Biomarker Testing Among Patients With Nonsquamous NSCLC Across Three Time Periods and Three Testing Platforms—A Brief Report Authors: Zhen (Jason) Fan, Roxana Tudor, Michael Cabanero, Lisa W Le, Mary Rabey, Jennifer H Law, Tracey Stockley, Shelley Kuang, Nicholas Meti, Andreas Fung, Kirstin Perdrizet, Kaitlin Chen, Janice JN Li, Noor Ghumman, Laura Ranich, Cuihong Wei, Peter Sabatini, Prodipto Pal, Ying-Han R Hsu, Hyang Mi Ko, Joerg Schwock, Ming S Tsao, Geoffrey Liu, Frances A Shepherd, Penelope Bradbury, Adrian Sacher, Lawson Eng, Natasha B Leighl Published: 2026 ABSTRACT AbstractBiomarker testing in stage III/IV nonsquamous NSCLC is critical in treatment decision-making. This ambispective review included 1028 patients across three time periods to evaluate the clinical impact of evolving genomic testing platforms. Cohort 1 (C1: 2015–2017) had reflex EGFR/ALK testing. Cohort 2 (C2: 2017–2020) used limited next-generation sequencing (NGS; 15 genes), ALK, and ROS1 testing. Cohort 3 (C3: 2020–2022) used comprehensive NGS (161 genes). Larger NGS panels identified more actionable genomic alterations, but increased biopsy-to-result turnaround times and fewer patients had complete molecular results at first medical oncology consultation leading to delays in treatment initiation.Keywords: Biomarkers, NSCLC, Precision medicine, Turnaround time Introduction Guidelines recommend molecular testing with broad-spectrum next-generation sequencing (NGS) at diagnosis for patients with advanced nonsquamous NSCLC to guide optimal treatment decisions.1 For example, use of targeted therapy in EGFR-mutant NSCLC has changed the landscape of metastatic lung cancer.2 In a historic patient cohort at our institution, only 21% had molecular test results before their first oncology consultation, with a median of 16 days from consult to treatment start versus 29 days among those for whom test results were not available.3 Furthermore, 19% of patients started cytotoxic chemotherapy when an actionable driver mutation was later identified.3 In the 2024 International Association for the Study of Lung Cancer survey, 43% of the respondents started treating patients before receiving molecular results, citing time, cost, and access among barriers to successful and timely biomarker testing.4 This contrasts the College of American Pathologists/International Association for the Study of Lung Cancer/Association for Molecular Pathology recommendations.5 At our institution, we have implemented reflex molecular profiling using NGS. Broad comprehensive assays lead to the identification of more actionable genomic alterations (AGAs) compared with single-gene assays. As testing panels evolve, the impact on turnaround time (TAT) and time to treatment has yet to be elucidated, particularly in a publicly funded health care system such as Canada. Therefore, we sought to explore the clinical impact of different molecular testing platforms on the frequencies of AGA detection and identify TAT and time to treatment in patients with advanced nonsquamous NSCLC. Materials and Methods Study Design and Patient Selection An ambispective review of patients with advanced nonsquamous NSCLC at the Princess Margaret Cancer Centre was conducted. Ethics approval was granted before the initiation of this study by the institutional research ethics board (UHN Board C). Informed consent was obtained from the patients. Patients with newly diagnosed stages III and IV (American Joint Committee on Cancer version 7) disease with successful molecular biomarker test results were included. C1 was identified retrospectively through the Princess Margaret Cancer Registry and included all pathology records signed out as lung cancer or suspicious for lung carcinoma, between January 2015 and January 2017 from the University Health Network molecular laboratory information system. C2 was prospectively identified between February 2017 and September 2020. C3 was identified prospectively between October 2020 and February 2022. Patients who only had pathology review at our institution were excluded. Data Collection Demographics including age, sex, ethnicity, cancer stage at diagnosis, and smoking status were collected. Relevant clinical dates, including date of tumor biopsy, diagnosis, biomarker test initiation and result return, first consultation with medical oncology (MO), and treatment start (cytotoxic, targeted therapy, immunotherapy, etc.), were abstracted from clinical notes and the molecular laboratory information system. Date of diagnosis was defined as the date of histologically confirmed cancer. Molecular Testing Protocols Patient samples in C1 underwent single-gene EGFR testing using real-time polymerase chain reaction to detect 30 recurrent hot spot mutations in exons 18 to 21 and exon 20 resistance mutations.6 C2 used TruSight Tumor 15 (Illumina, San Diego, CA) NGS to evaluate 15 clinically relevant cancer-associated genes.7 C3 used Oncomine Comprehensive Assay version 3 (Thermo Fisher Scientific, Waltham, MA) NGS to assess variants in 161 genes, including single-nucleotide variants, copy number variants, and gene fusions.8 In cases where NGS could not be performed, single-gene testing of EGFR, immunohistochemistry (IHC) for ALK (5A4 IHC), PD-L1 (22C3 IHC), and fluorescent in situ hybridization and IHC for ROS-1 (D4D6) were performed.9 Across all cohorts, genomic testing is requested reflexively by the pathologist after a pathologic diagnosis of lung cancer is made. Statistical Analysis Patient characteristics, TAT, and clinical outcomes were summarized using median and interquartile range for continuous variables and frequency and percentage for categorical variables. The nonparametric Wilcoxon-Mann-Whitney test was used to test differences in time from initial MO consult to treatment start. Cochran–Armitage test was used to test the trend in three time periods. SAS v.9.4 was used, and statistical significance level was set at 0.05. Results and Discussion Patient demographics and characteristics of all three cohorts are presented in Table 1. There were 177 patients in C1, 657 patients in C2, and 194 patients in C3 with stage III and stage IV diseases. Across all cohorts, there was Asian predominance and primarily adenocarcinoma histology with stage IV disease. Median age was comparable among the cohorts, with more females and never smokers observed in C1. The median time from biopsy to test request was 4 and 7 days in C1 and C2 to 3, respectively. Molecular reporting from request date was longer in C2 (18 d) and C3 (19 d) versus C1 (9 d). At our institution, samples are all processed in-house. Samples are batched, but this occurs for both NGS and single-gene testing. Bottlenecks likely occur at the pre-genomics laboratory step, as contemporary cohorts require 7 days before genomic testing is requested. However, NGS testing itself prolongs TAT, on the basis of the laborious analytical processing and annotations required. We did not evaluate workflow within our genomics laboratory, but we suspect that there are bottlenecks as well. This translates to a statistically significant difference in patients with molecular results at time of MO consultation and time to start treatment. However, our TAT still largely falls within the recommended TAT in the Canadian context.10 Further work can evaluate our institutional practices to identify the rates of failed genomic testing among all biopsy samples and workflow. Table 1Demographics and Disease Characteristics of Patients With Nonsquamous NSCLC and Molecular Alterations IdentifiedCohortC1C2C3Stage III/IV patients (n)177657194Median age (range), y70 (46–93)69 (19–97)69 (32–91)Female sex106/177 (60%)313/657 (48%)96/194 (49%)Asian ethnicity55/111 (50%)124/334 (37%)37/63 (58%)Never-smoker71/171 (42%)193/639 (30%)65/190 (34%)Pathology subtype Adenocarcinoma173/177 (98%)558/657 (85%)175/194 (90%) Other4/177 (2%)99/657 (15%)19/194 (10%)Stage (AJCC v.7) III54/177 (31%)194/657 (30%)56/194 (29%) IV123/177 (69%)463/657 (70%)138/194 (71%)Median (IQR), d Time from biopsy to test request Time from test request to report Time from biopsy to report4 (1–7)9 (6–12)14 (6–12)7 (4–11)18 (15–22)26 (21–32)7 (5–10.5)19 (15–24)27 (22–34)Patients with AGA identifiedEGFR Exon 19 del L858R Exon 20 ins Exon 18 alt T790M L861Q S768I ALK BRAF V600E ERBB2, exon 20/TMD/other KRAS G12C MET exon 14/amp/other NRG1 fusion RET fusion ROS1 fusionStage IV patients with AGA66/177 (37%)61 (34%)2130500235 (3%)NANANANANANA052/123 (42%)252/657 (38%)128 (19%)62517543036 (5%)8 (1%)NA71 (11%)1 (0.2%)NANA8 (1%)181/463 (39%)95/194 (49%)43 (22%)425268-136 (3%)5 (3%)5 (3%)18 (9%)13 (7%)1 (0.5%)2 (1%)3 (1.5%)73/138 (53%)Note: Genes marked as NA were not tested.AGA, actionable genomic alteration; AJCC, American Joint Committee on Cancer; IQR, interquartile range; NA, not available. Demographics and Disease Characteristics of Patients With Nonsquamous NSCLC and Molecular Alterations Identified Note: Genes marked as NA were not tested. AGA, actionable genomic alteration; AJCC, American Joint Committee on Cancer; IQR, interquartile range; NA, not available. More than 85% of patients saw MO in consultation, but further work is required to identify reasons for no MO consult (Table 2). In C1, 64% of stage IV patients had complete molecular results at the initial consult compared with C2 (31% stage IV) and C3 (14% stage IV). The median time from initial MO consult to treatment start (Table 3) was shorter if complete molecular results were available compared with those without complete molecular results (13 versus 25 d C1, p < 0.0001; 15 versus 25 d C2, p < 0.0001; 19 versus 29 d C3, p < 0.0001). Conversely, time