Note: This copy is for your personal, non-commercial use only. To order presentation-ready copies for
distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights.
ORIGINAL RESEARCH
Nicola Sverzellati, MD
Athol U. Wells, MD, FRACP
Sara Tomassetti, MD
Sujal R. Desai, MD
Susan J. Copley, MD
Zelena A. Aziz, MD
Maurizio Zompatori, MD
Marco Chilosi, MD
Andrew G. Nicholson, MD, MRCPath
Venerino Poletti, MD
David M. Hansell, MD, FRCP
1
From the Department of Clinical Sciences, Section of
Diagnostic Imaging, University of Parma Padiglione Barbieri,
University Hospital of Parma, V. Gramsci 14, 43100 Parma,
Italy (N.S.); Interstitial Lung Disease Unit (A.U.W.) and
Departments of Radiology (D.M.H.) and Pathology (A.G.N.),
Royal Brompton Hospital, London, England; Department of
Thoracic Diseases, Giovanni Battista Morgagni Hospital,
Forlì, Italy (S.T., V.P.); Department of Radiology, King’s
College Hospital, London, England (S.R.D.); Department of
Radiology, Hammersmith Hospital, Imperial NHS Trust, London, England (S.J.C.); Department of Radiology, the London
Chest Hospital, London, England (Z.A.A.); Department of
Radiology, Policlinico Sant’Orsola-Malpighi, Bologna, Italy
(M.Z.); and Department of Pathology, University of Verona,
Verona, Italy (M.C.). Received May 22, 2009; revision
requested July 12; final revision received September 18;
accepted September 29; final version accepted October
5. M.C. supported in part by European Union FP7 Health
Research grant HEALTH-F4-2008-202047. Address correspondence to N.S. (e-mail: [email protected] ).
q
Purpose:
To document the spectrum of misleading thin-section
computed tomographic (CT) diagnoses in patients with
biopsy-proved idiopathic pulmonary fibrosis (IPF).
Materials and
Methods:
This study had institutional review board approval, and
patient consent was not required. Three observers, blinded
to any clinical information and the purpose of the study,
recorded thin-section CT differential diagnoses and
assigned a percentage likelihood to each for a group of
123 patients (79 men, 44 women; age range, 27–82 years)
with various chronic interstitial lung diseases, including a
core group of 55 biopsy-proved cases of IPF. Patients with
IPF in the core group, in whom IPF was diagnosed as lowgrade probability (,30%) by at least two observers, were
considered to have atypical IPF cases, and the alternative
diagnoses were analyzed further.
Results:
Thirty-four (62%) of 55 biopsy-proved IPF cases were
regarded as alternative diagnoses. In these atypical IPF
cases, the first-choice diagnoses, expressed with high
degree of probability, were nonspecific interstitial pneumonia (NSIP; 18 [53%] of 34), chronic hypersensitivity
pneumonitis (HP; four [12%] of 34), sarcoidosis (three
[9%] of 34), and organizing pneumonia (one [3%] of
34); in eight (23%) of 34 cases, no single diagnosis
was favored by more than one observer. Frequent differential diagnoses, although not always the first-choice
diagnosis, were NSIP (n = 29), chronic HP (n = 23), and
sarcoidosis (n = 9).
Conclusion:
In the correct clinical setting, a diagnosis of IPF is not
excluded by thin-section CT appearances more suggestive
of NSIP, chronic HP, or sarcoidosis.
q
RSNA, 2010
RSNA, 2010
Radiology: Volume 254: Number 3—March 2010
n
radiology.rsna.org
957
n THORACIC IMAGING
Biopsy-proved Idiopathic
Pulmonary Fibrosis: Spectrum
of Nondiagnostic Thin-Section
CT Diagnoses1
THORACIC IMAGING: Idiopathic Pulmonary Fibrosis: Nondiagnostic Thin-Section CT Diagnoses
T
he term idiopathic pulmonary
fibrosis (IPF) is applied to diagnosis
in patients with a histologic and/or
computed tomographic (CT) pattern of
usual interstitial pneumonia (UIP) and
compatible clinical features (1). The
major reason for separating IPF from
the other idiopathic interstitial pneumonias is to rule out a disorder with
a prognosis worse than that of many
cancers (2). Familiarity with the typical
thin-section CT findings of IPF is important because these, in association with
a compatible clinical profile, are often
sufficient to allow a confident diagnosis of IPF without surgical biopsy (3).
