The term “unclassifiable interstitial lung disease” (unclassifiable ILD) was introduced in the American Thoracic Society/European Respiratory Society Consensus Classification of the Idiopathic Interstitial Pneumonias (IIP) in 2002 to encompass a subset of ILDs that cannot be classified within the confines of our current diagnostic framework (1). The paradoxical classification as “unclassifiable” results from either 1) inadequate or 2) discordant clinical, radiologic, and pathologic data, such that a specific ILD diagnosis is not possible (1). Rather than being a unique, cohesive entity, unclassifiable ILD is effectively a black box (Figure 1) constituted by cases that have slipped through our existing diagnostic algorithm. The mere existence of this category exposes the shortcomings of our current classification system. Succinctly put, as good as we are, we are not as good as we need to be. The fact that there has been no major progress in shrinking the size of the black box since its inception signals that we continue to fail our patients and the community. The nomenclature “unclassifiable” is indeed reminiscent of the term “nonspecific interstitial pneumonia”. Although nonspecific interstitial pneumonia is now a distinct clinical entity, when it was first described, it served as a wastebasket diagnosis similar to unclassifiable. For our patients and even many physicians, the naming of these two entities is itself nonspecific and understandably confusing.

Figure 1. The black box of unclassifiable ILD in the context of the most common ILD diagnoses. CTD-ILD = connective tissue disease-associated interstitial lung disease; HP: hypersensitivity pneumonitis; ILD = interstitial lung disease; IPF = idiopathic pulmonary fibrosis; NSIP = nonspecific interstitial pneumonia.

Download Figure | Download Powerpoint

[More] [Minimize]

In this edition of AnnalsATS, Guler and colleagues (pp. 854–863) share the results of their meta-analysis of the literature on unclassifiable ILD through the fall of 2017 (2). In an attempt to shed light on this elusive entity, they rigorously reviewed published studies and collated the highest-quality data with the lowest potential for bias. Using the resulting pooled data, the authors describe the definition, prevalence, and prognosis of unclassifiable ILD. The resulting message is clear and useful: this is an ill-defined entity with a heterogeneous prevalence depending on the definition used and whether surgical lung biopsy was performed or multidisciplinary discussion was held. Not unexpectedly, there is a lack of agreement on what unclassifiable means, as illustrated by the varied pathways to diagnosis. The reader is left with the impression that unclassifiable ILD is truly as nebulous as its name. However, there are notable findings that are worth mentioning. First and foremost, unclassifiable ILD is not rare and accounts for ∼12% of ILD (2). Even as the gold standard in ILD, a multidisciplinary discussion is not uniformly available and has its limitations. In this analysis, multidisciplinary discussion improved the diagnostic accuracy of ILD, but the prevalence of “unclassifiable group” only decreased modestly (2). Lastly, the prognosis of patients who receive a diagnosis of unclassifiable ILD is poor, with a 2-year survival of 70–76% and 5-year survival of 46–70% (2).

Unclassifiable ILD may represent the final common pathway of fibrotic lung disease, and regardless of the initial insult or cause, it may behave similarly. This is illustrated by the pooled survival statistics, which fall somewhere between published outcomes of idiopathic pulmonary fibrosis (IPF) and other fibrotic lung diseases (2). Although the “unclassifiable” designation is typically assigned after close review of high-resolution computed tomography with or without histopathology, there is likely additive value to incorporating a Bayesian approach with recognition of a heightened probability of IPF with increasing age, particularly once conditions such as connective tissue disease and chronic hypersensitivity pneumonitis are deemed unlikely. Thus, clinical context should be used as a tie-breaker in the formation of a presumptive unifying diagnosis in these difficult cases. Management should be based on the most probable diagnosis, acknowledging that this may change over time and can be revisited as new information arises (3). Along these lines, a recent diagnostic ontologic assignment of fibrotic ILD endorses levels of confidence for making a provisional diagnosis (4). This latitude may empower pulmonologists to provide a provisional diagnosis of IPF (with low or high confidence) and move toward shrinking the unclassifiable black box.

There are a number of reasons why making a diagnosis is important. First and foremost, it is important to patients. Patients who are given a diagnosis of unclassifiable ILD are left with the disappointment that not only do they have a serious disease that will likely cause their death, but also its pattern is so aberrant that it defies a meaningful nomenclature. This may conjure up feelings of abandonment because it implies that they are in a diagnostic desert without recourse to treatment. Although we counsel patients that surgical lung biopsy could be nondiagnostic, they are nonetheless under the impression that it is the last leg of a tripod on which to rest a unifying diagnosis. Patients can easily lose confidence and become disillusioned when their diagnosis changes from “we don’t know” to “unclassifiable” after they undergo an invasive test. Perhaps in our pursuit of diagnostic accuracy, the pendulum has swung too far.

Although it is semantics, classifying the unclassifiable matters for two reasons: prognosis and treatment. The inability to establish a clear diagnosis may result in inaction or delayed treatment. As we grapple with the question of treating these patients with immunosuppression versus antifibrotic therapy, we may inadvertently revert to a paralyzing “watch and wait” approach, which may deny patients the opportunity for early implementation of therapy. What if some of the unclassifiable cases are indeed IPF? Could these patients potentially benefit from antifibrotic therapy?

International guidelines for the diagnosis of IPF and the current published guidelines for chronic hypersensitivity pneumonitis seek to provide more guidance in the quest for an ILD diagnosis (4–6). The recently published Fleischner Society criteria for the radiographic diagnosis of usual interstitial pneumonitis seem to recognize the benefits of inclusivity and are more generous in assigning a diagnosis of IPF, but at the same time recognize and specify different degrees of confidence in the diagnosis (7). This change may result in a reduction in unclassifiable ILD and facilitate treatment with antifibrotic agents.

As time goes on, the continued discovery and use of molecular footprint patterns in ILD may result in less unclassifiable disease as well (8, 9). Diagnostic nomenclature may change to include more precise genomic fingerprinting. Indeed, categorization of ILDs may ultimately morph to molecular- or biomarker-based diagnoses, similar to those used in the oncology field. As this nascent field evolves, we will likely witness an evolution from a three-dimensional multidisciplinary discussion to a four-dimensional process, with molecular profiling providing the final piece of the puzzle.

The authors of this paper are to be commended for defining the scope of this issue, as recognition of the challenge is the first step in moving the needle forward. It is hoped that the collation of these data will provide a foundation for further research into this neglected and largely uncharted entity. What are we to do in the meantime? How should we counsel patients who may feel disenfranchised by the labeling of their disease as “unclassifiable”? At this point, aside from acknowledging the issue, we could also consider an alternate terminology that may convey that diagnosis is imperfect, and treatment and prognosis based on the most clinically likely diagnosis. “Evolving fibrotic lung disease” could be a fitting term that illustrates that not only is their disease dynamic, but also that the field itself is still evolving. This issue’s meta-analysis of unclassifiable ILD highlights the size of the black box and the need to shrink it…or at least paint it a different color.

References

Section:

ChooseTop of pageReferences <<CITING ARTICLES