Idiopathic nonspecific interstitial pneumonia: Waters remain murky…remember ABCDEFGHI

Right since the term was used in the context of an interstitial lung disease (ILD) by Katzenstein and Fiorelli in 1994, nonspecific interstitial pneumonia (NSIP) has been controversial and perplexing.[1] It was initially described as a histopathological entity characterised by uniform involvement of alveolar septae with variable amounts of chronic inflammation and/or fibrosis that did not demonstrate the features of usual interstitial pneumonia (UIP), desquamative interstitial pneumonia, or acute interstitial pneumonia.[1,2] Over the last 3 decades, the clinical profile, radiologic findings, disease associations, prognosis, and treatment of NSIP have been described.[3] However, compared to other common entities grouped under ILD or diffuse parenchymal lung disease, the use of the term NSIP incites more doubt than clarity, unbecoming of a definite disease entity. Classically, NSIP is a term given to a histologic (as above) or radiologic pattern (various combinations of lower lobe predominant ground-glass opacities, reticulation, subpleural involvement or sparing, and traction bronchiectasis [Figure 1a and b]). The clinical use of this term as a disease is considered appropriate only when the disease is idiopathic (iNSIP). Idiopathic NSIP has a prevalence of about 3 per 100,000 population.[4] Its idiopathic nature can be established after excluding the known 'causes', or more appropriately the underlying conditions and associations, of an NSIP pattern.[5] The acronym ABCDEFGHI may help in not missing any of the important 'causes'.Figure 1: Various appearances of the NSIP pattern. (a) Fibrotic NSIP pattern with subpleural involvement (black arrow); (b) Fibrotic NSIP pattern with subpleural sparing (white arrow) seen in only 20% of patients; (c) Bibasilar reticulation with traction bronchiectasis suggesting a probable UIP pattern; fibrotic NSIP may have a similar appearance especially after resolution of ground-glass opacities; (d and e) (same patient) Bibasilar reticulation with traction bronchiectasis; separation from probable UIP is easier in this case due to the presence of ground-glass opacities, axially uniform involvement, and the presence of subpleural sparing (open arrow). (f) NSIP with extensive consolidation, termed as an overlap of NSIP and OP patterns; some radiologists might name it as a pure OP pattern. A - Acute lung injury in the recovering phase B - Bronchiolitis obliterans organising pneumonia (now known as organizing pneumonia [OP]) of the relapsing type C - Connective tissue diseases D - Drug exposures E - Environmental exposures including smoke, organic and inorganic dusts, and others F - Familial pulmonary fibrosis G - Graft-versus-host disease H - Human immunodeficiency virus infection and haematologic conditions (Rosai-Dorfman disease, multisystem Castleman disease, myelodysplastic syndromes) I - Immunoglobulin G4-related disease (IgG4-RD) There are several contentious issues surrounding iNSIP in the clinical, radiologic, and pathologic contexts.[3] Connective tissue diseases (CTDs) are, by far, the most common underlying conditions associated with an NSIP pattern. A clear diagnosis of a CTD along with a typical radiologic NSIP pattern is possibly the only clinical situation where the association of a well-defined disease (i.e., the CTD) with the NSIP pattern is concrete. Alternatively, if there are only vague clinical clues to a rheumatologic disorder, the association of NSIP to this 'undifferentiated' CTD introduces obscurity and clinical heterogeneity. The term interstitial pneumonia with autoimmune features (IPAF) remains a proposed research term for such an ILD, but it has not evolved into a definite clinical diagnosis, as of today.[6] Similarly, clinical uncertainty exists when attributing NSIP to drug exposures, to tobacco smoking (the so-called smoking-related interstitial fibrosis), to sequelae of viral pneumonia (especially coronavirus disease 2019) and when associating serum IgG4 levels to NSIP in the absence of typical manifestations of IgG4-RD. Radiologically, the criteria for fibrotic NSIP and probable UIP overlap.[7] Both are characterised by bibasilar reticulation with traction bronchiectasis [Figure 1c]. Potential factors favoring NSIP over UIP in such situations could be subpleural sparing (present in only about 20% of the patients), an axially uniform distribution of the abnormalities, and extensive ground-glass opacities [Figure 1d and e].[8] In the absence of these features, the agreement amongst radiologists on UIP versus NSIP is not good at all.[9] To complicate this further, a radiologic NSIP pattern may transform into an apparent UIP pattern over time.[10] Finding patchy consolidation on the background of peripheral ground-glass opacities and reticulation elicits discordant impressions from radiologists ranging from OP to NSIP through NSIP-OP overlap [Figure 1f].[11] The permissible extent of air trapping on expiratory scans acceptable for a pure NSIP diagnosis without raising suspicion for hypersensitivity pneumonitis (HP) remains uncertain. A threshold of mosaic attenuation/air trapping less than three lobes has been suggested for UIP/idiopathic pulmonary fibrosis (IPF).[12] Pathologically, the size of the biopsied lung tissue is the single most important factor that may influence the confidence of a multidisciplinary team (MDT) in categorizing NSIP as an idiopathic and 'standalone' pattern.[13] While a large lung biopsy may show classic NSIP in one part, other areas may reveal prominent lymphoid follicles with germinal centres with or without cysts, bronchiolocentric inflammation with ill-formed granulomas or giant cells, finding of OP in >20% of the biopsied lung, or a definite and extensive UIP pattern. This may convince the MDT to consider lymphocytic interstitial pneumonia, CTD-ILD, HP, relapsing OP, and IPF, respectively, as multidisciplinary diagnoses, rather than iNSIP. With a progressive shift from surgical lung biopsies to overdependence on imaging alone and to lung cryobiopsies, the probability of missing important ancillary findings would increase, resulting in increasing mislabelling of the above ILD subtypes as iNSIP.[14] It is in the above context that we must critique the study by Wijayaratne and colleagues published in this issue, who have commendably added another original research article to the growing NSIP literature. In a retrospective chart review of 44 subjects with NSIP, they found 28 (64%) cases to be idiopathic compared to 16 (36%) subjects in whom the NSIP was considered to be secondary to known 'causes'. This study leaves much to be desired apart from the small sample size. Crucial details including the causes of secondary NSIP, exposures including tobacco smoke and organic dusts, and the laboratory tests performed to exclude secondary causes have not been presented. A mean age of 74 years and a preponderance (64%) of males amongst iNSIP subjects raise the possibility of misclassifying IPF as iNSIP. The omission of radiologic findings from the manuscript strengthens this doubt. Evidence exists that about 10% of iNSIP patients develop a CTD during their clinical course.[15] Had such follow-up data been presented by the authors, it could boost the certainty in the idiopathic or secondary nature of the NSIP. Due to these important deficiencies and the retrospective nature of the study, the numerical difference in the reported mortality of 'idiopathic' and 'secondary' NSIP should be taken with a pinch of salt. Future studies on NSIP must ensure that such critical details are not omitted.