Abstract
BACKGROUND:
Programmed cell death ligand 1 (PD-L1) expression by the 22C3 pharmDx companion assay has been validated in surgical specimens to support pembrolizumab treatment decisions for patients with non–small cell lung carcinoma (NSCLC). The aims of this study were 1) to assess the adequacy of cytologic specimens for PD-L1 evaluation and 2) to explore correlations of PD-L1 expression with clinicopathologic and molecular features.
METHODS:
The study cohort included 100 cytology specimens (fluid [n = 28] and fine-needle aspiration [n = 72]) and 165 surgical specimens (biopsy [n = 138] and resection [n = 27]). The PD-L1 immunohistochemistry 22C3 assay and staining assessment were performed according to the manufacturer’s instructions. PD-L1 expression was correlated with patients’ demographics, pathologic characteristics, and molecular alterations.
RESULTS:
One hundred forty-two specimens (53.6%) were positive for PD-L1 expression (≥1%). No statistically significant difference in PD-L1 expression was identified between cytologic (56.0%) and surgical specimens (52.1%). Seventy-four of 190 tested cases (38.9%) had genetic alterations. PD-L1 positivity was significantly more prevalent in cases with genetic alterations than in cases without genetic alterations. Furthermore, both PD-L1 positivity and high PD-L1 expression (≥50%) had statistically significant associations with Kirsten rat sarcoma viral oncogene homolog (KRAS) mutations. PD-L1 expression had no significant association with histologic phenotypes or other clinicopathologic features.
CONCLUSIONS:
The data indicate that cytologic specimens are comparable to surgical specimens for PD-L1 evaluation. The association of PD-L1 expression with KRAS mutations may have clinical relevance in selecting patients with NSCLC for immunotherapy.
Keywords: anaplastic lymphoma kinase (ALK), cytology, epidermal growth factor receptor (EGFR), Kirsten rat sarcoma viral oncogene homolog (KRAS), non-small cell lung carcinoma (NSCLC), programmed cell death ligand 1 (PD-L1)
INTRODUCTION
Most patients with lung cancer are diagnosed at an advanced stage, and this makes them not amenable to surgery. Traditional chemotherapy has been the first-line treatment for patients with advanced non–small cell lung carcinoma (NSCLC) without targetable molecular alterations.1 Recently, cancer immunotherapy has dramatically shifted the paradigm of lung cancer treatment, and immune checkpoint inhibitors targeting programmed cell death ligand 1 (PD-L1)/programmed cell death 1 (PD-1) have significantly improved the survival of patients with NSCLC.2–6 Furthermore, immunotherapy is currently the only personalized therapy available to patients with lung squamous cell carcinoma because molecular targeted therapy for squamous cell carcinoma has not yet been developed. PD-L1 expression levels have been correlated with PD-1/PD-L1 blockade immunotherapy response rates and survival outcomes in patients with NSCLC.2,3,5,7,8 A wide range of frequencies of PD-L1 expression (7%−82%) have been reported for NSCLCs of specific histologic types.9–16 One study reported that PD-L1 was expressed differently between adenocarcinoma (46%) and squamous cell carcinoma (61%),17 but such a difference was not found in other studies.13,18 Some studies have observed that high PD-L1 expression is associated with smoking, an advanced stage, squamous morphology, and molecular alterations, including epidermal growth factor receptor (EGFR), Kirsten rat sarcoma viral oncogene homolog (KRAS), and anaplastic lymphoma kinase (ALK) oncogenic mutations, whereas other studies have shown conflicting results.5,11,13,15,19–25 These conflicting data may be caused by differences in population cohorts, PD-L1 antibodies, PD-L1 expression cutoff values, specimen types (whole sections or tissue microarrays), and so forth.
Most studies have evaluated PD-L1 expression in surgical specimens, and studies of cytology specimens are limited.26–30 However, cytologic specimens are frequently the only specimens available, especially for patients at an advanced stage. Therefore, it is necessary to validate PD-L1 testing in cytologic specimens and to confirm the consistency between cytologic and surgical specimens.
In the current study, we investigated tumoral PD-L1 expression with a US Food and Drug Administration–approved PD-L1 immunohistochemistry (IHC) assay (Dako PD-L1 IHC 22C3 pharmDx) in both cytologic and surgical specimens from patients with NSCLC and its association with clinicopathologic features and molecular alterations.
MATERIALS AND METHODS
Patients and Specimens
After approval was obtained from the institutional review board at Ohio State University, a computer-based search was performed with our CoPath laboratory information system to retrieve consecutive patients with NSCLC and PD-L1 testing between January 2016 and June 2018. Clinical and pathologic characteristics were collected from the electronic medical record and pathology archive system. Clinical characteristics (eg, smoking history and clinical stages) were not fully available for some patients who had only biopsy procedures but were not treated at our institution. Diagnoses were rendered by subspecialized thoracic pathologists for surgical specimens and by cytopathologists for cytologic specimens. Special IHCs such as thyroid transcription factor 1, p40/p63, napsin A, and cytokeratin were used when it was necessary to confirm a diagnosis. All slides were reviewed to confirm the diagnosis and adequacy for PD-L1 testing (≥100 tumor cells on a slide). The study cohort included 100 cytology specimens (72 fine-needle aspiration [FNA] specimens and 28 fluid specimens) and 165 surgical specimens (138 biopsies and 27 resections). Each patient had only 1 PD-L1 test on either a cytologic specimen or a histologic specimen from either a primary lesion or a meta-static/recurrent lesion. The specimen with the most tumor content was chosen for PD-L1 testing. If there were multiple types of specimens with enough tumor content, the sequence of specimens for ordering PD-L1 testing was as follows: histologic specimens, FNA cytology specimens, and then fluid specimens.
Sample Processing
Fluids were processed as cytospin slides stained with Diff-Quik and/or Papanicolaou stain together with cell block sections stained with hematoxylin-eosin (H & E) stain. FNAs were processed as direct smears stained with Diff-Quik and/or Papanicolaou stain together with cell block sections stained with H & E stain. Cell blocks were prepared from pellets of a centrifuged cell suspension by the addition of plasma and thrombin to enmesh cellular materials fixed in 10% neutral buffered formalin. Surgical biopsy or resection specimens were embedded in paraffin after fixation in 10% neutral buffered formalin, and the slides were stained with H & E stain.
