Comparison of proton and photon therapy in stereotactic arrhythmia radioablation for ventricular tachycardia

Highlights

• •Proton plans reduced mean heart dose from 5.5 Gy to 3.6 Gy in arrhythmia treatment.
• •Mean lung dose decreased from 1.2 Gy to 0.6 Gy using proton radiation therapy.
• •Proton therapy spared more organs for laterally located ventricular targets.
• •All plans met organ dose limits for single-fraction cardiac radioablation.
• •Proton therapy may benefit management of non-malignant cardiac disease.

Abstract

Background and Purpose

Ventricular tachycardia (VT) is a life-threatening arrhythmia commonly treated with catheter ablation; however, some cases remain refractory. Stereotactic arrhythmia radioablation (STAR) offers a non-invasive alternative. While photon-based STAR is effective, proton therapy may improve dose conformity and spare critical organs at risk (OARs), including the heart itself. The aim of this study was to compare the dose-volume metrics between proton and photon therapy for VT.

Materials and Methods

We retrospectively analyzed 34 VT patients who received photon STAR. Proton STAR plans were generated using robust optimization in a commercial treatment planning system to deliver the same prescription dose of 25 Gy in a single fraction. Dose-volume metrics, including D₉₉, D₉₅, D_mean, and $D_{{0.03cm}^3}$, were extracted for critical OARs (heart, lungs, cardiac-chambers) and target. Shapiro-Wilk tests were used to assess normality, with paired t-tests or Wilcoxon signed-rank tests for statistical comparisons between modalities, with Bonferroni correction applied for multiple comparisons.

Results

Proton and photon plans achieved comparable target coverage, with CTV D₉₅ of 25.8 [21.6–28.7] Gy(RBE) vs. 27.2 [21.6–29.3] Gy (p < 0.001). Proton therapy significantly reduced OAR doses, including heart D_mean (3.6 ± 1.5 Gy(RBE) vs. 5.5 ± 2.0 Gy, p < 0.001) and lungs D_mean (0.6 [0.0–1.9] Gy(RBE) vs. 1.2 [0.2–2.6] Gy, p < 0.001), while maintaining optimal target coverage.

Conclusion

Proton therapy for STAR demonstrated significant potential for OARs sparing compared to photon therapy for VT, while maintaining equivalent target coverage. These findings highlight the potential of proton therapy to improve outcomes for VT patients.

1. Introduction

Ventricular tachycardia (VT) is a heart rhythm disorder caused by abnormal electrical signals in the lower chambers of the heart [1,2]. VT commonly occurs in individuals with underlying heart conditions, such as prior myocardial infarction [3]. Current treatment options for VT include implantable cardioverter defibrillators (ICDs), antiarrhythmic medications, and catheter ablation (CA), often used in combination to manage the condition. CA remains the gold standard for treating drug-refractory VT by targeting arrhythmogenic substrates with localized radiofrequency energy via intracardiac catheters [4,5]. However, CA is not curative for many patients, particularly those with inaccessible arrhythmogenic tissue or inadequate energy delivery across the myocardial wall. Additionally, the procedure is associated with high rates of complications and even mortality, particularly in patients with advanced heart failure or extensive comorbidities [6,7].

Recent advances in heart failure therapies have extended survival in patients with severe cardiac dysfunction. However, these patients frequently develop VT because of progressive myocardial disease. Treating VT in this population remains challenging due to the complexity and extent of arrhythmogenic substrates, as well as the basal and epicardial locations of VT origins that are challenging to access [8]. This has driven interest in exploring alternative, less invasive approaches for VT treatment. Stereotactic arrhythmia radioablation (STAR), a specialized application of stereotactic body radiation therapy (SBRT) [9], has emerged as a non-invasive alternative for patients with refractory VT who are poor candidates for catheter ablation. By delivering high-dose radiation to arrhythmogenic substrates, STAR can achieve electrical modulation without the procedural risks of invasive therapies. However, vascular and inflammatory effects induced by high-dose radiation remain poorly understood, necessitating further study [10,11]. Recent meta-analyses, including both clinical trial-focused [12] and broader cross-study [13], reported significant reductions in VT burden, with modest toxicity, but also noted high recurrence rates and substantial heterogeneity—underscoring the need for continued investigation. In parallel, efforts are underway to harmonize STAR practices and accelerate its clinical integration through international collaboration [14].

