Real- world analysis of the prognostic value of EGFR mutation detection in plasma ctDNA from patients with advanced non- small cell lung cancer

Background: The plasma sample has emerged as a promising surrogate sample for EGFR mutation detection in advanced non- small cell lung cancer (NSCLC). In clinical practice, whether EGFR variants in baseline plasma ctDNA of advanced NSCLC can predict prognosis in addition to guiding targeted therapy remains to be further explored. Material and Methods: In total, 315 NSCLC patients were retrospectively enrolled. EGFR mutation data from tissue detected by ARMS- PCR and paired plasma samples within 1 month of admission detected by SuperARMS


| INTRODUCTION
Lung cancer has been the worst global burden of cancer according to the statistics from World Health Organization's International Agency for Research on Cancer, with a high incidence and high mortality, among which non-small cell lung cancer (NSCLC) accounts for about 85% of all cases. 1,2 In recent years, with the development and clinical application of small molecule targeted drugs, the five-year survival rate has been dramatically improved in NSCLC panties. 3,4 Epidermal growth factor receptor (EGFR) is one of the key target, which mutations drive about 50% of NSCLC in Asia. 5 Advanced NSCLC patients with EGFR mutations occurring in exons 18-21 can dramatically benefit from EGFR tyrosine kinase inhibitors (TKIs) out of other drug therapies. 6,7 Since it is often difficult to obtain adequate tissue samples for EGFR testing in patients with advanced lung cancer, liquid biopsy has been the promising surrogate sample. 8,9 Liquid biopsy is an analytical diagnosis of diseases such as cancer by examining blood or other body fluids such as urine, saliva, pleural fluid, and cerebrospinal fluid. 10 Plasma circulating tumor DNA (ctDNA) refers to tumor-derived DNA fragments consisting of plasma-free DNA (cfDNA). 11 Multiple clinical studies have confirmed that the accuracy rate of plasma ctDNA EGFR mutation detection at baseline before treatment can reach 70%, and the guiding role of targeted drugs was consistent with tissue results. [12][13][14] Current studies have also shown that ctDNA is a noninvasive real-time biomarker that can provide prognostic and predictive information for monitoring treatment. 15,16 The prognostic value of ctDNA has been demonstrated in recent years in detecting minimal residual disease after surgery or treatment. [17][18][19] In addition, Plasma ctDNA has also been an indicator for tracking molecular residual disease (MRD) after radical treatment in patients with locally advanced lung cancer. [20][21][22][23] All of these studies confirmed the prognostic value of plasma ctDNA-based assessment of MRD, i.e., no mutation was detected in plasms ctDNA within 1 month after curative treatment, and patients were more likely to have a longer DFS. 17,20,21,24 However, few studies have demonstrated that plasma ctDNA mutation status can predict the prognosis of advanced lung cancer. In a pan-cancer analysis of immune checkpoint blockade, baseline ctDNA levels appeared prognostic value. 25 In addition, a prospective NSCLC study showed that both ctDNA detected at baseline diagnosis and residual ctDNA identified at the first post-treatment assessment were associated with poor prognosis. 15 From the perspective of dynamic monitoring, ctDNA clearance is a marker of good prognosis. Patients with at least one ctDNA clearance have significantly longer OS and PFS than those with no ctDNA clearance. 26 However, with different detection methods, ctDNAdetected results in therapeutic monitoring may also be inconsistent. Currently, among the various blood EGFR mutation detection techniques, SuperARMS technology is a highly sensitive and specific method suitable for non-invasive detection of advanced NSCLC patients with plasma EGFR mutation detection, the detection sensitivity of EGFR mutation detection in plasma samples can reach 0.2%, which is can effectively predict the efficacy of EGFR-TKIs and guide clinical medication. [27][28][29] In our study, we collected paired tissue and plasma samples to explore the consistency of plasma cfDNA EGFR variants. And further explored the prognostic value of ctDNA monitoring in advanced NSCLC therapies. In the present study, neither the molecular detection method including ARMS-PCR or SuperARMS-PCR nor the follow-up treatment regimens have been intervened. The assay performance of the reagents thus evaluated may be more suggestive for clinical practice. The factors that could affect the detection rate of ctDNA EGFR mutation were deeply analyzed. The data from this study showed that the SuperARMS can be used to detect plasma samples from NSCLC patients in early stages, whereas previous data reported in the literature almost all focus on samples from patients in advanced stages. 28,34 Second, in the evaluation of the correlation between baseline ctDNA mutation status and patient prognosis, patients treated with targeted therapy, and chemotherapy were included in the analysis. This also distinguished itself from previous reports in which targeted therapies have been the mainstay of research. 15,43 Conclusions: A plasma sample could be an alternative for a molecular test when tissue samples was unavailable. The SuperARMS-PCR detection method has high sensitivity in real-world clinical practice. Furthermore, in patients with stage IIIB-IV, baseline plasma ctDNA EGFR mutation positivity not only guides targeted therapy but also predicts a worse prognosis.

