
pneumonia has been identified in 50% of patients who die pleural, subpleural non-fibrotic, and subpleural fibrotic
with acute interstitial pneumonia.102 We have excluded ILAs should be recorded in the radiology report, as this
preclinical interstitial abnormalities identified during the information could be useful in predicting progression and
screening of individuals at high risk (eg, those with mortality (panel 4). Risk factors for ILAs include age,
rheumatoid arthritis, scleroderma, occupational exposure, cigarette smoking and other inhalational exposures, and
and familial interstitial lung disease) from the scope of this genetic markers. Although this Position Paper proposes a
Position Paper to simplify the approaches. However, future rational strategy that can help to identify when ILAs are
studies and discussions are needed to investigate the role likely to represent clinically significant ILD, future work is
of ILAs in preclinical interstitial abnormalities identified needed to determine the best approach to follow up ILAs
during screening of individuals at high risk.” in individuals in whom the evaluation is less definitive.
Further study is needed to determine the importance of We believe that establishing a common terminology for
incidentally found histological fibrosis to determine which communication and a clear under­ standing of current
cases are more likely to progress to clinically significant knowledge are important steps towards further advances
disease. In particular, there is an opportunity to clarify the in the multidisciplinary approach to ILAs.
effect of specific histological findings and cellular Contributors
subpopulations on long-term outcome. Additional related DAL developed and implemented the systematic search strategy.
questions include the natural history and biological cause CJR advised on the systematic search. All authors participated in the

734 www.thelancet.com/respiratory Vol 8 July 2020

 Position Paper

literature search. HH, GMH, LR, KKB, AUW, MR-J, AGN, MBB, DCC, 5 Seibold MA, Wise AL, Speer MC, et al. A common MUC5B
RSJE, CAP, KSL, YI, and DAL wrote the first draft of the Position promoter polymorphism and pulmonary fibrosis. N Engl J Med 2011;
Paper. All authors critically reviewed the manuscript and approved the 364: 1503–12.
final version, taking accountability for the work. The Document 6 Hunninghake GM, Hatabu H, Okajima Y, et al. MUC5B promoter
Development and Oversight Committee and the Executive Committee polymorphism and interstitial lung abnormalities. N Engl J Med
of the Fleischner Society approved the manuscript before submission 2013; 368: 2192–200.
to the journal. 7 Jin GY, Lynch D, Chawla A, et al. Interstitial lung abnormalities in a
CT lung cancer screening population: prevalence and progression
Declaration of interests rate. Radiology 2013; 268: 563–71.
HH reports grants from Canon Medical Systems and Konica Minolta; 8 Araki T, Putman RK, Hatabu H, et al. Development and progression
and personal fees from Canon Medical Systems and Mitsubishi of interstitial lung abnormalities in the Framingham Heart Study.
Chemical, outside the submitted work. GMH reports personal fees Am J Respir Crit Care Med 2016; 194: 1514–22.
from Boehringer Ingelheim, Gerson Lehrman Group, and Mitsubishi 9 Putman RK, Hatabu H, Araki T, et al. Association between
Chemical, outside the submitted work. LR reports personal fees from interstitial lung abnormalities and all-cause mortality. JAMA 2016;
Biogen and ImmuneWorks; and grants and personal fees from Roche, 315: 672–81.
Boehringer Ingelheim, Celgene, Nitto, FibroGen, Promedior, Pliant 10 Doyle TJ, Washko GR, Fernandez IE, et al. Interstitial lung
Therapeutics, Asahi Kasei, Toray, Bristol-Myers Squibb, RespiVant, and abnormalities and reduced exercise capacity.
CSL Behring, outside the submitted work. KKB reports grants from the Am J Respir Crit Care Med 2012; 185: 756–62.
National Institutes of Health; and advisory board participation for 11 Lederer DJ, Enright PL, Kawut SM, et al. Cigarette smoking is
Biogen, Blade, Boehringer Ingelheim, Galapagos, Galecto, Genoa, associated with subclinical parenchymal lung disease: the Multi-
Lifemax, MedImmune, the Open Source Imaging Consortium, Pliant, Ethnic Study of Atherosclerosis (MESA)-Lung study.
