Concordance between sequential transbronchial lung cryobiopsy and surgical lung biopsy in patients with diffuse interstitial lung disease

We compared pathology diagnosis and histological findings of cryobiopsy and surgical lung biopsy from identical patients in detail and found that cryobiopsy is useful for the recognition of UIP pattern, but may be cautious for other diseases. Findings such as organizing pneumonia may not be represented.

Transbronchial cryobiopsy (TBLC) is a relatively new method for obtaining lung tissue for diagnostic purposes [1]. It has been performed mainly for the diagnosis of lung endobronchial neoplasms and their genetic testing [2, 3]. In recent years, it has also been used for the pathological diagnosis of diffuse lung disease [4]. The method can be used to collect approximately 10–30 mm² of tissue samples which often cover enough areas to observe the primary lobules of the lung parenchyma [1, 5]. This is larger than the specimen collected using forceps biopsy [6]. These samples include relatively less artifacts such as nuclear crush and alveolar collapse than samples from transbronchial forceps biopsy and provide more information needed for making a pathological diagnosis of diffuse lung disease [7, 8]. In past studies, TBLC was diagnostic for diffuse lung disease in 70–80% of cases [4, 9‐12], which is higher than transbronchial lung forceps biopsy [5, 13]. TBLC is accurate and efficient especially in diagnosing Usual interstitial pneumonia (UIP) [14]. In addition, similar to surgical lung biopsy (SLB), improvement in diagnostic accuracy has been reported by multidisciplinary discussion (MDD) when making a diagnosis with TBLC samples [15, 16]. However, most studies did not compare TBLC and SLB in the same patients. Observational differences between TBLC and SLB caused by different sample sites and methods are important information in diagnostic confirmation. Recently, Romagnoli and Colby reported poor concordance of pathology diagnoses between TBLC and SLB [17]. However, at present, no study has performed TBLC and SLB on the same patients and compared their histopathological characteristics.

TBLC has been reported to be useful for the diagnosis of diffuse interstitial lung disease and its use has been increasing since 2014 [19]. However, no study has compared the histopathological findings and diagnosis of the same patient using TBLC and SLB. This is the first report in which 7 patients with diffuse pulmonary disease who underwent both TBLC and SLB were selected and their histopathological findings as well as diagnosis were compared.

TBLC and SLB diagnosis based on UIP guideline were consistent in 5 out of 7 cases. In the 2 cases with discordant results, the original diagnosis with TBLC was not contradicted but rather changed from indeterminate for UIP to probable UIP. There was no false positive diagnosis of UIP with TBLC. Although the specificity of TBLC may be inferior in UIP diagnosis, its sensitivity is determined to be satisfactory. In other words, if enough observational findings that indicate UIP are obtained using TBLC, a definitive diagnosis of UIP without SLB could possibly be provided.

Histopathological assessment included examination of 10 items. Dense fibrosis, peripheral distribution, and fibroblastic foci which are the diagnostic basis of UIP according to the official 2018 ATS/ERS/JRS/ALAT clinical practice guidelines for diagnosis of idiopathic pulmonary fibrosis [18] were largely consistent between TBLC and SLB diagnoses., discordant tendencies were noted in the assessment of diffuse lesions such as cellular IP, lesions around the airways such as PBM, and discrete lesions of the lung such as organized pneumonia. This may be causally related to the attributes of TBLC; disseminated lesions within the lung may not be collected using this method, fibrosis along the peripheral lung or other main lesions of the patient may not be represented in the collect specimen, and the small sample may over assess regions that are considered minor in SLB specimen. Pathologists must be aware of these biases which could lead to misdiagnosis. Because specimen do not contain tissues directly beneath the pleura, all 4 pathologists participating in this study agreed that lung tissue around the bronchovascular bundle should be recognized as peripheral lobule and make a decision based on their assessment.

Tomasetti et al. suggested that MDD improves the accuracy of IPF diagnosis with cryobiopsy, and therefore it is a critical element to diagnosis [15]. There was 1 case in our study which was difficult to be determined as IPF with cryobiopsy only and after MDD, CHP was more strongly suspected (case 2). MDD is thought to be highly significant in diagnosing ILD with TBLC.

SLB is generally performed in cases with low diagnostic confidence of cryobiopsy, which includes not only the uncertainty of histopathological diagnosis but also the inconsistency with HRCT and clinical findings. A case was encountered in this study in which UIP diagnosis with high confidence level was provided but SLB was also performed. The HRCT result of this case showed indeterminate for UIP pattern, and the patient was too young for IPF to be suspected [20, 21] and these inconsistencies resulted in additional SLB (Fig. 2).

Recently, Romagnoli et al. reported poor concordance of pathology diagnoses between TBLC and SLB [17]. Careful evaluation of the contents revealed there were nine UIP diagnosis with TBLC among which seven were also considered to have UIP pattern with SLB. Within the seven UIP pattern diagnoses with SLB, two were favored more as CHP over IPF by pathology, however, after multidisciplinary diagnosis, both cases were concluded to be IPF. Eventually, out of the nine cases of UIP with TBLC, six were diagnosed as IPF after multidisciplinary discussion. Their result is somewhat similar to our series in terms of having higher agreement in UIP judgement than other histologic patterns.

This study has some limitations. In particular, this was a single facility pathological study including only a small number of cases which were determined by MDD that TBLC was insufficient to make a definite diagnosis. Only seven cases in which patients underwent both TBLC and SLB were selected. Additional SLB is not necessary if the confidence level of TBLC diagnosis is high, so only those with low confidence TBLC diagnosis or whose TBLC diagnosis was inconsistent with clinical or imaging findings underwent SLB. Therefore, this may have some biases and many not truly reflect a true comparison of TBLC and SLB. A higher concordance is expected by comparing pathological findings and diagnosis in cases with highly confident TBLC diagnosis. Iftikhar IH et al. shows that pooled diagnostic yield, pooled sensitivity, and pooled specificity of TBLC were 83.7, 87, and 57%, respectively [22]. However, this report was a systematic review of past literature, and no document which examined TBLB and SLB in the same case was included. In terms of standardization of definitive diagnosis, it is important to collect cases from multiple facilities and study their concordance of histopathological assessments and diagnoses. Additionally, it is vital to collect evidence in terms of indication for TBLC and the need for additional SLB in order to develop diagnostic guidelines.

This study showed that TBLC is somewhat inferior in sensitivity yet relatively high in specificity for diagnosing UIP. It is suggested that if enough observational data is obtained from TBLC to determine UIP, definitive diagnosis could be possible without SLB. Meanwhile, additional SLB may be indicated for the diagnosis of other ILD such as NSIP and SR-ILD, since TBLC may be inferior in specificity in these cases.