Molecular Assays
Molecular assays were amplicon-based next-generation sequencing assays using either a lung carcinoma panel (22 genes) or a solid tumor mutation panel (48 genes). Genomic DNA isolated from tissue was profiled with a polymerase chain reaction–based AmpliSeq library on the Ion Chef and S5 sequencing platform (Thermo Fisher Scientific, Waltham, Massachusetts). An analysis of the neoplasm-associated variants included the use of hg19, Torrent Suite Software, and the GenomOncology platform. The lung carcinoma panel (22 genes) detected mutations in the following genes: AKT1, ALK, BRAF, CTNNB1, DDR2, EGFR, ERBB2, ERBB4, FBXW7, FGFR1, FGFR2, FGFR3, KRAS, MAP2K1, MET (excluding exon-skipping mutations), NOTCH1, NRAS, PIK3CA, PTEN, SMAD4, STK11, and TP53. The solid tumor mutation panel (48 genes) detected cancer-associated mutations in ABL1, AKT1, ALK, APC, ATM, BRAF, CDH1, CDKN2A, CSF1R, CTNNB1, EGFR, ERBB2, ERBB4, EZH2, FBXW7, FGFR1, FGFR2, FGFR3, GNA11, GNAS, GNAQ, HNF1A, HRAS, IDH1, IDH2, JAK2, JAK3, KDR, KIT, KRAS, MET, MLH1, NOTCH1, NPM1, NRAS, PDGFRA, PIK3CA, PTPN11, PTEN, RB1, RET, SMAD4, SMARCB1, SMO, SRC, STK11, TP53, and VHL.
Fluorescence in situ hybridization studies for ALK, ROS1, and RET translocations and MET amplification were performed on 4-μm-thick paraffin sections from tissue or cell blocks. ALK gene rearrangement was detected with the dual-color Vysis LSI ALK Break Apart Rearrangement Probe Kit (Abbott Molecular, Des Plaines, Illinois). ROS1 gene rearrangement was detected with the LSI ROS1 (Cen) Spectrum Green Probe (Abbott Molecular). RET gene rearrangement was detected with the Vysis LSI RET (Cen) Spectrum Green Probe (Abbott Molecular). MET amplification was detected with a dual-color probe at 7q31 along with a CEP probe for chromosome 7 centromere (ZytoLight SPEC MET/CEN 7 Dual Color Probe; ZytoVision GmbH, Bremerhaven, Germany). All probes were used according to the manufacturer’s instructions.
In general, molecular testing was ordered for patients with advanced NSCLC at our institution. Occasionally, molecular testing was also requested directly from clinicians for patients with nonadvanced disease.
PD-L1 IHC and Scoring
PD-L1 IHC was performed with Dako PD-L1 IHC 22C3 pharmDx (clone 22C3; Dako/Agilent Technologies, Carpinteria, California) on the Dako Autostainer Link 48 autostainer (Dako/Agilent Technologies) according to the manufacturer’s protocol. PD-L1 controls were run concurrently with test samples, including positive and negative cell line controls provided in the assay kit, an in-house tonsil control as a positive tissue control, and a negative control without a secondary antibody. Complete or incomplete membranous staining in tumor cells was considered specific, regardless of intensity. PD-L1 IHC was scored as the percentage of tumor cells with PD-L1 staining to total tumor cells from 0% to 100%. PD-L1 positivity was defined as at least 1% of tumor cells with staining, and high PD-L1 expression was defined as at least 50% of tumor cells with staining. All pathologists who scored PD-L1 IHCs had completed an online training module (the Dako PD-L1 IHC 22C3 pharmDx Interpretation Training Program, which is available at https://pdl122c3-learning.dako.com/). PD-L1 results were obtained from pathology reports. Representative PD-L1 IHCs with different staining scores are shown in Figures 1 (cytology specimens) and 2 (surgical specimens). PD-L1 testing was usually ordered for patients with advanced NSCLC at our institution. However, PD-L1 testing was also occasionally requested by clinicians for patients with nonadvanced disease.
Figure 1.
PD-L1 expression in representative cytology cases (fine-needle aspirations): (A,B) 1 case with negative PD-L1 expression, (C,D) 1 case with positive PD-L1 expression (≥1% but <50%), and (E,F) 1 case with high PD-L1 expression (≥50%). In panels A, C, and E, cell block sections with H & E staining are shown (×200); in panels B, D, and F, cell block sections with PD-L1 immunostaining with the 22C3 antibody are shown (×200). PD-L1 indicates programmed cell death ligand 1.
Figure 2.
PD-L1 expression in representative surgical specimens (computed tomography–guided core-needle biopsies): (A,B) 1 case with negative PD-L1 expression, (C,D) 1 case with positive PD-L1 expression (≥1% but <50%), and (E,F) 1 case with high PD-L1 expression (≥50%). In panels A, C, and E, H & E staining is shown (×200); in panels B, D, and F, PD-L1 immunostaining with the 22C3 antibody is shown (×200). PD-L1 indicates programmed cell death ligand 1.
Statistical Analysis
Categorical data were summarized as frequencies and percentages, and continuous variables were summarized as medians and ranges. To study the associations with PD-L1 (negative vs positive), Fisher’s exact test or the chi-square test was used for categorical variables. The Wilcoxon rank-sum test was used to compare continuous variables (ie, age). The PD-L1 expression level was tested for normality with the Kolmogorov-Smirnov test. To study the associations with the PD-L1 expression level, the Wilcoxon rank-sum test or the Kruskal-Wallis test (if there were more than 2 groups) was used because PD-L1 expression was not normally distributed. A Spearman correlation was used to study the association between PD-L1 expression and age. P values were not adjusted for multiplicity (ie, multiple molecular alterations). Variables with P < .05 were considered statistically significant.
RESULTS
Patient Clinical and Pathologic Characteristics
The study cohort consisted of 265 patients with lung carcinomas, including 148 adenocarcinomas, 83 squamous cell carcinomas, 18 NSCLCs, 6 adenosquamous cell carcinomas, and 10 other types (4 large cell neuroendocrine carcinomas, 4 combined carcinomas with NSCLC and a small cell component, 1 adenoid cystic carcinoma, and 1 neuroendocrine carcinoma). The median age was 66 years (range, 37–92 years). One hundred fifty-three (57.7%) were male, and 112 (42.3%) were female. Two hundred twenty-three patients (84.2%) were current/past smokers, and almost half of the patients had a clinical stage of IV. One hundred ninety cases underwent molecular analysis, and 74 showed mutations/rearrangements, including 6 in EGFR exon 19, 1 in EGFR exon 21, 56 in KRAS, 4 in MET, 6 in ALK, and 1 in RET (Table 1). Most clinical characteristics were similar between the cytologic and surgical specimen groups. The cytologic group showed more cases with advanced stages, but the histologic group showed more squamous cell carcinoma cases and more cases with KRAS mutations (Table 1).
TABLE 1.
Clinicopathologic Characteristics, PD-L1 Expression, and Molecular Alterations in 265 Lung Carcinomas
| Variable | Level | Total (n = 265) | Cytology (n = 100) | Surgical (n = 165) | Pa |
|---|---|---|---|---|---|
| Age, y | Median (range) | 66 (37–92) | 65 (37–92) | 66 (43–87) | NS |
| Sex, No. (%) | Male | 153 (57.7) | 58 (58.0) | 95 (57.6) | NS |
| Female | 112 (42.3) | 42 (42.0) | 70 (42.4) | ||
| Specimen type, No. (%) | Fluid | 28 (10.6) | 28 (28.0) | N/A | |
| FNA | 72 (22.3) | 72 (72.0) | |||
| Surgical biopsy | 138 (52.1) | 138 (83.6) | |||
| Surgical resection | 27 (10.2) | 27 (16.4) | |||
| Smoking, No. (%) | Yes | 223 (84.2) | 87 (87.0) | 136 (82.4) | NS |
| No | 24 (9.1) | 10 (10.0) | 14 (8.5) | ||
| N/A | 18 (6.8) | 3 (3.0) | 15 (9.1) | ||
| Stage, No. (%) | I | 28 (10.6) | 2 (2.0) | 26 (15.8) | <.001 |
| II | 31 (11.7) | 7 (7.0) | 24 (14.5) | ||
| III | 50 (18.9) | 21 (21.0) | 29 (17.6) | ||
| IV | 130 (49.1) | 62 (62.0) | 68 (41.2) | ||
| N/A | 26 (9.8) | 8 (8.0) | 18 (10.9) | ||
| Histologic diagnosis, No. (%) | ADC | 148 (55.8) | 74 (74.0) | 74 (44.8) | <.001 |
| SqCC | 83 (31.3) | 13 (13.0) | 70 (42.4) | ||
| NSCLC | 18 (6.8) | 9 (9.0) | 9 (5.5) | ||
| Adenosquamous | 6 (2.3) | 0 (0.0) | 6 (3.6) | ||
| Other | 10 (3.9) | 4 (4.0) | 6 (3.6) | ||
| PD-L1, No. (%) | Insufficient | 4b (1.5) | 4 (4.0) | 0 (0.0) | NS |
| Negative (<1%) | 119 (44.9) | 40 (40.0) | 79 (47.9) | ||
| Positive (≥1%) | 142 (53.6) | 56 (56.0) | 86 (52.1) | ||
| Highly positive (≥50%) | 78 (29.4) | 31 (31.0) | 47 (28.5) | ||
| Molecular alteration tested, | EGFR exon 19 | 6 (3.2) | 3 (3.3) | 3 (3.0) | NS |
| No. (%) | EGFR exon 21 | 1 (0.5) | 0 (0) | 1 (1.0) | NS |
| KRAS | 56 (29.5) | 20 (22.0) | 36 (36.4) | .031 | |
| MET | 4 (2.1) | 2 (2.2) | 2 (2.0) | NS | |
| ALK | 6 (3.2) | 4 (4.4) | 2 (2.0) | NS | |
| RET | 1 (0.5) | 0 (0) | 1 (1.0) | NS | |
| Negative | 116 (61.1) | 62 (68.1) | 54 (54.6) | NS | |
| Total with molecular test | 190 (100) | 91 (100) | 99 (100) | N/A |
Abbreviations: ADC, adenocarcinoma; ALK, anaplastic lymphoma kinase; EGFR, epidermal growth factor receptor; FNA, fine-needle aspiration; KRAS, Kirsten rat sarcoma viral oncogene homolog; N/A, not available; NS, not statistically significant; NSCLC, non–small cell lung carcinoma; PD-L1, programmed cell death ligand 1; SqCC, squamous cell carcinoma.
P values indicate the difference between cytology and surgical cases.
Three ADCs and one NSCLC were insufficient for PD-L1 immunohistochemistry.
IHC Quantification of PD-L1 Expression
Four of 100 cytology specimens were not adequate for a PD-L1 evaluation by IHC because of low cellularity (<100 tumor cells), and all 165 surgical specimens were adequate. One hundred forty-two cases (53.6%) were positive for PD-L1 expression (≥1%), and 78 cases (29.4%) showed high PD-L1 expression (≥50%). Cytology specimens and surgical specimens showed similar PD-L1–positive rates (56.0% vs 52.1%) and high PD-L1 expression rates (31.0% vs 28.5%; Table 1).
Association of PD-L1 Expression With Clinicopathologic Features and Molecular Alterations
First, we examined the association of PD-L1 positivity (≥1%) with clinicopathologic features and molecular alterations. As shown in Table 2, PD-L1 positivity was significantly associated with any molecular alteration and KRAS mutations but was not associated with other specific molecular mutations (eg, EGFR and ALK), histologic types, smoking, stage, age, or sex.
TABLE 2.
