Lung cancer is one of the most common malignancies in the world. According to the epidemiological statistics, lung cancer has ranked sixth among the top 20 major causes of death in the world, and is the malignant tumor with the highest morbidity and mortality among males (1–3). Only about 25% of all patients with non-small cell lung cancer (NSCLC) survived more than five years. The 5-year survival rate for patients with NSCLC was about 25%. The ratio was decreased to less than 6% for metastatic NSCLC even after cytotoxic chemotherapy (4). Up to now, the prognosis of NSCLC is mainly evaluated based on clinicopathological parameters such as TNM stage, pathological subtype and tumor differentiation degree (5). The 2020 edition of the National Comprehensive Cancer Network Guidelines (NCCN) recommends adjuvant chemotherapy after radical surgery for the patients with stage IB-IIIA NSCLC who have high risk for recurrence (6).

However, a significant proportion of patients with low risk of recurrence based on these prognostic indicators experienced tumor recurrence, while some patients with high risk of recurrence still remained recurrence-free after long-time follow-up. At the same time, the outcome of some patients with high recurrence risk was not improved by adjuvant therapy (7). Clinicians urgently need a more effective method to identify patients at high risk of tumor recurrence and who are likely to benefit from adjuvant therapy.

In recent years, liquid biopsy has been introduced into clinical practice as a valuable tool for early diagnosis of recurrence and metastasis of non-small cell lung cancer (NSCLC) (8–12). Circulating tumor DNA (ctDNA) mainly comes from the small fragments of DNA released by apoptotic or necrotic tumor cells in tumor tissues into the blood system and the DNA released by the lysis of circulating tumor cells or micrometastases in the blood (13). Plasma ctDNA analysis can increase the chance of identifying targeted mutations, and is considered as an alternative to tissue biopsy. It has been widely used for early diagnosis, prognostic stratification, disease monitoring and treatment response assessment in different cancer types (14–16). Liquid biopsy techniques based on plasma ctDNA analysis can monitor and track personalized disease-related markers at the molecular level, and can be used to detect molecular residual desease (MRD) after treatment (10, 17, 18). In the TRACERx study, Addosh et al (19) found that ctDNA test after surgery could detect tumor recurrence or metastasis earlier than imaging. Recent trials using personalized mutation detection or CAPP-seq have also demonstrated the potential of ctDNA for MRD detection and early recurrence prediction, especially in post-surgical patients awaiting adjuvant therapy (20, 21). These studies were helpful to evaluate the prognostic value of post-operative ctDNA status in NSCLC. In addition, previous studies mainly focused on detecting MRD after the completion of adjuvant therapy. Few reports have discussed about whether the MRD detection based on the definition of post-operative ctDNA status could guide post-operative adjuvant therapy for NSCLC patients. Also, the utility of dynamic changes in serial ctDNA to predict the risk of relapse during disease surveillance has not been well characterized.

This study collected tumor pathological diagnosis information, post-operative plasma ctDNA detection results, post-operative adjuvant therapy and recurrence status of 73 patients with NSCLC who received radical surgery. The role of MRD detection based on post-operative plasma ctDNA in the prediction of recurrence and metastasis was retrospectively analyzed. It is expected that this study could provide some guidance on personalized treatment according to the molecular characteristics of NSCLC patients.

There are correlations between plasma ctDNA and tumor burden, metabolism, proliferative indices, lymphovascular invasion, and the risk of disease recurrence/progression in patients with non-small cell lung cancer (NSCLC) (25). It is possible to use ctDNA profiling to quantify MRD, a disease state associated with medically occult residual disease that can be detected using ctDNA, and inferior disease-free survival (DFS) (26). Plasma ctDNA monitoring has become a new method for prognosis prediction and post-operative monitoring of NSCLC. At present, various methods have been developed for ctDNA detection in solid tumors (3, 21, 27). However, it is difficult to determine the proportion of ctDNA by direct cfDNA quantification. Indirect quantification is based on the sequencing of cfDNA and the abundance of tumor-specific mutations carried in cfDNA (28). One study of 230 lung cancer patients found a positive correlation between postoperative ctDNA and late recurrence rates (27). Based on plasma samples, researchers determined the correlation between ctDNA status and postoperative RFS of patients using ctDNA quantified by highest mutation allele fraction (MAF), ctDNA status has a greater effect on RFS than any one or combination of clinicopathologic risk factors. In this prospective cohort study, we examined 73 NSCLC patients treated with surgical resection with adjuvant therapy to evaluate the utility of ctDNA in disease monitoring and treatment determination.

