| dc.description.abstract | Carbon ion radiotherapy (CIRT) exhibits higher relative biological effectiveness (RBE), compared to photon and proton RT. However, there are substantial uncertainties regarding the clinical RBE of carbon ions. Therefore, prescription doses and organ at risk (OAR) dose constraints derived from experience with photon or proton RT may not be applicable to CIRT, and should preferably be derived and validated within the framework of this modality.
Two major approaches have been used for the clinical implementation of CIRT. Japanese centres typically use hypofractionated treatments (e.g. 16 fractions of 3.6 – 4.6 Gy [RBE]) in which prescription doses and OAR constraints initially were defined through carefully conducted dose-escalation trials at the National Institute of Radiological Sciences (NIRS, Japan). The NIRS clinical dose model, originally designed for a passively scattered carbon ion beam, is used to predict the RBE-weighted dose (DNIRS). More than 22.000 patients have been treated at Japanese centres, resulting in several publications addressing tolerance doses or dose constraints for various OARs following DNIRS optimized CIRT.
In contrast, CIRT at the Gesellschaft für Schwerionenforschung (GSI), Darmstadt, Germany, was initiated using moderately hypofractionated schedules (20-22 fractions of 3.0 - 3.5 Gy [RBE]) in which the Local effect model version I (LEM I) was used to predict the RBE. Trusting the LEM I to be sufficiently accurate, dose constraints derived from photon RT was applied. This strategy has been adopted by Heidelberg Ion Beam Therapy Center (HIT) and Marburg Ion Beam Center (MIT) in Germany.
When the National Center of Oncological Hadrontherapy (CNAO, Italy) started treating patients with LEM I optimized CIRT in 2012, it was decided to adopt the successful treatment approach developed at NIRS. However, comparative studies show that the LEM I predicts a 5-15% higher RBE in the spread out Bragg peak (SOBP) of a carbon ion beam, relative to the NIRS clinical dose model. In the entrance region, the RBE predicted by LEM I can be 60% higher. Consequently, dependent on the clinical indication, prescription doses at CNAO (reported in LEM I weighted dose, DLEM I) were increased by 5-15% relative to the prescription doses at NIRS (as reported in DNIRS). However, dose constraints to OARs were not adjusted accordingly. This was a cautious approach aimed at avoiding unexpected toxicity due to the lack of validated DLEM I constraints. However, it may lead to suboptimal target coverage.
The aim of this thesis was to explore and update DLEM I constraints for important organs at risk in the head and neck region. Anonymized data gathered from medical records of consenting patients treated within prospective protocols at CNAO were used in this project. Paper I focused on the event of carotid blowout. This rare, but often fatal, complication occurs more frequently after re-irradiation (re-RT) than after primary treatment. High rates (8-15%) have been seen in hypofractionated re-RT using stereotactic body RT (SBRT). Detailed analysis of the relation between the cumulative dose to the carotid artery and carotid blowout has never been published. At CNAO the current practice has been to avoid cumulative equivalent dose in 2 Gy fractions (EQD2) > 120 Gy (RBE) in the re-RT setting. Thus, the medical records of 96 patients re-irradiated with particles (protons: n=17, carbon ions: n=79) at CNAO were analyzed. We found one confirmed, and one possible case of carotid blowout, both amongst the patients receiving proton re-RT. There were no cases recorded among the patients receiving carbon ion re-RT. An actuarial rate of 2.7% for the event was acceptable compared to other studies. For 51% of the patients the cumulative dose to the carotid artery could be calculated. Cases occurred at cumulative EQD2 of 129 Gy (RBE) for the confirmed event, and 107 Gy (RBE) for the possible event. Unfortunately, these data were insufficient to conclude firmly on the relation of cumulative dose and the risk of carotid blowout.
The focus of Paper II and III was to derive more optimal DLEM I constraints for the optic nerve and brainstem applicable for hypofractionated CIRT. For Paper II we analyzed the toxicity outcome and dose distributions of 65 optic nerves in 38 patients treated with CIRT. Visual decline developed in 3 cases, at DLEM I|1%>71 Gy (RBE) and DLEM I|20% >68 Gy (RBE), thus far higher than the dose constraints adopted from NIRS, i.e. DNIRS|1%≤40 Gy (RBE) and DNIRS|20%≤28 Gy (RBE), respectively. The patient treatment plans were recalculated to DNIRS, and thus a dose translation model was obtained, showing that NIRS constraints of DNIRS│1%≤40 Gy (RBE) and DNIRS│20%≤28 Gy (RBE) corresponded to DLEM│1%≤50 Gy (RBE) and DLEM│20%≤40 Gy (RBE), respectively. However, due to uncertainties in the method, and to cohere with constraints used at GSI/HIT, a more moderate constraint escalation to DLEM│1%≤45 Gy (RBE)/DLEM│20%≤37 Gy (RBE) has been implemented in CNAO clinical routine since October 2018.
Paper III focused on the brainstem, in which CNAO has used the constraint DLEMI|1%<30 Gy (RBE), in accordance with the original constraint used at NIRS. Due to this conservative approach, CNAO has not even observed asymptomatic brainstem injury following CIRT. A recent dose response analysis from Gunma University Heavy Ion Medical Center (GHMC), Japan, revealed that asymptomatic brainstem lesions did not occur when DNIRS|0.1cm3 < 40 Gy (RBE) and DNIRS|0.7cm3 < 30 Gy (RBE). Making use of 30 of the treatment plans recalculated to DNIRS for Paper I, a dose translation model was derived also for these brainstem dose metrics, DNIRS constraints corresponded to DLEM I|0.7cm3<41 Gy (RBE) (95% CI: 38-44 Gy [RBE]) and DLEM I|0.1cm3<49 Gy (RBE) (95% CI: 46-52 Gy [RBE]). The value corresponding to the lower bound of the 95% CI’s were proposed as new constraints.
Overall, this thesis has contributed to knowledge of the risk of carotid blowout after particle re-RT. Moreover, we demonstrated a method to reliably calculate the cumulative dose to this OAR, which hopefully can inspire to future dose-response studies for this important endpoint. A novel approach of creating of a dose translation model to support the proposal of updated DLEM I constraints for the optic nerve and brainstem, proved useful, as detailed dose-response data for these OARs following DLEM I optimized CIRT is lacking. The new constraints can have an immediate clinical impact in regard to achieving more optimal treatments at CIRT centres applying the LEM I. Finally, the dose translation method can be used to compare and unify carbon ion treatments worldwide. | en_US |
