Highly sensitive genotyping techniques are useful to detect epidermal growth factor receptor (EGFR) mutations on lung cancer cytological samples, when these specimens feature only few neoplastic cells. This study aimed to validate digital PCR (dPCR) methodology on cytological material. In plasmid model system, dPCR allowed for the detection of a minimal percentage (1%) of EGFR mutant alleles. Cytological samples (n=30), with neoplastic cell percentage ranging from 10% to 80% and yielding a quantity of extracted DNA ranging from 1.75 to 60 ng/µL were selected. Results previously generated by fragment length and TaqMan assays (n=8 exon 19 deletions, n=2 L858R mutations and n=20 wild-type DNA) were compared with those obtained by dPCR. Data were highly concordant (96.6%). However, dPCR detected an additional L858R mutation that had been missed by TaqMan assay on a paucicellular smear. This mutation was confirmed by cloning PCR products and sequencing. Thus, dPCR can reliably be used to increase EGFR mutation detection rate on scarcely cellular lung cancer smears.