Extending STAR from photons to protons introduces additional considerations. Proton therapy leverages the Bragg peak to achieve precise dose delivery with minimal exit dose, potentially improving sparing of adjacent critical structures. Preclinical studies have explored the use of high-energy particles for arrhythmia ablation, including carbon ions and protons [[15], [16], [17], [18]]. While carbon ion therapy offers superior radiobiological effectiveness, its cost and limited availability make protons a more accessible alternative. A first-in-man case report demonstrated the feasibility and safety of proton STAR, achieving a significant reduction in VT burden in a patient with advanced heart failure and refractory VT [19]. Despite these promising results, challenges such as range uncertainties and motion management remain significant barriers to its widespread adoption.

The aim of this study was to evaluate the dose trade-offs between proton- and photon-based STAR for VT, focusing on target coverage, dose conformity, and sparing of critical organs at risk. By analyzing clinically derived photon STAR plans alongside retrospectively developed proton plans, this work aimed to guide clinical implementation and identify patient-specific factors that could influence modality selection in VT management.

2. Materials and Methods

2.1. Patient and target selection

Patients included in this retrospective dose-volume study were treated under compassionate-use approval from the institutional review board. Inclusion criteria required a confirmed diagnosis of refractory VT with documented recurrence despite at least two failed antiarrhythmic drugs, one prior radiofrequency ablation procedure, or failure of adjunctive therapies such as mechanical circulatory support or sympathetic blockade.

Table 1 provides an overview of the patient cohort, including demographics, comorbidities, and treatment-specific details. The median patient age was 64.5 years (range: 48–84 years), with 26 males and 8 females. Most patients were white (22 out of 34), and almost all (94 %) had an implantable ICD. The CTV had a median volume of 29.9 cm³ (range: 5.2–104.3 cm³), reflecting the variability in target sizes. Target locations were diverse, with 15 classified as lateral, 8 multi-zone, 7 septal, 2 basal, and 2 anterior.

Table 1.

Patient characteristics for the cohort undergoing stereotactic arrhythmia radioablation (STAR). Values are expressed as medians with ranges where applicable. ICD percentages represent the proportion of patients with implanted devices.

Characteristic	Value
Number of Patients	34
Age (years)	64.5 [48, 84]
Gender	26 Male/ 8 Female
Race	22 White/12 Others
Location	2 Anterior/ 2 Basal/ 15 Lateral/ 8 Multi-zone/ 7 Septal
Median CTV volume (cm3)	29.98 [5.2, 104.3]
Median PTV volume (cm3)	69.91 [20.5, 239.4]
ICD (%)	94

CTV: Clinical Target Volume, PTV: Planning Target Volume, ICD: Implantable Cardioverter-Defibrillator.

While LVEF values were not available for the cohort, most patients met clinical criteria for advanced heart failure. In this population, diminished myocardial contractility is commonly observed, which has been associated with reduced cardiac motion and limited dose-volume impact during STAR [20,21].

2.2. Electroanatomical mapping and imaging

Electroanatomical mapping data from EnSite Precision or CARTO systems were used to define the target. Myocardial scars were identified as regions of low voltage (<0.5 mV bipolar) and validated using gadolinium-enhanced magnetic resonance imaging (MRI) or contrast-enhanced computed tomography (CT). These datasets were fused to optimize target delineation and ensure comprehensive coverage of the arrhythmogenic substrate.

2.3. Treatment planning

Planning images were obtained using the average CT derived from four-dimensional CT (4DCT) scans to account for cardiac and respiratory motion. Electrocardiogram (ECG)-gated imaging was not performed, consistent with our institutional STAR workflow and prior experience in patients with advanced heart failure [22]. To ensure reproducibility, patient positioning was standardized using an SBRT wing board. Photon STAR plans were created using volumetric modulated arc therapy (VMAT) with 25 Gy prescribed in a single fraction. The internal target volume (ITV) was generated by contouring the target in all respiratory phases from 4DCT and computing the union of those contours. A planning target volume (PTV) was then derived by expanding the ITV with a margin to account for positional uncertainties [23]. Retrospective proton plans were developed using robust optimization in RayStation 2023B (RaySearch Lab., Stockholm, Sweden). To enable consistent comparisons, the photon-derived ITV was used as the internal CTV (iCTV) for proton planning to account for target motion. Robust optimization parameters were then set to reflect the same level of positional uncertainty as incorporated in the photon PTV margin. Relative biological effectiveness (RBE) of 1.1 was assumed for protons, consistent with current clinical practice [24,25]. Proton range uncertainty margins were set at ± 3.5 %, following institutional clinical guidelines. To ensure an equitable comparison, proton plans were scaled to match the mean dose (D_mean) to the CTV from the corresponding photon plans on a per-patient basis. This strategy facilitated direct comparison of organ-at-risk (OAR) doses while maintaining consistent target coverage. This approach aligns with prior multi-institutional efforts that explored various normalization strategies to enable fair comparisons across different irradiation techniques [26]. OAR dose constraints were established based on institutional protocols and AAPM TG-101 recommendations [27]. All proton plans used four beam angles. Gantry angles and beam weights were selected by experienced medical physicists based on target location and anatomic constraints, allowing for optimal trade-offs between conformity and OAR sparing.

2.4. Dose-Volume comparisons

Dose-volume histograms (DVHs) were analyzed to compare dose-volume parameters between photon and proton plans. Key target metrics included D₉₈, D₉₅, and V_25Gy to assess dose coverage and high-dose conformity for the CTV and PTV. Conformity index (CI) [28] at the 95 % isodose level was calculated to assess spatial precision, and homogeneity index (HI) was used to evaluate dose uniformity within the CTV. For OARs, analyzed metrics included mean dose (D_mean), maximum dose to the hottest 0.03 cm³ ($D_{{0.03cm}^3}$), V_5Gy, and V_25Gy to evaluate both low- and high-dose exposure. Specific emphasis was placed on critical thoracic structures, including the heart, lungs, esophagus, stomach, spinal cord, and cardiac chambers. Additional analyses were performed to assess dose to patients with ICDs, ensuring compliance with established safety thresholds. To contextualize potential toxicity risks, the results were further evaluated against single-fraction SBRT dose constraints outlined by Timmerman [29]. Subgroup analyses were also performed by target location to assess whether proton therapy offered differential OAR sparing based on anatomical site.

2.5. Statistical analysis

Normality of the dose-volume data was assessed using Shapiro-Wilk tests. Depending on the distribution, paired t-tests were used for normally distributed data, while Wilcoxon signed-rank tests were applied for non-normally distributed data. Multiple comparisons were adjusted using the Bonferroni correction method. Statistical significance was defined as p < 0.05. All analyses were conducted using Python (version 3.9.11) and SciPy (version 1.8.1).

3. Results

The dose-volume analysis demonstrated that photon therapy provided slightly better CTV coverage, with higher D₉₉ (26.5 [15.2–28.2] Gy for photons vs. 24.7 [14.9–27.7] Gy (RBE) for protons, p < 0.001) and D₉₅ coverage (27.2 [21.6–29.3] Gy vs. 25.8 [21.6–28.7] Gy (RBE), p < 0.001), and higher V_25Gy (100.0 [67.4–100.0] vs. 98.8 [66.6–100.0], p < 0.001). Full details are provided in Table 2.

Table 2.

Dose-Volume comparisons between proton and photon treatment plans for various target and OAR metrics. Values are reported as median [range] or mean ± SD based on normality, with clinically significant results highlighted.

Structure	DVH Metric	Photon	Proton	p-value
CTV	D99 (Gy)	26.5 [15.2–28.2]	24.7 [14.9–27.7]	<0.001
	D95 (Gy)	27.2 [21.6–29.3]	25.8 [21.6–28.7]	<0.001
	V25Gy (%)	100.0 [67.4–100.0]	98.8 [66.6–100.0]	<0.001
PTV	D99 (Gy)	23.4 [12.4–24.4]	22.4 [11.0–24.2]	0.03
	D95 (Gy)	24.9 [19.1–25.6]	24.1 [20.4–27.2]	0.42
	V25Gy (%)	94.2 [46.8–97.4]	87.9 [47.4–98.0]	0.001
Whole Heart with CTV	Dmean (Gy)	5.5 ± 2.0	3.6 ± 1.5	<0.001
	$D_{{0.03cm}^3}$(Gy)	31.5 ± 1.4	31.9 ± 1.8	1.00
Normal Heart without CTV	Dmean (Gy)	5.0 ± 1.9	3.0 ± 1.3	<0.001
	V5Gy (%)	26.7 [11.2–80.1]	17.4 [5.5–37.7]	<0.001
Lungs	Dmean (Gy)	1.2 [0.2–2.6]	0.6 [0.0–1.9]	<0.001
Right Lung	${\mathit{D}}_{{0.03\mathit{c}\mathit{m}}^3}$(Gy)	4.4 [1.2–16.8]	0.0 [0.0–10.2]	<0.001
Chest Wall	$D_{{0.03cm}^3}$(Gy)	16.7 [8.6–29.1]	16.6 [7.3–30.4]	1.00
Esophagus	Dmean (Gy)	1.1 [0.3–5.9]	0.0 [0.0–2.2]	<0.001
	${\mathit{D}}_{{0.03\mathit{c}\mathit{m}}^3}$(Gy)	6.4 [1.9–15.5]	0.1 [0.0–15.0]	<0.001
Spinal Cord	${\mathit{D}}_{{0.03\mathit{c}\mathit{m}}^3}$(Gy)	2.8 [1.0–7.8]	0.0 [0.0–7.3]	<0.001
Left Ventricle	Dmean (Gy)	7.6 ± 3.0	5.9 ± 2.6	<0.001
Right Ventricle	Dmean (Gy)	4.5 [0.7–13.9]	2.7 [0.0–12.5]	<0.001
Left Atrium	Dmean (Gy)	2.8 [0.2–6.8]	0.2 [0.0–2.3]	<0.001
Right Atrium	Dmean (Gy)	1.8 [0.2–5.6]	0.0 [0.0–2.1]	<0.001
Aorta	Dmean (Gy)	1.0 [0.1–8.1]	0.0 [0.0–4.4]	<0.001

D_XX: Dose received by XX% of the volume, V_YGy: Volume receiving at least Y Gy, D_mean: Mean dose,

$D_{{0.03cm}^3}$: Hottest 0.03 cm³ dose, CTV: Clinical Target Volume, PTV: Planning Target Volume.

All proton doses are reported as Gy (RBE), assuming a relative biological effectiveness (RBE) of 1.1 in accordance with ICRU 93 guidelines. Photon doses are reported in Gy.

Proton therapy demonstrated superior dose sparing for multiple OARs. For the heart, the mean dose was significantly lower with protons (3.6 ± 1.5 Gy (RBE)) compared to photons (5.5 ± 2.0 Gy, p < 0.001). Protons delivered significantly reduced D_mean to the lungs, individual cardiac chambers and spinal cord, with additional $D_{{0.03cm}^3}$ reductions observed in the spinal cord. Chest wall received comparable doses with protons for certain metrics, such as $D_{{0.03cm}^3}$ (16.7 [8.6–29.1] Gy vs. 16.6 [7.3–30.4] Gy (RBE), p = 1.0). The left ventricle D_mean, often implicated in VT substrates, was reduced from 7.6 ± 3.0 Gy with photons to 5.9 ± 2.6 Gy (RBE) with protons (p < 0.001). A detailed breakdown of these metrics is available in Table 2. Subgroup analysis revealed that proton therapy was particularly advantageous for laterally located VT targets, showing significant reductions in $D_{{0.03cm}^3}$ to the lungs and right atrium (Table 3).

