Making a 'Working Diagnosis' of Idiopathic Pulmonary Fibrosis

Aaron B. Holley, MD


August 15, 2019

Classification of interstitial lung disease (ILD) is challenging.[1] The most common and best studied ILD is idiopathic pulmonary fibrosis (IPF). Still, it makes up only one third of all ILD cases, and diagnosis is far from easy. Clinicians must integrate demographic, clinical, and CT features and decide whether surgical lung biopsy (SLB) would be helpful.[2] With expensive but effective therapies now available to treat IPF, making an accurate diagnosis is more important than ever.

How exactly are clinical and radiographic features used to decide whether SLB is needed to make an IPF diagnosis? After the clinician has excluded relevant exposures or medications as a cause for ILD and confirmed that there's no classifiable autoimmune disease present, the pattern on CT is assessed. According to the 2018 ATS/ERS guidelines on IPF diagnosis, if the CT shows a usual interstitial pneumonia pattern, a diagnosis of IPF can be made. In this case, the likelihood of histologic confirmation is between 90% and 100%, so SLB is unnecessary. If CT shows any other pattern, a multidisciplinary discussion is required and SLB is considered.

A Fleischner Society White Paper on IPF, also published in 2018, is slightly less rigid and allows for a "working diagnosis" of IPF to be made without usual interstitial pneumonia pattern on CT and without SLB.[3] A "working diagnosis" of IPF is defined as a "nondefinite diagnosis made with sufficient confidence to justify disease-specific therapy." At this time, it's less clear what the percentage likelihood of an actual IPF diagnosis (confirmed by histology) is when a working diagnosis is made. Furthermore, there is no general acceptance of what the likelihood threshold for IPF should be before SLB is considered unnecessary.

Authors of a manuscript published online in June in the American Journal of Respiratory and Critical Care Medicine (AJRCCM) attempted to attach likelihood thresholds to a working diagnosis of IPF.[4] Using a framework for diagnosing ILD that was published by Ryerson and colleagues 2 years ago,[5] they surveyed clinicians and asked them to attach a percentage likelihood for an IPF diagnosis using 60 ILD cases. Respondents to the survey were also asked, for each case, whether they would pursue SLB and what their initial treatment strategy would be.

Survey responses revealed that the majority of clinicians (63%) would prescribe antifibrotics without SLB when the likelihood for IPF was ≥ 70% but < 90%. By the Ryerson framework, this level of likelihood was labeled "provisional high confidence." Even when these provisional high confidence patients were considered low risk (age < 65 and diffusing capacity of the lungs for carbon monoxide > 40%), SLB wasn't pursued in 40.1% of patients. On the basis of these responses, the authors concluded that clinicians embrace the concept of a working diagnosis for IPF, and a diagnostic likelihood of 70% represents the threshold above which a majority would forego SLB. Another important, but not surprising, finding was that agreement on need for SLB was poor (κ = 0.15; interquartile range, 0.05-0.26). It got better among physicians who had access to interdisciplinary team meetings, but not by much.

You can infer what you wish from the survey findings. It does seem that clinicians have embraced the working diagnosis concept. Without access to a multidisciplinary discussion, however, it's not entirely clear that they should. Given that patient mortality was statistically equivalent when comparing definite IPF (90%-100% likelihood) to provisional high confidence IPF (70%-90% likelihood), though, perhaps the poor agreement on SLB is less important. In either case, standardizing an IPF diagnosis remains a struggle. The discussions prompted by the subtle differences between the ATS/ERS 2018 Guidelines and the Fleischner Society White Paper, along with the results from the AJRCCM survey, do bring us closer to a shared mental model for diagnosing IPF. We still have a way to go.

Follow Medscape on Facebook, Twitter, Instagram, and YouTube


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: