Vulnerable and Vital: A heart for radiotherapy: Aiming beyond the STARs

In this thesis, we have investigated the role of the heart in radiotherapy as an organ at risk but also as a target.
In the first research chapter, the heart was seen as an organ at risk during the treatment of lung cancer patients. We have researched the correlation between heart dose, specifically cardiac substructure dose, and the survival of lung cancer patients. Our focus was on the robustness of this research, particularly, the use of artificial intelligence (AI) for the automated contouring of substructures, taking into account the contouring inaccuracy, modelling using different underlying biological parameters, the use of models with the ability to handle multicollinearity, and the correction for patient characteristics. Finally, we found a robust correlation between the mean dose to the left atrium and overall survival.
The next chapter builds on the auto-contouring methods for cardiac substructures. The AI model from the first study is reevaluated for stereotactic arrhythmia radioablation/radiotherapy (STAR), a new treatment for patients with ventricular tachycardia (VT). This study introduces and evaluates a similar AI model specifically for VT patients. These models can reduce inter-observer variability and increase the reproducibility, which is essential for future treatment planning and multicentre studies. The results proved that modern auto-contouring methods are capable of generating consistent and usable clinical contours.
Subsequently, we analysed cardiorespiratory motion (CRM). Here, we focus on the influence of CRM on the dose distribution of STAR treatments. It was revealed to substantially impact the dose distribution, both within the target area and in surrounding structures. Especially, small cardiac substructures showed large dosimetric differences due to CRM. Different motion management techniques, such as breath-hold and tracking, were evaluated and showed variable effects on the accuracy of the planned dose. These findings highlighted the need for patient-specific motion analysis and tailored treatment strategies.
The final study comprises the largest clinical study of STAR for treatment-refractory VT to date. This non-invasive treatment option is a promising alternative for patients who couldn’t be helped using conventional therapies, such as medication or catheter ablation. The results of this study suggest that STAR can lead to a significant reduction in the number of VT episodes, with an acceptable safety profile. Simultaneously, it is stressed that the further standardisation of the treatment and long-term follow-up of patients is necessary to further investigate the durability of the treatment effect and the possible late side effects. The survival of the patient cohort is generally poor; however, this study shows that longer follow-up (past 24 months) seems feasible. Furthermore, the STAR treatment might be able to move earlier into the treatment cascade of VT patients, but this requires prospective studies.
In this thesis, we have researched the role of the heart, and specifically its substructures, in the field of radiotherapy. It was shown that the heart is a vulnerable and vital part. If radiotherapy is applied to the thoracic region, the heart has to be taken into account, whether the treatment is for cancer, arrhythmia, or something else in the future.