The characteristic CT features of IPF
are a reticular pattern with honeycomb
destruction in a subpleural and basal
distribution (4,5). Although the typical
thin-section CT pattern is highly predictive of a histologic diagnosis of UIP, the
characteristic thin-section CT features
of UIP are absent in as many as 30%
of patients (5–8). To our knowledge,
the full range of thin-section appearances of UIP has not been reported,
although it is known that some cases of
UIP mimic nonspecific interstitial pneumonia (NSIP) (7,9,10). The aim of this
study was to document the spectrum of
misleading thin-section CT diagnoses in
patients with biopsy-proved IPF.
Materials and Methods
This retrospective study had institutional review board approval at the
two centers that contributed cases, and
informed patient consent was waived.
We identified the study population by
reviewing the interstitial lung disease
databases of two teaching hospitals
(Royal Brompton Hospital [London,
England] and Morgagni Hospital [Forlì,
Sverzellati et al
Italy]) between January 1, 2003, and
December 31, 2006. The core study
group comprised consecutive patients
who had a combined clinical-radiologicpathologic diagnosis of IPF; this group
included 55 subjects (age range, 44–74
years; mean age, 59 years 6 6.2 [standard deviation]), with 39 men (age
range, 47–69 years; mean age, 59 years
6 5.6) and 16 women (age range, 44–74
years; mean age, 59 years 6 7.6). These
cases were reviewed by paired pathologists, and the diagnosis was confirmed
according to accepted histopathologic
criteria of UIP.
To ensure that observers participating in this study assessed the thin-section
CT appearance of the cases in the core
study group in a blinded fashion, two
other cohorts of patients with chronic
interstitial lung diseases were selected
randomly from our databases and mixed
with the core study group. These cohorts
comprised patients with IPF diagnosed
on the basis of clinical and thin-section
CT criteria. Included were 20 subjects
(age range, 43–82 years; mean age, 65.1
years 6 8.9), with 14 men (age range,
52–82 years; mean age, 67 years 6
7.5) and six women (age range, 43–72
years; mean age, 60 years 6 3.4). Also
included was a mixed group of subjects
with various chronic and fibrotic interstitial lung diseases that comprised 48
subjects (age range, 27–74 years; mean
age, 56 years 6 11.6), with 26 men (age
range, 27–74 years; mean age, 55 years
6 12.7) and 22 women (age range, 38–
71 years; mean age, 56 years 6 9.7). In
the mixed group, subjects had diseases
such as NSIP (n = 17), sarcoidosis (n = 6),
chronic hypersensitivity pneumonitis
(HP [n = 8]), desquamative interstitial
pneumonia (n = 5), fibrotic Langerhans cell histiocytosis (n = 4), organizing
Advance in Knowledge
n Thin-section CT findings in
patients with biopsy-proved idiopathic pulmonary fibrosis (IPF)
are often atypical and overlap
with those of other chronic interstitial lung diseases, particularly
nonspecific interstitial pneumonia
(NSIP).
958
Implication for Patient Care
n The diagnosis of IPF should not
be excluded when it is favored by
the clinical context but thinsection CT findings are compatible
with, particularly, NSIP, chronic
hypersensitivity pneumonitis, or
sarcoidosis.
pneumonia (n = 3), mixed NSIP and
organizing pneumonia (n = 3), and lymphoid interstitial pneumonia (n = 2).
Clinical data (eg, absence of previous environmental exposures and connective tissue disease) were reviewed
by two chest physicians (S.T., A.U.W.,
with 7 and 20 years of experience in
evaluating patients with interstitial lung
disease, respectively). Patient exclusion
criteria included coexistent infection,
cardiac failure, and acute exacerbation
of disease at the time of CT.
Histologic Evaluation
Lung biopsy specimens with a histologic
diagnosis of UIP were reviewed by paired
pathologists (M.C. and V.P., A.G.N. and
a nonauthor with 18–32 years of experience in lung pathologic findings). Decisions were made with consensus.
UIP was diagnosed by using the
American Thoracic Society and European Respiratory Society criteria (3).
