Articles

Multicentre evaluation of multidisciplinary team meeting

agreement on diagnosis in diffuse parenchymal lung disease:
a case-cohort study
Simon L F Walsh, Athol U Wells, Sujal R Desai, Venerino Poletti, Sara Piciucchi, Alessandra Dubini, Hilario Nunes, Dominique Valeyre,
Pierre Y Brillet, Marianne Kambouchner, António Morais, José M Pereira, Conceição Souto Moura, Jan C Grutters, Daniel A van den Heuvel,
Hendrik W van Es, Matthijs F van Oosterhout, Cornelis A Seldenrijk, Elisabeth Bendstrup, Finn Rasmussen, Line B Madsen, Bibek Gooptu,
Sabine Pomplun, Hiroyuki Taniguchi, Junya Fukuoka, Takeshi Johkoh, Andrew G Nicholson, Charlie Sayer, Lilian Edmunds, Joseph Jacob,
Maria A Kokosi, Jeffrey L Myers, Kevin R Flaherty, David M Hansell

Summary
Background Diffuse parenchymal lung disease represents a diverse and challenging group of pulmonary disorders. Lancet Respir Med 2016;
A consistent diagnostic approach to diffuse parenchymal lung disease is crucial if clinical trial data are to be applied 4: 557–65
to individual patients. We aimed to evaluate inter-multidisciplinary team agreement for the diagnosis of diffuse Published Online
May 11, 2016
parenchymal lung disease.
http://dx.doi.org/10.1016/
S2213-2600(16)30033-9
Methods We did a multicentre evaluation of clinical data of patients who presented to the interstitial lung disease See Comment page 529
unit of the Royal Brompton and Harefield NHS Foundation Trust (London, UK; host institution) and required Department of Radiology
multidisciplinary team meeting (MDTM) characterisation between March 1, 2010, and Aug 31, 2010. Only patients (S L F Walsh MD), King’s College
whose baseline clinical, radiological, and, if biopsy was taken, pathological data were undertaken at the host Hospital NHS Foundation
institution were included. Seven MDTMs, consisting of at least one clinician, radiologist, and pathologist, from Trust, London, UK
(S R Desai MD); Interstitial Lung
seven countries (Denmark, France, Italy, Japan, Netherlands, Portugal, and the UK) evaluated cases of diffuse Disease Unit (Prof A U Wells MD,
parenchymal lung disease in a two-stage process between Jan 1, and Oct 15, 2015. First, the clinician, radiologist, M A Kokosi MD) and
and pathologist (if lung biopsy was completed) independently evaluated each case, selected up to five differential Department of Radiology
diagnoses from a choice of diffuse lung diseases, and chose likelihoods (censored at 5% and summing to 100% in (Prof D M Hansell FRSM), Royal
Brompton and Harefield NHS
each case) for each of their differential diagnoses, without inter-disciplinary consultation. Second, these specialists Foundation Trust, London, UK
convened at an MDTM and reviewed all data, selected up to five differential diagnoses, and chose diagnosis (J Jacob FRCR); Department of
likelihoods. We compared inter-observer and inter-MDTM agreements on patient first-choice diagnoses using Diseases of the Thorax
Cohen’s kappa coefficient (κ). We then estimated inter-observer and inter-MDTM agreement on the probability of (Prof V Poletti MD) and
Department of Radiology
diagnosis using weighted kappa coefficient (κw). We compared inter-observer and inter-MDTM confidence of (S Piciucchi MD), GB Morgagni
patient first-choice diagnosis. Finally, we evaluated the prognostic significance of a first-choice diagnosis of Hospital, Forlì, Italy;
idiopathic pulmonary fibrosis (IPF) versus not IPF for MDTMs, clinicians, and radiologists, using univariate Cox Department of Surgical
regression analysis. Pathology, Morgagni
Pierantoni Hospital, Forlì, Italy
(A Dubini MD); Université Paris,
Findings 70 patients were included in the final study cohort. Clinicians, radiologists, pathologists, and the MDTMs Sorbonne Paris Cité, EA2363
assigned their patient diagnoses between Jan 1, and Oct 15, 2015. IPF made up 88 (18%) of all 490 MDTM first- Réponses cellulaires et
fonctionnelles à l’hypoxie,
choice diagnoses. Inter-MDTM agreement for first-choice diagnoses overall was moderate (κ=0·50). Inter-MDTM
Bobigny, France
agreement on diagnostic likelihoods was good for IPF (κw=0·71 [IQR 0·64–0·77]) and connective tissue disease- (Prof H Nunes MD,
related interstitial lung disease (κw=0·73 [0·68–0·78]); moderate for non-specific interstitial pneumonia (NSIP; Prof D Valeyre MD,
κw=0·42 [0·37–0·49]); and fair for hypersensitivity pneumonitis (κw=0·29 [0·24–0·40]). High-confidence diagnoses Prof P Y Brillet MD); Assistance
Publique-Hôpitaux de Paris,
(>65% likelihood) of IPF were given in 68 (77%) of 88 cases by MDTMs, 62 (65%) of 96 cases by clinicians, and in
Service de Pneumologie
57 (66%) of 86 cases by radiologists. Greater prognostic separation was shown for an MDTM diagnosis of IPF than (Prof H Nunes, Prof D Valeyre),
compared with individual clinician’s diagnosis of this disease in five of seven MDTMs, and radiologist’s diagnosis of Service de Radiologie
IPF in four of seven MDTMs. (Prof P Y Brillet), and Service
d’Anatomie Pathologique
(M Kambouchner MD), Hôpital
Interpretation Agreement between MDTMs for diagnosis in diffuse lung disease is acceptable and good for a diagnosis Avicenne, Bobigny, France;
of IPF, as validated by the non-significant greater prognostic separation of an IPF diagnosis made by MDTMs than Serviço de Pneumologia
the separation of a diagnosis made by individual clinicians or radiologists. Furthermore, MDTMs made the diagnosis (Prof A Morais MD) and
Pathology Department
of IPF with higher confidence and more frequently than did clinicians or radiologists. This difference is of particular
(C Souto Moura MD), Centro
importance, because accurate and consistent diagnoses of IPF are needed if clinical outcomes are to be optimised. Hospitalar São João, Porto,
Inter-multidisciplinary team agreement for a diagnosis of hypersensitivity pneumonitis is low, highlighting an urgent Portugal (J M Pereira MD);
need for standardised diagnostic guidelines for this disease. Faculdade de Medicina,
Universidade do Porto, Porto,
Portugal (Prof A Morais);
Funding National Institute of Health Research, Imperial College London.