K E Y W O R D S
EGFR mutation, NSCLC, plasma samples, prognosis, superARMS-PCR 2 | MATERIALS AND METHODS

| Patients and study design
From July 2016 to April 2021, 623 patients with both tissue and plasma EGFR mutation detection records were retrospectively screened in the Department of Pathology, Beijing Chest Hospital. Among them, the exclusion criteria were set as follows: (1) Patients who cannot be pathologically diagnosed with non-small cell lung cancer, (2) The interval detection time between tissue and plasma exceeded 1 month. Therefore, 315 pathologically diagnosed NSCLC patients were finally enrolled in this study. The patients' clinicopathological information, including age, gender, tumor-node-metastasis (TNM) stage, smoking history, EGFR molecular detection results of tissue and plasma samples, and treatment history were investigated and collected in the clinicopathological system and pathological information system of our hospital ( Figure 1). Of these 315 patients, ARMS-PCR was used for plasma samples from June 2016 to April 2018 and SuperARMS-PCR was used for plasma samples after April 2018 along with its approval by NMPA (National Medical Products Administration). Patients' survival follow-up data were analyzed using July 2021, data cut-off.

mutation detection
The EGFR mutation status of plasma samples was detected with ARMS or Super-ARMS EGFR Mutation Detection Kit (Amoy Diagnostics) according to the manufacturer's instructions. EGFR mutation status of the tumor tissues was detected using the ARMS EGFR Mutations Detection Kit (Amoy Diagnostics) according to the manufacturer's instructions.

| Statistics on progressionfree survival
Among the 315 enrolled patients, a total of 268 patients received chemotherapy or targeted therapy after plasma testing, and only 227 of 268 patients had relatively complete treatment follow-up information, among which 139 treatment naïve patients and 82 relapsed patients were in stage IIIB-IV. Patients' treatment efficacy data were collected and progression-free survival (PFS) was calculated. PFS differences were compared between patients with positive and negative ctDNA EGFR mutation in plasma before treatment, regardless of subsequent treatment regimens.
F I G U R E 1 Flowchart of the study. *Tissue or plasma samples have been detected more than one EGFR mutation, and only one of them was detected in both kinds of samples. **150 patients received first-or second-generation EGFR-TKIs and 77 patients received chemotherapy.

Patients with tissue and plasma samples N=623
Patients with tissue and plasma samples N=620

| Statistical analysis
The data were analyzed using SPSS software version 20 (IBM). The relationship between the two groups was evaluated using a standard chi-squared test, and p < 0.05 was considered statistically significant. The Kaplan-Meier was used for PFS analysis, and survival curves were compared using the log-rank test.  11.5%), G719X (8/156, 5.1%), 20-Ins (5/156, 3.2%), and L861Q (4/156, 2.6%). The EGFR mutation profile in detail is shown in Figure S1.
The clinicopathological characteristics affecting the consistency between tissue DNA EGFR detection and plasma ctDNA EGFR detection by SuperARMS were analyzed. The results showed that there were statistically significant differences in treatment status and pathological types. Eight of the 28 non-adenocarcinoma patients detected by SuperARMS had EGFR mutation, seven of which were positive both in tissue and plasma, one of which was only positive in tissue. The concordance in non-adenocarcinoma patients was obviously higher than that in adenocarcinoma (p = 0.02) ( Table S1). The  concordance in treatment naïve patients was significantly higher than that in relapse patients (p = 0.047) (Table S2,  Table 3). No statistical difference was observed in age, sex, smoking status, and tumor specimen stage (Table 3).