UIP was diagnosed histologically given
the presence of temporal heterogeneity
with nonuniform and variable interstitial changes, including intermingled
zones of established interstitial fibrosis,
inflammation, fibroblastic foci, honeycomb change, and normal lung coexisting in variable proportions (3). The
Published online
10.1148/radiol.0990898
Radiology 2010; 254:957–964
Abbreviations:
CI = confidence interval
HP = hypersensitivity pneumonitis
IPF = idiopathic pulmonary fibrosis
kw = weighted k
NSIP = nonspecific interstitial pneumonia
UIP = usual interstitial pneumonia
Author contributions:
Guarantors of integrity of entire study, N.S., M.Z., V.P.,
D.M.H.; study concepts/study design or data acquisition
or data analysis/interpretation, all authors; manuscript
drafting or manuscript revision for important intellectual
content, all authors; approval of final version of submitted
manuscript, all authors; literature research, N.S., M.Z.,
M.C., V.P.; clinical studies, N.S., A.U.W., S.T., S.R.D., S.J.C.,
M.Z., A.G.N., V.P., D.M.H.; statistical analysis, A.U.W., V.P.;
and manuscript editing, N.S., A.U.W., S.R.D., M.Z., M.C.,
V.P., D.M.H.
Authors stated no financial relationship to disclose.
radiology.rsna.org
n
Radiology: Volume 254: Number 3—March 2010
THORACIC IMAGING: Idiopathic Pulmonary Fibrosis: Nondiagnostic Thin-Section CT Diagnoses
diagnosis in patients with mixed chronic
and fibrotic cases was established at
each participating institution on the
basis of compatible clinical and histologic findings obtained by means of
surgical lung biopsy (for HP, NSIP, organizing pneumonia, desquamative interstitial pneumonia, and lymphoid interstitial pneumonia) or transbronchial
biopsy (for sarcoidosis).
CT Scanning Protocol
Several CT scanners were used for this
study: one 16–detector row (LightSpeed
16; GE Healthcare, Milwaukee, Wis),
one four–detector row (Somatom Volume Zoom; Siemens Medical Solutions,
Forchheim, Germany), one 64–detector
row (Somatom Sensation 64; Siemens
Medical Solutions), and one electronbeam (C-150L; Imatron, San Francisco,
Calif) scanner. Sections were obtained
with 1- or 1.5-mm collimation at 10-mm
intervals or volumetrically with multi–
detector row CT scanners with 0.6- or
1-mm collimation and 1-mm reconstruction. Scans were obtained with the
patient in the supine position and at
full inspiration and were reconstructed
by using a high-spatial-frequency algorithm. All images were viewed at window settings optimized for assessment
of lung parenchyma (window width,
1500–1600 HU; window level, 2500 to
2600 HU).
Image Evaluation
The thin-section CT images in 116 patients were transferred via CD-ROM to
three personal computers and were reviewed with Digital Imaging and Communications in Medicine viewing software
(DicomWorks, version 1.3; http://dicom.
online.fr/). For seven patients, hard-copy
images were reviewed. Images were reviewed independently by three thoracic
radiologists (Z.A.A., S.J.C., and S.R.D.,
with 5, 8, and 11 years of experience,
respectively, of interpreting thin-section
CT scans in patients with interstitial
lung disease). These observers worked
in three teaching hospitals (The London
Chest Hospital, [London, UK], Hammersmith Hospital [London, UK] and
King’s College Hospital [London, UK])
that did not provide any case included in
Radiology: Volume 254: Number 3—March 2010
n
Sverzellati et al
Table 1
Definition of the Diagnostic Probabilities Extracted by Combining the Observers’
Evaluations for Biopsy-proved IPF Cases
Diagnostic Probability
High
Intermediate
Low*
IPF Diagnosis
Grade 3 or higher according to at least two of three observers
(eg, 80%, observer 1; 100%, observer 2; 50%, observer 3)
Grade 3 or higher according to one observer, grade 2 according
to one observer, and grade 1 or lower according to one
observer (eg, 70%, observer 1; 40%, observer 2; 20%, observer 3)
Grade 2 according to at least two of three observers (eg, 40%,
observer 1; 55%, observer 2; 5%, observer 3)
Grade 1 or lower according to at least two of three observers (eg,
10%, observer 1; 0%, observer 2; 40%, observer 3)
Note.—Grade 4 = 100%, grade 3 = 70%–95%, grade 2 = 30%–65%, grade 1 = 5%–25%, grade 0 = not mentioned.
* Atypical IPF cases.
the study population. Observers had no
knowledge of clinical findings or details
of the patient population and were
not aware of the purpose of the study.
The observers were asked to list their
differential diagnoses (with no limit to
the number of possible diagnoses) and
to assign a likelihood to each diagnosis (to the nearest 5%, totaling 100%).