www.thelancet.com/respiratory Vol 4 July 2016 557

 Articles
ILD Center of Excellence
St Antonius Hospital, Division
Heart and Lungs, University,
Medical Centre Utrecht,
Netherlands
(Prof J C Grutters MD);
Department of Radiology
(D A van den Heuvel MD,
H W van Es MD) and
Department of Pathology
(M F van Oosterhout MD,
C A Seldenrijk MD), St Antonius
Hospital, Nieuwegein,
Netherlands; Department of
Respiratory Diseases and
Allergy (E Bendstrup PhD) and
Department of Radiology
(F Rasmussen MD) and
Department of Pathology
(L B Madsen PhD), Aarhus
University Hospital, Aarhus,
Denmark; Division of Asthma,
Allergy and Lung Biology,
King’s College London, London,
UK (B Gooptu PhD);
Department of Cellular
Pathology, University College
Hospital London, London, UK
(S Pomplun FRCPath);
Department of Respiratory
Medicine and Allergy, Tosei
General Hospital, Aichi, Japan
(H Taniguchi MD); Department
of Pathology, Nagasaki
University Graduate School of
Biomedical Sciences, Nagasaki,
Japan (Prof J Fukuoka MD);
Department of Radiology,
Kinki Central Hospital of
Mutual Aid Association of
Public School Teachers, Itami,
Japan (T Johkoh MD); National
Heart and Lung Institute,
Imperial College London,
London, UK
(Prof A G Nicholson DM,
L Edmunds FRCPath); Brighton
and Sussex University
Hospitals Trust, Brighton, UK
(C Sayer FRCR); and Department
of Pathology (Prof J L Myers MD)
and Department of Internal
Medicine, Division of
Pulmonary and Critical Care
Medicine (Prof K R Flaherty MD),
University of Michigan, Ann
Arbor, MI, USA
Correspondence to:
Dr Simon L F Walsh, Department
of Radiology, King’s College
Hospital NHS Foundation Trust,
London SE5 9RS, UK
slfwalsh@gmail.com
558
Research in context
Evidence before this study
We searched PubMed on Jan 17, 2014, using the search terms
“diffuse parenchymal lung disease”, “idiopathic pulmonary
fibrosis”, “idiopathic interstitial pneumonias”, “interobserver
agreement”, “diagnosis”, and “multidisciplinary team” for
articles written in English and published between Jan 1, 2000,
and Jan 1, 2014. We extended our search to include studies
published up to Dec 1, 2015, during the writing of the
manuscript. We identified seven key publications that were
pertinent to our study. Of these, we identified four studies of
observer agreement in the setting of diffuse parenchymal lung
disease. All of these seven studies, however, pre-dated the
latest American Thoracic Society and the European Respiratory
Society (ATS/ERS) update for the classification of idiopathic
interstitial pneumonias, and not all studies had evaluated
diagnostic agreement between multidisciplinary teams, but
rather focused on diagnostic agreement between individual
observers.
Added value of this study
Our study is the first evaluation, to our knowledge, of inter-
multidisciplinary team agreement for diagnosis in the setting
of diffuse parenchymal lung disease since the updated
Introduction
Diffuse parenchymal lung disease represents a diverse
and challenging group of pulmonary disorders with
varied prognoses and different management options.
A consistent diagnostic approach to these diseases is
essential if clinical trial data are to be reliably applied to
individual patients. With the 2014 licensing of two new
antifibrotic idiopathic pulmonary fibrosis (IPF) drugs
(pirfenidone1 and nintedanib2), accurate and consistent
diagnosis of IPF is of particular importance to achieve
clinical benefits for patients. In 2002, a joint statement by
the American Thoracic Society (ATS) and the European
Respiratory Society (ERS) on the classification of
idiopathic interstitial pneumonias advocated a multi-
disciplinary diagnostic approach, involving integration of
clinical, radiological, and, in cases for which lung biopsy
material is available, pathological data.3 This approach
has been emphasised by several studies4–7 in the past
12 years and was restated in the 2013 ATS/ERS update8 on
idiopathic interstitial pneumonia classification. Although
this recommendation specifically applies to idiopathic
interstitial pneumonia, a multidisciplinary approach has
been widely used as the diagnostic gold standard for
diffuse parenchymal lung disease in general.4,6 Several
studies4–6,9 have evaluated inter-observer agreement for
diagnosis in the setting of diffuse parenchymal lung
disease. However, most of these studies pre-date the
2013 ATS/ERS update,8 the 2011 joint ATS/ERS/Japanese
Respiratory Society/Latin American Thoracic Association
statement10 on the diagnosis and management of IPF,
2013 ATS/ERS classification of idiopathic interstitial
pneumonias and the 2011 ATS/ERS/Japanese Respiratory
Society/Latin American Thoracic Association guidelines for the
diagnosis and management of idiopathic pulmonary fibrosis.
Our study shows that diagnostic agreement between MDTMs is
non-significant compared with inter-observer agreement
between clinicians, radiologists, and pathologists in the setting
of diffuse parenchymal lung disease. Specifically in idiopathic
pulmonary fibrosis, MDTMs have a higher level of agreement
on diagnoses, assign diagnoses with higher confidence more
frequently, and provide diagnoses that have non-significant
greater prognostic separation than do clinicians or radiologists
in most cases.
Implications of all the available evidence
The differences in diagnostic certainty between MDTMs and
individual clinicians is of particular importance, because
accurate and consistent diagnoses of idiopathic pulmonary
fibrosis are needed if clinical outcomes are to be optimised. By
contrast, inter-multidisciplinary team agreement for a
diagnosis of hypersensitivity pneumonitis is low, highlighting
an urgent need for standardised diagnostic guidelines for
this disease.
and the availability of novel antifibrotic IPF drugs
(pirfenidone1 and nintedanib2), all of which might affect
diagnostic decisions. Furthermore, many of these studies
focused on individual observers rather than agreement
between multidisciplinary teams.4–6,9,11 In this study, we
aimed to evaluate the level of inter-multidisciplinary team
diagnostic agreement between seven international
centres for diagnosis of diffuse parenchymal lung disease.
Methods
Study design, patients, and multidisciplinary team
selection
For this case-cohort study we selected consecutive
patients who presented to the interstitial lung disease
unit of the Royal Brompton and Harefield NHS
Foundation Trust (London, UK; host institution) and
patients with challenging diagnosis had MDTM
characterisation, between March 1, 2010, and
Aug 31, 2010. Only patients who had all their clinical
investigations (serology, high-resolution CT, and, when
required, surgical lung biopsy) completed at the host
institution were included. Seven multidisciplinary
teams from seven different countries (Denmark,
France, Italy, Japan, Netherlands, Portugal, and the
UK), each with specialist expertise in the diagnosis and
management of diffuse parenchymal lung disease,
were invited and agreed to participate in the study. The
only prerequisite for participation in the study was that
each multidisciplinary team had to have a regular
multidisciplinary meeting (about once a week) for
www.thelancet.com/respiratory Vol 4 July 2016