ProMetic, Third Pole, Theravance, Three Lakes Partners, and Veracyte, Am J Respir Crit Care Med 2009; 180: 407–14.
outside the submitted work. AUW reports personal fees from Roche, 12 Podolanczuk AJ, Oelsner EC, Barr RG, et al. High attenuation areas
Boehringer Ingelheim, and Blade, outside the submitted work. on chest computed tomography in community-dwelling adults:
the MESA study. Eur Respir J 2016; 48: 1442–52.
AGN reports personal fees from Medical Quantitative Image Analysis,
Galapagos, Boehringer Ingelheim, and Roche, outside the submitted 13 Sack CS, Doney BC, Podolanczuk AJ, et al. Occupational exposures
and subclinical interstitial lung disease. Am J Respir Crit Care Med
work. RSJE reports grants from the National Heart, Lung, and Blood
2017; 196: 1031–39.
Institute, during the conduct of the work; personal fees from
14 Podolanczuk AJ, Oelsner EC, Barr RG, et al. High-attenuation areas
LeukoLabs, Boehringer Ingelheim, and Chiesi; and grants from
on chest computed tomography and clinical respiratory outcomes
Boehringer Ingelheim, outside the submitted work. RSJE is also a in community-dwelling adults. Am J Respir Crit Care Med 2017;
founder and co-owner of Quantitative Imaging Solutions, which is a 196: 1434–42.
company that provides image-based consulting and develops software 15 Tsushima K, Sone S, Yoshikawa S, Yokoyama T, Suzuki T, Kubo K.
to enable data sharing. CJR reports grants and personal fees from The radiological patterns of interstitial change at an early phase:
Boehringer Ingelheim and Hoffmann-La Roche, outside the submitted over a 4-year follow-up. Respir Med 2010; 104: 1712–21.
work. RGB reports grants from the National Institutes of Health, 16 Pompe E, de Jong PA, Lynch DA, et al. Computed tomographic
during the conduct of the work; and grants from the Alpha1 findings in subjects who died from respiratory disease in the
Foundation and the COPD Foundation, outside the submitted work. National Lung Screening Trial. Eur Respir J 2017; 49: 1601814.
JMG reports grants from Infinitt Healthcare and Dongkook 17 Ash SY, Harmouche R, Putman RK, et al. Clinical and genetic
Lifescience, outside the submitted work. CAP reports personal fees associations of objectively identified interstitial changes in smokers.
from Daiichi Sankyo and AstraZeneca, outside the submitted work. Chest 2017; 152: 780–91.
YI reports grants from the Japanese Ministry of Health, Labour, and 18 Sverzellati N, Guerci L, Randi G, et al. Interstitial lung diseases in a
Welfare, and the Japan Agency for Medical Research and Development; lung cancer screening trial. Eur Respir J 2011; 38: 392–400.
and other support from Boehlinger Ingelhem and Shionogi, outside 19 Hoyer N, Wille MMW, Thomsen LH, et al. Interstitial lung
the submitted work. DAL reports personal fees from Boehringer abnormalities are associated with increased mortality in smokers.
Ingelheim, Parexel, and Veracyte, outside the submitted work. Respir Med 2018; 136: 77–82.
DAL also has a pending patent: systems and methods for automatic 20 Oldham JM, Adegunsoye A, Khera S, et al. Underreporting of
detection and quantification of pathology using dynamic feature interstitial lung abnormalities on lung cancer screening computed
classification (US Patent 15/595,260). All other authors declare no tomography. Ann Am Thorac Soc 2018; 15: 764–66.
competing interests. 21 Aberle DR, Gamsu G, Ray CS, Feuerstein IM. Asbestos-related
pleural and parenchymal fibrosis: detection with high-resolution
Acknowledgments CT. Radiology 1988; 166: 729–34.
We are grateful for the help of Shandra Lee Knight (librarian at National 22 Remy-Jardin M, Remy J, Wallaert B, Bataille D, Hatron PY.