Furthermore, they were instructed to
assimilate all features present on the
CT images with the ability to manipulate the window setting. Specific diagnostic criteria for the interstitial lung
diseases were not provided, so the diagnoses were based on each observer’s
own experience and understanding of
the current CT literature. However, the
observers used the terminology of the
American Thoracic Society and European Respiratory Society classification
for the diagnosis of idiopathic interstitial
pneumonias (3).
Data and Statistical Analyses
Unadjusted k coefficients of agreement
were computed for the first-choice diagnosis in the entire study population and
in the cohort of patients with biopsyproved IPF. The weighted k (kw) coefficient of agreement was used to calculate
the observer variation for the estimation of the probability of IPF diagnosis
in the entire cohort and in the cohort
of patients with biopsy-proved IPF
between paired observers (n = 3). To
do this, the percentage likelihood given
to each diagnosis was assigned a grade
radiology.rsna.org
of 0 to 4, representing clinically useful
probabilities: grade 0, condition not
included in the differential diagnosis;
grade 1, unlikely (5%–25%); grade 2,
intermediate probability (30%–65%);
grade 3, high probability (70%–95%);
and grade 4, definite (100%). Observer
agreement was categorized as poor, fair,
moderate, good, or excellent according
to k values of less than 0.20, 0.21–0.40,
0.41–0.60, 0.61–0.80, and 0.81–1.00,
respectively (11).
Diagnostic Observation Evaluations
Diagnostic observations were evaluated
as individual observations, combined observations, and outcome in relation to
CT findings.
Individual observations.—The percentage likelihoods given by each observer to an IPF diagnosis for the core
study group (ie, the 55 biopsy-proved IPF
cases) were grouped according to grade.
Probability was high (ⱖ grade 3), intermediate (grade 2), or low (ⱕ grade 1).
Combined observations.—To extract the atypical IPF cases, we obtained
a unique estimation of the percentage
likelihood given to an IPF diagnosis by
combining the scores provided by the
three observers as shown in Table 1 .
If at least two of the three observers
favored the diagnosis of IPF at grade
3 or higher, IPF diagnostic probability
was high. If only one of three observers
favored the diagnosis of IPF at grade 3
or higher and at least one of the other
two observers favored the diagnosis
959
THORACIC IMAGING: Idiopathic Pulmonary Fibrosis: Nondiagnostic Thin-Section CT Diagnoses
Figure 1
Figure 2
Figure 1: Biopsy-proved IPF in 53-year-old woman. Thin-section CT pattern was
thought to be consistent with IPF by two observers, who scored UIP with 60% and
NSIP with 40% probability. Conversely, the third observer scored NSIP with 60% and
UIP with 40% probability. Transverse CT section through the lower zones shows a
peripheral reticular pattern with traction bronchiectasis but no honeycombing.
of IPF at grade 2 (for example, 70%,
observer 1; 40%, observer 2; 20%,
observer 3) or at least two of the
three observers favored the diagnosis
of IPF at grade 2 (for example, 40%,
observer 1; 55%, observer 2; 5%,
observer 3), IPF diagnostic probability was intermediate. If at least two of
three observers assessed the diagnosis
of IPF at grade 1 or lower, IPF diagnostic probability was low, and these
cases constituted the atypical IPF
group. The diagnostic probabilities in
the group of cases with a low probability of IPF (ie, atypical IPF) were then
examined to establish the first-choice
diagnoses by applying the same subclassification system (ie, first-choice
diagnosis with high or intermediate
probability and no single favored
first-choice diagnosis) and assessing
the differential diagnoses.
Outcome in relation to CT findings.—
To assess the longitudinal behavior of
atypical IPF, we recorded pulmonary function 1 year later (range, 9–15 months).
Functional deterioration was defined
as a decrease greater than 15% in
diffusing capacity of carbon monoxide
(12). All analyses were performed by
using commercially available software
960
Sverzellati et al
Figure 2: Biopsy-proved IPF in 47-year-old man that was interpreted as
high-probability (ⱖ70%) NSIP by all three observers. Transverse thin-section CT
scan obtained through the lower lungs shows a basilar peripheral predominant
reticular pattern with ground-glass opacity and traction bronchiectasis.
(Stata version 4; StataCorp, College
Station, Tex).