diffuse parenchymal lung disease in place with
consistent attendance by at least one clinician,
radiologist, and pathologist. For the retrospective
examination of clinically indicated data, the institutional
ethics review board of Royal Brompton and Harefield
NHS Foundation Trust waived the need for informed
patient consent.
Evaluation of cases
The evaluation of every case took place in two stages
between Jan 1, and Oct 15, 2015. At first, clinicians,
radiologists, and pathologists had to review the cases
independently without interspecialty consultation.
Clinicians had access to all the presenting clinical
information (age, sex, smoking history, history of
established connective tissue disease, symptoms
[including those suggestive of connective tissue disease],
autoantibody profile, exposure history, drugs at
presentation, bronchoalveolar lavage result [if done],
and angiotensin converting enzyme concentration [if
assessed]), pulmonary function tests, and high-resolution
CT (no access to the original high-resolution CT report or
the pathology result). Radiologists and pathologists had
access to only the age, sex, and smoking history for the
patient, and the high-resolution CT (radiologist) or
digitalised surgical lung biopsy slides (pathologist) taken
at presentation. Pathologists had access to all pathology
data that were available in the form of digitalised slides
(in .svs format), which were viewed using Aperio
ImageScope (version 12.3) viewing software. This digital
viewing application has all the imaging functionality
normally available to pathologists in routine clinical
practice and the host institution uses it to evaluate cases
referred from outside institutions for additional opinions.
For each patient, observers (clinicians, radiologists,
and pathologists) had to select up to five differential
diagnoses and provide a diagnostic likelihood (censored
at 5% increments and summing to 100% in each case)
from a drop-down menu of diffuse lung diseases
(appendix). The only stipulation was that diagnoses were
considered in the context of the current 2013 version of
the ATS/ERS classification and terminology for idiopathic
interstitial pneumonias.8
Once the cases had been reviewed independently, the
clinician, radiologist, and pathologist convened as a
multidisciplinary team to review the cases together and
provide up to five diagnoses with diagnostic likelihoods
(also censored at 5% increments and summing to
100% in each case). All clinical information supplied in
the first stage, including pulmonary function tests,
high-resolution CT at presentation and digitalised
surgical lung biopsy slides were available to the
multidisciplinary team.
In this study we aimed to evaluate inter-
multidisciplinary team agreement for the diagnosis of
diffuse parenchymal lung disease. To validate the
diagnosis made by the MDTMs at the seven centres
www.thelancet.com/respiratory Vol 4 July 2016
Articles
versus individual specialists, we compared the mortality
of each team’s diagnosis of IPF. We compared diagnoses
by separating the entire cohort into a binary IPF diagnosis
category (IPF [if first-choice diagnosis for a patient is
IPF] and not [if first-choice diagnosis is of another cause])
for each MDTM, clinician, and radiologist, based on
assigned diagnoses. Pathologist diagnoses are not shown
due to small number of cases (appendix). We calculated
the survival period for each patient from the date of
referral to the host institution to the minimum of date of
death, date the patient was last known to be alive, or
June 1, 2015 (end of the study period). We obtained vital
status for every patient on June 1, 2015, by evaluating
their electronic patient record.
Statistical analysis
We used Cohen’s kappa coefficient (κ) to evaluate
inter-observer and inter-MDTM agreement for patient’s
first-diagnosis. We used Cohen’s weighted-kappa
coefficient (κw) to evaluate inter-observer agreement and
inter-MDTM agreement for an estimation of the probability
of each diagnosis. We converted the percentage diagnostic
likelihood given for each diagnosis to a five point scale
(0–4), representing clinically useful probabilities (0 for
condition not included in the differential diagnosis; 1 for
low probability [5–25%]; 2 for intermediate probability
[30–65%]; 3 for high probability [70–95%]; and 4 for
pathognomonic [100%]). For example, if the differential
diagnoses given by an MDTM were IPF
(65% diagnostic likelihood), non-specific interstitial
pneumonia (NSIP; 25% diagnostic likelihood), and
hypersensitivity pneumonitis (10% diagnostic likelihood),
the probability grades would be 2 for IPF, 1 for NSIP, and
1 for hypersensitivity pneumonitis. We calculated weighted-
kappa coefficient values between paired observers (for
statements of inter-observer agreement), and between
paired multidisciplinary teams (for statements of inter-
MDTM agreement), which were expressed as median
(IQR) values for all unique combinations of pairs (21 for
seven observers or seven multidisciplinary teams). See Online for appendix
Weighting of the kappa coefficient allowed us to quantify
the extent of disagreement, by assigning greater emphasis
to large differences between scores. Weighted-kappa
coefficients were categorised as poor (0<κw≤0·20), fair
(0·20<κw≤0·40), moderate (0·40<κw≤0·60), good
(0·60<κw≤0·80), and excellent (0·80<κw≤1·00). This
approach has been used in previous investigations9,11 of
inter-observer agreement for diagnosis in diffuse lung
diseases.
Additionally, for each patient the first-choice diagnosis
was considered low confidence (diagnostic likelihood
<70%) or high confidence (diagnostic likelihood ≥70%).
These categories were based on the diagnostic likelihood
categories used to assess the clinical probability of
pulmonary embolism in the PIOPED study.12
We used univariate Cox regression analysis to identify
associations between mortality and MDTM, clinician,
559
 Articles

