Work

Postoperative Pleural Fluid Cell-Free DNA Captures Locoregional MRD That Stratifies Risk for Lung Cancer Relapse after Surgery

Year: 2024

Type: article

Abstract: Patients with resectable lung cancer may have the best chance of achieving remission. However, the results of the Lung-ART study suggest clinicians cannot predict which patients will benefit from adjuvant radiotherapy. Our group has previously shown that postoperative lymph drain fluid cell-free DNA (cfDNA) has superior sensitivity to detect locoregional molecular residual disease (MRD) compared to plasma and stratifies recurrence in oropharynx cancer patients. Similarly, we hypothesized circulating tumor DNA (ctDNA) in chest tube postoperative pleural fluid (PPF) will correlate with aggressive pathology and locoregional recurrence. We isolated formalin-fixed paraffin-embedded (FFPE) tumor tissue DNA and 24-hour postoperative pleural fluid (PPF) cfDNA from 43 lung cancer patients (88% NSCLC; 12% SCLC) and 4 t patients diagnosed with benign lung lesions. cfDNA was also isolated from 26 paired lung cancer plasmas. We prepared sequencing FFPE, plasma, and PPF sequencing libraries, using the Enzymatic Methyl-Seq and TruSight Oncology 500 library kits, to quantify ctDNA using tumor-informed mutations and differentially methylated regions (DMRs). Lastly ctDNA was compared to AJCC 8 th edition tumor (T) and nodal (N) stage, aggressive pathology such as visceral pleural invasion (VPI) and airway invasion (AI), and disease progression. After sequencing with TruSight, we censored patients with < 1 variant detected in FFPE tissue, to avoid MRD false negatives. 24 PPFs and 19 plasmas were included in downstream mutational analysis. Among the 19 patients with both analytes, PPF had a higher mean variant allele frequency (mVAF) (P = 0.029). There were no significant differences in PPF mVAF or mutational burden among different T stages, but interestingly, PPF from patients with stage N1-N2 disease had a significantly higher mutational burden (P = 0.007) and trended towards higher mVAF (P = 0.08) than N0 patients. VPI and AI were not associated with mutational ctDNA signal, but patients at least one high risk pathologic feature (N1-N2 stage, VPI, or AI) had significantly higher mutational burden (P = 0.046) and mVAF (P = 0.034) compared to patients with none. Though we were underpowered for Kaplan-Meier analysis, MRD positive PPF detected 100% of the locoregional relapses prior to second line treatment initiation and trended toward predicting locoregional disease-free survival (Hazard Ratio = 4.8, Log Rank P = 0.18). Ongoing cfDNA methylation analysis demonstrates that cancer associated DMRs learned from TCGA methylation microarray data can accurately separate healthy controls from adenocarcinoma and SCC PPF on principal component analysis. Similar to our work in HNC, we sensitively detected pleural fluid ctDNA early after surgery. Our assay strongly aligned with high-risk pathology and PPF MRD sensitively detected all locoregional recurrences. In the future, PPF biomarkers could help guide adjuvant radiotherapy.

Source: International Journal of Radiation Oncology*Biology*Physics

DOI: 10.1016/j.ijrobp.2024.07.1817