from biopsy to first MO consult was longer across all three cohorts in patients who had complete molecular data, compared with those who saw MO without complete molecular data. We suspect that varying triaging practices exist among staff oncologists and may influence this finding. For instance, certain staff oncologist may only accept the referral once complete genomic information is available thus leading to the longer TAT initially, with shorter TAT to starting treatment. Conversely, some may accept referrals regardless of genomic result completeness. Further work is required to evaluate our triaging process and the timing of referrals to our clinic by the provider who requested the initial biopsy. Furthermore, TAT in C3 within the cohort of patients without complete genomic data from biopsy to MO consult was likely driven by use of liquid biopsy, whereas tissue results remained pending. Compared with our historic cohort from 2011 to 2013, median TAT from oncology consultation to treatment initiation was improved from 16 days to 13 days in C1 if complete molecular results were available, but it remained comparable with contemporary cohorts.3 Table 2Proportion of Patients Who Saw MO and Patients Who Had Complete ResultsCohortC1C2C3MO consult requested156/172 (91%)556/657 (85%)166/185 (90%) Stage III47/53 (87%)163/194 (84%)49/56 (88%) Stage IV109/119 (92%)393/463 (85%)117/129 (91%)Complete molecular results available at initial MO consulta99/144 (69%)202/548 (37%)38/166 (23%) Stage III32/40 (80%)83/158 (53%)22/49 (45%) Stage IV67/104 (64%)119/390 (31%)16/117 (14%)Note: Complete = patients with a histologic diagnosis and full biomarker results at first MO consultation.MO, medical oncology.a12 patients in C1 and 8 patients in C2 saw MO but status of molecular results was unknown at initial consult. Proportion of Patients Who Saw MO and Patients Who Had Complete Results Note: Complete = patients with a histologic diagnosis and full biomarker results at first MO consultation. MO, medical oncology. 12 patients in C1 and 8 patients in C2 saw MO but status of molecular results was unknown at initial consult. Table 3TAT and Choice of First-Line Therapy in Patients With Stage IV DiseaseCohortC1C2C3Number of stage IV patients123463129Patients with complete molecular results at time of MO consult, median (IQR), d Time from biopsy to initial MO consult25 (18–33)28 (21–38)34 (13–42) Time from initial MO consult to treatment start13 (0–25)15 (5–28)19 (7–32) Time from biopsy to treatment start37 (25–56)44 (34–64)51 (48–74)Patients with incomplete molecular results at time of MO consult, median (IQR), d Time from biopsy to initial MO consulta10 (4–19)13 (7–21)8.5 (−7.5 to 14) Time from initial MO consult to treatment start25 (8–35)25 (15–43)29 (16–39) Time from biopsy to treatment start30 (17–55)39 (27–60)35 (20–45)First-line treatment Chemo26 (21%)93 (20%)7 (5%) IO4 (3%)51 (11%)18 (14%) Target42 (34%)110 (24%)44 (34%) Chemo/IO1 (1%)36 (8%)21 (16%) Target/IO or Target/Chemo04 (1%)3 (2%) No systemic therapy50 (41%)169 (37%)36 (28%)Note: Target = targeted therapy.Chemo, chemotherapy; IO, immunotherapy; IQR, interquartile range; MO, medical oncology; TAT, turnaround time.aNegative values indicate that some patients had tissue biopsy after the initial consult. TAT and Choice of First-Line Therapy in Patients With Stage IV Disease Note: Target = targeted therapy. Chemo, chemotherapy; IO, immunotherapy; IQR, interquartile range; MO, medical oncology; TAT, turnaround time. Negative values indicate that some patients had tissue biopsy after the initial consult. Overall, more patients in C1 started chemotherapy, fewer started immunotherapy, and similar numbers started targeted therapy across all cohorts. Targeted therapies are typically publicly funded. Our data did not delineate the specific therapies used; however, classical EGFR and ALK mutations remain the most common AGAs where tyrosine kinase inhibitors are used across the cohorts. The proportion of stage IV patients who did not receive any systemic therapy was significantly reduced in three time periods (41% C1 versus 37% C2 versus 28% C3, p = 0.03). This may be explained by the increasing use of immunotherapy in the first-line setting in patients with no AGA in recent years and a shift away from cytotoxic chemotherapy alone.11 Further work is needed to identify how many patients subsequently enrolled into clinical trials or sought treatments through non–publicly funded avenues given more AGAs were found with expanded panel genomic testing. Patient outcomes, including overall survival and disease-free survival, across the cohorts would be of interest, although they must be considered against the evolution of systemic treatment options as well. Plasma NGS has demonstrated to increase the rate of complete genotyping and