Feasibility study of cryobiopsy for practical pathological diagnosis of primary lung cancer including immunohistochemical assessment

Abstract

Background: Precision medicine in non-small cell lung cancer requires attainment of a sufficient amount of high-quality tumor tissue. Transbronchial cryobiopsy has emerged as a new diagnostic method for non-neoplastic lung disease with a better potential to assess morphology compared with conventional methods. However, the influence of cryobiopsy on specimen quality, particularly detection of protein expression, is unknown. We performed a comparative immunohistochemical study in specimens obtained by cryobiopsy versus conventional sampling to evaluate the feasibility of cryobiopsy for lung cancer diagnosis. Methods: Pairs of artificial biopsy specimens, collected using a cryoprobe or conventional scalpel, were obtained from 43 surgically resected primary lung tumors. Formalin-fixed, paraffin-embedded blocks were prepared in an ISO15189-certified laboratory. Immunohistochemical staining of thyroid transcription factor-1, p40, Ki67 and programmed death-ligand 1 (22C3) was performed. The H-scores for thyroid transcription factor-1 and p40, labeling index for Ki67 and tumor proportion score for programmed death-ligand 1 were assessed. Pearson's correlation coefficients between two sampling types were calculated. Results: The thyroid transcription factor-1 and p40 H-scores showed perfect correlations between the cryobiopsy and conventional scalpel-obtained specimens (R2 = 0.977 and 0.996, respectively). Ki67 labeling index and PD-L1 tumor proportion score also showed strong correlations between the two sample types (R2 = 0.896 and 0.851, respectively). Five cases (11.6%) exhibited differences in tumor proportion score category between sample types, potentially because of intratumoral heterogeneity. Conclusions: Immunohistochemical expression of certain tumor markers showed a high concordance between cryobiopsy and conventional scalpel sampling. Cryobiopsy is feasible for pathological diagnostics including PD-L1 evaluation.