Table 3.

Subgroup-wise comparison of maximum dose received by 0.03 cm³(${\mathit{D}}_{{0.03\mathit{c}\mathit{m}}^3}$) to selected organs-at-risk (OARs) between photon and proton plans. Values are reported as median [range] or mean ± SD based on normality. Bolded p-values indicate statistical significance (p < 0.05).

Structure	Location	Photon	Proton	p-value
Esophagus	Lateral	6.3 [1.9–13.2]	0.1 [0.0–14.7]	0.002
	Multi-zone	7.1 [4.2–15.5]	0.8 [0.0–15.0]	0.01
	Septal	4.6 [3.8–14.0]	0.0 [0.0–13.0]	0.02
Spinal Cord	Lateral	2.8 [1.0–7.8]	0.0 [0.0–7.3]	<0.001
	Multi-zone	2.9 [1.2–6.3]	0.0 [0.0–0.1]	0.01
	Septal	2.4 [1.8–6.0]	0.0 [0.0–0.0]	0.02
Lungs	Lateral	29.3 [3.9–31.9]	27.8 [0.0–32.3]	0.05
	Multi-zone	22.0 [3.9–29.0]	22.6 [0.0–27.2]	0.84
	Septal	19.1 ± 8.1	18.7 ± 9.3	0.71
Right Atrium	Lateral	6.1 [2.6–27.9]	0.0 [0.0–30.2]	<0.001
	Multi-zone	13.1 [2.6–27.9]	16.1 [0.0–28.3]	0.58
	Septal	17.8 [7.4–24.9]	21.2 [0.1–26.1]	0.69

All proton doses are reported as Gy (RBE), assuming a relative biological effectiveness (RBE) of 1.1 in accordance with ICRU 93 guidelines. Photon doses are reported in Gy.

When comparing OAR doses to single-fraction constraints, no violations were observed for esophagus ($D_{{5cm}^3}$ < 14.5 Gy), spinal cord ($D_{{0.25cm}^3}$ < 10 Gy),or stomach ($D_{{10cm}^3}$ < 13 Gy). This indicates that both photon and proton STAR plans adhered to conservative normal tissue dose limits, further supporting the safety of this approach.

Fig. 1 presents dose distributions and DVHs for a patient with CTV in the lateral region of the left ventricle. Both photon and proton plans achieved excellent target coverage, with CTV D₉₅ values of 27.1 Gy for photons and 26.1 Gy (RBE) for protons. The CI was identical for both modalities at 1.0, highlighting precise dose coverage of the target volumes. Despite comparable target coverage, proton therapy showed improved sparing of critical cardiac structures. The mean heart dose was lower with protons (3.7 Gy vs. 1.8 Gy (RBE)), accompanied by a notable reduction in low-dose heart exposure (V_5Gy: 20.9 % vs. 9.1 %). Additionally, proton therapy provided better sparing of the aorta, which was adjacent to the target. The mean dose to the aorta was reduced from 6.2 Gy with photons to 3.6 Gy (RBE) with protons. Similarly, high-dose exposure to the esophagus was greatly reduced with protons (esophagus $D_{{0.03cm}^3}$: 5.4 Gy vs. 0.36 cGy (RBE)). Two additional case examples are presented in the Supplementary Material to illustrate inter-patient variation in target location and OAR sparing.

Fig. 1.