Jewish Health, Denver, USA), who assisted with the development of the Pulmonary involvement in progressive systemic sclerosis:
systematic search strategy and did the literature search. We are also sequential evaluation with CT, pulmonary function tests, and
grateful for the librarians Myung-Ah Shim and Jaero Park for their bronchoalveolar lavage. Radiology 1993; 188: 499–506.
dedicated support in formatting the manuscript. Both librarians are 23 Doyle TJ, Hunninghake GM, Rosas IO. Subclinical interstitial lung
working at the Samsung Medical Information and Media Services of disease: why you should care. Am J Respir Crit Care Med 2012;
Samsung Medical Center located in Seoul, South Korea. No external 185: 1147–53.
source of funding was used in the writing of the manuscript or the 24 Salisbury ML, Lynch DA. Toward early identification of clinically
decision to submit for publication. relevant interstitial lung disease. Am J Respir Crit Care Med 2017;
196: 1368–69.
References
25 Raghu G, Remy-Jardin M, Myers JL, et al. Diagnosis of idiopathic
1 Schwarz MI, King TE Jr. Interstitial lung disease, 5th edn. Shelton,
pulmonary fibrosis. An official ATS/ERS/JRS/ALAT clinical practice
CT: People’s Medical Publishing House, 2011.
guideline. Am J Respir Crit Care Med 2018; 198: e44–68.
2 Hatabu H, Hunninghake GM, Lynch DA. Interstitial lung
26 Lynch DA, Sverzellati N, Travis WD, et al. Diagnostic criteria for
abnormality: recognition and perspectives. Radiology 2019;
idiopathic pulmonary fibrosis: a Fleischner Society White Paper.
291: 1–3.
Lancet Respir Med 2018; 6: 138–53.
3 Washko GR, Lynch DA, Matsuoka S, et al. Identification of early
27 Remy-Jardin M, Remy J, Boulenguez C, Sobaszek A, Edme JL,
interstitial lung disease in smokers from the COPDGene Study.
Furon D. Morphologic effects of cigarette smoking on airways and
Acad Radiol 2010; 17: 48–53.
pulmonary parenchyma in healthy adult volunteers: CT evaluation
4 Washko GR, Hunninghake GM, Fernandez IE, et al. Lung volumes and correlation with pulmonary function tests. Radiology 1993;
and emphysema in smokers with interstitial lung abnormalities. 186: 107–15.
N Engl J Med 2011; 364: 897–906.

www.thelancet.com/respiratory Vol 8 July 2020 735

 Position Paper

28 Remy-Jardin M, Remy J, Gosselin B, Becette V, Edme JL. 50 Nicholson AG, Kerr K, Gosney J. Standards and datasets for reporting
Lung parenchymal changes secondary to cigarette smoking: cancers. Dataset for histopathological reporting of lung cancer.
pathologic-CT correlations. Radiology 1993; 186: 643–51. September, 2018. https://www.rcpath.org/uploads/assets/265cdf74–
29 Putman RK, Gudmundsson G, Axelsson GT, et al. Imaging patterns 3376–40b0-b7d0e3ed8a588398/G048-Dataset-for-histopathological-
are associated with interstitial lung abnormality progression and reporting-of-lung-cancer.pdf (accessed June 11, 2020).
mortality. Am J Respir Crit Care Med 2019; 200: 175–83. 51 Kropski JA, Pritchett JM, Zoz DF, et al. Extensive phenotyping of
30 Putman RK, Gudmundsson G, Araki T, et al. The MUC5B promoter individuals at risk for familial interstitial pneumonia reveals clues
polymorphism is associated with specific interstitial lung to the pathogenesis of interstitial lung disease.
abnormality subtypes. Eur Respir J 2017; 50: 1700537. Am J Respir Crit Care Med 2015; 191: 417–26.
