Tumor-agnostic circulating cell-free DNA biomarkers for longitudinal minimal residual disease monitoring in NSCLC

Brief Report Tumor-agnostic circulating cell-free DNA biomarkers for longitudinal minimal residual disease monitoring in NSCLC Chen Yeh1, Shu-Ti Lin1, Pei-Sung Hsu2, Hung-Chih Lai3 1OncoDxRx, Los Angeles, Abstract: CA, USA. 2Division of Chest Medicine, Identification of cancer patients who are at ongoing risk of relapse remains of high priority to direct Department of Internal effective adjuvant therapy. Plasma circulating cell-free DNA (cfDNA) analyses offer an unprecedented approach to identify patients with minimal residual disease (MRD) earlier than Medicine, Shin Kong Wu clinical or radiological evidence, detecting drug resistance, progression and relapse with relatively Ho-Su Memorial Hospital, high sensitivity and specificity. Here we used longitudinal surveillance to profile cfDNA mutational Taipei, Taiwan. 3Division of landscape in NSCLC patients undergoing EGFR-TKI. We demonstrated how cutting-edge cfDNA- Hematology and Oncology, based NGS and digital qPCR technologies are defining new subgroups of patients with MRD for individually tailored treatment strategy much earlier, and provide new surrogate endpoints for Department of Internal early registration of these therapies. Medicine, Shin Kong Wu Ho-Su Memorial Hospital, Taipei, Taiwan Address for Keywords: circulating cell-free DNA, biomarker, minimal residual disease, EGFR-TKI, NSCLC correspondence: D espite the increasing adoption of Most significantly, cfDNA biomarkers such as Dr. Chen Yeh, personalized and precision medicine, actionable mutations, tumor mutational OncoDxRx, the response rate, resistance and relapse in burden, and microsatellite instability, are patients with advanced non-small-cell lung stratification indicators for targeted therapy Los Angeles, cancer (NSCLC) remains dismal (1). Liquid and immunotherapy. Reduction in cfDNA CA, USA. E‑mail: biopsy, in the precision oncology era, has levels and variant allele frequency (VAF) in the [email protected] dramatically revolutionized the management disease course have also been shown to be of NSCLC, potentially overcoming tissue predictive of treatment benefit that may Submission: 24‑03‑2023 biopsy limitations of single sampling at single supplement conventional imaging approaches Accepted: 21‑06‑2023 time point while providing clonal evolution (7). Moreover, the monitoring of longitudinal Published: 10‑07‑2023 insights as an unprecedented theranostic VAF using cfDNA can identify long-term technology (2-4). Specifically, circulating cell- responders among individuals with initial Access this article online free DNA (ctDNA), enriched and released radiologically stable disease in advanced from high-turnover tumor cells into the cancers (8). Taken together, cfDNA may aid in Quick Response Code: bloodstream of cancer patients, has emerged as the precise treatment of cancer and may help a non-invasive, faster, cheaper and real-time monitor patients’ responses to treatment both Website: standard of care for monitoring efficacy, during and after treatment. In parallel, www.thoracicmedicine.org resistance and recurrence in advanced NSCLC circulating cfDNA surveillance can potentially DOI: patients (5, 6). capture drug resistance earlier and provide 10.4103/atm.atm_520_21 Plasma cfDNA analysis can identify residual more opportunities for precision therapy for proliferating disease in adjuvant settings and patients. Clinical utility of cfDNA testing are estimate tumor load in metastatic settings. being explored, including clonal evolution under selection pressure and analysis of This is an open access journal, and articles are resistance pathways (8). In general, cfDNA can distributed under the terms of the Creative Commons Attribution‑NonCommercial‑ShareAlike 4.0 License, which How to cite this article: Yeh C, Lin S-T, Hsu P-S, allows others to remix, tweak, and build upon the work Lai H-C. Tumor-agnostic circulating cell-free non‑commercially, as long as appropriate credit is given and DNA biomarkers for longitudinal minimal the new creations are licensed under the identical terms. residual disease monitoring in NSCLC. Ann Thorac Med 2023;19:66‑71. For reprints contact: [email protected] 66 © 2023 Annals of Thoracic Medicine | Published by Wolters Kluwer ‑ Medknow