Dose distributions and dose-volume histograms (DVHs) comparing proton and photon plans for a patient with the clinical target volume (CTV) located within the left ventricle. (a) Axial view for the proton plan. (b) Axial view for the photon plan. Contours for the CTV (blue), heart (magenta), lungs (purple), spinal cord (brown), aorta (red) and esophagus (cyan) are overlaid. Thicker contour boundaries were used to distinguish structures from isodose lines. (c) DVHs illustrate superior sparing of the aorta, esophagus, and heart with proton therapy, while maintaining comparable CTV coverage. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

4. Discussion

Our study demonstrated that proton therapy improved dose-volume outcomes for STAR in patients with VT by leveraging the Bragg peak to spare critical OARs without compromising target coverage. These findings underscore the role of proton STAR in optimizing treatment for high-risk patients, particularly those with complex cardiac anatomy or significant comorbidities. Proton therapy demonstrated comparable target coverage to photons. However, significant reductions in mean and high-dose exposure to critical OARs were observed with protons. Subgroup analyses revealed that proton therapy was particularly advantageous for laterally located VT targets, offering superior sparing of adjacent OARs, including the cardiac chambers. Similar OAR sparing has been reported in thoracic oncology studies, including RTOG 1308 [30]. The precision and conformity of proton therapy make it particularly well-suited for STAR, where critical structures often lie near the target.

The ability of proton therapy to spare critical OARs has profound implications for minimizing toxicities in VT patients. Reducing esophageal and lung doses can decrease the risk of radiation-induced esophagitis and pneumonitis, particularly for patients with compromised baseline function. Sparing cardiac substructures may lower the risk of long-term cardiac dysfunction and arrhythmias. This is particularly relevant in patients with compromised cardiac health or prior radiation exposure, where minimizing cumulative dose to the heart is critical. Our findings are consistent with prior studies in thoracic oncology [[31], [32], [33], [34]] which demonstrated reductions in cardiac and pulmonary doses with proton therapy compared to photon-based SBRT. These reductions are particularly important in VT patients with structural heart disease.

While STAR is designed to modulate arrhythmogenic substrates, care must be taken to minimize inadvertent dose to adjacent cardiac structures, as emerging data suggest that high-dose radiation to certain cardiac substructures could potentially exacerbate arrhythmias in vulnerable patients [[35], [36], [37], [38]]. Recent STAR-specific studies have raised concern about dose to cardiac valves and coronary arteries, which may contribute to functional decline or arrhythmogenic risk [39]. These structures were not contoured in our study, which reflects current clinical practice and the lack of consensus on their delineation, as highlighted by the STOPSTORM benchmark study [40]. Auto-contouring tools for these regions are emerging but remain limited in accuracy, particularly for smaller structures [41].Detailed dose analysis was included for individual cardiac chambers, including the atria and ventricles, which are frequently involved in VT pathophysiology. Guidelines for contouring finer cardiac substructures are also being developed by international efforts such as the STOPSTORM consortium and represent a valuable direction for future work [42].

Interestingly, recent work suggests that moderate-dose radiation (∼5 Gy) may have therapeutic effects on cardiac remodeling in heart failure models and in STAR patients [43]. These findings raise the important question of whether cardiac sparing is always desirable and support the need for substructure-specific planning approaches.

Patient selection remains a critical consideration for maximizing the benefits of proton therapy. Studies have demonstrated that patients with pre-existing cardiac disease or targets near critical structures, such as the heart or esophagus, derive the greatest benefit from protons [44]. Pediatric patients with VT represent a unique and challenging population due to their increased susceptibility to late toxicities and longer life expectancy. A case report demonstrated the feasibility of STAR in an 11-year-old with refractory VT and severe dilated cardiomyopathy [45]. Despite multiple failed CA attempts and intolerance to antiarrhythmic drugs, photon-based STAR successfully reduced the VT burden without significant adverse effects. Proton therapy could further improve outcomes in such cases by reducing dose to surrounding tissues [[46], [47], [48], [49]].

Range and relative biological effectiveness (RBE) uncertainties remain a challenge in proton therapy, particularly where critical structures lie near the distal edge of the proton beam. While protons offer superior OAR sparing, fewer beams and variable RBE near the Bragg peak can increase high-dose exposure in some tissues [[50], [51], [52]]. This is especially relevant for STAR, where small positional shifts may impact adjacent structures. LET-based optimization and RBE-adaptive planning represent promising strategies to reduce these risks [53,54].

