Association of Circulating Tumor DNA Testing Before Tissue Diagnosis With Time to Treatment Among Patients With Suspected Advanced Lung Cancer

Key Points Question Can we improve time to treatment using circulating tumor DNA (ctDNA) genotyping before tissue diagnosis among patients with suspected advanced lung cancer? Findings In this nonrandomized clinical trial, 150 patients with suspected advanced lung cancer underwent ctDNA testing during initial diagnostic workup; 90 patients had tissue confirmation of advanced nonsquamous non–small cell lung cancer. Median time to treatment was 39 days vs 62 days for a reference cohort, with faster turnaround time for genotyping results. Meaning The use of plasma ctDNA testing before tissue diagnosis among patients with suspected advanced lung cancer may expedite biomarker testing and accelerate time to treatment.


LIST OF ABBREVIATIONS/TERMINOLOGY 72
Study Design: This is a prospective single arm, non-therapeutic, minimally invasive study Study Duration: This study will begin recruiting in November 2020. Accrual will occur over 12-18 months with study follow-up completing by April 2023.

Main Criteria for Inclusion/Exclusion:
Inclusion Criteria: Patients referred to the UHN Lung RAMP program with radiologic evidence of advanced disease (incurable stage III or IV) will be eligible, if the following criteria are met: 1.
Ability to provide written informed consent; 5.
Diagnostic biopsy and molecular profiling ordered or planned. Patients remain eligible even if biopsy or tumour testing later fails or is deemed not feasible.
Concurrent active malignancy except for localized non-melanomatous skin cancer or non-invasive cervical cancer. Any previous cancer (excluding NSCLC) must have been treated more than 2 years prior to study entry with no current evidence of active disease.

Objectives:
Primary: 1. To compare time to treatment initiation in advanced NSCLC patients for those that had a liquid biopsy compared to patients referred in the previous 12 months that meet the eligibility criteria. Exploratory: 6. To model the impact of using liquid biopsy ("blood-first") to permit diagnostic cancer tissue sparing. 7. To explore treatment outcomes using liquid biopsy including response and progression-free survival with targeted therapy, and association with time to treatment initiation.

NSCLC and Targeted Therapy 77
Lung cancer is the most frequently diagnosed cancer worldwide and the leading cause of 78 cancer-related mortality in both men and women [1]. Time to diagnosis can vary widely, as this is a complex and multi-dimensional variable. We 120 have learned that prolonged time from symptoms to treatment decision significantly reduces 121 the chance for a patient to access precision medicine, and in some cases, any treatment. The 122 "wait time" to treatment from the patient's point of view is a period of uncertainty that feeds 123 fear and anxiety [13]. It also has a detrimental impact on patient outcomes, with fewer 124 patients able to access effective therapy and consequently reduced survival. 125

Liquid Biopsy and Next-Generation Sequencing 126
Liquid biopsies are simple, non-invasive blood tests to detect circulating tumour DNA (ctDNA) 127 and have been shown to be non-inferior to tumour tissue genotyping in lung cancer [14]. 128 Next-generation sequencing (NGS) of ctDNA using a hybrid capture approach detects 129 aberrations multiple targetable genes simultaneously, including mutations, fusions and copy 130 number variation. This helps to determine the oncogenes potentially driving the growth of a 131 patient's malignancy and their potential to benefit from targeted therapy, however liquid 132 biopsy is faster and easier to obtain than standard tissue biopsy [