Yeh, et al.: Circulating cell-free DNA biomarkers for MRD in NSCLC provide a detailed landscape of emerging mechanisms of overcome these constraints. The dPCR drug resistance. method distributes PCR reactions into millions Compelling evidence also suggested that cfDNA can be of water-oil emulsion droplets. The advantages used as a prognostic predictor of relapse risk via analysis of of dPCR are its outstanding sensitivity minimal residual disease (MRD). MRD refers to residual for absolute mutant quantification down to a tumor cells or biomarkers in the body after local or single copy. The method has the disadvantage systemic cancer treatment, and its revival promotes tumor of detecting only known sequences and metastasis and recurrence (which is also described as analyzing only a limited number of mutations. measurable residual disease or molecular residual disease). Although dPCR has excellent sensitivity and Recently, the use of cfDNA analysis to identify MRD in detection capabilities for various stages of cancer, their solid tumors before clinical or radiographic disease MRD applicability is restricted since these methods recurrence has demonstrated significant therapeutic can only identify known mutations (10). promise (9). Besides, MRD positivity by cfDNA analysis was associated with a poor prognosis in patients with a Case Reports malignant tumor. Because the number of remaining cancer The primary application of cfDNA assay in early-stage cells is likely to be small that they may not cause any signs NSCLC treatment is its ability to identify MRD after or symptoms, and they may even be undetectable by primary tumor resection, thus enabling accurate risk conventional techniques. The commonly used high- assessment and adjuvant therapy. Adjuvant treatment sensitive MRD detection techniques include quantitative may be avoided in the future for a significant PCR (qPCR), digital PCR (dPCR), and next-generation proportion of cfDNA-negative individuals who are sequencing (NGS). Among them, NGS, as an emerging deemed high-risk. Second, cfDNA clearance may serve MRD detection technique, is gaining increasing attention as an endpoint in adjuvant trials to assess the and clinical application. effectiveness of treatment, allowing for shorter follow-up NGS is a high-throughput technology that can identify de times and smaller sample sizes. Third, confirming novo genetic alterations. As more therapeutically relevant adjuvant treatment duration based on cfDNA clearance molecular targets become available, NGS becomes more will aid in reducing excessive toxicity. As a result, the important in oncology. Although genome-wide sequencing cfDNA assay has enormous promise for speeding up the may provide more detailed genomic information, cfDNA development of NSCLC adjuvant therapies. Taken NGS in clinical application usually employs hybrid capture together, circulating cfDNA surveillance during adjuvant or amplicon-based panels to provide clinically relevant therapy could provide opportunities for genome-based information with lower cost and higher coverage depth. therapy before rapid NSCLC progression. Analytical sensitivity is limited by low levels of cfDNA in Hence, additional data validating the role of cfDNA in the blood and sequencing artifacts. More clinical predicting and monitoring clinical outcomes in the first- investigation of novel NGS approaches is required to line setting of NSCLC are warranted. In this real- Figure 1: Longitudinal cfDNA monitoring can detect MRD earlier than CT imaging. (A) The timeline for patient #1 clinical course from treatment initiation until last follow up date with cfDNA analyses was shown. The blue bars represented the total EGFR copy numbers detected by digital qPCR, including wild-type alleles. The green, red and grey curves indicated CEA levels, EGFR L858R, and T790M allele frequencies, respectively. The grey closed circles showed the number of somatic driver mutations detected by comprehensive NGS at the corresponding time points. Afatinib and osimertinib treatment initiation were also marked by arrows. (B) The radiographic and bone scan images for patient #1 following afatinib and subsequent osimertinib treatment were shown. 67

Yeh, et al.: Circulating cell-free DNA biomarkers for MRD in NSCLC world study, we prospectively evaluated longitudinal low allele frequency (AF). After one cycle, the EGFR plasma samples to investigate the potential of cfDNA L858R mutation were no longer detectable. Repeat biomarker kinetics as an early indication of cfDNA assessments following the third and fourth cycles therapeutic efficacy and predictor of disease state in of first line therapy detected no somatic mutations, and patients with NSCLC undergoing standard first-line imaging after completion of the fourth cycle showed a treatments. Our MRD solution is being designed to partial response. Forty-six weeks after completion of first utilize a tumor-agnostic approach with NGS and line therapy, cfDNA analysis by digital qPCR showed digital qPCR. This approach identifies somatic genetic recurrence of the EGFR L858R