Results
Individual Observations
Thin-section CT appearances in the
core study group were interpreted as
high-probability IPF in 20 (36%), 13
(23%), and nine (16%) of 55 patients
by the three observers. One observer
interpreted cases as intermediate IPF
more frequently (18 [33%] of 55) than
did the other two observers (seven
[13%] and four [7%] of 55). The firstchoice thin-section CT diagnoses in
low-probability IPF cases (biopsy proved)
are shown in Table 2. NSIP was the most
frequent first-choice diagnosis, reported
in 20 (71%) of 28 cases (95% confidence
interval (CI): 53%, 85%), in 21 (55%)
of 38 cases (95% CI: 40%, 70%), and
in 17 (61%) of 28 cases (95% CI: 42%,
76%). The other first-choice diagnoses were chronic HP, sarcoidosis, and
organizing pneumonia (Table 2). IPF
was not given as a differential diagnosis
by individual observers in 21 (75%)
of 28 cases (95% CI: 56%, 87%), 28
(74%) of 38 cases (95% CI: 58%, 85%),
and in 19 (68%) of 28 cases (95% CI:
49%, 82%).
Combined Observations
Diagnoses of IPF were reported with
high, intermediate, and low probability
in 15 (27%), six (11%), and 34 (62%)
of 55 patients, respectively. In cases
with the diagnosis of IPF expressed
with intermediate probability, the major
differential diagnosis was NSIP (n = 5)
or chronic HP (n = 1) (Fig 1).
First-choice diagnoses with their
grades of probability and the associated
frequent differential diagnoses are summarized for atypical IPF cases in Table 3.
In cases with atypical IPF, NSIP was the
first-choice diagnosis in 18 (53%) of 34
cases (95% CI: 37%, 69%), with high
probability in 16 cases and intermediate probability in two cases (Fig 2). No
single diagnosis was favored (eg, NSIP,
75%, observer 1; IPF, 50%, observer 2;
chronic HP, 90%, observer 3) in eight
(23%) of 34 cases (95% CI: 13%, 40%).
Chronic HP, sarcoidosis, and organizing
pneumonia were also reported as highprobability first-choice diagnoses in four
(12%) (95% CI: 5%, 27%), three (9%)
(95% CI: 3%, 24%), and one (3%) of 34
cases, respectively (Figs 3–5). Frequent
radiology.rsna.org
n
Radiology: Volume 254: Number 3—March 2010
THORACIC IMAGING: Idiopathic Pulmonary Fibrosis: Nondiagnostic Thin-Section CT Diagnoses
Sverzellati et al
Figure 3
Figure 3: Biopsy-proved IPF in 59-year-old man that was interpreted as high-probability chronic HP by all three observers. Transverse thin-section CT scans at levels of (a) aortic arch and (b) lung bases show bilateral patchy areas of ground-glass opacity superimposed on fine reticulation with minimal subpleural honeycombing
(curved arrow). Some lobules appear to be of relatively decreased attenuation, suggesting air trapping (straight arrows).
Table 2
First-Choice Diagnoses in Low-Probability IPF Cases Expressed by Each Observer in
Subgroup of Patients with Biopsy-proved IPF (n = 55)
First-Choice Diagnosis
in Low-Probability IPF
NSIP
Chronic HP
Sarcoidosis
Organizing pneumonia
Indeterminate
Observer 1, 28
Observations
Observer 2, 38
Observations
Observer 3, 28
Observations
20
4
3
1
0
21
4
10
2
1
17
8
0
2
1
Note.—IPF was not in the differential diagnosis in 21, 28, and 19 observations for observers 1, 2, and 3, respectively.
Table 3
Combined Results of Three Observers for Assessing Thin-Section CT Appearance in
Patients with Low-Probability Biopsy-proved IPF (n = 34)
First-choice Diagnosis
Frequent Differential Diagnoses
NSIP (n = 18): high probability (n = 16)
and intermediate probability (n = 2)
Chronic HP (n = 4): all high probability
Sarcoidosis (n = 3): all high probability
IPF (n = 12) and chronic HP (n = 11)
IPF (n = 3) and NSIP (n = 3)
NSIP (n = 2), chronic HP (n = 2),
unclassifiable (n = 2)
NSIP (n = 1)
Chronic HP (n = 6), sarcoidosis
(n = 6), NSIP (n = 5), and IPF (n = 4)
Organizing pneumonia (n = 1): high probability
No single favored diagnosis (n = 8)
(k = 0.39 [95% CI: 0.34, 0.44]) agreement in the whole study population.