| Correlation between baseline plasma ctDNA EGFR mutation status and prognosis in advanced NSCLC patients
In 139-treatment naïve IIIB-IV patients, patients with baseline ctDNA EGFR-negative had a significantly longer PFS than those of EGFR-positive ( In 82 IIIB-IV patients who relapsed after treatment, there was no significant difference in PFS between baseline ctDNA EGFR-negative and EGFR-positive patients regardless of treatment regimen (11.0 months vs. 9.0 months, p = 0.1) ( Figure 2D). Fifty-eight of 82 patients with tissue EGFR-positive received first-or second-generation EGFR-TKIs treatment and 24 of 82 patients with tissue EGFRnegative received chemotherapy. The mPFS of ctDNA EGFR-negative and EGFR-positive patients in the targeted therapy group was 14.0 months vs. 11.0 months (p = 0.1), and that was 6.5 months vs. 6.5 months (p = 0.5) in the chemotherapy group ( Figure 2E,F).

| DISCUSSION
EGFR-TKI has been widely used in the treatment of advanced NSCLC, but the precondition of EGFR-TKI application is whether EGFR mutation has been determined. Genetic testing based on tumor tissue is considered the gold standard method, However, it was reported that only 18% of advanced NSCLC patients have sufficient tumor specimens for full tissue genotyping of all eight guidelinerecommended genomic biomarkers, thus new alternative tissue samples are needed for genetic testing in NSCLC

