Primerjava rezultatov programov za analizo sevalnega izocentra pri linearnem pospeševalniku : magistrsko delo

Virant, Bogdan (avtor), Peterlin, Primož (mentor)

Uvod: Sevalni izocenter je presečišče osi rotacij gantrija, kolimatorja in mize, ki pa se zaradi več dejavnikov ne srečajo v isti točki, ampak znotraj sfere. Na velikost sevalnega izocentra vplivajo: mehanska toleranca ležajev, upogib strukture gantrija in nepopolna poravnanost sredine nosilca večlistnega kolimatorja z rotacijsko osjo kolimatorja ali osi vrtenja mize in kolimatorja. Namen: Želeli smo ugotoviti, ali programa PIPSpro in Pylinac enako izračunata velikost sevalnega izocentra na osnovi zvezdastih posnetkov. Metode dela: Na devetih linearnih pospeševalnikih smo naredili zvezdaste posnetke za kolimator, gantrij in mizo. Za vsako sliko smo uporabili tri kolimirana polja v obliki reže. Pri gantriju in kolimatorju je bila razlika med koti 120°, pri mizi pa 60°. Dobljene posnetke smo analizirali s programoma PIPSpro in Pylinac ter izračunali razlike radijev včrtanih krogov. Rezultati: Oba programa sta računala podobno, saj so bili povprečne vrednosti in standardni odkloni za vse aparate podobni. Pri gantriju in kolimatorju je bila razlika manjša od 0,02 mm, nekoliko večja razlika pa se je pojavila pri mizi, kjer je v povprečju znašala 0,11 mm, pri standardnem odklonu pa 0,06 mm. Po dodatni obdelavi, ko smo iste slike večkrat analizirali iz različnih začetnih točk, so rezultati programa PIPSpro pokazali podobna odstopanja, kot smo jih imeli med programoma. S programom Pylinac smo slike obdelali na dva načina; pri prvem smo začetne točke spreminjali za eno slikovno točko, pri drugem pa smo jih razmaknili za deset slikovnih točk. V obeh primerih smo dobili podobne rezultate, saj se položaji centrov krožnih profilov niso veliko razlikovali. Kljub majhnim razlikam v položaju središča krožnice krožnega profila je bila razlika v velikosti včrtanih krogov znatna. Pri PIPSpro so bila odstopanja velikosti včrtanih krogov večja kot pri Pylinacu, kar je verjetno posledica večjih razlik v položaju središč krožnic krožnih profilov. Razprava in zaključek: Rezultati kažejo, da sta si programa med seboj primerljiva in da kljub medsebojnim razlikam računata enako. Na vrednost včrtanega kroga je vplivala začetna točka analize, ki je bila pri vseh slikah naključna, kar se kaže tudi v tem, da so bile razlike radijev včrtanih krogov med programoma normalno porazdeljene.

• Jezik: Slovenski jezik
• Ključne besede: Pylinac, PIPSpro, zagotavljanje kakovosti, zvezdasti posnetek
• Vrsta gradiva: Magistrsko delo/naloga
• Organizacija: ZF - Zdravstvena fakulteta
• Leto izida: 2019
• PID: 20.500.12556/RUL-112712
• COBISS.SI-ID: 5725803
• Datum objave v RUL: 07.11.2019

Sekundarni jezik

• Jezik: Angleški jezik
• Naslov: Comparing the results of two programs for the radiation isocenter analysis of a linear accelerator : masterʹs thesis

Izvleček

Introduction: The radiation isocenter is the intersection of the rotation axes of the gantry, the collimator and the couch, which, due to several factors, do not meet in the same point but within the sphere. The size of the radiation isocenter is influenced by mechanical tolerance of bearings, flexion of the gantry structure, and incomplete alignment of the center of the multi-leaf collimator carriage with the rotary axis of the collimator or of the rotary axis of the couch and the collimator. Purpose: We wanted to determine whether both used software, PIPSpro and Pylinac, equally calculated the size of the radiation isocenter on the basis of star shot images. Methods: We made star shots for the collimator, the gantry and couch on 9 linear accelerators. For each image, we used 3 collimated slit-shaped fields. The difference between the angles of the gantry and the collimator was 120 ° and of the couch it was 60 °. To analyze the images that we made and to calculate the radii of the incircles we used the PIPSpro and Pylinac software. Results: Both software programs calculated similarly, as the mean values and standard deviations were similar for all linear accelerators. The difference with the gantry and the collimator was smaller than 0.02 mm, while at the couch it was slightly bigger; in average it was 0.11 mm and at standard deviation it came to 0.06 mm. After further processing, when we analyzed the same image multiple times from different starting points, the results of PIPSpro software showed similar deviation as we had between the software. With Pylinac, we processed the images in two ways; first, we changed the starting points by 1 pixel and then we separated them by 10 pixels. In both cases, similar results were obtained since the positions of the circular profile centers did not differ much. Despite small differences in the position of the circular profile circle center, the difference in the size of incircles was significant. Deviations of the magnitude of the incircles more significant with PIPSpro than with Pylinac, which was probably due to the bigger differences in the position of circular profiles circle centers. Discussion and conclusion: The results show that both software programs are comparable and that in spite of differences between them, they still calculate in the same way. The value of the incircle was influenced by the starting point of the analysis, which was random in all the images, which was also reflected in the fact that the differences of the incircle radii were normally distributed between both software programs.

• Ključne besede: Pylinac, PIPSpro, quality assurance, star shot