181 Dynamics of methylated cell-free DNA in the urine of non-small cell lung cancer patients by including all available patient samples (n=138), and 2) random sampling to compute the statistical difference between methylation levels of cases and controls when only one urine sample would have been collected. Linear mixed-effects models were fitted as described above with subject as a random effect to account for repeated measures in the patient group. Likewise, differences in methylation levels between patients and controls were tested as described above. Final models are available in the Supplementary material. Random sampling was conducted according to (33). Briefly, only one urine sample of each patient was randomly selected from the six available samples to compare the methylation levels of each marker between patients and controls, which was repeated 100 times in total. A median p-value was computed to summarize the outcome of 100 rounds of random sampling. Statistical analyses were performed in R (v.3.6.1) and Rstudio (v.1.1.463). For statistical tests, p-values <0.05 were considered statistically significant. R packages Linear mixed models were computed using the ‘lme4’ package (30), combined with the Companion to Applied Regression ‘car’ package (34) for statistical testing. Normalization of the cfDNA concentration data was performed according to the ‘bestNormalize’ package (35). The ‘sjPlot’ package (36) was used to test model assumptions and extract model summaries. Boxplots and conditional scatterplots were computed using the ‘ggplot2’ package (37) and ‘lattice’ package (38), respectively. RESULTS Patient characteristics A total of 23 patients with NSCLC were included, of which relevant clinical and pathological features are presented in Table 1. The median age at diagnosis was 69 (range 65-75) and nine patients were female. The patient cohort covered the major histological subtypes of NSCLC, with TNM stages ranging from IIb to IVb. The majority of patients were current or former smokers. Variation in cfDNA concentration Variation during the day and between days Total cfDNA concentrations of all urine samples (n=138) were quantified by Qubit and compared within and between days by a linear mixed model approach. The cfDNA concentrations measured across the six different time points are shown in Figure 1. No significant differences were found between the morning, afternoon, and evening, 7
RkJQdWJsaXNoZXIy MjY0ODMw