Lung Rapid Assessment and Management Program 170
The Lung Rapid Assessment and Management Program (Lung RAMP) is the University 171 Health Network Lung Diagnostic Assessment Program (DAP) supported by Cancer Care 172 Ontario. Lung RAMP aims to quickly and appropriately assess, diagnose and manage 173 patients with presumed lung cancer in the shortest possible timeframe. The Lung RAMP 174 Program has significantly improved the care of lung cancer patients at the University Health 175 Network and in Ontario. What was previously an uncoordinated individual referral process is 176 now a coordinated system capable of providing timely, efficient, coordinated care for patients. 177 The referring physician calls 1-866-LUNG-911 and provides important patient information that 178 avoids delays or repetition of diagnostic tests. The patient is called within 24 hours with a 179 diagnostic and/or treatment plan. The diagnostic workup and referral to the relevant oncology 180 team is expedited, in order to minimize wait times. However, the program has been a victim of 181 its own success, growing from ~350 referrals per year to over 900 per year, with few or no 182 additional resources allocated ( Figure 2A). Most of these patients (~2/3), are referred with 183 advanced disease and are not surgical candidates ( Figure 2B), but still require pathologic 184 diagnosis and staging. Prior to the COVID-19 pandemic, wait times for imaging were less 185 than 28 days, while wait times for biopsy ranged from 28 to 32 days ( Figure 3). These have 186 both lengthened significantly during the pandemic. In addition, the time from biopsy to 187 molecular profiling results was a mean of 28 days pre-COVID but this has also lengthened. 188 Thus, patients can wait approximately 3 months, or potentially longer during the pandemic 189 and recovery, in order to have complete lung cancer staging, a pathologic diagnosis and all 190 molecular results available in order to start treatment. Given the current dearth of resources 191 allocated for biopsy slots and pathologists, the system requires an innovative approach to 192 shorten current wait times, accelerate time to treatment and promote access to personalized 193 medicine for more of Ontario's patients with lung cancer. 194

Impact of the COVID-19 Pandemic 195
Clinical activity throughout Canadian hospitals, including at the University Health Network,196 has been markedly restricted to protect patients and providers during the  pandemic. will accelerate patient access to precision medicine, through allowing rapid molecular 205 diagnosis with plasma ctDNA testing in the face of growing wait times for routine imaging, 206 tumour biopsy, pathologic diagnosis and molecular testing. The potential for a "blood-first" 207 approach to decrease the number of invasive lung biopsies will not only decrease risks for 208 patients, but also for health care providers. The specialist teams that perform these invasive 209 procedures are at high risk for COVID-19 transmission during bronchoscopy and/or 210 interventional biopsy, including transthoracic sampling (image-guided biopsies). 211

Rationale for the study 212
The purpose of this trial is to prospectively assess the utility of blood-based next generation These results suggest that initial biomarker assessment using ctDNA rather than tissue 226 ("blood-first"), reserving tissue for PD-L1 IHC and reflex testing when ctDNA is negative for 227 any known oncogenic driver mutations, will improve the biomarker discovery rate, turn-around 228 time and decrease time to treatment. The net result will be to increase the number of patients 229 with newly diagnosed advanced NSCLC that will receive guideline complete biomarker testing 230 and be able to access a precision medicine approach to their cancer therapy [14]. The 231 acceleration of time to diagnosis and treatment is expected to favorably impact outcomes in 232 this population. 233 Based on these promising results, in this study we will analyze if the "blood-first" approach 234 can accelerate time to treatment compared with the standard diagnostic pathway including 235 tissue genotyping. 236 237

STUDY OBJECTIVES 238
This study will assess the utility of liquid biopsy to accelerate time to treatment for selected 239 patients with radiographic evidence of advanced lung cancer (Figure 1). 240

METHODS 269
This is a prospective single arm, non-therapeutic, minimally invasive study which will 270 conducted at the University Health Network (UHN). One hundred and fifty patients will be 271 accrued over an estimated 12-18 months. Based on current Lung RAMP referral patterns and 272 demographics, one third will be never smokers (target subgroup accrual N=40). 273 Eligible patients will be identified through the weekly Lung RAMP MCC and contacted by the 274 study coordinator. Consenting patients would undergo liquid biopsy (plasma ctDNA testing) in 275 addition to standard of care imaging, tumour biopsy and tumour tissue molecular profiling 276 through Lung RAMP. Patients with non-diagnostic tumour biopsies or insufficient tumour 277 tissue for molecular profiling would also be eligible to participate. 278 Consenting patients would undergo liquid biopsy during their first or next onsite visit to UHN 279 (e.g. imaging, standard of care blood tests or medical assessment). Whenever possible, the 280 liquid biopsy would be added to a planned blood draw for standard of care, avoiding 281 additional venipuncture for the patient. A total of 4 Streck tubes of blood (40 mL total) will be 282 drawn. Samples will be de-identified and labelled with a study code. 283