There was good to excellent agreement
between paired observers on the probability of IPF (kw = 0.73 [95% CI: 0.55,
0.91] between observer 1 and observer
2; kw = 0.73 [95% CI: 0.56, 0.90] between observer 1 and observer 3; kw =
0.82 [95% CI: 0.65, 0.99] between observer 2 and observer 3) and moderate
to good agreement in the 55 patients
with biopsy-proved IPF (kw = 0.59 [95%
CI: 0.33, 0.85], 0.62 [95% CI: 0.36,
0.88], and 0.74 [95% CI: 0.48, 0.99],
respectively).
Outcome in Relation to CT Findings
At 1-year follow-up, patients in 23 of
34 atypical IPF cases were examined;
there were two (6%) deaths, and nine
(26%) patients were not available. The
mean changes in diffusing capacity of
carbon monoxide were 215% 6 30.7.
Eleven (48%) of 23 cases showed a
decrease in diffusing capacity of carbon
monoxide greater than 15%.
Discussion
differential diagnoses, although not always the first-choice diagnosis, included
NSIP (n = 29), chronic HP (n = 23), and
sarcoidosis (n = 9). IPF was included in
the differential diagnosis in 19 cases.
Radiology: Volume 254: Number 3—March 2010
n
Observer Variation
There was moderate (k = 0.45 [95%
CI: 0.32, 0.58]) agreement for the firstchoice diagnosis in the cohort of patients
with biopsy-proved IPF and only fair
radiology.rsna.org
Our findings highlight the danger of
excluding the diagnosis of IPF on the
basis of thin-section CT appearances
alone. IPF may be misdiagnosed at
thin-section CT as NSIP, chronic HP,
961
THORACIC IMAGING: Idiopathic Pulmonary Fibrosis: Nondiagnostic Thin-Section CT Diagnoses
Sverzellati et al
Figure 4
Figure 4: Biopsy-proved IPF in 58-year-old woman. Two of three observers
diagnosed sarcoidosis with 100% probability, whereas the other observer
split the diagnosis between chronic HP (55% probability) and NSIP (45%
probability). Transverse thin-section CT scans at levels of (a) trachea
and (b) bronchus intermedius show reticular pattern composed of irregular
intralobular lines and subpleural nodularity (straight arrows) and some
patchy honeycombing (curved arrow). (c) Transverse section through lung
bases demonstrates relative sparing of lower zones with limited subpleural
reticulation and patchy ground-glass opacity.
or sarcoidosis. NSIP was the most frequent first-choice diagnosis in cases not
identified as IPF at thin-section CT, and
this diagnosis was made with high levels
of confidence. In a study of 98 patients,
Flaherty et al (7) showed that 26 (35%)
of 73 patients with UIP at biopsy had a
thin-section CT appearance more akin
to that of NSIP. Elliot et al (13) reported
an NSIP-like pattern in 21 (44%) of 48
readings in patients with a histologic diagnosis of UIP. In a study in which the
thin-section CT findings of various idiopathic interstitial pneumonias were compared in 92 patients, 15 of 40 readings
of false-negative diagnoses in patients
with IPF were misdiagnosed as NSIP
(14). Furthermore, in the study by Silva
et al (15), thin-section CT findings in patients with IPF were interpreted as definite NSIP in eight (32%) of 25 patients.
However, in many patients with atypi962
cal IPF in our study group, thin-section
CT appearances did not resemble NSIP.
Chronic HP was the first-choice diagnosis in 12% of such patients and
was the second most favored differential diagnosis. It has previously been
documented that there is overlap between the thin-section CT findings of
chronic HP and those of either NSIP
or UIP (16,17). In a study in which the
accuracy of thin-section CT in distinguishing chronic HP from IPF and NSIP
was assessed, seven (15%) of 46 readings
of false-negative diagnoses in patients
with IPF were misdiagnosed as chronic
HP (16).
Although the finding of one IPF case
misdiagnosed as organizing pneumonia
with a high level of confidence is, at
first sight, surprising, the presence of
areas of consolidation partly covering
the underlying fibrotic features of UIP
may have misled the observers. In a
study in which thin-section CT findings
were evaluated in 129 patients who had
various interstitial pneumonias, UIP
was incorrectly diagnosed as organizing
pneumonia in six (8%) of 70 readings
(18). In the study of Sumikawa et al (9),
a number of IPF cases with CT findings
suggestive of an alternative diagnosis
displayed airspace consolidation.
