Analysis of circulating cell-free tumor DNA (cftDNA) has emerged as a specific and sensitive blood-based approach to detect epidermal growth factor receptor (EGFR) mutations in non-small cell lung cancer (NSCLC) patients. Still, there is some debate on what should be the preferential clinical method for plasma-derived cftDNA analysis. We tested 31 NSCLC patients treated with anti-EGFR tyrosine kinase inhibitors (TKIs), at baseline and serially during therapy, by comparing three methodologies in detecting EGFR mutations (L858R, exon 19 deletion, and T790M) from plasma: scorpions-amplification refractory mutation system (ARMS) methodology by using EGFR Plasma RGQ PCR Kit-QIAGEN, peptide nucleic acid (PNA) clamp and PANA RealTyper integration by using PNAClamp EGFR-PANAGENE, and digital real time PCR by using QuantStudio 3D Digital PCR System-Thermo Fisher Scientific. Specificity was 100% for all three mutations, independently from the platform used. The sensitivity for L858R (42.86%) and T790M (100%) did not change based on the method, while the sensitivity for Del 19 differed markedly (Scorpion-ARMS 45%, PNAClamp 75%, and Digital PCR 85%). The detection rate was also higher (94.23%) as measured by Digital PCR, and when we monitored the evolution of EGFR mutations over time, it evidenced the extreme inter-patient heterogeneity in terms of levels of circulating mutated copies. In our study, Digital PCR showed the best correlation with tissue biopsy and the highest sensitivity to attain the potential clinical utility of monitoring plasma levels of EGFR mutations.