Development and validation of a methylation-specific droplet digital PCR multiplex for lung cancer detection

Abstract

Biomarkers are increasingly used in cancer management, including lung cancer. The use of circulating tumour DNA (ctDNA) detection has attracted significant interest as a non-invasive, highly specific, and sensitive strategy. In this study, we developed and validated a methylation-specific droplet digital PCR (ddPCR) multiplex assay with five tumour-specific methylation markers identified by in silico analysis for lung cancer detection across various clinical settings. The performance of the ddPCR multiplex was validated in tissue and plasma in cohorts of healthy controls and patients with both non-metastatic and metastatic disease by examining sensitivity, specificity, and marker dynamics. Furthermore, two different cut-off methods to determine ctDNA-status and their effects on sensitivity and specificity were examined. In non-metastatic disease, the ddPCR multiplex showed ctDNA-positive rates of 38.7% and 46.8%, respectively, with the two cut-off methods. In metastatic cases, these rates increased to 70.2% and 83.0%. Higher sensitivities were observed for small cell lung cancer and squamous cell carcinoma. Analyses of longitudinal samples from patients with metastatic disease undergoing treatment indicated a potential use in prognostication and treatment guidance. We present a robust, cost-effective approach to lung cancer detection. Moving forward, its performance should be evaluated and validated in larger patient cohorts across a range of clinical scenarios, exploring its full potential.

Keywords: Lung cancer; Multiplex; ctDNA; ddPCR.

Fig. 1

Linearity of 6-plex ddPCR. Linear correlation between expected and measured methylation ratios for each assay in a six-point two-fold dilution of Universal Methylated Human DNA Standard (Zymo Research) in a constant background of 5600 copies of bisulfite-converted genomic DNA from whole blood. (a) HOXA9, (b) MCIDAS, (c) OTX1, (d) SP9, and (e) TFAP2B were all tested in duplex with ALB and in the 6-plex ddPCR.

Fig. 2

Methylation levels of the five cancer-specific markers included in the 6-plex ddPCR in normal, benign, and lung tumour tissue. Boxplots show the ratio of detected copies/µl of each marker, (a) HOXA9, (b) MCIDAS, (c) OTX1, (d) SP9, and (e) TFAP2B, and the reference gene (ALB). The statistical significance between each group was defined using Mann-Whitney U tests (ns: p > 0.05, *: p ≤ 0.05, **: p ≤ 0.01, ***: p ≤ 0.001, ****: p ≤ 0.0001).

Fig. 3

Limit of blank in plasma from non-cancer patients (n = 60). Limit of blank were set for (a) HOXA9, (b) MCIDAS, (c) OTX1, (d) SP9, and (e) TFAP2B using two different methods: a set ≥ 95% specificity for each marker (blue dashed line) and with Youden’s J statistics (red dashed line).

Fig. 4

Percentage of ctDNA-positive non-cancer patients (n = 60) and cancer patients with non-metastatic (n = 62) or metastatic disease (n = 47) according to the number of markers above established limits of blanks (LoB) with (a) a set of ≥ 95% specificity and (b) Youden’s J statistics.

Fig. 5

Percentage of ctDNA-positive cancer patients with non-metastatic (n = 62) or metastatic disease (n = 47) according to histological subtypes above established limits of blanks (LoBs) with (a) a set of ≥ 95% specificity and (b) Youden’s J statistics. Samples were considered ctDNA positive if they had ≥ 2 markers above LoB. SCC: squamous cell carcinoma, AC: adenocarcinoma, SCLC: small cell lung cancer. Other includes low differentiated carcinomas, carcinoid tumours, large cell neuroendocrine carcinomas, and sarcomatoid carcinomas.

Fig. 6

Kaplan-Meier plots showing survival curves of NSCLC patients undergoing immunotherapy according to ctDNA status. Patients were divided into three groups: Stable, consisting of patients with insignificant changes or where < 5 markers showed a significant decrease. Decrease, consisting of patients with significant decreases in all markers, and Increase, consisting of patients with a significant increase in at least one marker. (a) Progression-free survival and (b) overall survival of patients based on change from baseline to the 2nd treatment cycle. (c) Progression-free survival and (d) overall survival of patients based on changes from baseline to the 4th treatment cycle.