concomitant with EGFR alterations in cfDNA extracted from the patient’s T790M mutation. Osimertinib was then administered plasma. The MRD tests that we developed are accordingly, and qPCR showed an excellent molecular intended for patients diagnosed with solid tumor response with the AF of both EGFR L858R and T790M malignancies to detect ctDNA before, during, and mutations diminished (Fig. 1A). At week 72, imaging after cancer treatment. Such information could be started to show progressive disease with clinical used for guidance of adjuvant therapy decisions and/ worsening and the development of left rib and sacrum or for monitoring of cancer recurrence, in conjunction progression (Fig. 1B). Most importantly, the with other clinicopathological findings. To determine comprehensive 500-gene NGS panel detected elevated if cfDNA-based technology can be used to monitor a numbers of somatic driver mutations earlier than digital patient’s response to systemic therapy, we analyzed qPCR, blood CEA levels and radiographic progression serial plasma samples from a cohort of 8 patients (Fig. 1A). In this case, cfDNA-based NGS testing captured during and after therapy. We detected a marked disease progression that was at least 3 months earlier increase in mutation abundance that preceded than EGFR-focused digital qPCR and computed radiographic evidence of disease progression in 7 tomography (CT) imaging. Moreover, digital qPCR alone patients. The representative cases of how changes can’t obtain the full landscape of tumor clonal evolution, detected in cfDNA preceded radiographic progression thus could miss the best opportunity for MRD detection of disease during a patient’s clinical course were and effective therapies. highlighted below. Case 2 Case 1 Patient #2, a 65 years-old female non-smoker, was Patient #1, a 59 years-old female non-smoker, was diagnosed with stage IV metastatic NSCLC with diagnosed with stage IV NSCLC with bone EGFR L858R mutation, and initially received first metastasis. Prior to initiation of first line EGFR-TKI line EGFR-TKI gefitinib with a good molecular with afatinib, we detected a EGFR L858R mutation at response measured by NGS, digital qPCR and blood Figure 2: Liquid biopsy cfDNA analysis is a valuable diagnostic companion to imaging in MRD surveillance for disease recurrence. (A) The timeline for patient #2 clinical course from treatment initiation until last follow up date with cfDNA analyses was shown. The blue bars represented the total EGFR copy numbers detected by digital qPCR, including wild-type alleles. The green, red and grey curves indicated CEA levels, EGFR L858R, and T790M allele frequencies, respectively. The grey closed circles showed the number of somatic driver mutations detected by comprehensive NGS at the corresponding time points. Gefitinib and osimertinib treatment initiation were also marked by arrows. (B) The radiographic images for patient #2 at diagnosis and following osimertinib treatment were shown. 68 Annals of Thoracic Medicine ‑ Volume 19, Issue 1, July‑September 2023

Yeh, et al.: Circulating cell-free DNA biomarkers for MRD in NSCLC CEA (Fig. 2A). At week 40 post diagnosis, local biopsies, and the difficulties of repeated sampling, the recurrence of the tumor was noted on CT scan (Fig. use of longitudinal cfDNA profiling to assess patient 2B) and digital qPCR showed EGFR L858R and responses and MRD could have significant clinical T790M mutations, and second line osimertinib benefits. In this pilot study with a cohort of total eight therapy was initiated. Plasma cfDNA analysis 8 patients, the combined NGS and digital qPCR methods weeks after initiation of second line therapy qPCR were able to detect actionable EGFR mutations (L858R, detected no tumor associated mutations. However, T790M, 19Del) in cfDNA with perfect concordance nearly six months after initiating second line therapy (100%, n = 49/49), and provides preliminary evidence the patient’s same disease-associated EGFR mutations of how cfDNA can reveal MRD in EGFR-TKI treated reappeared (L858R and T790M), yet blood CEA levels NSCLC patients. The number of somatic driver were decreased. Repeat cfDNA NGS analysis revealed mutation in cfDNA analysis by comprehensive NGS a 3-6 fold increase in the numbers of somatic driver was found to be a reliable, robust and early biomarker mutations (Fig. 2A). Surveillance imaging also for MRD; in contrast, EGFR-only digital qPCR, though showed progressive disease. This case illustrates an highly sensitive, has missed relevant clones beyond example where liquid biopsy cfDNA testing was a EGFR and thus was not enough for MRD detection. valuable diagnostic companion to imaging in MRD The MRD panel should include both ubiquitous