31 Chua F, Desai SR, Nicholson AG, et al. Pleuroparenchymal 52 Rosas IO, Ren P, Avila NA, et al. Early interstitial lung disease in
fibroelastosis: a review of clinical, radiological and pathological familial pulmonary fibrosis. Am J Respir Crit Care Med 2007;
characteristics. Ann Am Thorac Soc 2019; 16: 1351–59. 176: 698–705.
32 Watanabe K, Ishii H, Kiyomi F, et al. Criteria for the diagnosis of 53 Raghu G, Collard HR, Egan JJ, et al. An official ATS/ERS/JRS/ALAT
idiopathic pleuroparenchymal fibroelastosis: a proposal. statement: idiopathic pulmonary fibrosis: evidence-based guidelines
Respir Investig 2019; 57: 312–20. for diagnosis and management. Am J Respir Crit Care Med 2011;
33 Hida T, Nishino M, Hino T, et al. Traction bronchiectasis/ 183: 788–824.
bronchiolectasis is associated with interstitial lung abnormality 54 American Thoracic Society. Idiopathic pulmonary fibrosis:
mortality. Eur J Radiol 2020; 129: 109073. diagnosis and treatment. International consensus statement.
34 Otani H, Tanaka T, Murata K, et al. Smoking-related interstitial American Thoracic Society (ATS), and the European Respiratory
fibrosis combined with pulmonary emphysema: computed Society (ERS). Am J Respir Crit Care Med 2000; 161: 646–64.
tomography-pathologic correlative study using lobectomy 55 Sack C, Vedal S, Sheppard L, et al. Air pollution and subclinical
specimens. Int J Chron Obstruct Pulmon Dis 2016; 11: 1521–32. interstitial lung disease: the Multi-Ethnic Study of Atherosclerosis
35 Watanabe Y, Kawabata Y, Kanauchi T, et al. Multiple, thin-walled (MESA) air-lung study. Eur Respir J 2017; 50: 1700559.
cysts are one of the HRCT features of airspace enlargement with 56 Rice MB, Li W, Schwartz J, et al. Ambient air pollution exposure
fibrosis. Eur J Radiol 2015; 84: 986–92. and risk and progression of interstitial lung abnormalities:
36 Katzenstein AL, Mukhopadhyay S, Zanardi C, Dexter E. Clinically the Framingham Heart Study. Thorax 2019; 74: 1063–69.
occult interstitial fibrosis in smokers: classification and significance 57 Fingerlin TE, Murphy E, Zhang W, et al. Genome-wide association
of a surprisingly common finding in lobectomy specimens. study identifies multiple susceptibility loci for pulmonary fibrosis.
Hum Pathol 2010; 41: 316–25. Nat Genet 2013; 45: 613–20.
37 Kawabata Y, Hoshi E, Murai K, et al. Smoking-related changes in 58 Noth I, Zhang Y, Ma SF, et al. Genetic variants associated with
the background lung of specimens resected for lung cancer: idiopathic pulmonary fibrosis susceptibility and mortality:
a semiquantitative study with correlation to postoperative course. a genome-wide association study. Lancet Respir Med 2013; 1: 309–17.
Histopathology 2008; 53: 707–14. 59 Fingerlin TE, Zhang W, Yang IV, et al. Genome-wide imputation
38 Travis WD, Costabel U, Hansell DM, et al. An official American study identifies novel HLA locus for pulmonary fibrosis and
Thoracic Society/European Respiratory Society statement: update of potential role for auto-immunity in fibrotic idiopathic interstitial
the international multidisciplinary classification of the idiopathic pneumonia. BMC Genet 2016; 17: 74.
interstitial pneumonias. Am J Respir Crit Care Med 2013; 188: 733–48. 60 Allen RJ, Porte J, Braybrooke R, et al. Genetic variants associated
39 Kawabata Y, Takemura T, Hebisawa A, et al. Desquamative with susceptibility to idiopathic pulmonary fibrosis in people of
interstitial pneumonia may progress to lung fibrosis as European ancestry: a genome-wide association study.
characterized radiologically. Respirology 2012; 17: 1214–21. Lancet Respir Med 2017; 5: 869–80.
