This copy is for personal use only. To order printed copies, contact [email protected] 1849 CHEST IMAGING American Thoracic Society– European Respiratory Society Classification of the Idiopathic Interstitial Pneumonias: Advances in Knowledge since 20021 Nicola Sverzellati, MD, PhD David A. Lynch, MB In the updated American Thoracic Society–European Respira- David M. Hansell, MD, FRCP, FRCR tory Society classification of the idiopathic interstitial pneumonias Takeshi Johkoh, MD, PhD (IIPs), the major entities have been preserved and grouped into Talmadge E. King, Jr, MD (a) “chronic fibrosing IIPs” (idiopathic pulmonary fibrosis and id- William D. Travis, MD iopathic nonspecific interstitial pneumonia), (b) “smoking-related IIPs” (respiratory bronchiolitis–associated interstitial lung disease Abbreviations: H-E = hematoxylin-eosin, and desquamative interstitial pneumonia), (c) “acute or subacute IIP = idiopathic interstitial pneumonia, IPF = IIPs” (cryptogenic organizing pneumonia and acute interstitial idiopathic pulmonary fibrosis, NSIP = nonspe- cific interstitial pneumonia, RB-ILD = respi- pneumonia), and (d) “rare IIPs” (lymphoid interstitial pneumonia ratory bronchiolitis–associated interstitial lung and idiopathic pleuroparenchymal fibroelastosis). Furthermore, it disease, UIP = usual interstitial pneumonia has been acknowledged that a final diagnosis is not always achiev- RadioGraphics 2015; 35:1849–1872 able, and the category “unclassifiable IIP” has been proposed. The Published online 10.1148/rg.2015140334 diagnostic interpretation of the IIPs is often challenging because Content Codes: other diseases with a known etiology (most notably, connective tis- 1From the Section of Diagnostic Imaging, Depart- sue disease and hypersensitivity pneumonitis) may show similar ment of Surgical Sciences, University of Parma, morphologic patterns. Indeed, more emphasis has been given to the Via Gramsci 14, 43100 Parma, Italy (N.S.); De- partment of Radiology, National Jewish Health, integration of clinical, computed tomographic (CT), and patholog- Denver, Colo (D.A.L.); Department of Radiol- ic findings for multidisciplinary diagnosis. Typical CT-based mor- ogy, Royal Brompton Hospital, London, England (D.M.H.); Department of Radiology, Kinki Cen- phologic patterns are associated with the IIPs, and radiologists play tral Hospital of Mutual Aid Association of Public an important role in diagnosis and characterization. Optimal CT School Teachers, Hyogo, Japan (T.J.); Depart- quality and a systematic approach are both pivotal for evaluation of ment of Medicine, University of California–San Francisco, San Francisco, Calif (T.E.K.); and IIP. Interobserver variation for the various patterns encountered in Department of Pathology, Memorial Sloan Ket- the IIPs is an issue. It is important for radiologists to understand tering Cancer Center, New York, NY (W.D.T.). Received December 9, 2014; revision requested the longitudinal behavior of IIPs at serial CT examinations, espe- January 22, 2015, and received February 12; ac- cially for providing a framework for cases that are unclassifiable or cepted February 17. For this journal-based SA- in which a histologic diagnosis cannot be obtained. CME activity, the authors D.A.L. and D.M.H. have provided disclosures (see p 1868); all other ©RSNA, 2015฀•฀radiographics.rsna.org authors, the editor, and reviewers have disclosed no relevant relationships. Address correspon- dence to N.S. (e-mail: [email protected]). See discussion on this article by McLoud (pp 1871–1872). Introduction ©RSNA, 2015 The American Thoracic Society–European Respiratory Society clas- sification of the idiopathic interstitial pneumonias (IIPs) underwent SA-CME LEARNING OBJECTIVES revision in 2013 (1). This update was not designed as a stand-alone After completing this journal-based SA-CME document but as a supplement to the previous 2002 IIP classifica- activity, participants will be able to: tion, which defined key individual diseases and the best diagnostic ■ Explain optimal-quality CT requirements approach to them (2). Pathologic patterns previously formed the basis for the evaluation of patients with IIP. for the classification of IIP subtypes, whereas in the update, more em- ■ Distinguish fibrotic from nonfibrotic CT phasis is given to the integration of clinical, computed tomographic features of the IIPs. (CT), and pathologic findings for multidisciplinary diagnosis (1). ■ Describe various longitudinal CT changes that may be seen in patients with any IIP. See www.rsna.org/education/search/RG. 1850 November-December 2015 radiographics.rsna.org cases that are unclassifiable or in which a histo- TEACHING POINTS logic diagnosis cannot be obtained (Table 2) (1). ■ A systematic approach to CT of the chest in subjects suspect- Importantly, imaging will play an ever-increasing ed of having IIP entails the evaluation of image quality, the role in monitoring IIP and in refining the diag- precise description of specific disease features with standard terminology, and the determination of the distribution of dis- nosis of chronic fibrosing lung diseases as their ease in the axial and craniocaudal planes. The most important disease behavior becomes evident. next step is to determine whether the CT features permit a The authors of the updated IIP classification high-confidence diagnosis of UIP. In cases that do not meet document reviewed MEDLINE publications strict criteria for UIP, the presence of traction bronchiectasis, relevant to key imaging questions for the years architectural distortion, or volume loss usually permits distinc- tion of fibrosing from nonfibrosing interstitial pneumonia. 2001–2011. However, there was inadequate ■ Terms such as UIP pattern and NSIP pattern are helpful in dem- space in the document to address several im- onstrating that one is discussing the radiologic pattern rather portant imaging-related issues, for example, the than the associated clinical-pathologic syndrome. Indeed, it is key CT principles for the evaluation of IIP, the important to appreciate that the CT patterns, especially those fundamental CT features of each entity, and the of the chronic fibrosing IIPs, may encompass several clinical- longitudinal behavior of these features. The pur- radiologic-pathologic entities; for example, NSIP and UIP pat- terns may be found in hypersensitivity pneumonitis and in pose of this review article is to expand the lit- association with connective tissue diseases. erature search through June 2014, concentrating ■ The primary role of CT is to separate chronic fibrosing lung on the CT aspects of IIP that were not covered diseases with a UIP pattern from those with non-UIP lesions, in depth in the 2013 classification but are nev- including those with findings associated with other IIPs (eg, ertheless important to radiologists tasked with RB-ILD, lymphoid interstitial pneumonia, and idiopathic pleu- tackling this difficult group of disorders. First, roparenchymal fibroelastosis). the principles of radiologic IIP assessment are ■ CT appearances of idiopathic pleuroparenchymal fibroelasto- reviewed. Then the four IIP groups are covered: sis are distinctive, with irregular pleural thickening and “tags” in the upper zones that merge with fibrotic changes in the (a) chronic fibrosing IIPs, (b) smoking-related subjacent lung, associated with evidence of substantial upper IIPs, (c) acute or subacute IIPs, and (d) rare lobe volume loss (architectural distortion, traction bronchiec- IIPs. Finally, the category of unclassifiable IIP is tasis, and hilar elevation). discussed. ■ The 2013 IIP consensus statement acknowledges, as did the 2002 consensus statement, that a final diagnosis is not always Key Principles for achievable. Radiologic Assessment of IIP CT is pivotal for the evaluation of IIP. Optimal- quality CT requires the use of thin-section (,2- In the updated classification, the major entities mm) reconstruction and high-spatial-resolution have been preserved and grouped into (a) “chronic reconstruction. Images should be obtained dur- fibrosing IIPs” (idiopathic pulmonary fibrosis [IPF] ing a full inspiration to total lung capacity (3). and idiopathic nonspecific interstitial pneumonia Despite the risk of increasing the radiation dose [NSIP]), (b) “smoking-related IIPs” (respira- delivered to patients, volumetric CT acquisition tory bronchiolitis–associated interstitial lung with multidetector CT is generally preferred to disease [RB-ILD] and desquamative interstitial noncontiguous imaging (4). Importantly, volu- pneumonia), and (c) “acute or subacute IIPs” metric CT acquisition and derived multiplanar (cryptogenic organizing pneumonia and acute reconstructions (a) allow improved character- interstitial pneumonia). In addition, idiopathic ization of patchy disease; (b) allow better delin- pleuroparenchymal fibroelastosis has been newly eation of disease extent; (c) simplify analysis of included and classified together with lymphoid disease distribution (5); (d) allow identification interstitial pneumonia in a fourth group, “rare of ancillary findings such as nodules or bron- IIPs” (Table 1). In the updated classification, chiectasis; (e) facilitate differentiation between it has again been acknowledged that in a sub- honeycombing and traction bronchiectasis, set of cases, a final diagnosis cannot always be which may be crucial to diagnose the UIP pat- achieved despite extensive multidisciplinary dis- tern, as shown in Figures 1 and 2 (6); and (f) cussion, and such cases are considered “unclas- optimize follow-up CT matching for any com- sifiable”