and surveillance for disease recurrence. heterogeneous genes, as well as clinically relevant driver mutations, such that heterogeneity of the tumor Case 3 could be well covered. Current assessments of Patient #3, a 79 years-old male has history of newly- treatment response and disease progression in NSCLC diagnosed left central lung cancer with brain/liver/ relied on clinical examination and the presence of MRD bone/right adrenal metastases. Profiling of the on radiological imaging. The detection of residual patient’s tumor DNA by different methodologies disease, recurrence or development of drug resistance gave discordant EGFR results, however, cfDNA based on cfDNA analysis may precede such digital qPCR and NGS detected an actionable EGFR assessments, thereby permitting earlier intervention. exon 19 deletion (19Del). Following treatment with Although at present, detailed tumor genome profiling the first line gefitinib, this 19Del gradually fell to requires the direct sequencing of tumor biopsy, our undetectable levels in serial samples collected up to 15 weeks after diagnosis (Fig. 3A), suggesting that the patient had EGFR TKI-sensitive disease. CT imaging after treatment completion and 3 months after completion of therapy also showed stable disease (Fig. 3B). In such case, the absence of detectable 19Del in longitudinal cfDNA samples could serve as the MRD-negative baseline for follow-up surveillance and may help guide clinical decisions. Serial cfDNA analyses thereby can reliably detect post-treatment MRD in patients with advanced stage of NSCLC, detecting residual or recurrent disease earlier than standard radiographic imaging. Discussion Figure 3: The absence of detectable biomarkers in longitudinal cfDNA analyses can For surveillance after intervention, cfDNA positivity serve as the MRD-negative baseline for follow-up surveillance. (A) The timeline for accurately predicted patients with progressive patient #3 clinical course from treatment initiation until last follow up date with cfDNA diseases. Comprehensive cfDNA alteration profiles analyses was shown. The blue bars represented the total EGFR copy numbers via NGS offer a reliable strategy for evaluating MRD, detected by digital qPCR, including wild-type alleles. The green, red and grey curves with high consistency with imaging findings. It was indicated CEA levels, EGFR exon 19 deletion, and L858R allele frequencies, able to detect the presence of tumors before imaging, respectively. The grey closed circles showed the number of somatic driver mutations and it is superior to serum biomarkers, such as CEA. detected by comprehensive NGS at the corresponding time points. Gefitinib Moreover, it has the potential to accurately identify treatment initiation was also marked. (B) The radiographic images for patient #3 MRD in advance and may predict prognostic before and after first line EGFR-TKI treatment were shown. outcomes for relapse-free survival and overall survival. Given the sampling limitations associated with tumor 69

Yeh, et al.: Circulating cell-free DNA biomarkers for MRD in NSCLC study here strongly supported the development of therapy trials. More studies are required to validate the non-invasive techniques including cfDNA analysis, clinical effectiveness of cfDNA biomarkers and further for tumor profiling, real-time monitoring of tumor enhance the sensitivity of cfDNA analysis. Plasma evolution, and the use of cfDNA in the detection of cfDNA assay standards should be established to ensure MRD that may in turn inform recurrence risk and the reproducibility of the results. In summary, the treatment strategies. In addition, longitudinal cfDNA application of cfDNA-based MRD analysis is of great profiling could help to advance precision benefit in providing clinical decision support and immunotherapy and provide more opportunities for enhancing patient survival outcomes in the era of early intervention in patients at high risk of disease precision medicine. progression. Despite promising preliminary results, obstacles exist Ethics approval to the widespread clinical application of cfDNA-based This study was approved by the Institutional assay for MRD monitoring. In plasma, cfDNA levels Research Board of Shin Kong Wu Ho-Su Memorial tend to be variable and low, resulting in a variable Hospital, Taipei, Taiwan, IRB 20190109R. detection threshold. In addition, negative results may be due to low copy number detection rather than the Acknowledgment absence of cfDNA. The limited sensitivity of the This study was supported in part by the Shin Kong Wu cfDNA analysis is a critical challenge, particularly in Ho-Su Memorial Hospital (Grant No patients with resected early-stage cancer, when 2019SKHADR003). plasma