Clinical dosimetry in molecular radiotherapy

Kayal Gunjan


Bardies Manuel, (UToulouse),

SCK•CEN Mentor

Struelens Lara
+32 14 33 28 85

SCK•CEN Co-mentor

Saldarriaga Vargas Clarita
+32 14 33 88 14

Expert group

Research in Dosimetric Applications

PhD started


Short project description

Clinical dosimetry in molecular radiotherapy is reaching a mature stage. Patient-specific dosimetry allow a major paradigm shift in the administration of molecular radiotherapy, from a “one size fits all” approach, where all patient receive the same activity, to real personalized medicine where administered activity is assessed specifically for each patient.

Radiopharmaceutical dosimetry basically comprises two steps:

  • First, the determination of the spatial distribution of the radiopharmaceutical in the patient as a function of time needs to be determined, usually from quantitative scintigraphic imaging.
  • Second, absorbed dose calculation must be performed using this distribution, with regards to propagating media, emitted radiation and their interactions within the patient.

Global clinical dosimetry accuracy relies on the accuracy of each of these steps. So far, procedures developed and used in the clinics differ in approach and sophistication, resulting in a large heterogeneity of implemented dosimetric approaches.

The main issue of current clinical dosimetry is that no standard operating procedures do yet exist.

One of the difficulties is how to evaluate the accuracy of the complete chain that leads from scintigraphic imaging to absorbed dose calculation. The absorbed dose within the patient’s tissues cannot be experimentally evaluated in situ. Therefore, in 2008 the project DosiTest ( was initiated by INSERM (Centre de Recherches en Cancérologie de Toulouse) and IRIT (Institut de Recherche en Informatique de Toulouse) with the aim to evaluate the impact of the various steps that contribute to the realization of a dosimetric study, by means of a virtual multicentric intercomparison based on Monte Carlo modelling.



Previous work performed within DosiTest led to its proof of concept.

DosiTest relies on a computing architecture, named “TestDose”, that generates scintigraphic images and reference absorbed dose calculations for a reference anthropomorphic model and radiopharmaceutical pharmacokinetics. TestDose initially included an analytic image simulator, and absorbed dose calculations were performed with MCNPX or EGS4. More recent developments led to the integration of Gate ( in TestDose, both for scintigraphic image generation and absorbed dose calculation.

The DosiTest project is currently in a ‘beta’ phase, in which 4 clinical centres are enrolled, to simulate the context of clinical planar dosimetry (sequential whole-body imaging). The clinical situation is that of PRRT (Peptide Receptor RadioTherapy), with 177Lu-Octreotate. A virtual patient was generated, on the base of the ICRP 110 reference computational model. A hypothetical pharmacokinetics derived from clinical data was integrated in TestDose, thus providing for time-activity curves in all relevant organs and tissues of the virtual patient.

A lot of work still needs to be done in the different steps involved within the project to reach the final aim of proposing a reference methodology applicable in a clinical context.

In its current stage, TestDose is adapted to planar imaging. Even though SPECT modelling is feasible, there is no possibility to generate images produced by hybrid cameras.

The challenge is to integrate in TestDose a CT image generator that can produce attenuation maps for the various devices currently available in the clinics. This represents a variety of equipment and methodologies to give account of hybrid systems installed in nuclear medicine departments. This will allow implementing DosiTest in a context of SPECT-based dosimetry, which can be now considered as “state of the art”.

The second challenge consists in including several clinical centres willing to participate in the “virtual intercomparison”. Each participating centre will have to determine the data type needed to perform a dosimetric study as how they would perform it locally: number and times of acquisition, mode of dosimetric calculation, etc. The gamma camera will be modelled (and validated) and scintigraphic images and additional data will be modelled in TestDose according to the local protocol, in order to obtain a dataset equivalent to what would have been acquired on a real patient. The participating centre will then perform image quantification and activity determination on the simulated scintigraphic images and the dosimetric study will be performed using local resources. The results will be benchmarked against the reference dosimetry generated from the reference model.       

From the comparison exercise with reference data, including different centres, the expected results are to identify the critical steps in a dosimetric study. The impact will be evaluated for the different steps involved, such as temporal sampling, acquisition mode and implemented corrections, ROI definition and absorbed dose determination.