and radiologist diagnoses in terms of IPF versus not IPF. Role of the funding source
Reported hazard ratios (HRs) are for diagnosis of IPF The funders of the study had no role in study design,
versus not IPF. We tested the assumptions of proportional data collection, data analysis, data interpretation, or
hazards by visual inspection of the log–log plot of writing of the report. The corresponding author had full
survival, comparison of the Kaplan-Meier observed access to all of the data in the study and had final
survival curves with the Cox predicted curves for the responsibility for the decision to submit for publication.
same variable, and graphical and formal analysis of
Schoenfeld residuals (analysis not shown). Results are Results
reported as HRs, 95% CIs, and p values, and are We identified 113 consecutive new patient referrals, who
graphically shown as Kaplan-Meier survival curves. We required local MDTM characterisation, from the clinical
completed statistical analyses using Stata (version 12; database of the host institute between March 1, 2010, and
StataCorp, College Station, Texas). Aug 31, 2010. We excluded 43 (38%) of 113 referrals on the
basis that their initial work-up high-resolution CT scan
(29 patients), lung function (four patients), or surgical
Patients (n=70)
lung biopsy (ten patients) were completed by the referring
Age (years) 60·9 (15·5) institution (appendix). The remaining 70 (62%) of
Sex 113 patient referrals were included as the final study
Men 24 (34%) cohort. Basic patient demographics are in table 1. 13 (19%)
Women 46 (66%) of 70 patients had an established diagnosis of a connective
Smoking tissue disease (seven patients had systemic sclerosis,
Never 38 (54%) three had rheumatoid arthritis, two had Sjögren’s
Previously smoked 24 (34%) syndrome, and one had mixed connective tissue disease)
Current 8 (11%) at the time of presentation to the host institution, and
Established connective tissue disease history 13 (19%) 22 (31%) of 70 cases underwent surgical lung biopsy at the
Biopsy taken host institution (table 1). In patients in whom a surgical
Yes 22 (31%) lung biopsy was not done (table 1), a confident diagnosis
No 48 (69%) had been made by the host institution without the need
DLco (% predicted) 44·8 (14·5%) for a surgical lung biopsy sample. Vital status was known
FEV1 (% predicted) 73·0 (20·5%) for all patients at the end of the study period (June 1, 2015).
FVC (% predicted) 79·0 (19·6%) 70 patients resulted in the assignment of 490 first-
choice MDTM diagnoses (70 patients evaluated by
Where appropriate, data are mean (SD) and n (%). DLco=diffusing capacity of the
lungs for carbon monoxide. FEV1=forced expiratory volume in 1 s. FVC=forced
seven MDTMs). First-choice diagnoses are shown in
vital capacity. table 2. The NSIP and organising pneumonia overlap with
interstitial lung disease category was combined with the
Table 1: Patient demographics and characteristics
NSIP category. The four most prevalent first-choice
diagnoses (connective tissue disease-related interstitial
lung disease, IPF, hypersensitivity pneumonitis, and
Number of first-choice NSIP) were the focus of subsequent analyses. The
diagnoses (n=490) remaining diagnosis categories whose frequency was less
Connective tissue disease-related interstitial 146 (30%) than 10% of the total number of first-choice diagnoses
lung disease (other, sarcoidosis, drug-related interstitial lung
Idiopathic pulmonary fibrosis 88 (18%) disease, occupational lung disease, pleuroparenchymal
Hypersensitivity pneumonitis 46 (9%) fibroelastosis, and organising pneumonia) were combined
Idiopathic non-specific interstitial pneumonia 42 (9%) into an others diagnosis category. The final diagnosis
Unclassifiable interstitial lung disease 38 (8%) categories were connective tissue disease-related
Other 37 (8%) interstitial lung disease, IPF, idiopathic NSIP,
Sarcoidosis 20 (4%) hypersensitivity pneumonitis, and others.
Drug-related interstitial lung disease 18 (4%) Inter-observer agreement (between clinicians,
Smoking-related interstitial lung disease 16 (3%) radiologists, and pathologists) and inter-MDTM
Occupational lung disease 14 (3%) agreement for first-choice diagnosis are listed in table 3.
Pleuroparenchymal fibroelastosis 10 (2%) Overall inter-MDTM agreement for first-choice diagnosis
Organising pneumonia or non-specific 8 (2%) in patients was moderate (table 3). Inter-MDTM
interstitial pneumonia overlapping disease agreement for a first-choice diagnosis of IPF or of
Organising pneumonia 7 (1%) connective tissue disease-related interstitial lung disease
was good, but for a first-choice diagnosis of idiopathic
Table 2: First-choice diagnoses given by seven multidisciplinary teams
NSIP or hypersensitivity pneumonitis was fair (table 3).
for 70 cases of diffuse lung disease
In the subgroup analysis of patients in whom a surgical

560 www.thelancet.com/respiratory Vol 4 July 2016

 Articles

lung biopsy was not done (table 1), overall inter-MDTM

Clinicians (κ) Radiologists (κ) Pathologists (κ) MDTM (κ)
agreement for diagnosis was moderate (table 3).
Inter-MDTM agreement for first-choice diagnoses of IPF, Total No Total No Total No Total No
(n=70) biopsy (n=70) biopsy (n=70) biopsy (n=70) biopsy
connective tissue disease-related interstitial lung disease, (n=48) (n=48) (n=48) (n=48)
and hypersensitivity pneumonitis in these patients were
Overall total 0·45 0·50 0·33 0·31 0·31 ¨ 0·50 0·57
all greater than the first-choice diagnoses of all patients
Idiopathic pulmonary 0·59 0·71 0·46 0·42 0·46 ¨ 0·60 0·70
(table 3). fibrosis total
Inter-MDTM agreement on the diagnostic likelihood of Non-specific interstitial 0·19 0·19 0·25 0·25 0·23 ¨ 0·25 0·25
IPF and connective tissue disease-related interstitial lung pneumonia total
disease was good, but moderate for idiopathic NSIP, and Connective tissue 0·57 0·62 0·10 0·11 0·22 ¨ 0·64 0·73
fair for hypersensitivity pneumonitis diagnoses (table 4). disease-related interstitial
lung disease total
Subgroup analysis of inter-MDTM agreement on the
Hypersensitivity 0·25 0·38 0·27 0·22 0·20 ¨ 0·24 0·31
estimation of diagnostic likelihood of IPF in patients pneumonitis total
without lung biopsy was good (κw=0·78, IQR 0·74–0·83).
Overall inter-observer agreement between clinicians MDTM=multidisciplinary team meeting.
for first-choice diagnosis was moderate and was fair Table 3: Unweighted kappa values (κ) for clinicians, radiologists, pathologists, and inter-
between radiologists and between pathologists (table 3). multidisciplinary team meeting agreement on individual diagnoses of diffuse parenchymal lung disease
Agreement between clinicians on the probability of a
diagnosis of IPF or connective tissue disease-related
interstitial lung disease was higher than agreement on Clinicians (κw) Radiologists (κw) Pathologists (κw) MDTM (κw)
the probability of a diagnosis of idiopathic NSIP or Idiopathic pulmonary 0·72 (0·67–0·76) 0·60 (0·46–0·66) 0·58 (0·45–0·66) 0·71 (0·64–0·77)
hypersensitivity pneumonitis (table 4). Agreement fibrosis
between radiologists or pathologists was highest for the Connective tissue 0·76 (0·70–0·78) 0·17 (0·08–0·31) 0·21 (0·06–0·36) 0·73 (0·68–0·78)
probability of a diagnosis of IPF compared with disease-related
interstitial lung disease
agreement on the probability of a diagnosis of connective
Non-specific interstitial 0·31 (0·27–0·41) 0·32 (0·26–0·41) 0·30 (0·00–0·53) 0·42 (0·37–0·49)
tissue disease-related interstitial lung disease, idiopathic
pneumonia
NSIP, or hypersensitivity pneumonitis (table 4).
Hypersensitivity 0·42 (0·30–0·47) 0·35 (0·29–0·43) 0·26 (0·10–0·45) 0·29 (0·24–0·40)
At the time of patient selection, 13 (19%) of 70 patients pneumonitis
had an established diagnosis of a connective tissue
disease, meaning any diagnosis of interstitial lung Data are median (IQR). MDTM=multidisciplinary team meeting.

