• attività non uniforme
• dosimetria interna
• internal dosimetry
• non-uniform activity
• probabilità del controllo del tumore
• targeted radionuclide therapy
• terapia radiometabolica
• tumor control probability

Internal dosimetry of non-uniform activity distributions: a simple method for the calculation of the absorbed dose distribution at voxel level in radionulcide therapy

Radiotherapy is the medical use of ionizing radiations to control or kill clonogenic cells. In particular in the targeted radionuclide therapy the radioactivity is administred to the patient internally. Non-uniform radioactivity within the target lesions and critical organs constitutes an important limitation for dosimetric estimates in patients treated with tumor-seeking radiopharmaceuticals. On the other hand, the Tumor Control Probability (TCP) and the Normal Tissue Complication Probability (NTCP) are heavily affected by the distribution of the radionuclide in the treated organ/tissue. In this thesis an easy-to-apply method for calculating the absorbed dose at voxel level is described; this new method takes into account a non-uniform radioactivity distribution in the target lesion/organ.

This method is based on the macroscopic S-values, i.e., the S-values calculated for the various organs by Monte Carlo simulations, as defined in the classical MIRD approach and reported in the OLINDA/EXM software, on the number of voxels and on the raw-count 3D array in the target lesion/organ. Starting from these parameters, the only mathematical operation required to calculate the internal absorbed dose at the voxel level, is to multiply the raw-count 3D array for a scalar value, thus avoiding the CPU intensive 3D convolution of the classical MIRD algorithm.

A comparison between the new method and the MIRD approach, fully described in the MIRD Pamphlet No. 17, using S-values at the voxel level, was performed considering four different spherical, homogeneous, mathematical phantoms. This comparison shows a fairly good agreement between the two methods for 131I and ⁹⁰Y. In addiction, the performance of the new algorithm was tested for 3D internal dose calculations for four patients affected by hepatic cancer and treated with ⁹⁰Y nano-spheres.

Voxel dosimetry is becoming more and more important when performing therapy with tumor-seeking radiopharmaceuticals. The method presented in this thesis does not require actual calculation of the S-values at the voxel level and thus allows to bypass the mathematical problems linked to convolution of 3D arrays and to the size of voxels. The results obtained with the new and simplified method and the possibility to use this method also for other radionuclides commonly employed in therapy are discussed.

This thesis is organized as follows:
Chapter 1 Foundations about therapy with radionuclides: a brief history of nuclear medicine is reported, the concept of ionizing radiations is explained and their applications are illustrate;
Chapter 2 Effect of ionizing radiations on human tissues: basic concepts of radiobiology and internal absorbed dose calculations: the definition of the absorbed dose is introduced and the worldwide accepted radiobiologic linear quadratic (LQ) model is illustrate; the main equation for determining the internal absorbed dose is provided and the MIRD algorithm is explained; the way to practically determine all the factors which contribute to the average target internal absorbed dose evaluation and the errors introduced with improper measurements are discussed;
Chapter 3 Internal dosimetry of non-uniform activity distributions: the definition of Tumor Control probability is introduced and its derivation from the Poisson or the binomial statistics is illustrate; the concept of Dose-Volume Histogram (DVH) and its practical utility in radiotherapy are discussed; the three computational approaches actually available for performing 3D dosimetry in a target lesion/organ with a non-uniform radioactivity distribution are presented: dose point-kernel convolution , direct Monte Carlo radiation transport , MIRD approach by using S-values at voxel level .
Chapter 4 A semplified method for the calculation of internal 3D absorbed-dose distributions: theoretical considerations about the new, simplified method are presented; statistical uncertainties on internal absorbed dose 3D distribution estimates are discussed; performance of the simplified method was tested by comparing it with the MIRD method for four different mathematical phantoms and for four patients affected by hepatic cancer; the results obtained are discussed;
Chapter 5 Conclusions: the results obtained with the new and simplified method and the possibility to use this method in future works are discussed.