Introduction et objectif : La prescription des traitements en radiothérapie est basée sur l'analyse de la répartition de dose calculée par le TPS. Un changement d'algorithme de calcul doit être précédé d'une analyse dosimétrique complète, afin que les impacts cliniques soient maitrisés. Nous présentons une méthodologie de mise en œuvre d'un nouveau TPS. Matériel et méthodologie : Nous avons utilisé 5 algorithmes de calcul de dose. Nous avons comparé 6 plans de traitement avec des configurations identiques : patient, énergie, balistique. Nous avons comparé 12 localisations tumorales : 5 poumons, 1 œsophage, 1 sein, 3 ORL, 1 encéphales et 1 prostate. Le principe de méthodologie est basé sur deux critères d'analyse : 1.Critère d'analyse dosimétrique : nous avons classé les outils d'analyse en 3 catégories : analyse liée à la dose de traitement, analyse liée à la distribution de dose et analyse liée à la répartition de la dose 2.Critère d'analyse statistique : nous avons considéré que nous avons 5 séries de mesure liée à 5 algorithmes. Nous avons considéré les valeurs dosimétriques calculées par l'ancien algorithme comme valeurs de référence. Le test de comparaison statistique utilisé était un Wilcoxon pour série apparié avec un seuil de signification de 5% et un intervalle de confiance à 95%. Résultats et discussion : Nous avons trouvé des écarts pour tous les paramètres comparés dans cette étude. Ces écarts dépendent de la localisation de la tumeur et de l'algorithme de calcul. La maîtrise statistique des résultats nous permet, d'une part de diagnostiquer et interpréter les écarts dosimétriques observés et d'autre part, de déterminer si les écarts sont significatifs. Conclusion : Nous proposons une méthodologie qui permet de quantifier d'éventuels écarts dosimétriques lors d'un changement d'algorithme. L'analyse statistique permet de s'assurer que les résultats sont significatifs. / Background and purpose The validation of a treatment plan is based on the analysis of dose distributions. The dose distributions are calculated by algorithms implanted in TPS. So, the changing of an algorithm must be preceded by a complete dosimetric analysis in order to provide a method for controlling the clinical impact of this change. We present in this study the methodology used for implementing a new TPS in our clinic. Materials and methods We used five algorithms for dose calculation: Clarkson, PBC, Batho Power Law, modified Batho and EqTAR. We compared six treatment plans with identical configurations: 2 plans without heterogeneity correction and 4 with density correction. We have compared nine tumours locations: 5 lungs, 1 oesophagus, 1 breast, 3head and neck, 1 brain and 1 prostate. We compared the following parameters: monitors units, HDV, isodoses, covering index, index of homogeneity, conformity index, geometric index and gamma index for 2D and 3D. We analyzed the results using a statistical evaluation and the method plot box. Results and discussion The gamma index 3D and histogram gamma generated for a CT slice can be used to compare various algorithms and radiotherapy plans. We found a difference in all parameters compared when the algorithm is changed. For example, we found a 5% difference in monitors units and 7% in dose for the pulmonary cancer case, when we change from PBC to EqTAR .This may leads to an increased of 30% in the complication rate. The statistical evaluation serves as a rapid interpretation and diagnostic of dosimetric differences and allows the determination of the significance of these differences. Conclusion We proposed a methodology that allows the quantification of dosimetric variation during the change of calculation algorithm in radiotherapy. This methodology provides a valuable technique for quantitative comparison of various algorithms and radiotherapy plans

Kohn-Sham density functional theory (KS-DFT) is one of the most widely used electronic structure methods. It is used to find the various properties of atoms, molecules, clusters, and solids. In principle, results for these properties can be found by solving self-consistent one-electron Schrödinger-like equations based on density functionals for the energy. In practice, the density functional for the exchange-correlation contribution to the energy must be approximated. The accuracy of practical DFT depends on the choice of density functional approximation (DFA) and also on the electron density produced by the DFA. The SCAN(strongly constrained and appropriately normed) functional developed by Sun, Ruzsinszky, and Perdew is the first meta-GGA (meta-generalized gradient approximation) functional that is constrained to obey all 17 known exact constraints that a meta-GGA can. SCAN has been found to outperform most other functionals when it is applied to aqueous systems. However, density-driven errors (energy errors occurring from an inexact density produced by a DFA) hinder SCAN from achieving chemical accuracy in some systems, including water. Density-corrected DFT (DC-DFT) can alleviate this shortcoming by adopting a more accurate electron density which, in most applications, is the electron density obtained at the Hartree-Fock level of theory, due to its relatively low computational cost. In the second chapter, calculations to determine the accuracy of the HF-SCAN functional for water clusters are performed. The interaction and binding energies of water clusters in the BEGDB and WATER27 data sets are computed, and then the spurious charge transfer in deprotonated, protonated, and neutral water dimer is interpreted. The density-corrected SCAN (DC-SCAN) functional elevates the accuracy of SCAN toward the CCSD(T) limit, not only for the neutral water clusters but also for all considered hydrated ion systems (to a lesser extent). In the third chapter, the barrier heights of the BH76 test set are analyzed. Three fully non-local proxy functionals (LC-ωPBE, SCAN50%, and SCAN-FLOSIC) and their selfconsistent proxy densities are used. These functionals share two important points of similarity to the exact functional. They produce reasonably accurate self-consistent barrier heights and their self-consistent total energies are nearly piecewise linear in fractional electron number. Somewhat-reliable cancellation of density - and functional-driven errors for the energy has been established. / Physics