40 Yousem SA. Respiratory bronchiolitis-associated interstitial lung 61 Armanios MY, Chen JJ, Cogan JD, et al. Telomerase mutations in
disease with fibrosis is a lesion distinct from fibrotic nonspecific families with idiopathic pulmonary fibrosis. N Engl J Med 2007;
interstitial pneumonia: a proposal. Mod Pathol 2006; 19: 1474–79. 356: 1317–26.
41 Vassallo R, Jensen EA, Colby TV, et al. The overlap between 62 Cronkhite JT, Xing C, Raghu G, et al. Telomere shortening in
respiratory bronchiolitis and desquamative interstitial pneumonia familial and sporadic pulmonary fibrosis. Am J Respir Crit Care Med
in pulmonary Langerhans cell histiocytosis: high-resolution CT, 2008; 178: 729–37.
histologic, and functional correlations. Chest 2003; 124: 1199–205. 63 Hobbs BD, Putman RK, Araki T, et al. Overlap of genetic risk
42 Ryu JH, Colby TV, Hartman TE, Vassallo R. Smoking-related between interstitial lung abnormalities and idiopathic pulmonary
interstitial lung diseases: a concise review. Eur Respir J 2001; fibrosis. Am J Respir Crit Care Med 2019; 200: 1402–13.
17: 122–32. 64 Diaz de Leon A, Cronkhite JT, Yilmaz C, et al. Subclinical lung
43 Craig PJ, Wells AU, Doffman S, et al. Desquamative interstitial disease, macrocytosis, and premature graying in kindreds with
pneumonia, respiratory bronchiolitis and their relationship to telomerase (TERT) mutations. Chest 2011; 140: 753–63.
smoking. Histopathology 2004; 45: 275–82. 65 Putman RK, Hunninghake GM, Dieffenbach PB, et al. Interstitial
44 Flaherty KR, Fell C, Aubry MC, et al. Smoking-related idiopathic lung abnormalities are associated with acute respiratory distress
interstitial pneumonia. Eur Respir J 2014; 44: 594–602. syndrome. Am J Respir Crit Care Med 2017; 195: 138–41.
45 Remy-Jardin M, Edme JL, Boulenguez C, Remy J, Mastora I, 66 Nishino M, Cardarella S, Dahlberg SE, et al. Interstitial lung
Sobaszek A. Longitudinal follow-up study of smoker’s lung with abnormalities in treatment-naïve advanced non-small-cell lung
thin-section CT in correlation with pulmonary function tests. cancer patients are associated with shorter survival. Eur J Radiol
Radiology 2002; 222: 261–70. 2015; 84: 998–1004.
46 Konopka KE, Myers JL. A review of smoking-related interstitial 67 Armstrong HF, Podolanczuk AJ, Barr RG, et al. Serum matrix
fibrosis, respiratory bronchiolitis, and desquamative interstitial metalloproteinase-7, respiratory symptoms, and mortality in
pneumonia: overlapping histology and confusing terminology. community-dwelling adults. Am J Respir Crit Care Med 2017;
Arch Pathol Lab Med 2018; 142: 1177–81. 196: 1311–17.
47 Miller ER, Putman RK, Vivero M, et al. Histopathology of interstitial 68 Copley SJ, Wells AU, Hawtin KE, et al. Lung morphology in the
lung abnormalities in the context of lung nodule resections. elderly: comparative CT study of subjects over 75 years old versus
Am J Respir Crit Care Med 2018; 197: 955–58. those under 55 years old. Radiology 2009; 251: 566–73.
48 Hung YP, Hunninghake GM, Miller ER, et al. Incidental 69 Kadoch M, Kitich A, Alqalyoobi S, et al. Interstitial lung
nonneoplastic parenchymal findings in patients undergoing lung abnormality is prevalent and associated with worse outcome in
resection for mass lesions. Hum Pathol 2019; 86: 93–101. patients undergoing transcatheter aortic valve replacement.