cfDNA levels are low. False negatives are inevitable due to the influence of biological variables Conflicts of interest such as low DNA-shedding tumor, and hidden micro- There are no conflicts of interest. metastasis. Comprehensive NGS panels covering a wide range of genomic alterations, larger sample References volume, serial testing, fragment size selection, and sequencing noise suppression may enhance assay 1. Siegel RL, Miller KD, Fuchs HE, Jemal A. Cancer statistics. CA sensitivity (11-15). Standardized and streamlined Cancer J Clin. 2022;72:7–33. methods for cfDNA collection, storage, and analysis are essential to ensure the widespread utilization of 2. Rolfo C, Mack P, Scagliotti GV, Aggarwal C, Arcila ME, Barlesi F, cfDNA technology in clinical MRD detection. Even et al. Liquid biopsy for advanced NSCLC: A consensus statement though our real-world patient study has provided from the International Association for the Study of Lung Cancer. convincing evidence supporting the role of cfDNA in J Thorac Oncol. 2021;16:1647–1662. MRD management, the study only included a small proportion of participants and lacked validation 3. Lone SN, Nisar S, Masoodi T, Singh M, Rizwan A, Hashem S, cohorts. Given the limited samples and observational et al. Liquid biopsy: A step closer to transform diagnosis, findings, further large-scale randomized, controlled prognosis and future of cancer treatments. Mol Cancer. trials are required to verify and clarify the clinical 2022;21:79. usefulness of cfDNA in MRD. 4. Fernandes MGO, Sousa C, Reis JP, Cruz-Martins N, Moura CS, Conclusion Guimarães S, et al. Liquid biopsy for disease monitoring in As more real-world evidence accumulates, it is non-small cell lung cancer: The link between biology and the becoming clear that cfDNA can be used as a clinic. Cells. 2021;10:1912. biomarker for MRD detection and treatment decision- making. The advancement of ultra-sensitive cfDNA 5. Zhou X, Li C, Zhang Z, Li DY, Du J, Ding P, et al. Kinetics of tests could significantly improve patient outcomes. plasma cfDNA predicts clinical response in non-small cell lung Accordingly, cfDNA-based MRD detection could cancer patients. Sci Rep. 2021;11:7633. become an indispensable part of cancer diagnostics and therapeutics. Serial postoperative cfDNA 6. Siravegna G, Marsoni S, Siena S, Bardelli A. Integrating liquid analyses can (i) provide more accurate risk biopsies into the management of cancer. Nat Rev Clin Oncol. stratification for recurrence in addition to 2017;14:531–548. pathological staging; (ii) be used to adjust the intensity and duration of adjuvant treatment 7. Cabel L, Proudhon C, Romano E, Girard N, Lantz O, Stern depending; (iii) predict the effectiveness of adjuvant MH, et al.. Clinical potential of circulating tumor DNA in treatment and enhance the efficiency of adjuvant patients receiving anticancer immunotherapy. Nat Rev Clin Oncol. 2018;15(10):639–50. 8. Zhang Q, Luo J, Wu S, Si H, Gao C, Xu W, et al. Prognostic and predictive impact of circulating tumor DNA in patients with advanced cancers treated with immune checkpoint blockade. Cancer Discov. 2020;10(12):1842–53. 9. Peng Y, Mei W, Ma K, Zeng C. Circulating tumor DNA and minimal residual disease (MRD) in solid tumors: Current horizons and future perspectives. Front Oncol. 2021;11:763790. 10. Franczak C, Filhine-Tresarrieu P, Gilson P, Merlin JL, Au L, Harle A. Technical considerations for circulating tumor DNA detection in oncology. Expert Rev Mol Diagn. 2019;19(2):121- 70 Annals of Thoracic Medicine ‑ Volume 19, Issue 1, July‑September 2023

Yeh, et al.: Circulating cell-free DNA biomarkers for MRD in NSCLC 35. 11. Yang J, Gong Y, Lam VK, Shi Y, Guan Y, Zhang Y, et al. Deep sequencing of circulating tumor DNA detects molecular residual disease and predicts recurrence in gastric cancer. Cell Death Dis. 2020;11(5):346. 12. Leal A, van Grieken NCT, Palsgrove DN, Phallen J, Medina JE, Hruban C, et al. White blood cell and cell-free DNA analyses for detection of residual disease in gastric cancer. Nat Commun. 2020;11(1):525. 13. Cai Z, Chen G, Zeng Y, Dong X, Li Z, Huang Y, et al. Comprehensive liquid profiling of circulating tumor DNA and protein biomarkers in long-term follow-up patients with hepatocellular carcinoma. Clin Cancer Res. 2019;25(17):5284– 94. 14. Razavi P, Li BT, Brown DN, Jung B, Hubbell E, Shen R, et al. High-intensity sequencing reveals the sources of plasma circulating cell-free DNA variants. Nat Med. 2019;25(12):1928–37. 15. Cristiano S, Leal A, Phallen J, Fiksel J, Adleff V, Bruhm DC, et al. Genome-wide cell-free DNA fragmentation in patients with cancer. Nature. 2019;570(7761):385–9. Annals of Thoracic Medicine ‑ Volume 19, Issue 1, July‑September 2023 71