Association of PD-L1 Positivity (Cutoff = 1%) With Clinicopathologic and Molecular Features (n = 261)
| Variable | Level | PD-L1-Negative: <1% (n = 119) | PD-L1-Positive: ≥1% (n = 142) | P |
|---|---|---|---|---|
| Sex, No. (%) | Female | 48 (40) | 64 (45) | .442 |
| Male | 71 (60) | 78 (55) | ||
| Age, y | Median (range) | 66 (44–88) | 65.5 (37–92) | .829 |
| Histologic type, No. (%) | ADC | 66 (55) | 79 (56) | .232a |
| SqCC | 37 (31) | 46 (32) | ||
| NSCLC | 8 (7) | 10 (7) | ||
| Adenosquamous | 1 (1) | 5 (4) | ||
| Other | 7 (6) | 2 (1) | ||
| Smoking, No. (%) | N/A | 8 (7) | 9 (6) | .640 |
| No | 12 (10) | 12 (8) | ||
| Yes | 99 (83) | 121 (85) | ||
| Stage, No. (%) | N/A | 12 (10) | 13 (9) | .327 |
| 16 (13) | 12 (8) | |||
| II | 15 (13) | 16 (11) | ||
| III | 18 (15) | 32 (23) | ||
| IV | 58 (49) | 69 (49) | ||
| Molecular alterations tested, No. (%) | N/A | 45 (38) | 45 (32) | |
| Negative | 51 (43) | 47 (33) | .007 | |
| Positive | 23 (19) | 50 (35) | ||
| EGFR exon 19, No. (%) | N/A | 43 (36) | 41 (29) | |
| Negative | 73 (61) | 98 (69) | 1a | |
| Positive | 3 (3) | 3 (2) | ||
| EGFR exon 21, No. (%) | N/A | 43 (36) | 41 (29) | 1a |
| Negative | 76 (64) | 100 (70) | ||
| Positive | 0 (0) | 1 (1) | ||
| KRAS, No. (%) | N/A | 42 (35) | 46 (32) | |
| Negative | 59 (50) | 58 (41) | .024 | |
| Positive | 18 (15) | 38 (27) | ||
| MET, No. (%) | N/A | 43 (36) | 46 (32) | |
| Negative | 75 (63) | 93 (65) | .631a | |
| Positive | 1 (1) | 3 (2) | ||
| ALK, No. (%) | N/A | 43 (36) | 41 (29) | |
| Negative | 75 (63) | 96 (68) | .239a | |
| Positive | 1 (1) | 5 (4) | ||
| RET, No. (%) | N/A | 45 (38) | 46 (32) | 1 |
| Negative | 74 (62) | 95 (67) | ||
| Positive | 0 (0) | 1 (1) |
Abbreviations: ADC, adenocarcinoma; ALK, anaplastic lymphoma kinase; EGFR, epidermal growth factor receptor; KRAS, Kirsten rat sarcoma viral oncogene homolog; N/A, not available; NSCLC, non–small cell lung carcinoma; PD-L1, programmed cell death ligand 1; SqCC, squamous cell carcinoma.
Fisher’s exact test.
Next, we examined the association of high PD-L1 expression (≥50%) with clinicopathologic features and molecular alterations. As shown in Table 3 (PD-L1 with ≥50% expression vs PD-L1 with 1%−49% expression) and Table 4 (PD-L1 with ≥50% expression vs PD-L1 with <50% expression), high PD-L1 expression was significantly associated with KRAS mutations but was not associated with other molecular mutations, histologic types, smoking, stage, age, or sex.
TABLE 3.
Association of High PD-L1 Expression (Cutoff = 50%) With Clinicopathologic and Molecular Variables (n = 142)
| Variable | Level | PD-L1: 1%−49% (n = 64) | PD-L1: ≥50% (n = 78) | P |
|---|---|---|---|---|
| Sex, No. (%) | Female | 28 (44) | 36 (46) | .866 |
| Male | 36 (56) | 42 (54) | ||
| Age, y | Median [IQR] (range) | 67.5 [60–72.3] (38.8–92.2) | 64.6 [59–71.6] (37.1–86.7) | .898 |
| Histologic type, No. (%) | ADC | 34 (53) | 45 (58) | .931a |
| SqCC | 22 (34) | 24 (31) | ||
| NSCLC | 4 (6) | 6 (8) | ||
| Adenosquamous | 3 (5) | 2 (3) | ||
| Other | 1 (2) | 1 (1) | ||
| Smoking, No. (%) | N/A | 3 (5) | 6 (8) | .364 |
| No | 7 (11) | 5 (6) | ||
| Yes | 54 (84) | 67 (86) | ||
| Stage, No. (%) | N/A | 6 (9) | 7 (9) | .284 |
| I | 8 (13) | 4 (5) | ||
| II | 7 (11) | 9 (12) | ||
| III | 11 (17) | 21 (27) | ||
| IV | 32 (50) | 37 (47) | ||
| Molecular alterations tested, No. (%) | N/A | 23 (36) | 22 (28) | |
| Negative | 23 (36) | 24 (31) | .197 | |
| Positive | 18 (28) | 32 (41) | ||
| EGFR exon 19, No. (%) | N/A | 19 (30) | 22 (28) | |
| Negative | 44 (69) | 54 (69) | 1a | |
| Positive | 1 (2) | 2 (3) | ||
| EGFR exon 21, No. (%) | N/A | 19 (30) | 22 (28) | |
| Negative | 44 (69) | 56 (72) | .446a | |
| Positive | 1 (2) | 0 (0) | ||
| KRAS, No. (%) | N/A | 23 (36) | 23 (29) | |
| Negative | 30 (47) | 28 (36) | .027 | |
| Positive | 11 (17) | 27 (35) | ||
| MET, No. (%) | N/A | 22 (34) | 24 (31) | |
| Negative | 40 (63) | 53 (68) | .579a | |
| Positive | 2 (3) | 1 (1) | ||
| ALK, No. (%) | N/A | 19 (30) | 22 (28) | |
| Negative | 43 (67) | 53 (68) | 1a | |
| Positive | 2 (3) | 3 (4) | ||
| RET, No. (%) | N/A | 22 (34) | 24 (31) | .438a |
| Negative | 41 (64) | 54 (69) | ||
| Positive | 1 (2) | 0 (0) |
Abbreviations: ADC, adenocarcinoma; ALK, anaplastic lymphoma kinase; EGFR, epidermal growth factor receptor; IQR, interquartile range; KRAS, Kirsten rat sarcoma viral oncogene homolog; N/A, not available; NSCLC, non–small cell lung carcinoma; PD-L1, programmed cell death ligand 1; SqCC, squamous cell carcinoma.
Fisher’s exact test.
TABLE 4.
Association of High PD-L1 Expression (Cutoff = 50%) With Clinicopathologic and Molecular Variables (n = 261)
| Variable | Level | PD-L1: <50% (n = 183) | PD-L1: ≥50% (n = 78) | P |
|---|---|---|---|---|
| Sex, No. (%) | Female | 76 (42) | 36 (46) | .490 |
| Male | 107 (58) | 42 (54) | ||
| Age, y | Median [IQR] (range) | 66.2 [59–72.5] (38.8–92.2) | 64.6 [59–71.6] (37.1–86.7) | .776 |
| Histologic type, No. (%) | ADC | 100 (55) | 45 (58) | .805a |
| SqCC | 59 (32) | 24 (31) | ||
| NSCLC | 12 (7) | 6 (8) | ||
| Adenosquamous | 4 (2) | 2 (3) | ||
| Other | 8 (4) | 1 (1) | ||
| Smoking, No. (%) | N/A | 11 (6) | 6 (8) | .326 |
| No | 19 (10) | 5 (6) | ||
| Yes | 153 (84) | 67 (86) | ||
| Stage, No. (%) | N/A | 18 (10) | 7 (9) | |
| I | 24 (13) | 4 (5) | .080 | |
| II | 22 (12) | 9 (12) | ||
| III | 29 (16) | 21 (27) | ||
| IV | 90 (49) | 37 (47) | ||
| Molecular alterations tested, No. (%) | N/A | 68 (37) | 22 (28) | |
| Negative | 74 (40) | 24 (31) | .008 | |
| Positive | 41 (22) | 32 (41) | ||
| EGFR exon 19, No. (%) | N/A | 62 (34) | 22 (28) | |
| Negative | 117 (64) | 54 (69) | 1a | |
| Positive | 4 (2) | 2 (3) | ||
| EGFR exon 21, No. (%) | N/A | 62 (34) | 22 (28) | |
| Negative | 120 (66) | 56 (72) | 1a | |
| Positive | 1 (1) | 0 (0) | ||
| KRAS, No. (%) | N/A | 65 (36) | 23 (29) | |
| Negative | 89 (49) | 28 (36) | .002 | |
| Positive | 29 (16) | 27 (35) | ||
| MET, No. (%) | N/A | 65 (36) | 24 (31) | |
| Negative | 115 (63) | 53 (68) | 1a | |
| Positive | 3 (2) | 1 (1) | ||
| ALK, No. (%) | N/A | 62 (34) | 22 (28) | |
| Negative | 118 (64) | 53 (68) | .382a | |
| Positive | 3 (2) | 3 (4) | ||
| RET, No. (%) | N/A | 67 (37) | 24 (31) | |
| Negative | 115 (63) | 54 (69) | 1a | |
| Positive | 1 (1) | 0 (0) |
Abbreviations: ADC, adenocarcinoma; ALK, anaplastic lymphoma kinase; EGFR, epidermal growth factor receptor; IQR, interquartile range; KRAS, Kirsten rat sarcoma viral oncogene homolog; N/A, not available; NSCLC, non–small cell lung carcinoma; PD-L1, programmed cell death ligand 1; SqCC, squamous cell carcinoma.
Fisher’s exact test.
Finally, we analyzed PD-L1 expression levels in groups with different clinicopathologic features or molecular alterations. The median PD-L1 expression levels and ranges in different groups categorized by sex, smoking status, stage, age, histologic types, or molecular alterations are summarized in Table 5. PD-L1 expression levels showed a significant difference only between groups with KRAS mutations and without KRAS mutations (78% vs 40%).
TABLE 5.
Association of PD-L1 Expression Levels With Clinicopathologic and Molecular Variables (n = 142)
| Variable | Level | No. | PD-L1 Expression, Median (Range) | P |
|---|---|---|---|---|
| Sex | Female | 64 | 0.50 (0.01–0.99) | .995 |
| Male | 78 | 0.50 (0.01–0.99) | ||
| Age | Spearman correlation | 142 | 0.007 | .935 |
| Histologic type | ADC | 79 | 0.50 (0.01–0.99) | .398 |
| SqCC | 46 | 0.50 (0.01–0.99) | ||
| NSCLC | 10 | 0.80 (0.30–0.95) | ||
| Adenosquamous | 5 | 0.20 (0.10–0.90) | ||
| Other | 2 | 0.26 (0.02–0.50) | ||
| Smoking | No | 12 | 0.25 (0.02–0.80) | .115 |
| Yes | 121 | 0.50 (0.01–0.99) | ||
| Stage | I | 12 | 0.25 (0.02–0.99) | .142 |
| II | 16 | 0.55 (0.02–0.95) | ||
| III | 32 | 0.60 (0.05–0.99) | ||
| IV | 69 | 0.50 (0.01–0.99) | ||
| Molecular alterations tested | Negative | 47 | 0.50 (0.02–0.99) | .119 |
| Positive | 50 | 0.60 (0.02–0.99) | ||
| EGFR exon 19 | Negative | 98 | 0.50 (0.02–0.99) | .596 |
| Positive | 3 | 0.55 (0.05–0.60) | ||
| EGFR exon 21 | Negative | 100 | 0.50 (0.02–0.99) | .113 |
| Positive | 1 | 0.02 (0.02–0.02) | ||
| KRAS | Negative | 58 | 0.40 (0.02–0.99) | .003 |
| Positive | 38 | 0.78 (0.02–0.99) | ||
| MET | Negative | 93 | 0.50 (0.02–0.99) | .659 |
| Positive | 3 | 0.20 (0.10–0.90) | ||
| ALK | Negative | 96 | 0.50 (0.02–0.99) | .448 |
| Positive | 5 | 0.50 (0.02–0.80) | ||
| RET | Negative | 95 | 0.50 (0.02–0.99) | .745 |
| Positive | 1 | 0.40 (0.40–0.40) |
Abbreviations: ADC, adenocarcinoma; ALK, anaplastic lymphoma kinase; EGFR, epidermal growth factor receptor; KRAS, Kirsten rat sarcoma viral oncogene homolog; NSCLC, non–small cell lung carcinoma; PD-L1, programmed cell death ligand 1; SqCC, squamous cell carcinoma.
