Gamma-ray lines from cosmic sources provide unique isotopic information, since they originate from energy level transitions in the atomic nucleus. Gamma-ray telescopes explored this astronomical window in the past three decades, detecting radioactive isotopes that have been ejected in interstellar space by cosmic nucleosynthesis events and nuclei that have been excited through collisions with energetic particles. Astronomical gamma-ray telescopes feature standard detectors of nuclear physics, but have to be surrounded by effective shields against local instrumental background, and need special detector and/or mask arrangements to collect imaging information. Due to exceptionally-low signal/noise ratios, progress in the field has been slow compared with other wavelengths. Despite the difficulties, this young field of astronomy is well established now, in particular due to advances made by the Compton Gamma-Ray Observatory in the 90ies. The most important achievements so far concern: short-lived radioactivities that have been detected in a couple of supernovae (56Co and 57Co in SN1987A, 44Ti in Cas A), the diffuse glow of long-lived 26Al that has been mapped along the entire plane of the Galaxy, several excited nuclei that have been detected in solar flares, and, last but not least, positron annihilation that has been observed in the inner Galaxy since the 70ies. High-resolution spectroscopy is now being performed: since 2002, ESAs INTEGRAL and NASAs RHESSI, two space-based gamma-ray telescopes with Ge detectors, are in operation. Recent results include: imaging and line shape measurements of e e annihilation emission from the Galactic bulge, which can hardly be accounted for by conventional sources of positrons; 26Al emission and line width measurement from the inner Galaxy and from the Cygnus region, which can constrain the properties of the interstellar medium; and a diffuse 60Fe gamma-ray line emission which appears rather weak, in view of current theoretical predictions. Recent Galactic core-collapse supernovae are studied through 44Ti radioactivity, but, apart from Cas A, no other source has been found; this is a rather surprising result, assuming a canonical Galactic supernova rate of ̃1/50 years. The characteristic signature of 22Na-line emission from a nearby ONeMg novae is expected to be measured during INTEGRALs lifetime.