Sarcoidosis rarely may produce
thin-section CT appearances resembling those of IPF (19), but we report
the contrary situation in which IPF
mimicked sarcoidosis. Sarcoidosis was
recorded as a highly confident firstchoice diagnosis in three patients with
IPF and was among the differential diagnoses given in six cases.
The observation that 34 (62%) of
55 patients with biopsy-proved IPF
were regarded as having alternative
radiology.rsna.org
n
Radiology: Volume 254: Number 3—March 2010
THORACIC IMAGING: Idiopathic Pulmonary Fibrosis: Nondiagnostic Thin-Section CT Diagnoses
Sverzellati et al
Figure 5
Figure 5: Biopsy-proved IPF in 44-year-old woman that was interpreted by the three observers as 60%, 70%, and 80% probability for organizing pneumonia.
(a) Transverse thin-section CT scan at level of bronchus intermedius shows patchy subpleural consolidation (straight arrows) with peripheral reticular opacity and
honeycombing (curved arrow). (b) Transverse scan at lung bases shows focal consolidation on the right (straight arrow) on a background of ground-glass opacity with
some microcystic honeycombing (curved arrow).
diagnoses by the observers participating this study is a function of the bias
of performing surgical lung biopsies in
patients who lack typical thin-section
CT findings of UIP. As compared with
results from previous studies in which
the atypical thin-section CT findings in
IPF were assessed (7–10,13,14), the
greater variability of diagnoses in atypical IPF cases in our study may have several explanations. In the previous studies, observers were asked to limit their
differential diagnoses to the subgroups
of the idiopathic interstitial pneumonias
(7,13,14). In contrast, the observers in
our study were asked to list their diagnoses, with no limit to the type of diagnoses and without the provision of any
clinical information or knowledge of the
purpose of the study. Furthermore, an
advantage of our study was that cases
were not limited to idiopathic interstitial
pneumonias because the study cohort
included a mixed group of other chronic
interstitial lung diseases, thereby more
closely mirroring the patient population
encountered in clinical practice.
We have shown that patients with
atypical IPF had a progressive course
that does not differ from that of patients with typical IPF. In keeping with
our finding of a significant functional
decrease at 1 year in 48% of the atypical IPF cases, results from a drug trial
Radiology: Volume 254: Number 3—March 2010
n
that included 182 patients with IPF with
thin-section CT appearances consistent
with UIP showed a similar decrease
in 45% of cases (20). This observation is important because decreases in
diffusing capacity of carbon monoxide
of 15% or more from baseline across
12 months are associated with an increased risk of death in patients with
IPF (12,21). Our findings indicate that,
in patients with inexorably progressive
interstitial fibrosis but with thin-section
CT appearances suggestive of a diagnosis other than IPF, the likelihood is that
the diagnosis is IPF.
Our study had several limitations.
We could not define the overall frequency with which thin-section CT findings are atypical for IPF because this
was not a population-based study. Notably, no radiologic-histopathologic correlations were undertaken to establish
whether some of the variability of thinsection CT appearances of IPF cases
was caused by specific histopathologic
features, such as foci of organizing
pneumonia in a background of UIP or
discordant UIP and NSIP features in
specimens obtained at two different levels. The lack of a detailed assessment of
individual thin-section CT findings and
their distribution precluded a more complete characterization of these cases,
which might have been helpful to better
radiology.rsna.org
understand the observers’ choices of diagnoses. Some selection bias may exist
because the study cohort included only
patients referred to specialized centers
that are more likely to include complex
or atypical cases. Furthermore, experienced observers working in teaching
hospitals may be more likely than community-based radiologists to consider a
diagnosis other than IPF (22).
In conclusion, the results of our
study show that patients with IPF may
have thin-section CT appearances other
than the classic description. Because
an IPF diagnosis is important from a
prognostic standpoint, careful attention
should be given to the diagnostic possibility of IPF in such cases that show the
longitudinal behavior of IPF.
References
1. Kim DS, Collard HR, King TE Jr. Classification and natural history of the idiopathic
interstitial pneumonias. Proc Am Thorac
Soc 2006;3(4):285–292.
2. Maher TM, Wells AU, Laurent GJ. Idiopathic pulmonary fibrosis: multiple causes
and multiple mechanisms? Eur Respir J
2007;30(5):835–839.
3. American Thoracic Society; European Respiratory Society. American Thoracic Society/
European Respiratory Society International
Multidisciplinary Consensus Classification of
963
THORACIC IMAGING: Idiopathic Pulmonary Fibrosis: Nondiagnostic Thin-Section CT Diagnoses
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(2):277–
304.[Published correction appears in Am J
Respir Crit Care Med 2002;166(3):426.].
4. Mueller-Mang C, Grosse C, Schmid K,
Stiebellehner L, Bankier AA. What every
radiologist should know about idiopathic
interstitial pneumonias. RadioGraphics 2007;
27(3):595–615.
5. Lynch DA, Travis WD, Müller NL, et al.
Idiopathic interstitial pneumonias: CT features. Radiology 2005;236(1):10–21.
6. Hunninghake GW, Zimmerman MB,
Schwartz DA, et al. Utility of a lung biopsy
for the diagnosis of idiopathic pulmonary fibrosis. Am J Respir Crit Care Med
2001;164(2):193–196.
7. Flaherty KR, Thwaite EL, Kazerooni EA,
et al. Radiological versus histological diagnosis in UIP and NSIP: survival implications. Thorax 2003;58(2):143–148.
8. Lynch DA, Godwin JD, Safrin S, et al.
High-resolution computed tomography in
idiopathic pulmonary fibrosis: diagnosis
and prognosis. Am J Respir Crit Care Med
2005;172(4):488–493.
9. Sumikawa H, Johkoh T, Colby TV, et al.
Computed tomography findings in pathological usual interstitial pneumonia: rela-
964
Sverzellati et al
tionship to survival. Am J Respir Crit Care
Med 2008;177(4):433–439.
10. MacDonald SL, Rubens MB, Hansell DM,
et al. Nonspecific interstitial pneumonia and
usual interstitial pneumonia: comparative
appearances at and diagnostic accuracy of
thin-section CT. Radiology 2001;221(3):
600–605.
11. Coblentz CL, Babcook CJ, Alton D,
Riley BJ, Norman G. Observer variation
in detecting the radiologic features associated with bronchiolitis. Invest Radiol
1991;26(2):115–118.
12. Flaherty KR, Mumford JA, Murray S, et al.
Prognostic implications of physiologic and
radiographic changes in idiopathic interstitial pneumonia. Am J Respir Crit Care
Med 2003;168(5):543–548.
13. Elliot TL, Lynch DA, Newell JD Jr, et al.
High-resolution computed tomography features of nonspecific interstitial pneumonia
and usual interstitial pneumonia. J Comput
Assist Tomogr 2005;29(3):339–345.
14. Sumikawa H, Johkoh T, Ichikado K, et al.
Usual interstitial pneumonia and chronic
idiopathic interstitial pneumonia: analysis
of CT appearance in 92 patients. Radiology
2006;241(1):258–266.
15. Silva CI, Müller NL, Hansell DM, Lee KS,
Nicholson AG, Wells AU. Nonspecific interstitial pneumonia and idiopathic pulmonary
fibrosis: changes in pattern and distribution of disease over time. Radiology 2008;
247(1):251–259.
16. Silva CI, Müller 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(1):288–297.
17. Zompatori M, Calabrò E, Chetta A, Chiari
G, Marangio E, Olivieri D. Chronic hypersensitivity pneumonitis or idiopathic pulmonary fibrosis? Diagnostic role of high
resolution computed tomography (HRCT).
Radiol Med 2003;106(3):135–146.
18. Johkoh T, Müller NL, Cartier Y, et al. Idiopathic interstitial pneumonias: diagnostic
accuracy of thin-section CT in 129 patients.
Radiology 1999;211(2):555–560.
19. Padley SP, Padhani AR, Nicholson A,
Hansell DM. Pulmonary sarcoidosis mimicking cryptogenic fibrosing alveolitis on
CT. Clin Radiol 1996;51(11):807–810.
20. Demedts M, Behr J, Buhl R, et al. Highdose acetylcysteine in idiopathic pulmonary
fibrosis. N Engl J Med 2005;353(21):2229–
2242.
21. Latsi PI, du Bois RM, Nicholson AG, et al.
Fibrotic idiopathic interstitial pneumonia:
the prognostic value of longitudinal functional trends. Am J Respir Crit Care Med
2003;168(5):531–537.
22. 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(10):1054–1060.
radiology.rsna.org
n
Radiology: Volume 254: Number 3—March 2010