DISCUSSION
PD-L1 IHC assessment is important for selecting patients with NSCLC for PD-1/PD-L1 blockage immunotherapy. Most clinical trials have evaluated PD-L1 IHC on surgical specimens. However, cytology specimens are often the only materials available, especially for patients at an advanced stage. Limited studies have compared PD-L1 IHC on cytologic specimens and surgical specimens, and most of them had small study cohorts with PD-L1 IHC not performed as part of routine clinical care.26–30 Our study addressed the comparability of PD-L1 IHC on cytologic specimens and surgical specimens in a large cohort of clinical cases (265 clinical cases, including 100 cytologic specimens and 165 surgical specimens). Consistent with previous publications, the findings of the current study confirm that PD-L1–positive rates and high-expression rates are comparable between cytologic and surgical specimens, and this suggests that cytologic specimens are sufficient for PD-L1 IHC evaluation. However, cytopathologists should always be cautious when interpreting PD-L1 staining on cytology specimens because of the presence of immune cells, which can be stained with PD-L1 and are easily misinterpreted as PD-L1–positive tumor cells.31 Therefore, PD-L1 staining on cytology specimens should be evaluated with reference to H & E–stained cell block sections or complementary IHC-stained slides.
PD-L1 expression has been observed to correlate with some clinicopathologic features, including smoking,5,11,19,20,32 an advanced tumor stage,11 and squamous cell carcinoma histology.11,17,21,32–34 However, other studies have failed to demonstrate such associations.13,18,35,36 In our study cohort, we did not observe a significant association of PD-L1 expression with squamous cell carcinoma histology or any other clinical features, including age, sex, smoking, and stage. Although most clinical characteristics were similar between the cytologic and surgical specimen groups in our cohort, there were some differences between these 2 groups: more cases with advanced stages in the cytologic group and more squamous cell carcinoma cases in the histologic group. The observation of similar PD-L1 expression between these 2 groups further suggests a lack of association between PD-L1 expression and advanced stages or squamous cell carcinoma.
PD-L1 expression does not always correlate with a response to PD-1/PD-L1 blockage immunotherapy in some patients with NSCLC.2 Recently, other biomarkers have been investigated to predict the immunotherapy response. The most promising one is the tumor mutation burden (TMB).37,38 In addition, recent studies have demonstrated that TMB and PD-L1 expression are independent variables, and a composite of TMB and PD-L1 test further enriched NSCLC patients who benefited from PD-1/PD-L1 blockage immunotherapy.39,40
An association between PD-L1 and molecular mutations in NSCLC has been investigated with conflicting results. Some studies have documented an association between PD-L1 expression and molecular alterations, including EGFR mutations,13,15,21–24 KRAS mutations,5,17,34,41 and ALK translocations,25 whereas other studies have failed to find such correlations.10,18,32,35,42,43 Our study did demonstrate a significant correlation of PD-L1 expression (PD-L1 positivity, high PD-L1 expression, and expression levels) with KRAS mutations but not with any other mutations. One study investigated the underlining mechanism regulating PD-L1 expression in NSCLCs with KRAS mutations and found that mitogen-activated protein kinase signaling along with signal transducer and activator of transcription 3 was crucial for PD-L1 expression.41 In our study cohort, the surgical specimens with molecular testing (n = 99) showed more cases with KRAS mutations than the cytologic specimens with molecular testing (n = 91). There was no significant difference in PD-L1 expression between all surgical specimens (with or without molecular testing) and all cytologic specimens (with or without molecular testing), but the PD-L1–positive rate was significantly higher in surgical specimens with molecular testing than in cytologic specimens with molecular testing; this suggests an association between PD-L1 expression and KRAS mutations.
The vast majority of the surgical specimens in our cohort were core biopsies. Although a good concordance between biopsy and resection specimens was observed,44 other studies have suggested significant intratumoral heterogeneity of PD-L1 expression,45 with discordance between biopsy and resection specimens.46
A recent lung cancer molecular testing guideline recommends universal molecular testing for patients with advanced-stage lung cancer with an adenocarcinoma component. Institutions can either offer a comprehensive cancer panel (including the genes EGFR, ALK, ROS1, BRAF, MET, ERBB2, KRAS, and RET) or offer targeted testing for the genes in the must-test category (EGFR, ALK, and ROS1) for all appropriate patients and then offer an expanded panel containing BRAF, MET, ERBB2, and RET after performing a single-gene KRAS test to exclude KRAS mutations for patients who are suitable candidates for clinical trials.47 The US Food and Drug Administration has approved immunomodulatory therapies as second-line agents for patients with advanced NSCLC and as first-line therapy for patients with NSCLC with a high level of PD-L1 expression (≥50%) and an absence of EGFR mutations or ALK rearrangements. The finding of an association of PD-L1 expression and KRAS mutations suggests that PD-L1 testing should be performed for patients with KRAS-mutated NSCLC because immunotherapy is the only potentially personalized therapy in these patients.
Our study had a few limitations. First, the study cohort was collected retrospectively, and only cases with PD-L1 testing were included; this might have resulted in a selection bias because PD-L1 testing was usually ordered for patients at an advanced stage at our institution. Inasmuch as squamous cell carcinomas are known to have a paucity of therapeutic options, it is possible that over time clinicians have developed a bias and/or practice to order PD-L1 testing for patients with squamous cell carcinoma more often because it adds another line of treatment. This selection bias might have caused some significant differences between histologic and cytologic specimens, including a higher rate of squamous cell carcinoma diagnoses in histologic specimens in comparison with cytologic specimens and a higher rate of KRAS mutations in histologic specimens in comparison with cytologic specimens. Another possible explanation for a higher rate of adenocarcinoma diagnoses in cytologic specimens is that lung adenocarcinomas are more often located peripherally, whereas squamous cell carcinomas are more often located centrally. A minimally invasive procedure such as FNA is usually preferred for peripheral lung lesions because of the high risk of pneumothorax and bleeding. Second, some groups such as those with EGFR or ALK alterations had a small number of cases, and this limited the statistical analysis power in these groups. Similarly, lower EGFR or ALK alteration rates in the current study cohort in comparison with the general population might be caused by this selective bias because patients with EGFR or ALK alterations usually received EGFR- or ALK-targeted therapy instead of immunotherapy.
In conclusion, our data demonstrate that cytologic specimens are comparable to surgical specimens for PD-L1 evaluation. The association between PD-L1 expression and KRAS mutations may have clinical relevance in selecting patients with NSCLC for immunotherapy.
FUNDING SUPPORT
Ping Mei is partially supported by the Science and Technology Planning Project of Guangdong Province (2017ZC0252), the Medical Scientific Research Foundation of Guangdong Province (A2018174), and the Science and Technology Planning Project of Guangzhou (Guangdong, China).
Footnotes
CONFLICT OF INTEREST DISCLOSURES
The authors made no disclosures.