Inaccuracies in substrate delineation can significantly impact the efficacy of STAR with reports of repeated SBRT due to imprecise initial targeting [55]. These observations emphasize the need for advanced substrate identification strategies, such as co-registration of electroanatomical maps with CT imaging, to ensure precise targeting. Segment-specific motion modeling may help inform individualized ITV margins and improve cardiac target localization in STAR [56]. Although many VT patients have reduced cardiac motion due to low LVEF, motion management strategies such as 4D dose calculation or gating [57] may be warranted in younger or less compromised patients. Incorporating auto-segmentation algorithms could further enhance reproducibility and accuracy in treatment planning, particularly for complex or intramural substrates. Unlike photon therapy, proton dose deposition is highly sensitive to tissue heterogeneity and motion, making accurate motion modeling critical.

Emerging technologies, such as MR-guided STAR, offer additional avenues for improving treatment precision. Recent studies have highlighted the feasibility of integrating real-time motion tracking and adaptive planning into STAR workflows, even in patients with implantable devices [58,59]. Advanced MR techniques, such as T1-based scar mapping, may further refine STAR planning and have shown feasibility in preclinical models [60]. These advances could enhance target coverage and OAR sparing by accounting for dynamic changes in cardiac and respiratory motion during treatment.

One notable limitation of proton therapy is the potential impact of secondary neutrons on ICDs and pacemakers. Transient malfunctions, such as power-on resets, have been reported in ICDs exposed to neutron scatter during proton therapy, with an approximate 1 in 50 Gy probability of occurrence [[61], [62], [63]]. While these events were infrequent and did not result in permanent damage, proactive strategies, including real-time monitoring and adherence to AAPM TG-203 guidelines, are essential to ensuring patient safety during proton STAR [64]. Further work is needed to quantify and mitigate neutron exposure to ICDs, with Monte Carlo-based modeling offering a potential avenue to improve dose estimation and refine treatment planning. This would allow for a more comprehensive risk assessment and the development of strategies to minimize secondary neutron effects in future proton STAR applications.

Ongoing clinical trials, such as the Mayo Clinic study (NCT04392193) and the TOVEL study (NCT06769451), will provide valuable insights into the feasibility and efficacy of proton STAR for VT. These studies complement preclinical research that has demonstrated the potential of protons to deliver conformal and homogeneous doses to the target while sparing surrounding OARs. As data from these trials become available, they will likely shape future clinical practice and support the broader adoption of proton STAR.

Future research should also prioritize long-term follow-up to evaluate the impact of reduced OAR doses on survival and toxicity. Multi-institutional collaborations and larger cohort studies are essential to validate our findings and assess their generalizability. Additionally, incorporating cardiac substructure contouring and advanced RBE models into treatment planning could further refine proton STAR for both pediatric and adult patients.

This study is limited by its retrospective design and small sample size, which may introduce selection bias. Prospective studies with larger cohorts and multi-institutional collaboration are needed to validate our findings and assess their impact on clinical outcomes. Furthermore, long-term follow-up is essential to evaluate the effects of reduced OAR doses on late toxicities and survival.

Our study demonstrated the potential advantage of proton therapy in sparing OARs while maintaining optimal target coverage. These results support the adoption of safer and more effective treatment strategies for VT patients and establish a foundation for utilizing proton STAR in clinical practice. With superior OAR sparing and the potential for dose escalation, proton therapy is a compelling option for managing complex cardiac arrhythmias, particularly in pediatric and high-risk adult populations. Ongoing advancements in imaging, planning, and delivery technologies, combined with prospective clinical validation, will further define the role of proton STAR in this evolving field.

5. Author Responsible for statistical analysis

Keyur D. Shah, PhD (kshah41@emory.edu).

Funding

This research is supported in part by the National Institutes of Health under Award Number R01CA272991, R01EB032680, R37CA272755 and U54CA274513.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article: