Correlation of PD-L1 Expression on Tumour Cells Between Diagnostic Biopsies and Surgical Specimens of Lung Cancer in Real Life With Respect to Biopsy Techniques and Neoadjuvant Treatment

Abstract

Purposes: Programmed death-ligand 1 (PD-L1) testing is performed mainly on biopsy specimens in patients with advanced lung cancer. It is questionable whether the small amount of tissue analysed in biopsies may represent the true PD-L1 expression of a tumour.Methods: In this retrospective study, PD-L1 expression on tumour cells derived from bronchoscopy brush cytology, endobronchial-ultrasound guided transbronchial needle aspiration (EBUS-TBNA), endobronchial biopsy, transbronchial biopsy (TBB) and computed tomography (CT)-guided transthoracic biopsy was compared to the PD-L1 expression of the corresponding surgical resection in lung cancer patients with regard to neoadjuvant treatment in-between.Results: A quantitative comparison between the diagnostic biopsy of the primary tumour with corresponding resected surgical specimens in a total of 113 lung cancer patients (60% male, mean age 65 ± 9 years) revealed a statistically significant correlation of PD-L1 expression on tumour cells (r=0.58, p<0.001), for patients without neoadjuvant treatment in-between and for patients who underwent neoadjuvant treatment (both p<0.001). Using a cut-off value of ≥50% PD-L1 TPS for comparing the biopsy samples and resected specimens, the concordance rate was 78% with a Cohen´s Kappa of 0.45.Conclusion: A statistically significant concordance for PD-L1 expression on tumour cells between biopsies from primary lung tumour and resected specimen was found, but of uncertain clinical accuracy. The use of a cut-off value of ≥50% PD-L1 TPS resulted only in a moderate agreement. Therefore, the interpretation of the PD-L1 determined form biopsy specimens status should only be considered with caution for treatment decisions.

PD‐L1 immunohistochemical testing: A review with reference to cytology specimens

“…These studies have been associated with kappas ranging from 0.31, 34 0.38, 35 and 0.39 36 consistent with fair agreement to 0.68, 37 0.91, 38 0.82, 39 and 0.73 40 associated with substantial agreement. Many studies have not published formal kappa statistics for agreement but have published percentage agreements varying from 74.1% to 97.3% 13,18–20,24,41–56 . The concordance rates appear to be impacted by precise specimen type and antibody clone used.…”

“…Overall, cytologic specimens appear to be acceptable specimens for PD‐L1 analysis especially when formalin fixed paraffin embedded cell block specimens are evaluated using the TPS for evaluation 9–13,20,22,24,34–70 …”

“…A large number of studies have investigated PD‐L1 expression in various cytologic preparations. Comparisons have been made among cytology specimens, small biopsy cores and surgical resection specimens 10,12,13,17–52 . Most studies have shown good correlation of PD‐L1 assays performed on cell block specimens with those performed on small biopsies and resection specimens.…”

“…Many studies have not published formal kappa statistics for agreement but have published percentage agreements varying from 74.1% to 97.3%. 13,[18][19][20]24,[41][42][43][44][45][46][47][48][49][50][51][52][53][54][55][56] The concordance rates appear to be impacted by precise specimen type and antibody clone used. The majority of these studies have shown moderate to substantial agreement between PD-L1 assessment of histologic and cytologic specimens as defined by Cohen.…”