1. A Novel Approach to Detect Programed Death Ligand 1 (PD-L1) Status and Multiple Tumor Mutations Using a Single Non–Small-Cell Lung Cancer (NSCLC) Bronchoscopy Specimen 
Amanda Vannitamby, Shona Hendry, Tanvi Makadia, Janine Danks, John Slavin, Louis Irving, Daniel Steinfort, Steven Bozinovski 
The Journal of Molecular Diagnostics 
Volume 21, Issue 2, Pages (March 2019)
DOI: /j.jmoldx
Copyright © 2019 American Society for Investigative Pathology and the Association for Molecular Pathology Terms and Conditions

2. Figure 1
MMP-9:TIMP3 ratio differentiated control (Con) and malignant endobronchial ultrasound (EBUS) brushings. After rapid on-site examination confirmation, MMP9 and TIMP3 transcript levels were determined by quantitative real-time RT-PCR in benign and malignant EBUS bronchoscopy brush specimens. A and B: MMP9 levels were significantly increased in non–small-cell lung cancer (NSCLC) specimens (A; 23-fold increase), whereas TIMP3 levels were significantly reduced (B; threefold reduction; P = 0.0004). C: The ratio of MMP-9:TIMP3 expression was significantly increased in the NSCLC specimens relative to the controls (137-fold increase). D: Receiver operating characteristic curve analyses of the MMP-9:TIMP3 ratio demonstrated its accuracy to differentiate benign and NSCLC specimens by generating an area under the curve value of 0.98 (95% CI, 0.9–1.0; P < 0.0001). n = 13 benign EBUS bronchoscopy brush specimens (A) n = 15 malignant EBUS bronchoscopy brush specimens (A). ∗∗∗P < 0.001, ∗∗∗∗P <  (t-test).
The Journal of Molecular Diagnostics  , DOI: ( /j.jmoldx )
Copyright © 2019 American Society for Investigative Pathology and the Association for Molecular Pathology Terms and Conditions

3. Figure 2
Assessment of oncogenic driver gene expression in endobronchial ultrasound (EBUS) specimens by quantitative real-time RT-PCR (RT-qPCR). Further analysis of the same EBUS specimens by TaqMan RT-qPCR was performed to evaluate gene expression changes of programed death ligand 1 (PD-L1), MET, and PTEN in non–small-cell lung cancer (NSCLC). A: PD-L1 transcript levels were significantly higher in the malignant EBUS specimens. B: MET gene expression was not significantly altered in the NSCLC biopsies; however, elevated transcript levels that are indicative of MET gene amplification were detected in two NSCLC specimens (>eightfold increase). C: Similarly, PTEN transcript levels were not significantly different in the NSCLC specimens; however, 4 of 15 were reduced by at least twofold relative to the median control (Con) levels. ∗P < 0.05 (t-test).
The Journal of Molecular Diagnostics  , DOI: ( /j.jmoldx )
Copyright © 2019 American Society for Investigative Pathology and the Association for Molecular Pathology Terms and Conditions

4. Figure 3
MMP9:TIMP3 transcript ratio was also elevated in the biopsy cohort. Snap-frozen lung biopsies were used to assess MMP9 transcript levels by quantitative real-time RT-PCR (RT-qPCR) in non–small-cell lung cancer (NSCLC) specimens, nonmalignant controls (Con), and adjacent tumor-free controls matched to adenocarcinoma (Adeno) patients. A and B: MMP9 transcript levels were significantly increased in NSCLC specimens relative to unmatched (A; 7.9-fold increase) and matched (B; 24.9-fold increase) controls. C: MMP9 levels were significantly elevated in both adenocarcinomas and squamous cell carcinoma (SCC), and there was no difference in expression between the tumor histologic types. D and E: Snap-frozen lung biopsies were also used to assess TIMP3 transcript levels, which were elevated in the NSCLC specimens relative to unmatched (D, fivefold reduction) and matched (E) controls. F: Comparison of TIMP3 expression across the histologic tumor types revealed significant reduction in both Adeno and SCC biopsies, with a further significant reduction detected in SCC. n = 90 NSCLC specimens; n = 10 nonmalignant control specimens; n = 10 adjacent tumor-free controls matched to Adeno patients. ∗∗P < 0.01, ∗∗∗P < 0.001, and ∗∗∗∗P <  (t-test); ††P < 0.01 (Wilcoxon's paired t-test); ‡‡P < 0.01, ‡‡‡P < 0.001, and ‡‡‡‡P <  (Kruskal Wallis test).
The Journal of Molecular Diagnostics  , DOI: ( /j.jmoldx )
Copyright © 2019 American Society for Investigative Pathology and the Association for Molecular Pathology Terms and Conditions