T A B L E 3
Clinicopathological characteristics affecting the consistency between tissue and plasma in clinical practice. 30,31 Based on real-world retrospective data, this study analyzed the concordance rate of EGFR mutation detection in paired tissue samples and blood samples. Based on the highly sensitive SuperARMS-PCR method to detect plasma ctDNA EGFR mutation, the concordance rate with tissue detection was as high as 79.5%. 32 In this study, we further analyzed the correlation between baseline plasma ctDNA EGFR mutation status and prognosis in advanced patients. Stratified analysis of the treatment status of patients, compared to the intention-to-treat population, treatment naïve patients with baseline plasma EGFR-negative would have better PFS which was slightly higher than previously reported. 33 A trend for longer PFS in baseline ctDNA EGFR-negative patients was also found in relapsed patients. So precise stratification of patients' treatment status can better benefit clinical patients. And since these data were obtained from real clinical practice, it has greater significance for clinical diagnosis and treatment guidance. Two plasma ctDNA EGFR detection methods were included in this study, ARMS-PCR and SuperARMS-PCR, because the SuperARMS-PCR method was available after the year 2018. 34 SuperARMS-PCR is more sensitive for mutation detection in cfDNA due to its modified primers and DNA polymerase. 35,36 In the present study, based on the results of ARMS-PCR detection, the sensitivity of plasma ctDNA EGFR mutation detection was only 59.7%, which was lower than those detection results of SuperARMS-PCR (68.4%) as well as the published data. 37 The superiority of SuperARMS-PCR in detecting ctDNA EGFR mutation was more prominent in plasma samples of I-IIIA patients (the detection sensitivity of SuperARMS-PCR was 33.3%, while the detection sensitivity of ARMS-PCR was 0%). This further illustrated the advantages of SuperARMS in the detection of plasma samples with low abundance. 38 Advanced patients who have progressed on therapy, especially those with EGFR-TKIs therapy, often carry the T790M resistance mutation, which was more frequently detected in liquid biopsies than in tissues. 39,40 In the present study, the sensitivity and specificity of ctDNA mutation detection rate by SuperARMS in EGFR TKIs relapsed patients were 76.2% and 60.0%, which were 67.1% and 100.0% in treatment naïve patients. The Relapsed patients had higher sensitivity but worse specificity for mutation detection in ctDNA compared to newly diagnosed patients, resulting in a lower concordance rate than the latter due to novel mutation sites in relapsed patients. Two relapsed patients with T790M co-mutation was detected in plasma but not in tissue. [39][40][41] The specificity for relapsed patients in this study in the SuperARMS-PCR detection group was 75%, also due to the fact that T790M was only detected in plasma samples in two patients. The detection sensitivity of SuperARMS-PCR ctDNA EGFR in relapsed patients was lower than that of ARMS-PCR, but there was no statistical difference, which may be caused by the low number of samples between the two groups. To further analyze the factors affecting the detection rate of EGFR mutation in SuperARMS-PCR ctDNA, the patient's pathological subtype was an influencing factor. In this study, the consistent rate of ctDNA EGFR detected by SuperARMS-PCR in patients with non-adenocarcinoma was higher than that in patients with adenocarcinoma. This phenomenon has also been mentioned in other studies. 20,42 In patients with advanced lung cancer, what is the prognostic value of baseline blood ctDNA mutation status in addition to guiding targeted drugs? Patients from the FLAURA study with undetectable plasma EGFR mutations at baseline had better EGFR-TKI efficacy than patients with detectable plasma EGFR mutations at baseline. 13 The first is the EURTAC trial, the study mentioned that compared with patients with L858R mutation both in tissue and plasma, median OS was longer in those the mutation was detected in a tissue but not in plasma (median OS, 27.7 vs 13.7 months). 43 In the AURA3 study, PFS was prolonged in patients with tissue T790M-positive and plasma T790M-negative results compared with patients with plasma T790Mpositive results both in osimertinib (median, 12.5 vs 8.3 months) and platinum-pemetrexed groups (median, 5.6 vs 4.2 months). 44 Gu et al. enrolled 117 acquired TKI resistance advanced NSCLC patients. It was found that patients with T790M detected in plasma had poor OS compared to that T790M un-detected (median OS, 26.9 months vs. NA). 45 Elizabeth Fabre et al. also found that baseline ctDNA positivity was associated with reduced OS (median, 13.6 vs.21.5 months) and poor PFS (median, 4.9 vs 10.4 months), those detected by NGS. 15 These clinical studies confirmed that both treatments naïve and relapsed patients had a worse prognosis if their baseline ctDNA EGFR mutation was detected.
In our study, the median duration of follow-up was 15.0 months (IQR 0.33-140.0). Since the patients in this study were treated with chemotherapy or targeted monotherapy, and the first-generation or second-generation EGFR TKIs was used in targeted therapy, the median follow-up time of 15 months could track more than 50% of patients' relapse. 46 For treatment-naïve patients, regardless of whether they received targeted therapy (mPFS 7.0 months vs.14.0 months, p = 0.004) or chemotherapy (mPFS 4.0 months vs. 7.0 months, p = 0.14), patients with baseline ctDNA EGFR mutation had worse PFS. No difference (p = 0.1) in the PFS was observed in the chemotherapy group because of the few number of ctDNA EGFR-positive patients (N = 4). Among the relapsed patients who received targeted therapy, patients with plasma ctDNA EGFR-positive before treatment also had worse PFS compared with patients with plasma ctDNA EGFRnegative (11.0 months vs. 14.0 months, p = 0.1). The patients with detectable plasma ctDNA mutation at baseline maybe have a high tumor load, caused by spatial heterogeneity of the tumor, leading to a poor clinical outcome. 47 There were several limitations in this study. First, due to the limited sample size, a more detailed analysis of the detection rate of EGFR mutation sites has no statistical value, so a stratified analysis of EGFR classic mutation sites and rare mutation sites was not performed. Second, when analyzing the prognostic value of baseline ctDNA mutation status, PFS was used as the study endpoint rather than OS due to the limited follow-up time. The survival status of these patients will be followed to obtain OS data. Finally, more patients receiving chemotherapy and immunotherapy will be included in follow-up studies to observe the correlation between baseline ctDNA EGFR mutation status and patient prognosis.

| CONCLUSIONS
This study confirmed that plasma samples can be used as a surrogate for molecular testing when tissue samples were not available in real-world studies, especially in patients with advanced NSCLC. Compared with the ARMS-PCR detection method, the SuperARMS-PCR detection method had higher sensitivity for the detection of ctDNA EGFR mutation in plasma. And in patients with stage IIIB-IV, baseline plasma ctDNA EGFR mutation positivity not only guides targeted therapy but also predict a worse prognosis.