The majority of the world hazelnut crop is roasted, thus developing a unique aroma that depends on the cultivar used and on the roasting conditions applied. Although several studies have investigated the volatile fraction of different cultivars and have correlated the data with overall sensory profiles, studies establishing a correlation between key odorants among the bulk of odorless volatiles and the respective aroma profiles are not yet available. On the basis of recently published stable isotope dilution assays (SIDAs) using comprehensive two-dimensional gas chromatography-time-of-flight mass spectrometry (GC×GC-TOF-MS), differences in concentrations of key odorants in different hazelnut cultivars roasted under defined conditions were monitored and compared with sensory data obtained by projective mapping, aroma profile analysis, and triangle tests. The results showed that the aroma-active compounds 2-acetyl-1-pyrroline, 2-propionyl-1-pyrroline, 5-methyl-(E)-2-hepten-4-one, 2,3-diethyl-5-methylpyrazine, 3,5-dimethyl-2-ethylpyrazine, and 2-furfurylthiol are appropriate marker odorants to differentiate the various nut aromas. In particular, the appreciated roasty, nutty aroma of optimally roasted hazelnuts was developed if both 5-methyl-(E)-2-hepten-4-one and 3-methyl-4-heptanone were >450 μg/kg, whereas the sum of the two 2-acyl-1-pyrrolines and two pyrazines should not exceed 400 μg/kg to avoid an over-roasted smell. Such a desired aroma can be obtained for each cultivar, but obviously specific roasting times, temperatures, and roasting techniques had to be applied.