Dissertation committee:
Παντελής Καραΐσκος, Καθηγητής, Ιατρική Σχολή ΕΚΠΑ, Επιβλέπων
Καθηγητής Ιωάννης Σεϊμένης, Καθηγητής, Ιατρική Σχολή ΕΚΠΑ
Σοφία Κόττου, Καθηγήτρια, Ιατρική Σχολή ΕΚΠΑ
Παναγιώτης Παπαγιάννης, Αναπληρωτής Καθηγητής, Ιατρική Σχολή ΕΚΠΑ
Βασίλειος Κουλουλίας, Αναπληρωτής Καθηγητής, Ιατρική Σχολή ΕΚΠΑ
Ευάγγελος Παντελής, Επίκουρος Καθηγητής, Ιατρική Σχολή ΕΚΠΑ
Κωνσταντίνος Λουκάς, Επίκουρος Καθηγητής, Ιατρική Σχολή ΕΚΠΑ
Summary:
Stereotactic radiosurgery with Volumetric Modulated Arc Therapy (VMAT) combined with image guidance (IGRT) has been increasingly implemented in applications of stereotactic radiosurgery/radiotherapy. More specifically, single isocenter SRS – VMAT technique with multiple non coplanar arcs can efficiently deliver plans of high dose conformity to the target volumes, while ensuring the required radioprotection of the adjacent healthy tissues and overall, the treatment time is short. However, the technique demonstrates increased sensitivity to geometric uncertainties. The scope of the thesis is to assess the geometric uncertainties involved in the modern radiotherapy SRS – VMAT technique for the treatment of complex cases, such as multiple brain metastases. To that end, the effect of the geometric uncertainties on dose distributions and plan quality indices, clinically used for plan evaluation and acceptance, is quantified and investigated. Firstly, the effect of the targets charcteristics (size, number) on singe-isocenter SRS-VMAT efficiency was examined. For limited number of targets (≤6), it was noticed that the plan quality indices were improved, for increasing diameter of target up to 2 cm. When the target size is quite large (2 cm in diameter targets), the mlc leaves adjust better according to the shape of target and contribute in higher conformity. It was also confirmed that the plan quality indices (dose conformity, target coverage) deteriorated and an increase in low dose bath was noticed with: i) target number increase (>6), ii) target size increase (up to 2 cm, maximum diameter), and iii) two or more targets sharing multiple pairs of leaves resulting in enhancement of island blocking effect. The examined sources of geometric uncertainty are the translation and rotational setup errors as well as the MLC leaf positional uncertainty. The results showed that translational errors ≥1mm could lead to severe underdosage of targets (target indices deterioration <10%). The underdosage seems to increase with: i) increasing degree of uncertainty, ii) reducing target size (<1cc) and iii) target number increase (>6). Regarding rotational errors investigation, it was noted that target distance form the isocenter is a factor that plays major role in dosimetric and geometric impact of rotational errors on targets. Rotational errors are geometrically expected to greatly influence targets in large distance from the isocenter (up to 6.5 cm, for the investigated single-isocenter SRS-VMAT technique). For comparison purposes, two-isocenter technique was also included in rotational errors investigation. Including two isocenters resulted in shorter target distances from the nearest isocenter (up to ~4 cm) compared to target distances of single-isocenter technique. Therefore, the potential benefit of reduced dosimetric impact of rotational errors on targets located at distance ≤4cm was assessed. The results revealed that spatial displacements >1mm of the targets occurred, when rotational errors ≥1° impacted targets, at distances of the order of ~ 4cm for both isocenter techniques. It was also confirmed that two-isocenter technique which led to shorter target distances from the nearest isocenter, resulted in reduced dosimetric impact of rotational errors on targets and adjacent OARs. More specifically, it was noticed that rotational errors ≤1° did not influence significantly the target dosimetry (indices changes <5%), when two-isocenter technique was used˙ while for single-isocenter technique, only rotational errors of the order of 0.5° were tolerated in terms of causing insignificant impact on target dosimetry. Moreover, the results showed that small targets (≤1cc) are more susceptible to rotational errors, for both techniques. Rotational errors, according to their direction, could impact significantly OARs dosimetry, for OARs laying in close proximity to target(s), leading to maximum dose limit (Dmax) exceeding the limit value Dmax even for the smaller rotational errors (0.5°), with observed changes of Dmax in the range of 10%-50%, for all simulated rotational errros. However, exceeding the dose limit occurred less frequently and in a smaller degree for two-isocenter technique. Regarding the investigation of the dosimetric impact of leaf positional uncertainty on the quailty of SRS-VMAT technique, it was found that leaf positional uncertainties ≥ 0.3mm could significantly deteriorate target dosimetry, leading to target coverage reduction >10%. Moreover it was noted that leaf positional uncertainties ≥ 0.2mm could also impact considerably the dosimetry of OARs in the vicinity of targets, leading to mean changes of Dmax in the range of 10%-30%, and the dose limit was exceeded in many cases. A linear relationship between the change in dose and the leaf positional uncertainty was also noticed (12%/mm). For all the investigated sources of uncertainty, it was noticed that small targets (<1cc) were more sensitive to them.
Keywords:
Radiotherapy, Radiosurgery, SRS-VMAT, Brain metastasis, Single isocentre
