A modern approach to the surgical treatment of early breast carcinoma requires intraoperative localisation of non-palpable lesions and assessment of the lymph node status. Localisation of breast lesions can be achieved by intratumoural injection of a small amount of radiotracer and intraoperative use of a gamma probe (i.e. radioguided occult lesion localisation, or ROLL). Assessment of the lymph node status is possible by means of the sentinel node approach. To date, two different radiopharmaceuticals have been used for localisation of tumour and sentinel node. We now propose the use of a single nanocolloidal tracer (Nanocoll, with a particle size of less than 80 nm) which is labelled with technetium-99m for simultaneous performance of ROLL and sentinel node identification. The aim of this study was to evaluate the feasibility of this approach, which should be easier and more practical than the dual-tracer injection method. We have employed this new technique in 73 patients with non-palpable, cytologically diagnosed breast cancer and non-palpable axillary lymph nodes. In all patients the radiocolloid, in a total volume of 0.3-0.4 cc, was injected under sonographic or stereotactic guidance. Half of the dose was injected intratumourally and half superficially, but very close to the tumour. Because of the slow lymphatic flow in the breast, Nanocoll must be injected some time before surgery in order to enable adequate migration to the axilla. We injected colloid in the afternoon before surgery (16-23 h before the start of the operation, with an average interval of 18 h). An average dose of 130 MBq (range 110-150) was injected in order to have about 10 MBq of radioactivity when surgery commenced. Lymphoscintigraphy was performed after 15-19 h, with an average interval of 17 h. The procedure was always successful in permitting the localisation of occult breast lesions. Lesions were always localised at the first attempt, and were always contained within the surgical margins. Histological examination revealed all 73 resected lesions to be malignant: there were 64 cases of infiltrating carcinoma and nine of intraductal carcinoma. All breast lesions were therefore confirmed to be early breast cancer. We achieved sentinel node localisation in 71 out of 73, either at scintigraphy or with the intraoperative probe; in two patients, radiopharmaceutical migration was absent. Lymphoscintigraphy showed only axillary drainage in 52 cases, only internal mammary chain (IMC) drainage in nine cases, and combined axillary and IMC drainage in eight cases. In two cases, lymphoscintigraphy suggested the sentinel node was located inside the same breast (intramammary lymph node). All the visualised sentinel nodes were biopsied except for four that were localised in the IMC. Histological examination of the nodes showed metastases in 20 cases: in 15 cases there were micrometastases, and in five, macrometastases. In conclusion, this study has demonstrated the feasibility of the proposed procedure. Simultaneous performance of ROLL and sentinel node localisation using a single tracer represents a useful and practicable choice in the management of early breast cancer.