REFERENCES
- 1.Schiller JH, Harrington D, Belani CP, et al. Comparison of four chemotherapy regimens for advanced non–small-cell lung cancer. N Engl J Med. 2002;346:92–98. [DOI] [PubMed] [Google Scholar]
- 2.Brahmer J, Reckamp KL, Baas P, et al. Nivolumab versus docetaxel in advanced squamous-cell non–small-cell lung cancer. N Engl J Med. 2015;373:123–135. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Borghaei H, Paz-Ares L, Horn L, et al. Nivolumab versus docetaxel in advanced nonsquamous non–small-cell lung cancer. N Engl J Med. 2015;373:1627–1639. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Topalian SL, Hodi FS, Brahmer JR, et al. Safety, activity, and immune correlates of anti–PD-1 antibody in cancer. N Engl J Med. 2012;366:2443–2454. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Garon EB, Rizvi NA, Hui R, et al. Pembrolizumab for the treatment of non–small-cell lung cancer. N Engl J Med. 2015;372:2018–2028. [DOI] [PubMed] [Google Scholar]
- 6.Reck M, Rodriguez-Abreu D, Robinson AG, et al. Pembrolizumab versus chemotherapy for PD-L1–positive non–small-cell lung cancer. N Engl J Med. 2016;375:1823–1833. [DOI] [PubMed] [Google Scholar]
- 7.Herbst RS, Baas P, Kim DW, et al. Pembrolizumab versus docetaxel for previously treated, PD-L1–positive, advanced non–small-cell lung cancer (KEYNOTE-010): a randomised controlled trial. Lancet. 2016;387:1540–1550. [DOI] [PubMed] [Google Scholar]
- 8.Fehrenbacher L, Spira A, Ballinger M, et al. Atezolizumab versus docetaxel for patients with previously treated non–small-cell lung cancer (POPLAR): a multicentre, open-label, phase 2 randomised controlled trial. Lancet. 2016;387:1837–1846. [DOI] [PubMed] [Google Scholar]
- 9.Ilie M, Falk AT, Butori C, et al. PD-L1 expression in basaloid squamous cell lung carcinoma: relationship to PD-1+ and CD8+ tumor-infiltrating T cells and outcome. Mod Pathol. 2016;29:1552–1564. [DOI] [PubMed] [Google Scholar]
- 10.Yang CY, Lin MW, Chang YL, Wu CT, Yang PC. Programmed cell death-ligand 1 expression in surgically resected stage I pulmonary adenocarcinoma and its correlation with driver mutations and clinical outcomes. Eur J Cancer. 2014;50:1361–1369. [DOI] [PubMed] [Google Scholar]
- 11.Velcheti V, Schalper KA, Carvajal DE, et al. Programmed death ligand-1 expression in non–small cell lung cancer. Lab Invest. 2013; 94:107–116. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Konishi J, Yamazaki K, Azuma M, Kinoshita I, Dosaka-Akita H, Nishimura M. B7-H1 expression on non–small cell lung cancer cells and its relationship with tumor-infiltrating lymphocytes and their PD-1 expression. Clin Cancer Res. 2004;10:5094–5100. [DOI] [PubMed] [Google Scholar]
- 13.D’Incecco A, Andreozzi M, Ludovini V, et al. PD-1 and PD-L1 expression in molecularly selected non–small-cell lung cancer patients. Br J Cancer. 2014;112:95–102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Boland JM, Kwon ED, Harrington SM, et al. Tumor B7-H1 and B7-H3 expression in squamous cell carcinoma of the lung. Clin Lung Cancer. 2013;14:157–163. [DOI] [PubMed] [Google Scholar]
- 15.Azuma K, Ota K, Kawahara A, et al. Association of PD-L1 over-expression with activating EGFR mutations in surgically resected nonsmall-cell lung cancer. Ann Oncol. 2014;25:1935–1940. [DOI] [PubMed] [Google Scholar]
- 16.Lin C, Chen X, Li M, et al. Programmed death-ligand 1 expression predicts tyrosine kinase inhibitor response and better prognosis in a cohort of patients with epidermal growth factor receptor mutation–positive lung adenocarcinoma. Clin Lung Cancer. 2015;16:e25–e35. [DOI] [PubMed] [Google Scholar]
- 17.Chan AWH, Tong JHM, Kwan JSH, et al. Assessment of programmed cell death ligand-1 expression by 4 diagnostic assays and its clinicopathological correlation in a large cohort of surgical resected non–small cell lung carcinoma. Mod Pathol. 2018;31:1381–1390. [DOI] [PubMed] [Google Scholar]
- 18.Rangachari D, VanderLaan PA, Shea M, et al. Correlation between classic driver oncogene mutations in EGFR, ALK, or ROS1 and 22C3–PD-L1 ≥50% expression in lung adenocarcinoma. J Thorac Oncol. 2017;12:878–883. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Calles A, Liao X, Sholl LM, et al. Expression of PD-1 and its ligands, PD-L1 and PD-L2, in smokers and never smokers with KRAS-mutant lung cancer. J Thorac Oncol. 2015;10:1726–1735. [DOI] [PubMed] [Google Scholar]
- 20.Takada K, Okamoto T, Shoji F, et al. Clinical significance of PD-L1 protein expression in surgically resected primary lung adenocarcinoma. J Thorac Oncol. 2016;11:1879–1890. [DOI] [PubMed] [Google Scholar]
- 21.Zhang M, Li G, Wang Y, et al. PD-L1 expression in lung cancer and its correlation with driver mutations: a meta-analysis. Sci Rep. 2017;7:10255. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Tang Y, Fang W, Zhang Y, et al. The association between PD-L1 and EGFR status and the prognostic value of PD-L1 in advanced non–small cell lung cancer patients treated with EGFR-TKIs. Oncotarget. 2015;6:14209–14219. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Akbay EA, Koyama S, Carretero J, et al. Activation of the PD-1 pathway contributes to immune escape in EGFR-driven lung tumors. Cancer Discov. 2013;3:1355–1363. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Chen N, Fang W, Zhan J, et al. Upregulation of PD-L1 by EGFR activation mediates the immune escape in EGFR-driven NSCLC: implication for optional immune targeted therapy for NSCLC patients with EGFR mutation. J Thorac Oncol. 2015;10:910–923. [DOI] [PubMed] [Google Scholar]