Liquid Biopsy 284
Consenting patients will undergo peripheral blood draw (approximately 40 mL) collected in 285 Streck™ tubes or kits provided by Inivata. Two tubes (20 mL) will be shipped to Inivata (North 286 Carolina, USA) for InVisionFirst ® Lung profiling in real time. The remaining 2 tubes will be 287 sent to the Advanced Molecular Diagnostics Laboratory (AMDL) at Princess Margaret Cancer 288 Centre for nucleic acid extraction and mutation profiling using the Oncomine™ Pan-Cancer 289 Cell-Free Assay (ThermoFisher). This panel can detect lung tumour-derived clinically-relevant 290 SNVs, small indels and fusions. 291

Data Collection 292
Molecular profiling results from both tumour and ctDNA will be recorded, and incremental 293 actionable genomic targets will be identified for each method. Any complications from either 294 the liquid or tumour biopsy will be captured, as well as any failure of tissue or blood sampling, 295 molecular profiling and repeat biopsies. 296 Turnaround time for molecular results for liquid and tumour tissue biopsy, time to treatment 297 initiation, treatment received and outcomes including tumour response, progression-free and 298 overall survival will be collected. 299 Patients will be asked to complete the EQ-5D-5L at baseline and at 3 months to assess 300 quality of life. Patients may complete the EQ-5D-5L survey by phone or by mail if they do not 301 have a routine (in person) doctor's visit scheduled. Patients that are unable to complete the 302 EQ-5D-5L (literacy, physical issues, unavailable translation) are still eligible to participate. 303 Patient visits and procedures during the diagnostic work up will be collected. Costs of both 304 diagnostic strategies will be captured using time in motion studies, direct costs from UHN and 305 list costs as appropriate. These will be presented as a cost-consequence analysis of liquid 306 biopsy vs. standard of care tissue biopsy and molecular profiling. 307 308

SELECTION OF SUBJECTS 309
Patients referred to the Lung RAMP program will be eligible if they are deemed to have 310 advanced lung carcinoma by imaging. These patients (all cases) are currently discussed at 311 the weekly Lung RAMP Multidisciplinary Cancer Conference (MCC) that includes thoracic 312 surgery, interventional respirology, radiology, radiation and medical oncology representatives. 313 Patients reviewed at MCC who meet the inclusion criteria will be approached to participate in 314 the study as part of their ongoing diagnostic work up. 315

Inclusion Criteria: 316
Patients referred to the Lung RAMP program with radiologic evidence of advanced disease 317 (incurable stage III or IV) will be eligible, if the following criteria are met: 318 or non-invasive cervical cancer. Any previous cancer (excluding NSCLC) must 329 have been treated more than 2 years prior to study entry with no current evidence 330 of active disease. 331 332

Schedule of Procedures and Observations 334
All participants must provide written, signed, informed consent using the latest approved 335 version of the Institutional Research Ethics Board informed consent form (ICF). A copy of the 336 signed ICF will be given to the subject. The original will be kept on file in study records. Study 337 overview is shown in Figure 4. 338

Screening Period 339
The following procedures and assessments must be completed prior to pre-study 340 assessment: 341 • MCC review and confirmation of eligibility 342 • Informed Consent 343

Pre-Study Period (can be same day as screening) 344
• Blood collection into 4 Streck cell preservation tubes (40 mL) 345 • Baseline quality of life assessment using the EQ-5D-5L 346

Study Period 347
Patients will proceed with standard of care diagnostic work up, imaging, tumour biopsy and 348 tumour tissue molecular profiling through Lung RAMP. There may be some variability in 349 timing of scans and tests. 350

•
Repeat quality of life assessment will be administered at 3 months (12 weeks +/-4 351 weeks) from study blood collection using the EQ-5D-5L 352 • Molecular profiling results from both tumour and ctDNA will be recorded 353 • Patients will be treated as per standard of care based on tissue and/or blood 354 genotyping results 355 • Time to treatment initiation and patient visits/procedures will be recorded 356