disease in these cases would most likely be classified as
connective tissue disease-related interstitial lung disease.
To investigate whether in these 13 cases there was high
agreement for this diagnosis resulted in an increase in
agreement on non-connective tissue disease diagnoses
and, particularly, affected agreement on a diagnosis of
IPF, we did a post-hoc subgroup analysis in the 57 (81%)
of 70 cases who did not have an established diagnosis of
connective tissue disease. In this analysis, although
inter-MDTM agreement for a first-choice diagnosis of
connective tissue disease-related interstitial lung disease
decreased (from κ=0·57 to κ=0·42), no significant change
was observed in inter-MDTM agreement for a first-
choice diagnosis of IPF (κ=0·58), idiopathic NSIP
(κ=0·24), or hypersensitivity pneumonitis (κ=0·23;
appendix).
347 (70%) of 490 first-choice multidisciplinary team
diagnoses were made with high confidence (appendix).
Median prevalence of first-choice multidisciplinary team
diagnoses made with high confidence was 67·1%
(IQR 54·3–88·8). Median prevalence of first-choice
diagnoses made with high confidence was 58·9%
(IQR 52·9–71·4) by clinicians, 68·6% (35·7–85·7) by
radiologists, and 72·7% (59·1–81·8) by pathologists
(appendix). In the subgroup analysis of the 48 (69%) of
70 patients who did not undergo a surgical lung biopsy,
237 (71%) of 336 first-choice multidisciplinary team
Table 4: Weighted kappa values (κw) for estimation of diagnostic likelihood for individual diagnoses of
diffuse parenchymal lung disease
diagnoses were made with high confidence. In this
subgroup, median prevalence of first-choice diagnoses
made with high confidence were 60·4% (37·5–75·0) by
clinicians, 66·6% (39·6–83·3) by radiologists, and 68·7%
(IQR 52·8–87·5) by the multidisciplinary teams.
For the diagnosis of IPF, supportive non-significant
increases in diagnostic confidence by MDTMs (68 [77%]
of 88) were shown, compared with clinicians (62 [65%] of
96) or radiologists (57 [66%] of 86; p=0·23). In the
22 (31%) of 70 cases that underwent surgical lung biopsy
(therefore a total of 154 diagnoses by seven pathologists),
15 (10%) of 154 cases were assigned a first-choice
diagnosis of IPF, of which 12 (8%) were assigned with
high confidence.
We reviewed the 15 cases who the pathologists gave a
first-choice diagnosis of IPF to ascertain whether in
patients who had had a surgical lung biopsy, the final
MDTM diagnosis was usually IPF. In six of 15 cases,
despite the pathologist giving a first-choice diagnosis of
IPF, the final MDTM first-choice diagnosis was not IPF.
Furthermore, in only two of 15 cases was IPF not already
suggested by either the clinician or radiologist in the
MDTM (appendix).

www.thelancet.com/respiratory Vol 4 July 2016 561

 Articles

In the univariate Cox regression analysis, the Discussion

multidisciplinary distinction between IPF and other We have shown that an acceptable level (based on κ>0·40
diagnoses (not IPF) had non-significant increases is deemed clinically acceptable) of diagnostic agreement
toward greater prognostic separation (as judged by exists between multidisciplinary teams in the setting of
HR p values; all seven MDTMs had a significant diffuse parenchymal lung disease. Additionally, we
separation) than the clinician distinction (in showed that this agreement was validated by the non-
five of seven teams) or the radiologist distinction (in significant increases towards greater prognostic
four of seven teams; table 5). The figure shows the separation of an IPF diagnosis made by multidisciplinary
Kaplan-Meier survival curves for the categorisation of teams than by individual clinicians or radiologists.
first-choice diagnosis as IPF or not IPF in cases (by at Furthermore, MDTMs make the diagnosis of IPF with
least four of seven MDTMs, four of seven clinicians, high confidence more frequently than do clinicians or
and four of seven radiologists). The same analysis for radiologists.
pathologists’ diagnosis of IPF was not significant for Since the publication of the ATS/ERS 2002 consensus
five of seven pathologists, probably because of the statement on the classification of idiopathic interstitial
small subgroup size (22 patients) and low prevalence pneumonias, multidisciplinary evaluation of diffuse
of IPF in this subgroup (in 15 of 154 diagnoses parenchymal lung disease has been widely used as the
pathologists’ first-choice diagnoses were IPF; diagnostic gold standard.3,8 This diagnostic approach has
appendix). been partly investigated in several settings. Flaherty and
colleagues4 examined the formulation of diagnosis in a
cohort of diffuse parenchymal lung diseases against
Clinicians (HR, 95% CI, Radiologists (HR, 95% CI, MDTM (HR, 95% CI, p value)
p value) p value) inter-observer agreement and diagnostic confidence in
one multidisciplinary team (consisting of three clinicians,
Team 1 2·09 (0·90–4·86, p=0·085) 2·80 (1·17–6·73, p=0·021) 2·67 (1·21–6·02, p=0·016)
two radiologists, and two pathologists) and showed that
Team 2 2·95 (1·33–6·59, p=0·008) 4·08 (1·84–9·04, p=0·001) 3·44 (1·54–7·68, p=0·003)
diagnostic agreement between observers improved with
Team 3 3·75 (1·65–8·51, p=0·002) 2·78 (1·11–6·97, p=0·030) 5·30 (2·26–12·41, p<0·001)
successive integration of clinical, radiological, and
Team 4 3·34 (1·38–8·00, p=0·007) 4·49 (1·71–12·29, p=0·003) 3·99 (1·49–10·66, p=0·006)
pathological data. In a second study,6 Flaherty and
Team 5 2·03 (0·87–4·69, p=0·100) 2·58 (1·08–6·21, p=0·033) 2·61 (1·12–6·06, p=0·025)
colleagues expanded on these findings by showing
Team 6 4·14 (1·72–9·97, p=0·002) 2·11 (0·91–4·89, p=0·082) 3·36 (1·40–8·07, p=0·007)
higher levels of agreement between academic physicians,
Team 7 2·96 (1·43–6·55, p=0·007) 1·28 (0·53–3·06, p=0·583) 2·43 (1·09–5·41, p=0·030)
radiologists, and pathologists for diagnosis in diffuse
Results for the multidisciplinary team meetings (MDTMs), clinicians, and radiologists are based on the whole patient lung disease than with their community counterparts.
cohort (n=70). HR=hazard ratio. Later in 2008, Thomeer and colleagues5 reported in
Table 5: Univariate Cox regression analysis for mortality according to clinician, radiologist, and MDTM
a cohort of patients included in an IPF trial, a high
diagnoses of idiopathic pulmonary fibrosis versus not idiopathic pulmonary fibrosis accuracy for a clinical diagnosis of this disease made
by six respiratory physicians from different European
centres. A limitation of these studies is that
100 not all evaluated the agreement between different
multidisciplinary teams for diagnoses,4 one focused
specifically on the diagnosis of IPF5 and all studies pre-
Cumulative survival proportion (%)

