Diagnostic tumour samples are mandatory for morphologic and molecular diagnosis of non-small cell lung cancer (NSCLC) to establish the best therapeutic approach. In the presence of small tumour tissue sample, the pathologist needs to make responsible choices to achieve a correct diagnosis and save material for subsequent molecular evaluations. Nevertheless, in some instances, the diagnostic process can lead to tissue depletion. The automated Idylla epidermal growth factor receptor (EGFR) mutation test has been developed to rapidly process formalin-fixed paraffin-embedded (FFPE) pathologic material, without previous DNA extraction. This study aimed to test whether this platform is suitable for the reuse of H&E, immunohistochemistry (IHC) and fluorescence in situ hybridisation (FISH) diagnostic slides. A training set of 19 FFPE tissues with known EGFR status was revaluated on H&E slides. Fourteen of them were also tested using IHC and FISH treated specimens. An additional series of 25 H&E, IHC or FISH slides of NSCLC cases tested for EGFR mutation at an external institution was blindly assessed as a validation cohort. Combining the two sets, 32 of 32 classical ex19dels and p.L858R were correctly identified. Three uncommon mutations (p.G719X, p.L861Q and ex20ins) were also detected. Four discrepancies were related to rare ex19del/ins not included in the Idylla list of detectable mutations. Two p.T790M variants were missed on one FFPE and two H&E slides but were detected using IHC and FISH sections from the same FFPE blocks. The Idylla EGFR mutation test is highly reliable using differently treated tumour specimens and should be validated in larger studies. © Author(s) (or their employer(s)) 2022. No commercial re-use. See rights and permissions. Published by BMJ.