Discontinuation of Study and Study Follow-Up 357
Patients will be followed for a minimum of 12 months and up to 2 years. Patients will be 358 discontinued from the study when the first of any of the following events occurs: 359 This is a single arm, minimally invasive non-therapeutic study conducted at a single centre. 377 The time to treatment decision (T LB ) in the study cohort by liquid (T L ) and tissue biopsy (T B ) is 378 measured from the date of referral to the earliest date of receiving a liquid or tissue biopsy 379 report indicating actionable genomic aberrations, or T LB = min (T L, T B ). The time to treatment 380 decision using tissue biopsy alone (T B ) will be collected in a chart-review comparison cohort 381 (patients referred in the previous 12 months that meet the eligibility criteria). The time to 382 treatment decision by liquid biopsy (T LB ) vs by tissue biopsy alone (T B ) will be compared. 383

Statistical Analyses 384
The primary objective is to compare time to treatment initiation in advanced NSCLC patients 385 for those that had a liquid biopsy compared to patients referred in the previous 12 months that 386 meet the eligibility criteria. • To evaluate concordance between liquid and tissue for identification of actionable 397 targets within the study cohort; and, 398 • To model cost-effectiveness of upfront use of liquid biopsy to triage advanced NSCLC 399 to medical oncology compared to current standard of care costs. 400 Descriptive analysis will be used to describe treatments received and disease outcomes 401 (confirmed or unconfirmed response by RECIST 1.1), including subgroup analyses of 402 actionable genomic alterations (such as EGFR). The Kaplan-Meier method will be used to 403 describe progression-free, time to treatment failure and overall survival in the study cohort. 404 Patients will be followed for a minimum of 12 months and up to 24 months. 405 Change scores in health-related qualify of life and patient utility (EQ5D-5L) between baseline 406 and 3 months will be calculated and summarized for those with actionable genomic 407 aberrations identified in liquid biopsy and those requiring standard of care tissue diagnosis 408 and profiling. 409 A cost-effectiveness model will be developed comparing the initial use of liquid biopsy versus 410 the current standard of tissue biopsy and profiling from the perspective of the Canadian 411 healthcare system for the horizon of the study period. Study data will be used as model 412 inputs, as well as costs (current CAD) from UHN and published list prices. Additional inputs 413 will be derived from published literature and expert opinion as required. Sub-analysis of 414 resource utilization with each approach will be performed to estimate COVID exposure risk to 415 patients and healthcare providers. In addition, the potential of liquid biopsy to spare tumour 416 tissue will also be derived. 417

Sample Size Calculation 418
It is estimated that ~175 patients/year with advanced lung cancer are referred to the UHN 419 Lung RAMP program for diagnostic work up. Based on current Lung RAMP data, at least 90% 420 of these will have NSCLC subtype, (<10% small cell carcinoma or non-lung cancer 421 pathology), and an additional 5% will decline participation or be ineligible. Thus 150 eligible 422 patients will be seen per year, and the majority will have non-squamous subtype (~85% the comparison group referred in the prior 12 months. Assuming a 4-6 week reduction in time 432 to treatment, an equal group size (N1=N2), a standard deviation of 4 for the liquid biopsy 433 group and an unequal standard deviation of 5 for the comparison group to account for 434 variability in getting successful tumour molecular testing, we will achieve a power >80%, at 435

Study Period
Repeat EQ-5D-5L at 12 +/-4 weeks Molecular results Treatment per SOC Treatment outcomes Record of visits/procedures  Comparing to a group from the previous 12 months (N2) and assuming a reduction in 4-6 weeks (δ), we will achieve a power of 614 >80% with this sample size. N1: Number of patients with actionable lung cancer targets in study (group 1). N2: Number of 615 patients with actionable lung cancer targets referred in the preceding 12 months (group 2). σ1: Standard Deviation of group 1; 616 σ2: Standard Deviation of group 2. δ: Difference in time from initiation of referral to treatment (in weeks) between group 1 and 617 group 2. 618 619