80 Non-IPF survival date the latest 2013 ATS/ERS idiopathic interstitial

pneumonia update.4–6,8,9 To our knowledge this study is
60 the first to evaluate inter-MDTM agreement for diagnosis
in diffuse lung disease since the 2013 ATS/ERS update
on the classification of idiopathic interstitial pneumonias.
40
With the 2014 licensing of two new drugs1,2 for
IPF survival
treatment of IPF in mild to moderate disease, early and
20 MDTMs accurate diagnosis of IPF is crucial. Our study showed
Clinicians that inter-MDTM agreement for the diagnosis of IPF is
Radiologists
0
good, with clinicians having only marginally lower levels
0 500 1000 1500 2000 of agreement than MDTMs for this diagnosis. On
Follow-up time (days) subgroup analysis in patients without surgical lung
biopsy, inter-MDTM agreement and inter-observer
Figure: Kaplan-Meier of survival differences between patients assigned a diagnosis of idiopathic pulmonary
fibrosis and those assigned other diagnoses (not idiopathic pulmonary fibrosis) agreement between clinicians for the diagnosis of IPF
The cohort of 70 patients was separated into cases for which at least four of seven multidisciplinary team improved and were, again, almost the same. As evidence-
meetings (MDTMs) assigned a first-choice diagnosis of idiopathic pulmonary fibrosis (IPF; blue curve; hazard based diagnostic criteria for IPF are now clearly defined
ratio 6·26, 95% CI 2·72–14·33, p<0·0001), at least four of seven clinicians assigned a first-choice diagnosis of IPF
and are relatively easily applied to many patients with
(red curve; 4·43, 1·94–10·01, p<0·0001), and at least four of seven radiologists assigned a first-choice diagnosis of
IPF (green curve; 3·76, 1·61–8·75, p=0·002); or into other diagnoses (not IPF). For number of patients at risk see this disease, the near parity of agreement between
appendix. MDTMs and between clinicians is not surprising.8,10 For

