Transbronchial Lung Cryobiopsy for Diagnosing ILD

A team of researchers recently examined the diagnostic accuracy of a novel, less-invasive approach to sampling lung tissue—transbronchial lung cryobiopsy (TBLC)—for diagnosing interstitial lung disease (ILD) compared with the accuracy of surgical lung biopsy (SLB).1

Pulmonology Consultant connected with lead researcher Lauren K. Troy, PhD, who is a respiratory specialist in the Department of Respiratory and Sleep Medicine at the Royal Prince Alfred Hospital in Australia, to learn more about her team’s study and its findings.

PULM CON: You and your colleagues conducted the COLDICE (Cryobiopsy versus Open Lung biopsy for the Diagnosis of Interstitial Lung Disease allianCE) study, which investigated the diagnostic agreement between TBLC and SLB. To start, can you give us an overview of TBLC and how it differs from SLB?

Lauren Troy: TBLC is a novel procedure for obtaining lung tissue. The technique is performed by passing a probe through the working channel of a bronchoscope out to the lung periphery. The probe tip is rapidly cooled through the Joule-Thomson effect, and cryoadhesion occurs between the tip and adjacent lung tissue. A tissue sample is extracted by a firm pull, then removal of the probe and bronchoscope from the airway.

Clearly, this approach differs from the surgical biopsy, which is performed via incisions through the chest wall. TBLC generates small tissue samples (usually 5-10 mm in longest axis), compared with much larger SLB specimens (about 10 cm in longest axis). Recently, TBLC has been used for obtaining histopathology for the diagnosis in ILD. The diagnostic yield for identifying a histopathologic pattern is reportedly about 80% for TBLC, compared with 95% for SLB. The risk of adverse outcomes is substantially lower with TBLC than with SLB.

PULM CON: Can you tell us more about the COLDICE study and how it came about?

LT: With increasing utilization of TBLC in clinical practice, particularly across Europe, we wondered how accurate the TBLC findings were. The SLB is well-established as the reference standard for ILD histopathology diagnosis, and indeed international ILD guidelines are derived from descriptions of SLB specimens. We recognized that diagnostic accuracy of TBLC could only be determined by analyzing paired TBLC and SLB from the same patients. The COLDICE study was established to answer the question of diagnostic accuracy of TBLC. We recruited patients who were deemed to require a biopsy to aid the diagnosis of their ILD. Subjects generally had mild but progressive disease; establishing a clear diagnosis would impact their treatment and outcomes.

All participants underwent TBLC and SLB, sequentially under the same anesthetic. Tissue samples were then independently reviewed by blinded pathologists and then submitted to a multidisciplinary discussion (MDD) for final clinical diagnosis. Our co-primary endpoints were agreement between the histopathological diagnoses made by the pathologists, and agreement for MDD diagnoses between TBLC and SLB. The MDD component was critical for the study given that ILD diagnosis relies on the integration of clinical, radiological, and (when available) histopathological data.

PULM CON: Let’s talk more about your findings now. What were the histopathologic and diagnostic agreements between TBLC and SLB?

LT: We recruited 65 subjects to the study, across 9 Australian centers. We found that the histopathologic agreement between paired TBLC and SLB was good, at 70.8% (weighted κ 0.70, 95% CI, 0.55–0.86). For MDD diagnoses, the diagnostic agreement was 76.9% (κ 0.62, 95% CI, 0.47–0.78). In particular, when diagnostic confidence at MDD was high for TBLC (as was the case in 39 of 65 participants, 60%), the majority (37 of 39 participants, 95%) were concordant with SLB diagnoses. In the 26 of 65 participants (40%) with low confidence or unclassifiable TBLC diagnoses, SLB reclassified 6 (23%) to alternative high-confidence or definite MDD diagnoses. In other words, the SLB provided additional useful information in only 1 out of 10 participants in the study.

From a safety perspective, mild to moderate airway bleeding occurred in 14 participants (22%) due to TBLC, similar to the reported rates. The 90-day mortality was 2% (1 of 65 participants), following acute exacerbation of idiopathic pulmonary fibrosis.

PULM CON: Taken together, what do these findings mean for clinical practice and how ILD might be diagnosed in the future?

LT: These findings reveal an acceptable diagnostic accuracy for TBLC and, importantly, clinical utility is demonstrated. With a more favorable safety profile, it is possible that TBLC will supplant SLB as the first-line biopsy procedure for ILD patients requiring lung tissue for diagnosis. In centers with training and experience in TBLC, it is likely that many patients will avoid SLB altogether. Additionally, small tissue samples may be able to yield other diagnostic data not yet in routine clinical use. For example, a molecular classifier that enables RNA sequencing of transbronchial lung biopsies was shown to predict certain ILD patterns with the incorporation of machine learning. With further development of such techniques, we may soon be using a combination of histopathology, immunohistochemistry, and molecular analysis for ILD diagnosis, much like the way lung cancer diagnostics have evolved.

PULM CON: What is the overall key take-home message from your study?

LT: The key take-home message from the COLDICE study is that cryobiopsy is a reasonable first-line alternative to SLB for ILD diagnosis among patients where histopathology is required. In the broader sense, our findings confirm the importance of the MDD, both with decision making around the need for biopsy and with the final clinical diagnosis.



  1. Troy LK, Grainge C, Corte TJ, et al; Cryobiopsy versus Open Lung biopsy in the Diagnosis of Interstitial lung disease alliance (COLDICE) Investigators. Diagnostic accuracy of transbronchial lung cryobiopsy for interstitial lung disease diagnosis (COLDICE): a prospective, comparative study. Lancet Respir Med. 2020;8(2):171-181.