- 25.Ota K, Azuma K, Kawahara A, et al. Induction of PD-L1 expression by the EML4-ALK oncoprotein and downstream signaling pathways in non–small cell lung cancer. Clin Cancer Res. 2015;21:4014–4021. [DOI] [PubMed] [Google Scholar]
- 26.Heymann JJ, Bulman WA, Swinarski D, et al. PD-L1 expression in non–small cell lung carcinoma: comparison among cytology, small biopsy, and surgical resection specimens. Cancer Cytopathol. 2017;125:896–907. [DOI] [PubMed] [Google Scholar]
- 27.Ilie M, Juco J, Huang L, Hofman V, Khambata-Ford S, Hofman P. Use of the 22C3 anti–programmed death-ligand 1 antibody to determine programmed death-ligand 1 expression in cytology samples obtained from non–small cell lung cancer patients. Cancer Cytopathol. 2018;126:264–274. [DOI] [PubMed] [Google Scholar]
- 28.Noll B, Wang WL, Gong Y, et al. Programmed death ligand 1 testing in non–small cell lung carcinoma cytology cell block and aspirate smear preparations. Cancer Cytopathol. 2018;126:342–352. [DOI] [PubMed] [Google Scholar]
- 29.Russell-Goldman E, Kravets S, Dahlberg SE, Sholl LM, Vivero M. Cytologic-histologic correlation of programmed death-ligand 1 immunohistochemistry in lung carcinomas. Cancer Cytopathol. 2018;126:253–263. [DOI] [PubMed] [Google Scholar]
- 30.Stoy SP, Rosen L, Mueller J, Murgu S. Programmed death-ligand 1 testing of lung cancer cytology specimens obtained with bronchoscopy. Cancer Cytopathol. 2018;126:122–128. [DOI] [PubMed] [Google Scholar]
- 31.Thunnissen E, de Langen AJ, Smit EF. PD-L1 IHC in NSCLC with a global and methodological perspective. Lung Cancer. 2017; 113:102–105. [DOI] [PubMed] [Google Scholar]
- 32.Igawa S, Sato Y, Ryuge S, et al. Impact of PD-L1 expression in patients with surgically resected non–small-cell lung cancer. Oncology. 2017;92:283–290. [DOI] [PubMed] [Google Scholar]
- 33.Li J, Chen Y, Shi X, et al. A systematic and genome-wide correlation meta-analysis of PD-L1 expression and targetable NSCLC driver genes. J Thorac Dis. 2017;9:2560–2571. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Ye L, Leslie C, Jacques A, Mesbah Ardakani N, Amanuel B, Millward M. Programmed death ligand-1 expression in non–small cell lung cancer in a Western Australian population and correlation with clinicopathologic features. Mod Pathol. 2019;32:524–531. [DOI] [PubMed] [Google Scholar]
- 35.Cooper WA, Tran T, Vilain RE, et al. PD-L1 expression is a favorable prognostic factor in early stage non–small cell carcinoma. Lung Cancer. 2015;89:181–188. [DOI] [PubMed] [Google Scholar]
- 36.Kim MY, Koh J, Kim S, Go H, Jeon YK, Chung DH. Clinicopathological analysis of PD-L1 and PD-L2 expression in pulmonary squamous cell carcinoma: comparison with tumor-infiltrating T cells and the status of oncogenic drivers. Lung Cancer. 2015;88:24–33. [DOI] [PubMed] [Google Scholar]
- 37.Hellmann MD, Ciuleanu TE, Pluzanski A, et al. Nivolumab plus ipilimumab in lung cancer with a high tumor mutational burden. N Engl J Med. 2018;378:2093–2104. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Hellmann MD, Callahan MK, Awad MM, et al. Tumor mutational burden and efficacy of nivolumab monotherapy and in combination with ipilimumab in small-cell lung cancer. Cancer Cell. 2018;33:853–861.e4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Rizvi H, Sanchez-Vega F, La K, et al. Molecular determinants of response to anti–programmed cell death (PD)-1 and anti–programmed death-ligand 1 (PD-L1) blockade in patients with non–small-cell lung cancer profiled with targeted next-generation sequencing. J Clin Oncol. 2018;36:633–641. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Hellmann MD, Nathanson T, Rizvi H, et al. Genomic features of response to combination immunotherapy in patients with advanced non–small-cell lung cancer. Cancer Cell. 2018;33: 843–852.e4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Sumimoto H, Takano A, Teramoto K, Daigo Y. RAS-mitogen-activated protein kinase signal is required for enhanced PD-L1 expression in human lung cancers. PLoS One. 2016;11:e0166626. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Inoue Y, Yoshimura K, Mori K, et al. Clinical significance of PD-L1 and PD-L2 copy number gains in non–small-cell lung cancer. Oncotarget. 2016;7:32113–32128. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Cha YJ, Kim HR, Lee CY, Cho BC, Shim HS. Clinicopathological and prognostic significance of programmed cell death ligand-1 expression in lung adenocarcinoma and its relationship with p53 status. Lung Cancer. 2016;97:73–80. [DOI] [PubMed] [Google Scholar]
- 44.Kitazono S, Fujiwara Y, Tsuta K, et al. Reliability of small biopsy samples compared with resected specimens for the determination of programmed death-ligand 1 expression in non–small-cell lung cancer. Clin Lung Cancer. 2015;16:385–390. [DOI] [PubMed] [Google Scholar]
- 45.McLaughlin J, Han G, Schalper KA, et al. Quantitative assessment of the heterogeneity of PD-L1 expression in non–small-cell lung cancer. JAMA Oncol. 2016;2:46–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Ilie M, Long-Mira E, Bence C, et al. Comparative study of the PD-L1 status between surgically resected specimens and matched biopsies of NSCLC patients reveal major discordances: a potential issue for anti–PD-L1 therapeutic strategies. Ann Oncol. 2016;27: 147–153. [DOI] [PubMed] [Google Scholar]
- 47.Lindeman NI, Cagle PT, Aisner DL, et al. Updated molecular testing guideline for the selection of lung cancer patients for treatment with targeted tyrosine kinase inhibitors: guideline from the College of American Pathologists, the International Association for the Study of Lung Cancer, and the Association for Molecular Pathology. J Thorac Oncol. 2018;13:323–358. [DOI] [PubMed] [Google Scholar]