In nuclear medicine, an increasing number of radiolabelled agents are under investigation for future use in diagnostic imaging and for applications in radionuclide therapy. All these studies require large amounts of human data to allow for statistical comparisons with existing and well established diagnostic or therapeutic methodologies. In the framework of a good clinical practice environment, clinical trials should be carried out according to international guidelines and regulations as described in the Declaration of Helsinki. Studies involving ionizing radiation, as is the case in nuclear medicine, require special consideration to comply with the ALARA (as low as reasonably achievable) principle. Special publications of the International Commission of Radiological Protection and the World Health Organization deal with this topic in medical research. From the legislation point of view, the 97/43/EURATOM Directive represents the reference to clinical research using ionizing radiation within the European Union. In order to keep the radiation dose of (healthy) volunteers as low as possible, predictive dosimetry studies based on in-vivo animal biokinetics are essential. On the other hand, patients included in dose-escalation radionuclide therapy trials should be monitored individually with respect to dosimetry of the tumour and the critical organs. In this paper the importance and methodology of contemporary patient dosimetry in diagnostic and therapeutic nuclear medicine research are reviewed. It is concluded that reliable dosimetry is essential in performing scientific clinical studies according to the principle of good clinical practice.