5. Figure 4
Receiver operating characteristic (ROC) curve analyses of matrix metalloproteinase (MMP)-9 and tissue inhibitor of metalloproteinase03 (TIMP3) in the biopsy cohort. A: The MMP9:TIMP3 ratio was significantly elevated in the non–small-cell lung cancer (NSCLC) specimens relative to the control (Con) levels (38-fold increase versus control). B: Similarly, this ratio was elevated in the malignant specimens matched to adjacent control tissue collected from the same individual. To evaluate the diagnostic potential of MMP9 and TIMP3 transcript levels in the NSCLC specimens relative to control specimens, a ROC curve analysis was performed. C and D: MMP-9 alone generated an area under the curve (AUC) value of (C; 95% CI, 0.76–0.95; P < 0.0001) and TIMP3 generated an AUC of (D; 95% CI, 0.79–0.95; P < 0.0001). E: The highest AUC of (95% CI, 0.86–1.0; P < 0.0001) was generated by using the MMP9:TIMP3 ratio. ∗∗∗∗P <  (t-test); ††P < 0.01 (Wilcoxon's paired test).
The Journal of Molecular Diagnostics  , DOI: ( /j.jmoldx )
Copyright © 2019 American Society for Investigative Pathology and the Association for Molecular Pathology Terms and Conditions

6. Figure 5
Programed death ligand 1 (PD-L1) transcript levels correlate with PD-L1 staining. A: PD-L1 transcript levels in the biopsy archival specimens were assessed by quantitative real-time RT-PCR (RT-qPCR). Levels were not significantly different relative to control (Con) biopsies. B: With the use of a twofold increase cutoff value, 2 of 50 adenocarcinoma (Adeno) biopsies (4%) and 840 squamous cell carcinoma (SCC) biopsies (20%) displayed elevated PD-L1 transcript levels. MET transcript levels were not significantly different across the groups; however, when using a twofold increase cutoff value, 12 of 50 Adeno biopsies (24%) and 2 of 40 SCC biopsies (5%) expressed higher levels of MET transcript. C: PTEN transcript was significantly reduced in Adeno and SCC biopsies, and, when using a twofold decrease cutoff value, 15 of 50 Adeno biopsies (30%) and 18 of 40 SCC biopsies (45%) expressed reduced PTEN transcript levels. D: A representative image of matching formalin-fixed, paraffin-embedded tumor slides obtained from biopsies derived from the SCC patients who were subjected to PD-L1 immunohistochemistry (IHC) staining. E: PD-L1 staining across the entire section was compared with PD-L1 transcript levels in matching SCC specimens, whereby a strong positive association was observed (Spearman r = 0.803, P < 0.0001). F: PD-L1 transcript levels were significantly higher in tumors with >1% tumor cell membrane staining. G: Receiver operating characteristic (ROC) curve analysis generated an area under the curve (AUC) of 0.857, and with the use of a cutoff of >0.423-fold increase, PD-L1 mRNA expression was 69% sensitive and 92% specific in identifying tumors with >1% PD-L1 tumor staining. H: With the use of a higher threshold of <50% versus >50% PD-L1–positive tumor cell membrane staining, PD-L1 transcript levels were significantly higher in tumors with >50% staining. I: ROC curve analysis generated an AUC of 0.948, and with the use of a cutoff of >0.878-fold increase, PD-L1 mRNA expression was 100% sensitive and 91% specific in identifying tumors with >50% PD-L1 tumor staining. n = 40 SCC patients (D). ∗P < 0.05 and ∗∗∗∗P <  (Kruskal Wallis test); ††††P <  (U-test). Original magnification, ×20 (D).
The Journal of Molecular Diagnostics  , DOI: ( /j.jmoldx )
Copyright © 2019 American Society for Investigative Pathology and the Association for Molecular Pathology Terms and Conditions

7. Figure 6
A flow chart depicting the streamlined approach for comprehensive molecular profiling of bronchial specimens. A new approach to enable large-scale molecular and mutational profiling of a single endobronchial ultrasound (EBUS) bronchial specimen is shown. A single EBUS specimen (brush or biopsy) was collected in stabilization buffer after rapid onsite evaluation of the malignant site. Nucleic acid integrity was preserved for adequate isolation and subsequent assessment of MMP9, TIMP3, and programed death ligand 1 (PD-L1) transcript levels through multipanel quantitative real-time RT-PCR (RT-qPCR), which can be achieved in a 24-hour time frame. After confirmation of the malignant cell content by using the MMP9:TIMP3 ratio, genomic DNA from the same specimen was suitable for multipanel–targeted next-generation sequencing to potentially assess the mutational burden and therefore guide the clinical use of targeted immunotherapies.
The Journal of Molecular Diagnostics  , DOI: ( /j.jmoldx )
Copyright © 2019 American Society for Investigative Pathology and the Association for Molecular Pathology Terms and Conditions