49 Schneider F, Butnor KJ, Beasley MB, et al. Protocol for the Respir Med 2018; 137: 55–60.
examination of resection specimens from patients with primary 70 Im Y, Park HY, Shin S, et al. Prevalence of and risk factors for
non-small cell carcinoma, small cell carcinoma, or carcinoid tumor pulmonary complications after curative resection in otherwise
of the lung. August, 2019. https://documents.cap.org/protocols/cp- healthy elderly patients with early stage lung cancer. Respir Res 2019;
thorax-lung-resection-19–4100.pdf (accessed June 11, 2020). 20: 136.

736 www.thelancet.com/respiratory Vol 8 July 2020

 Position Paper

71 Iwasawa T, Okudela K, Takemura T, et al. Computer-aided 87 Kim GB, Jung KH, Lee Y, et al. Comparison of shallow and deep
quantification of pulmonary fibrosis in patients with lung cancer: learning methods on classifying the regional pattern of diffuse lung
relationship to disease-free survival. Radiology 2019; 292: 489–98. disease. J Digit Imaging 2018; 31: 415–24.
72 Araki T, Dahlberg SE, Hida T, et al. Interstitial lung abnormality in 88 Bermejo Peláez D, Ash SY, Washko GR, San José Estépar R,
stage IV non-small cell lung cancer: a validation study for the Ledesma-Carbayo MJ. Classification of interstitial lung abnormality
association with poor clinical outcome. Eur J Radiol Open 2019; patterns with an ensemble of deep convolutional neural networks.
6: 128–31. Sci Rep 2020; 10: 338.
73 Yamaguchi S, Ohguri T, Ide S, et al. Stereotactic body radiotherapy 89 Mathai SK, Humphries S, Kropski JA, et al. MUC5B variant is
for lung tumors in patients with subclinical interstitial lung disease: associated with visually and quantitatively detected preclinical
the potential risk of extensive radiation pneumonitis. Lung Cancer pulmonary fibrosis. Thorax 2019; 74: 1131–39.
2013; 82: 260–65. 90 Park SO, Seo JB, Kim N, et al. Feasibility of automated
74 Li F, Zhou Z, Wu A, et al. Preexisting radiological interstitial lung quantification of regional disease patterns depicted on high-
abnormalities are a risk factor for severe radiation pneumonitis in resolution computed tomography in patients with various diffuse
patients with small-cell lung cancer after thoracic radiation therapy. lung diseases. Korean J Radiol 2009; 10: 455–63.
Radiat Oncol 2018; 13: 82. 91 Lee SM, Seo JB, Oh SY, et al. Prediction of survival by texture-based
75 Brahmer JR, Lacchetti C, Schneider BJ, et al. Management of automated quantitative assessment of regional disease patterns on
immune-related adverse events in patients treated with immune CT in idiopathic pulmonary fibrosis. Eur Radiol 2018; 28: 1293–300.
checkpoint inhibitor therapy: American Society of Clinical 92 Park HJ, Lee SM, Song JW, et al. Texture-based automated
Oncology clinical practice guideline. J Clin Oncol 2018; quantitative assessment of regional patterns on initial CT in
36: 1714–68. patients with idiopathic pulmonary fibrosis: relationship to decline
76 Puzanov I, Diab A, Abdallah K, et al. Managing toxicities in forced vital capacity. AJR Am J Roentgenol 2016; 207: 976–83.
associated with immune checkpoint inhibitors: consensus 93 Yoon RG, Seo JB, Kim N, et al. Quantitative assessment of change
recommendations from the Society for Immunotherapy of Cancer in regional disease patterns on serial HRCT of fibrotic interstitial
(SITC) Toxicity Management Working Group. pneumonia with texture-based automated quantification system.
J Immunother Cancer 2017; 5: 95. Eur Radiol 2013; 23: 692–701.
77 Nakanishi Y, Masuda T, Yamaguchi K, et al. Pre-existing interstitial 94 Wu X, Kim GH, Salisbury ML, et al. Computed tomographic
lung abnormalities are risk factors for immune checkpoint biomarkers in idiopathic pulmonary fibrosis. The future of
inhibitor-induced interstitial lung disease in non-small cell lung quantitative analysis. Am J Respir Crit Care Med 2019; 199: 12–21.
cancer. Respir Investig 2019; 57: 451–59. 95 American College of Radiology. Lung-RADS version 1.1. 2019.