Several previous studies have shown that positive plasma ctDNA after surgical resection is associated with poor prognosis in lung cancer patients (20, 25, 29, 30). ctDNA serves as a robust biomarker for postsurgical and post-adjuvant chemotherapy risk stratification and early detection of recurrence in NSCLC (31). According to Wu, the prognostic value of MRD detection in patients with NSCLC after definitive surgery was confirmed, especially in those with longitudinal undetectable MRD (32), subgroup analysis suggested that adjuvant therapy may be unnecessary for patients with undetectable MRD. During surgery, ctDNA analysis can detect recurrence and MRD risk stratification of NSCLC. Adjuvant treatment is an independent factor in recurrence-free survival in MRD-positive population (33), Investigators confirmed MRD is a strong predictor for disease relapse, and whether receiving adjuvant therapies is an independent factor for RFS in the MRD-positive population. In this retrospective study, we examined 73 NSCLC patients undergoing radical surgical resection to assess the utility of ctDNA in disease surveillance and treatment determination. In our study, there were several situations: 4 patients who were ctDNA-negative 1-3 days after surgery, but their lessions recurred during subsequent monitoring. Among these 4 patients, one died of acute renal failure in 1 year after surgery, and 2 patients had ctDNA detected in blood tests after recurrence. One patient with brain metastasis did not detect ctDNA at any time point before and after recurrence, or even during the entire post-operative monitoring period. That could be explained by the fact that the blood brain barrier may deter ctDNA from brain lesions releasing into the circulation (31). Another reason for the negative ctDNA test may be the heterogeneity of the primary tumor, as the mutated gene sequence of the recurrent tumor may not be included in the sequencing region of the tumor biopsy and the range of test sequences used in this study.

The majority of patients in our study cohort were in stage I (58.9%). In contrast, most of the previous studies mainly enrolled patients with stage IB and the above stages (26, 34). Our cohort may more closely reflect the real-world distribution of NSCLC patients undergoing radical surgical treatment (32, 35). Currently, tumor recurrence is monitored mainly by imaging techniques. However, some studies have demonstrated the value of longitudinal ctDNA analysis to predict recurrence earlier than imaing (36–38). Our results suggest that post-operative ctDNA-positive is a strong predictor of NSCLC recurrence, and patients with post-operative positive MRD are more likely to relapse than patients with negative MRD. Multivariate Cox analysis also showed that MRD status was an independent factor of RFS. These results suggest that post-operative MRD status can be a powerful indicator of patient prognosis stratification.

In this study, most patients in stage IB and II-III received ADT, while most patients in stage IA did not. Therefore, it is acceptable that adjuvant therapy is associated with shorter RFS across the entire cohort (average ADT PFS 27.3 months vs. average non-ADT PFS 55.7 months). However, in subgroup analysis, two subgroups showed opposite trends: for MRD-positive patients, ADT is favorable to longer RFS, while for MRD-negative patients who received ADT had shorter RFS. The results held true when the analysis was further limited to a subgroup of stage IB-III patients. These results clearly indicate that MRD testing can be a powerful tool in screening patients for ADT.

To sum up, the key findings of this study includes: (1) ctDNA can be used as a biomarker for the prognosis of NSCLC after radical surgery; (2) Early prediction of tumor recurrence can be made by ctDNA-based MRD detection after NSCLC surgery; (3) MRD positive patients will benefit from adjuvant therapy, while adjuvant therapy is not beneficial or even harmful to MRD-negative patients. These findings suggested that ctDNA quantification is a more sensitive monitoring technique than traditional imaging.

There are still some limitations to our study. First, our study is limited by a modest sample size and an absence of validation cohorts. However, the results of our study are consistent with those of others that have shown the potential clinical use of ctDNA analysis as a prognostic predictor. Future studies are still needed to further validate whether ctDNA detection can be adopted in routine clinical care. Secondly, the adoption of negative results needs further verification. Multiple factors may lead to false negative results of ctDNA. Thirdly, Some patients failed to complete the blood test in time for some reasons, resulting in missing genetic test results before and after recurrence.

In conclusion, in this retrospective study, we demonstrate that post-operative ctDNA testing can effectively detect MRD and identify NSCLC patients at high risk of recurrence. Adjuvant therapy based on MRD grouping can achieve greater survival benefit for patients with NSCLC. Our findings reveal the potential clinical use of postoperative ctDNA-based MRD testing in NSCLC patients.

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.

CL and XZ conceived the study. CL and XZ provided project management and supervision. XZ, YZ and SZ provided or facilitated the accrual of patient samples, pathology, and/or clinical data. PY and SW performed bioinformatics and genomic analyses. XZ performed statistical analyses. XZ wrote the original draft, with input from all authors. All authors contributed to the article and approved the submitted version.