562 www.thelancet.com/respiratory Vol 4 July 2016


example, in a patient with classic appearances for usual
interstitial pneumonia on high-resolution CT, a rapidly
progressive disease course, and no identifiable triggers,
multidisciplinary discussion is unlikely to alter a
clinician’s initial impression of IPF. Our findings suggest
that formal multidisciplinary input might not be
necessary in every case of suspected IPF if expert clinical
evaluation is available, which could be of particular
relevance to centres with restricted access to appropriate
radiology or pathology expertise.13
A strength of our findings is that the greater agreement
on multidisciplinary diagnosis is mirrored by the
non-significant improvement in prognostic separation
of a multidisciplinary distinction between IPF and other
diagnoses compared with the clinicians’ or radiologists’
distinction. To show this, we selected patients from
2010 data when they first presented to the host institution,
to allow an analysis of 5-year survival. An added
advantage of this approach was that these patients were
referred to the Royal Brompton and Harefield NHS
Foundation Trust in a pre-antifibrotic drug era, therefore
mortality differences between patients with IPF and
those without this disease were not confounded by
antifibrotic therapy. The fact that non-significant
improvements are present but inconclusive might reflect
powering limitations. Additionally, this inconsistency
might indicate that multidisciplinary discussion adds
prognostic value in the subgroup of patients in which
there is significant diagnostic uncertainty.
By contrast with those with IPF, many patients with
non-IPF interstitial lung diseases are not covered by
evidence-based diagnostic guidelines, therefore their
diagnosis is driven by clinical reasoning and analysis of
all available data in a multidisciplinary setting. In these
situations, a level of disagreement between MDTMs is
predictable and borne out by the poor level of inter-MDTM
agreement in the current study for diagnoses of NSIP
and hypersensitivity pneumonitis.8,14–16 The diagnosis of
hypersensitivity pneumonitis is well known to be
challenging because it relies on an array of data, which
are not definitive when considered in isolation.15
Furthermore, at least on high-resolution CT, patterns of
NSIP, usual interstitial pneumonia, or organising
pneumonia might be the only expressions of this
disease.16–19 The low MDTM agreement for the diagnosis
of hypersensitivity pneumonitis in our study highlights
an urgent need for international consensus on what
hypersensitivity pneumonitis actually is.
Our use of the weighted kappa to investigate inter-
MDTM and inter-observer agreement on diagnostic
probabilities is similar to other studies9 of diagnostic
agreement, but warrants particular consideration. We used
the weighted kappa converting the diagnostic likelihoods
to a five-point probability scale, which enabled examination
of the range of diagnostic likelihoods from minimal
likelihood to pathognomonic. Since exclusion of IPF in
diagnosis is as important as making a diagnosis of IPF,
www.thelancet.com/respiratory Vol 4 July 2016
Articles
our methodology has allowed us to show that there is good
agreement on the likelihood of IPF and, as previously
stated, reflects consistent application of established
diagnostic guidelines for this disease. By contrast, MDTM
agreement on the likelihood of hypersensitivity
pneumonitis was low, reinforcing the view that MDTMs
were unclear on how this diagnosis is made.
Two separate observations from our study warrant
further discussion. First, 13 patients had an established
diagnosis of a connective tissue disease at presentation.
However, after MDT evaluation, a connective tissue
disease diagnosis was constructed in an additional
eight cases or more by five of the seven MDTMs based on
presenting clinical symptoms and serology. Separation of
patients with idiopathic interstitial pneumonia from
those with connective tissue disease-related interstitial
lung disease can be challenging—some patients present
with subtle clinical features or serological abnormalities,
suggesting an autoimmune process, but do not meet
established criteria for a specific connective tissue
disease.20–22 In 2015, an ERS/ATS task force was formed to
establish consensus on how to classify these patients,
and a set of diagnostic criteria has been suggested.23–26
After removal of patients with an established diagnosis of
connective tissue disease, our subgroup analysis showed
acceptable levels of agreement on connective tissue
disease-related interstitial lung disease (by contrast
with hypersensitivity pneumonitis), underlying the
importance of reaching an international consensus on
disease definitions. Furthermore, the high frequency of
connective tissue disease-related interstitial lung disease
diagnoses made in our study highlights the importance
of formal rheumatology input within the MDTM, which
might include face-to-face rheumatological consultation
with patients suspected of having an undiagnosed
connective tissue disease.
Second, dynamic exchanges of clinical, radiological,
and pathological information between experts in a
multidisciplinary process has previously been suggested
to result in higher-confidence diagnoses.4 However, in
our study the proportion of high-confidence diagnoses
(≥70%) assigned after MDTM evaluation did not increase
compared with the proportion of high-confidence
diagnoses assigned by the individual components of the
MDTM. In most MDTMs, high-confidence diagnoses
were more frequently assigned by radiologists and
pathologists than were MDTM diagnoses or diagnoses
assigned by clinicians. In our study, radiologists and
pathologists did not have access to clinical information,
so their interpretation was based almost entirely on
pattern recognition, which might conceivably result in
more confident, but not necessarily more accurate,
diagnoses. Although one benefit of the multidisciplinary
process is that diagnoses might be challenged and must
be publicly defended, it is possible that extra discussion
creates additional difficulties in some cases that initially
seem straightforward when evaluated by individuals in
563
 Articles
isolation. However, we must highlight that in the specific
case of IPF, MDTMs made the diagnosis of IPF with
high confidence more frequently than did clinicians or
radiologists.
Our methodology has some limitations. Since patients
were selected from a pre-antifibrotic drug era, IPF was
possibly not as prevalent in our study cohort as it would
be now at most referral centres, which expend more time
in their evaluation of patients for approval for antifibrotic
treatment. However, as discussed, the inclusion of
patients who had not had antifibrotic drugs allowed us
to evaluate the veracity of MDTM diagnosis for IPF
against the outcomes. Second, unlike a real-world
multidisciplinary process, no observers had face-to-face
consultations with patients and therefore did not have
the benefit of obtaining a clinical history or doing
a physical examination first-hand. In difficult to
characterise patients, a clinician’s impression might be
influenced by direct contact with the patient in ways that
are not easily and objectively quantified. Equally
important, but also difficult to assess, is the effect of
individual personalities on MDTM diagnosis. Arguably, a
consensus diagnosis at multidisciplinary evaluation
sometimes might reflect the strongest voice in the room
rather than represent true agreement between
participants. Lastly, the wide range of proportions of
high-confidence diagnoses reported between the
different MDTMs in our study might reflect cultural
influences on the dynamics of multidisciplinary
evaluation. These less tangible influences could be the
focus of further studies.
In conclusion, our study showed that diagnostic
agreement between MDTMs is higher than inter-
observer agreement between clinicians, radiologists, and
pathologists in the setting of diffuse parenchymal lung
disease. In particular, inter-MDTM agreement for a
diagnosis of IPF is good and validated by the non-
significant greater prognostic separation of an IPF
diagnosis made by multidisciplinary teams than by
individual specialists. We showed low levels of inter-
MDTM agreement for NSIP and hypersensitivity
pneumonitis, which for hypersensitivity pneumonitis
might be in relation to the absence of evidence-based
guidelines to diagnose this disease.
Contributors
SLFW, AUW, SRD, and DMH had the study concept and did data
analysis, and manuscript writing. SLFW, AUW, SRD, AGN, CS, LE, JJ,
MAK, and DMH did data collection. AGN and LE did evaluation of
digitalised pathology data. JLM and KRF did manuscript review and
editing. All remaining authors made up the seven multidisciplinary
teams from Denmark (EB, FR, LBM), France (HN, DV, PYB, MK), Italy
(VP, SP, AD), Japan (HT, JF, TJ), Netherlands (JCG, DAvdH, HWvE,
MFvO, CAS), Portugal (AM, JMP, CSM), and the UK (SRD, BG, SP).
Declaration of interests
AUW has received personal fees from Intermune, Roche, Bayer, and
Gilead. HN is an investigator for clinical trials by Intermune, Roche,
Boehringer Ingelheim, Sanofi, and Centocor. DV has received personal
fees from Roche, Intermune and Boehringer Ingelheim. HT has
received personal fees from Abbott Japan Co Ltd, Actelion
564
Pharmaceuticals Japan Ltd, Ashai Kasei Pharmaceutical Corporation,
Astellas Pharmaceutical Incorporated, AstraZeneca, Bayer, Boehringer
Ingelheim, and Chugai Pharmaceuticals. JF has received personal fees
from Astellas Pharmaceutical Incorporated, Pathology Institute
Corporation, Chugai Pharmaceuticals, and Sakura Finetek Japan. AGN
has received personal fees from Sanofi, Intermune, Boehringer
Ingelheim, and Actelion. KRF has received personal fees from
Boehringer Ingelheim, Genentech, Ikaria, Immuneworks, Veracyte,
Roche, Gilead, Biogen, Afferent, Aeolus, and Pharmakea; and has
received grants from Boehringer Ingelheim, Genentech, Roche, and
Afferent. DMH has received personal fees from Boehringer Ingelheim,
Sanofi, AstraZeneca, Roche, and GlaxoSmithKline. The other authors
declare no competing interests.
Acknowledgments
We thank the National Institute of Health Research Respiratory Disease
Biomedical Unit (Royal Brompton and Harefield NHS Foundation Trust,
London, UK) and Imperial College London (London, UK). DMH is the
recipient of a National Institute of Health Research Senior Investigator
Award.
References
1 King TE Jr, Bradford WZ, Castro-Bernardini S, et al. A phase 3 trial
of pirfenidone in patients with idiopathic pulmonary fibrosis.
N Engl J Med 2014; 370: 2083–92.
2 Richeldi L, du Bois RM, Raghu G, et al. Efficacy and safety of
nintedanib in idiopathic pulmonary fibrosis. N Engl J Med 2014;
370: 2071–82.
3 American Thoracic Society, European Respiratory Society.
American Thoracic Society/European Respiratory Society
international multidisciplinary consensus classification of the
idiopathic interstitial pneumonias. This joint statement of the
American Thoracic Society (ATS), and the European Respiratory
Society (ERS) was adopted by the ATS board of directors, June 2001
and by the ERS Executive Committee, June 2001.
Am J Respir Crit Care Med 2002; 165: 277–304.
4 Flaherty KR, King TE Jr, Raghu G, et al. Idiopathic interstitial
pneumonia: what is the effect of a multidisciplinary approach to
diagnosis? Am J Respir Crit Care Med 2004; 170: 904–10.
5 Thomeer M, Demedts M, Behr J, et al. Multidisciplinary
interobserver agreement in the diagnosis of idiopathic pulmonary
fibrosis. Eur Respir J 2008; 31: 585–91.
6 Flaherty KR, Andrei AC, King TE Jr, et al. Idiopathic interstitial
pneumonia: do community and academic physicians agree on
diagnosis? Am J Respir Crit Care Med 2007; 175: 1054–60.
7 Tomassetti S, Piciucchi S, Tantalocco P, Dubini A, Poletti V.
The multidisciplinary approach in the diagnosis of idiopathic
pulmonary fibrosis: a patient case-based review. Eur Respir Rev 2015;
24: 69–77.
8 Travis WD, Costabel U, Hansell DM, et al. An official American
Thoracic Society/European Respiratory Society statement: update of
the international multidisciplinary classification of the idiopathic
interstitial pneumonias. Am J Respir Crit Care Med 2013;
188: 733–48.
9 Aziz ZA, Wells AU, Hansell DM, et al. HRCT diagnosis of diffuse
parenchymal lung disease: inter-observer variation. Thorax 2004;
59: 506–11.
10 Raghu G, Collard HR, Egan JJ, et al. An official ATS/ERS/JRS/ALAT
statement: idiopathic pulmonary fibrosis: evidence-based guidelines
for diagnosis and management. Am J Respir Crit Care Med 2011;
183: 788–824.
11 Walsh SL, Calandriello L, Sverzellati N, Wells AU, Hansell DM,
UIP Observer Consort. Interobserver agreement for the ATS/ERS/
JRS/ALAT criteria for a UIP pattern on CT. Thorax 2016; 71: 45–51.
12 PIOPED Investigators. Value of the ventilation/perfusion scan in
acute pulmonary embolism. Results of the prospective investigation
of pulmonary embolism diagnosis (PIOPED). JAMA 1990;
263: 2753–59.
13 Jo HE, Corte TJ, Moodley Y, et al. Evaluating the interstitial lung
disease multidisciplinary meeting: a survey of expert centres.
BMC Pulm Med 2016; 16: 22.
14 Travis WD, Hunninghake G, King TE Jr, et al. Idiopathic nonspecific
interstitial pneumonia: report of an American Thoracic Society
project. Am J Respir Crit Care Med 2008; 177: 1338–47.
www.thelancet.com/respiratory Vol 4 July 2016

