Lung Cryobiopsy Versus Surgical Lung Biopsy for Diagnosing Fibrotic Interstitial Lung Disease: Still Looking for the “Perfect” Trial

Interstitial lung diseases (ILDs) comprise a heterogeneous group of conditions whose diagnosis requires the integration of clinical, serological, high-resolution computed tomographic, and possibly histological findings within a dynamic multidisciplinary discussion (MDD). Although this work-up leaves a considerable proportion (up to one-third) of patients without a definite or high-confidence diagnosis, the number of surgical lung biopsies (SLBs) declined steadily over the past years, with real-life experiences demonstrating a rate of SLB in ILDs of only approximately 5% (1). Moreover, a recent study showed that even ILD experts would prescribe an antifibrotic drug in a nonnegligible percentage of patients with a provisional low-confidence diagnosis of idiopathic pulmonary fibrosis (IPF), without the need for a biopsy (2).

In the last decade, the introduction of transbronchial lung cryobiopsy (TBCB) has provided clinicians with a less-invasive method to obtain relatively large, well-preserved histologic specimens that could help make informed decisions in patients with unclassifiable fibrotic ILDs after MDD (3). Studies showed that TBCB, included in the most recent update of the international clinical practice guidelines on IPF as a viable diagnostic alternative to SLB (4), increases the diagnostic confidence in the setting of MDD (5, 6) and has prognostic value (7). However, the European Respiratory Society guidelines on the role of TBCB in the diagnosis of ILDs demonstrated that the quality of the scientific evidence for the outcomes assessed is very low and that there is a need for well-designed studies in different procedure-related domains (3). One of the areas in which additional information would be welcome is the comparative diagnostic accuracy of TBCB and SLB, the latter being still considered the gold standard, at both the histopathologic and the MMD levels, as conflicting results emerged from the three studies designed with that specific aim and published up to now (8–10). Favorable data on TBCB safety and comparable diagnostic accuracy versus SLB could theoretically increase the use of TBCB in unclassifiable ILD cases after MDD and could expand its future use in settings such as that of the interstitial lung abnormalities, in which our knowledge is still poor.

In this issue of the Journal (pp. 1612–1619), Fortin and colleagues report the results of a Canadian multicenter prospective study (Canadian ICE [CAN-ICE]), in which patients with an unclassifiable or low-confidence diagnosis of chronic diffuse ILD after local MDD underwent TBCB and SLB in a single operative session (11). Histology was evaluated in a blinded fashion by three pathologists, and cases were then discussed by three independent multidisciplinary ILD teams, first considering the results of the TBCB (TBCB-MDD) and subsequently those of the SLB (SLB-MDD). The outcomes included within- and between-center diagnostic agreement between TBCB and SLB, both at the histopathologic and MDD levels. A post hoc subgroup analysis of within-center diagnostic agreement for TBCB-MDD and SLB-MDD in patients with an SLB-MDD diagnosis of IPF and fibrotic hypersensitivity pneumonitis (fHP) was also conducted. The results demonstrated a moderate TBCB-MDD/SLB-MDD diagnostic agreement (61.7%, k = 0.46), with most discordant cases being related to a TBCB-MDD diagnosis of IPF being classified as fHP at SLB-MDD. The between-center agreement, assessed for the first time in this study, was very low (k = 0.29) for TBCB-MDD, whereas it was substantial for SLB-MDD (k = 0.71). Importantly, the underlying ILD remained unclassifiable after TBCB and SLB in 13.3% and 6.7% of the cases, respectively.

The key message emerging from this study is the limited ability of TBCB to reliably distinguish fHP from IPF. Unfortunately, the differential diagnosis between IPF and fHP is a difficult one per se (12), with one-third of cases that cannot be classified confidently even when the results of an SLB are available (13). Interestingly, in the CHILL study, a final diagnosis of fHP was established in eight patients, of whom five had a usual interstitial pneumonia pattern at the pathologic examination of the SLBs (11). It is therefore unsurprising that the diagnosis of fHP has the lowest interobserver agreement among all the ILDs (14).

The results of the CAN-ICE study should be interpreted in the context of some methodological limitations. Besides the small sample size, the low number of TBCBs obtained per patient, and the possible recall of clinical and imaging details for paired cases at MDD, the limited experience of the MDD members and pathologists with TBCB, acknowledged by the authors, is especially worth noting. Digging into the data (Table S3 of the study) (11), the proportion of histological samples labeled as nondiagnostic, or yielding a histologic diagnosis of unclassifiable ILD or a low-confidence diagnosis of a specific pattern, was markedly higher for TBCB (68.3%) than for SLB (36.7%), with notable between-center differences. This finding might have significantly affected the overall diagnostic accuracy of TBCB and might partly explain the much lower between-center agreement for TBCB-MDD versus SLB-MDD. The COLDICE study, in fact, demonstrated a strong correlation between the confidence levels of the pathologists for TBCB interpretation and the confidence level of the subsequent MDD diagnosis (Spearman’s R = 0.64; P < 0.0001) (15).

All things considered, the CAN-ICE study provides a useful glimpse of what might be the results of a widespread diffusion of TBCB to centers lacking a specific clinical and, especially, pathological background with the procedure. Such a scenario is feared by both ILD experts and interventional pulmonologists with clinical and research experience with TBCB (3, 4). Although limiting the use of TBCB to expert centers would currently make it virtually impossible to offer this option to many patients with unclassifiable ILD after MDD, this would undoubtedly prevent a high number of noncontributory and possibly dangerous invasive procedures.