Co‐mutational assessment of circulating tumour DNA (ctDNA) during osimertinib treatment for T790M mutant lung cancer

Abstract Osimertinib is designed to target the secondary resistant EGFR T790M mutant and has shown outstanding clinical efficacy. However, the prognostic prediction of osimertinib patients is a big problem in clinical practice. The resistance mechanism of osimertinib is also not fully understood. NGS and a 1021 gene capture panel were used to analyse the somatic mutation profile of thirty‐six lung adenocarcinoma patients' serial ctDNA samples. Progression‐free survival of subgroup patients is analysed. Patients harbour TP53 mutations and patients with higher TMB value in pre‐treatment samples showed a shorter PFS. Moreover, compared to CT evaluation, ctDNA changes generally correlated with treatment responses in most patients. Novel resistance mechanisms are discovered including EGFR mutations and alternative activation pathway. Our results implied a broad potential of ctDNA as an adjuvant tool in practical clinical management of NSCLC patients. ctDNA could help with clinical practice during osimertinib treatment, regarding monitoring tumour response, detecting development of heterogeneity, identifying potential resistance mechanisms, predicting treatment efficacy and patient outcome.

In this study, we enrolled thirty-six EGFR-mutant advanced lung adenocarcinoma patients, who were confirmed T790M-positive based on tumour tissue genotyping assay, and investigated the clinical implications of serial analysis using a 1021 gene capture panel of ctDNA during osimertinib treatment. Our results implied a broad potential of ctDNA as an adjuvant tool in practical clinical management.

| Study population
A total of 36 metastatic lung adenocarcinoma patients with acquired EGFR T790M mutations were enrolled from National Cancer Center/ Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College. The study was reviewed and approved by Ethics Committee of this institution. All patients provided written contents according to ethical regulations. For each patient, a volume of 10 mL blood was collected before osimertinib treatment, at individual response assessment time-point and after disease progression. Response assessments were performed every 3 months during treatment. Progression-free survival (PFS) was measured from the beginning of osimertinib treatment to progression confirmed by computer tomography (CT) scan according to RECIST 1.1,17 or treatment cessation, or death.

| Targeted NGS and data processing
cfDNA extraction and library construction were performed according to previous publication. 18 After the removal of terminal adaptor sequences and low-quality data, reads were mapped to the reference human genome (hg19) and aligned using BWA (0.7.12-r1039). MuTect2 (3.4-46-gbc02625) was employed to call somatic small insertions and deletions (InDels) and single nucleotide variants (SNVs). Contra (2.0.8) was used to detect copy number variations. All final candidate variants were manually verified with the integrative genomics viewer browser.

| Pyclone and TMB analysis
PyClone was used to analyse the clonal population structure of ctDNA collected serially from each patient. 20 The copy number information of each single nucleotide variation (SNV) was used as input. Variants located in the cluster with greatest mean cancer cell fraction (CCF) were defined as clonal and the rest were subclonal.
TMB of blood (bTMB) was analysed using SNVs (non-synonymous only) and indels at allele frequencies of ≥0.5%. The cut-off value for bTMB-H and bTMB-L is defined as 9 mutations/Mb, approximately the median value in our assay experience.
To better characterize the co-occurring genetic prognostic features, the patients were divided into two subgroups according to a We then explored the PFS in patient subgroups divided by EGFR L858R/19del genotype, T790M, TP53 status and additional drivers. The patients with 19del, T790M-positive, T790M-subclonal and the subgroup negative for additional driver gene mutations a shorter median PFS than the other subgroups ( Figure S2), but the differences were not statistically significant (all P > .05).
However, patients with suppressor gene mutations or TP53 mutations alone had significantly worse outcomes than those without these mutations (P = .04; Figure 2A Figure 2F). These results suggest that TP53 and the bTMB in pre-treatment plasma samples were two potentially clinically useful prognostic features in osimertinib treatment patients.

| Therapeutic response monitoring using serial ctDNA
Twelve patients experienced disease progression, 3 of whom had PR and 9 had stable disease (SD) at the first response assessment.
Eleven of these 12 patients had pre-treatment ctDNA detectable at baseline. In order to analyse the correlation between the change of ctDNA level and the change of tumour size in the patients, we defined the difference between the frequency of the ctDNA gene ctDNA failed to reflect disease progression in a non-target lesion in one patient, which may due to the small tumour size ( Figure S3) and subsequent low ctDNA infiltration rate. All acquired mutations at PD are presented in Figure 5. The three patients who had PFS < 3 months had more acquired mutations than others ( Figure S4).

| Resistance mechanism in twelve progressed patients indicated by paired ctDNA
All EGFR family mutations found in the 12 progressed patients are annotated in schematic models in Figure 6. Six of these EGFR muta-

| D ISCUSS I ON
Compared to tissue biopsy, liquid biopsy, especially ctDNA sequencing, allows comprehensive and longitudinal monitoring of tumour F I G U R E 5 Acquired mutations during treatment in eleven patients who have both positive pre-treatment and positive after treatment ctDNA results. P036 is not shown due to the lack of after treatment ctDNA data. Alteration types are represented by the same colours as listed in Figure 1.

| CON CLUS IONS
Our study explored the potential applications of serial ctDNA analysis in the management of advanced lung adenocarcinoma patients.
ctDNA could help with clinical practice during osimertinib treatment, regarding monitoring tumour response, detecting development of heterogeneity, identifying potential resistance mechanisms, predicting treatment efficacy and patient outcome.

ACK N OWLED G EM ENTS
This research was primary funded by grants from CAMS Innovation

CO N FLI C T O F I NTE R E S T
Sha Wang, Lianpeng Chang, Yanfang Guan and Xin Yi are current employees of Geneplus-Beijing. All other authors declare no competing interest.

AUTH O R S ' CO NTR I B UTI O N S
YKS contributed to study design and conception, data analysis and interpretation. All authors were involved in data acquisition. All authors had full access to all the date in the study and contributed to the writing of the report, reviewed it for intellectual content and approved the submitted version.

DATA AVA I L A B I L I T Y S TAT E M E N T
Some information was provided in the supplementary files, and all other relevant data could be obtained from the corresponding authors of this study.