15 Lacasse Y, Selman M, Costabel U, et al. Clinical diagnosis of
hypersensitivity pneumonitis. Am J Respir Crit Care Med 2003;
168: 952–58.
16 Silva CI, Muller NL, Lynch DA, et al. Chronic hypersensitivity
pneumonitis: differentiation from idiopathic pulmonary fibrosis
and nonspecific interstitial pneumonia by using thin-section CT.
Radiology 2008; 246: 288–97.
17 Lacasse Y, Girard M, Cormier Y. Recent advances in hypersensitivity
pneumonitis. Chest 2012; 142: 208–17.
18 Churg A, Sin DD, Everett D, Brown K, Cool C. Pathologic patterns
and survival in chronic hypersensitivity pneumonitis.
Am J Surg Pathol 2009; 33: 1765–70.
19 Silva CI, Churg A, Muller NL. Hypersensitivity pneumonitis:
spectrum of high-resolution CT and pathologic findings.
AJR Am J Roentgenol 2007; 188: 334–44.
20 Fischer A, du Bois R. Interstitial lung disease in connective tissue
disorders. Lancet 2012; 380: 689–98.
21 Tzelepis GE, Toya SP, Moutsopoulos HM. Occult connective tissue
diseases mimicking idiopathic interstitial pneumonias. Eur Respir J
2008; 31: 11–20.
www.thelancet.com/respiratory Vol 4 July 2016
Articles
22 Mittoo S, Gelber AC, Christopher-Stine L, Horton MR, Lechtzin N,
Danoff SK. Ascertainment of collagen vascular disease in patients
presenting with interstitial lung disease. Respir Med 2009;
103: 1152–58.
23 Fischer A, West SG, Swigris JJ, Brown KK, du Bois RM.
Connective tissue disease-associated interstitial lung disease:
a call for clarification. Chest 2010; 138: 251–56.
24 Vij R, Noth I, Strek ME. Autoimmune-featured interstitial lung
disease: a distinct entity. Chest 2011; 140: 1292–99.
25 Kinder BW, Collard HR, Koth L, et al. Idiopathic nonspecific
interstitial pneumonia: lung manifestation of undifferentiated
connective tissue disease? Am J Respir Crit Care Med 2007;
176: 691–97.
26 Fischer A, Antoniou KM, Brown KK, et al, ERS/ATS Task Force on
Undifferentiated Forms of CTD-ILD. An official European
Respiratory Society/American Thoracic Society research statement:
interstitial pneumonia with autoimmune features. Eur Respir J
2015; 46: 976–87.
565