Dermatomyositis (DM) is a major clinical subset of autoimmune myositis (AIM). The characteristic DM rash (Gottron papules, heliotrope rash) and perifascicular atrophy at skeletal muscle biopsy are regarded as specific features for this diagnosis. However, new concepts are challenging the current definition of DM. A modified Bohan and Peter classification of AIM was proposed in which the core concept was the inclusion of the diagnostic significance of overlap connective tissue disease features. In this clinical classification, a DM rash in association with myositis in the absence of overlap features indicates a diagnosis of pure DM. However, overlap features in association with myositis allow a diagnosis of overlap myositis (OM), irrespective of the presence or absence of the DM rash. Perifascicular atrophy may be present in both pure DM and OM. Recently, the presence of perifascicular atrophy in myositis without a DM rash was proposed as diagnostic of a novel entity, adermatopathic DM. We conducted the present study to evaluate these new concepts to further differentiate pure DM from OM.Using the modified Bohan and Peter classification, we performed a follow-up study of a longitudinal cohort of 100 consecutive adult French Canadian patients with AIM, including 44 patients with a DM phenotype, defined as a DM rash, and/or DM-type calcinosis, and/or the presence of perifascicular atrophy on muscle biopsy. A detailed evaluation was performed for overlap features, the extent and natural history of the DM rash, adermatopathic DM, DM-specific and overlap autoantibodies by protein A immunoprecipitation on coded serum samples, and associations with cancer and survival.Two distinct subsets were identified in patients with a DM phenotype: pure DM (n = 24) and OM with DM features, or OMDM (n = 20). In pure DM, the DM rash was a dominant finding. It was the first disease manifestation, was always present at the time of myositis diagnosis, and was associated with a high cutaneous score and chronicity. Concurrent heliotrope rash and Gottron papules (positive predictive value [PPV] 91%), as well as the V-sign and/or shawl sign (PPV 100%), were diagnostic of pure DM. Anti-Mi-2, anti-MJ, and anti-p155 autoantibodies were present in 50% of pure DM patients and were restricted to this subset (PPV 100%). Cancer was present in 21% of pure DM patients. The 15-year survival was excellent (92%).In contrast, in patients with OMDM, the first manifestation was proximal muscle weakness or other skeletal muscle-related complaints. The DM rash appeared at diagnosis or at follow-up, was associated with a low cutaneous extent score and was transient. Adermatopathic DM, which was absent in pure DM, was highly predictive (PPV 100%) of OMDM. Overlap autoantibodies (including anti-Jo-1, anti-PL-7, anti-PM-Scl, anti-U1RNP, and/or anti-U5-RNP) were found in 70% of OMDM patients. OMDM was not associated with cancer, but the 15-year survival was significantly decreased (65%).Perifascicular atrophy occurred as commonly in OMDM (n = 6/20, 30%) as in pure DM (n = 4/24, 17%) patients. These 6 OMDM patients had adermatopathic DM at myositis diagnosis, and only 1 of them developed a DM rash at follow-up, emphasizing the lack of specificity of perifascicular atrophy for pure DM.In conclusion, using the modified Bohan and Peter classification of AIM allowed identification of OMDM, a new clinical subset of OM. Furthermore, identification of OMDM allowed recognition of pure DM as a new entity that was distinct from OMDM or from OM without DM features. However, the absolute specificity of a DM rash and perifascicular muscle atrophy for the diagnosis of pure DM was lost. The distinctive clinical manifestations and autoantibody profiles presented are proposed as diagnostic criteria to differentiate pure DM from OMDM.