The aim of the study was to control the in vivo localisation of implanted cells in cell-based therapies. Labelling cells with (111)indium-oxine is one of the most interesting methods proposed. We evaluated this method in the setting of autologous osteoblast implantation in nonunion fractures. An in vitro study of osteoblasts was conducted after (111)indium-oxine labelling. Radioactivity retention and viability, proliferation and the ability to produce alkaline phosphatase were evaluated in a seven-day culture. In vivo labelling of implanted osteoblastic cells was conducted during a therapeutic trial of atrophic nonunion fractures, with the leakage outside the nonunion site and local uptake evolution at four, 24 and 48 hour being studied. The mean labelling efficiency for osteoprogenitors was 78.8 ± 4.6 %. The intracellular retention was 89.4 ± 2.1 % at three hours and 67.3 ± 4.7 % at 18 hours. The viability assessed at three hours was 93.7 ± 0.6 %. After seven days of culture, morphology and alkaline phosphatase staining were similar for both labelled and unlabelled control cells, although the proliferation rate was decreased in the labelled cells. Some local intraosseous leakage was observed in four of 17 cases. All patients showed uptake at the injection site, with four having no other uptake. Four patients showed additional uptake in the bladder, liver and spleen, while 11 patients had additional uptake in the lungs in addition to the bladder, liver and spleen. The activity ratios (injection site/body) were 48 ± 28 % at four hours, 40 ± 25 % at 24 hours and 35 ± 25 % at 48 hours. After correcting for decay, the activity within the injection site was 82 ± 15 % at 24 hours and 69 ± 11 % at 48 hours compared with the activity measured at four hours. No relationship was found between uptake and radiological bone repair. The (111)indium-oxine labelling appears to be a good method for monitoring the behaviour of the osteoblastic cells after their implantation in atrophic nonunion fractures.