br overexpression have varied in di erent publications
overexpression have varied in diﬀerent publications, and up to date no consensus exists [2,6–10]. The role of c-MET as a predictive biomarker is still unclear. A recent surgical cohort study in Asian NSCLC patients (n = 311, of whom 151 received adjuvant chemotherapy), showed that overall survival (OS) was significantly prolonged in the c-MET positive cases who received platinum-based adjuvant chemotherapy, but this finding needs to be verified in larger cohorts . However, in a study by Cappuzzo et al. in early stage NSCLC, MET gene amplification was a negative prognostic biomarker for OS . Another study done in 689 patients with early stage disease who were treated with surgery, failed to show any prognostic value of MET amplification or cMET over-expression. In this study the prevalence of MET overexpression and MET amplification was 17% and 2.4% respectively .
An Asian randomized phase 3 trial with erlotinib with or without the c-MET inhibitor tivantinib (ARQ197) in EGFR wild-type non-squa-mous NSLCL failed to show any survival benefit with the combination, and was terminated prematurely due to higher incidence of interstitial lung disease in the tivantinib arm . Another phase 3 study with a
Corresponding author at: Department of Oncology, Karolinska University Hospital/Karolinska Institute, 17176, Stockholm, Sweden. E-mail address: [email protected] (G. Tsakonas).
similar design in Caucasian population was terminated early at the interim analysis due to futility . Both of these studies were done in an unselected population regarding c-MET expression. Onartuzumab, a fully humanized recombinant monoclonal antibody binding to the ex-tracellular domain of c-MET, failed to show any survival benefit when combined with erlotinib compared to single erlotinib in a phase 3 randomized trial in previously treated stage IIIB or IV NSCLC de-termined to be c-MET positive (≥ 50% of tumor cells with IHC scores of 2+ [moderate] or 3+ [strong] levels of MET) . Furthermore, no eﬀect of onartuzumab/erlotinib combo was shown in exploratory analyses using MET FISH and gene expression. Although the above mentioned trials failed to show any benefit, several smaller trials and case series have shown eﬃcacy of c-MET tyrosine kinase inhibitors or monoclonal Ferrostatin1 targeting c-MET, mainly in MET exon 14 mu-tated NSCLC, but these findings need verification in larger randomized trials [9,15].
The aim of our study was to identify a clinically significant IHC cut-oﬀ for c-MET protein expression in patients with early stage NSCLC, and investigate its role as a potential prognostic or predictive (re-garding adjuvant chemotherapy) biomarker.
2.1. Patient population
We designed a retrospective cohort study, consisting of 725 patients with surgically removed NSCLC. Patients received surgical treatment from 1/1/1995 to 30/12/2010 (Table 1)
Demographic data were collected retrospectively. We collected in-formation about physician´s evaluation of performance status (PS) at the time of diagnosis, age at diagnosis, histology, received adjuvant treatment, surgical stage of disease, smoking status and gender. The TMA cohort was based on diagnostic tissue from NSCLC patients op-erated at the Uppsala University Hospital between 1995 and 2010 and histopathological data for parts of this cohort has been reported pre-viously .
Immunohistochemistry (IHC) was conducted in tissue microarrays (TMA) from lung tumors and healthy tissue adjacent to the tumor, using a specific antibody against human c-MET (MET PharmDx). The same surgical specimen was used for the preparation of healthy tissue and tumour TMA. Haematoxylin-eosin stained slides were reviewed by two pathologists (PM, JB) and tumor areas to be included in the TMA were marked. No major discrepancies regarding IHC scoring between the two pathologists were observed. The TMA was constructed using a manual tissue arrayer (MTA-1, Beecher Instruments, Sun Prairie, CA), essen-tially as previously described [17,18]. All tumours were included in duplicates (2 x 1 mm tissue cores). Four-micrometer sections were cut
All patients 725
3 Missing survival date
43 Missing cMET
2 Missing smoking status
2 Missing ALK mutation status
4 Missing histology and exclusion of
18 Exclusion of stage IV at diagnosis
72 Summary of all exclusions Final study cohort 653
cMET: cellular Mesenchymal Epithelial Transition factor, ALK: Anaplastic
from the TMA blocks, mounted on adhesive slides and baked in 60 °C for 45 min. IHC staining was quantified using H-scores (range 0–300), which incorporate staining intensity (range 0–3) and the percentage of positively-stained tumor cells (range 0–100%). IHC analysis of c-MET was performed in the whole study population, and the average H-score of 2 separate core biopsies for each tumour was used. 682 patients had an average H-score and were included in the analysis.