78 Choi B, Kawut SM, Raghu G, et al. Regional distribution of high https://www.acr.org/-/media/ACR/Files/RADS/Lung-RADS/
attenuation areas on CT in the multi-ethnic study of atherosclerosis LungRADSAssessmentCategoriesv1-1.pdf?la=en (accessed
(MESA). Am J Respir Crit Care Med 2018; 197: A1616. June 11, 2020).
79 Kliment CR, Araki T, Doyle TJ, et al. A comparison of visual and 96 Yoon SH, Hong J, Hwang EJ, et al. Significant abnormalities other
quantitative methods to identify interstitial lung abnormalities. than lung cancer in Korean lung cancer CT screening.
BMC Pulm Med 2015; 15: 134. J Korean Soc Radiol 2019; 80: 837–48.
80 Uppaluri R, Hoffman EA, Sonka M, Hunninghake GW, 97 Lee J, Lim J, Kim Y, et al. Development of protocol for Korean Lung
McLennan G. Interstitial lung disease: a quantitative study using Cancer Screening Project (K-LUCAS) to evaluate effectiveness and
the adaptive multiple feature method. Am J Respir Crit Care Med feasibility to implement National Cancer Screening Program.
1999; 159: 519–25. Cancer Res Treat 2019; 51: 1285–94.
81 Bartholmai BJ, Raghunath S, Karwoski RA, et al. Quantitative 98 Richeldi L, du Bois RM, Raghu G, et al. Efficacy and safety of
computed tomography imaging of interstitial lung diseases. nintedanib in idiopathic pulmonary fibrosis. N Engl J Med 2014;
J Thorac Imaging 2013; 28: 298–307. 370: 2071–82.
82 Depeursinge A, Vargas A, Platon A, Geissbuhler A, Poletti PA, 99 King TE Jr, Bradford WZ, Castro-Bernardini S, et al. A phase 3 trial
Müller H. Building a reference multimedia database for interstitial of pirfenidone in patients with idiopathic pulmonary fibrosis.
lung diseases. Comput Med Imaging Graph 2012; 36: 227–38. N Engl J Med 2014; 370: 2083–92.
83 Christodoulidis S, Anthimopoulos M, Ebner L, Christe A, 100 Flaherty KR, Wells AU, Cottin V, et al. Nintedanib in progressive
Mougiakakou S. Multisource transfer learning with convolutional fibrosing interstitial lung diseases. N Engl J Med 2019; 381: 1718–27.
neural networks for lung pattern analysis. 101 Distler O, Highland KB, Gahlemann M, et al. Nintedanib for
IEEE J Biomed Health Inform 2017; 21: 76–84. systemic sclerosis-associated interstitial lung disease. N Engl J Med
84 Humphries SM, Yagihashi K, Huckleberry J, et al. Idiopathic 2019; 380: 2518–28.
pulmonary fibrosis: data-driven textural analysis of extent of 102 Araya J, Kawabata Y, Jinho P, Uchiyama T, Ogata H, Sugita Y.
fibrosis at baseline and 15-month follow-up. Radiology 2017; Clinically occult subpleural fibrosis and acute interstitial
285: 270–78. pneumonia a precursor to idiopathic pulmonary fibrosis?
85 Ash SY, Harmouche R, Ross JC, et al. The objective identification Respirology 2008; 13: 408–12.
and quantification of interstitial lung abnormalities in smokers.
Acad Radiol 2017; 24: 941–46.
© 2020 Elsevier Ltd. All rights reserved.
86 Jun S, Kim N, Seo JB, Lee YK, Lynch DA. An ensemble method for
classifying regional disease patterns of diffuse interstitial lung
disease using HRCT images from different vendors. J Digit Imaging
2017; 30: 761–71.

www.thelancet.com/respiratory Vol 8 July 2020 737