to accelerate time to genotyping results and treatment initiation.12, 13, 14 In our contemporary cohorts, some patients received plasma NGS testing typically on trials or alternate non–publicly funded routes. There are an increasing number of assays, including NGS or multiplex droplet-digital polymerase chain reaction–based methods, and TAT is progressively shortening. Targeted therapy is now indicated in the nonmetastatic setting; thus, biomarker testing is critical across all stages of lung cancer treatment.15 Therefore, the implementation of plasma NGS to facilitate seeing patients sooner remains to be addressed, especially integrating these tests in a publicly funded system. CRediT Authorship Contribution Statement Zhen (Jason) Fan: Conceptualization, Data curation, Investigation, Writing – original draft. Roxana Tudor: Conceptualization, Data curation, Investigation, Writing – review & editing. Michael Cabanero: Conceptualization, Resources, Data curation. Lisa Le: Conceptualization, Methodology, Writing – original draft, Formal analysis. Mary Rabey: Data curation, Writing – review & editing, Project administration. Jennifer Law: Data curation, Resources, Writing – review & editing, Project administration. Tracey Stockley: Resources. Shelly Kuang: Investigation, Data curation, Writing – review & editing. Nicholas Meti: Investigation, Data curation, Writing – review & editing. Andrea Fung: Investigation, Data curation, Writing – review & editing. Kirstin Perdrizet: Investigation, Data curation, Writing – review & editing. Kaitlin Chen: Investigation, Data curation, Writing – review & editing. Janice Li: Investigation, Data curation, Writing – review & editing. Noor Ghumman: Investigation, Data curation, Writing – review & editing. Laura Ranich: Resources. Cuihong Wei: Resources. Peter Sabatini: Resources. Prodipto Pal: Resources. Ying-Han R. Hsu: Resources. Hyang Mi Ko: Resources. Joerg Schwock: Resources. Ming Tsao: Resources, Writing – review & editing. Geoffrey Liu: Resources. Frances Shepherd: Resources, Writing – review & editing. Penelope Bradbury: Resources. Adrian Sacher: Resources. Lawson Eng: Resources. Natasha Leighl: Conceptualization, Methodology, Investigation, Resources, Supervision, Writing – review & editing. Disclosure Disclosure: Dr. Kuang reports receiving payment/honoraria for lectures/presentations/speakers/bureaus/etc. from Merck, AstraZeneca, and Janssen. Dr. Hsu reports providing consulting for Merck and having participation on a data safety monitoring board from AstraZeneca. Dr. Tsao reports receiving grants from AstraZeneca, Bayer, and Sanofi; receiving payment/honoraria for lectures/presentations/speakers/bureaus/etc. from Sanofi and AbbVie; having leadership/fiduciary role for board/society/committee/advocacy from AbbVie, Amgen, AstraZeneca, Bayer, Bristol Myers Squibb, Daiichi Sankyo, and Lilly. Dr. Liu reports receiving grants from AstraZeneca, Boehringer Ingelheim, Roche, and Takeda; receiving payment/honoraria for lectures/presentations/speakers/bureaus/etc. from AbbVie, Amgen, AstraZeneca, Bayer, Bristol Myers Squibb, Jazz Pharmaceuticals, Merck, Novartis, Pfizer, Roche Canada, and Takeda; having leadership/fiduciary role for board/society/committee/advocacy from AstraZeneca, Novartis, Pfizer, Roche Canada, and Takeda. Dr. Shepherd reports receiving grants from AstraZeneca, Bristol Myers Squibb, Lilly, Pfizer, and Roche Canada; receiving payment/honoraria for lectures/presentations/speakers/bureaus/etc. from AstraZeneca, Daiichi Sankyo, Merck Serono, and Takeda; and having leadership/fiduciary role for board/society/committee/advocacy from AstraZeneca and Lilly. Dr. Bradbury reports having leadership/fiduciary role for board/society/committee/advocacy from Mirati Therapeutics, AbbVie, and Bristol Myers Squibb. Dr. Sacher reports receiving grants from AstraZeneca, Bristol Myers Squibb, and Genentech and other financial/nonfinancial interests from AstraZeneca and Genentech. Dr. Leighl reports receiving grants from Amgen, AstraZeneca, BMS, Eli Lilly, GlaxoSmithKline, Janssen, MSD, Novartis, Pfizer, Roche, Takeda, Guardant Health, and Inivata/Neogenomics; receiving support for attending meetings/travel from AstraZeneca, MSD, Roche, Janssen, Sanofi, Guardant Health, Pfizer, and DavaOncology; and serving on the data safety monitoring board (uncompensated) of Mirati Therapeutics and Daiichi Sankyo. The remaining authors declare no conflict of interest. Acknowledgments This work was supported by the Princess Margaret Cancer Registry, Pfizer Canada ULC, and the Princess Margaret Cancer Foundation. Footnotes Cite this article as: Fan ZJ, Tudor R, Cabanero M, et al. Biomarker testing among patients with nonsquamous NSCLC across three time periods and three testing platforms—a brief report. JTO Clin Res Rep 2026;7:101012 References