Abstract
Background/Aim: Although the effects of varying heart doses on overall survival (OS) in curative thoracic radiotherapy have been investigated, their impact in palliative settings remains underexplored. This study aimed to examine the impact of heart dose on OS in patients with bone metastases treated with palliative radiotherapy over a three-year follow-up period.
Materials and Methods: This study included 303 patients who underwent palliative radiotherapy for bone metastases between 2013 and 2022. The primary endpoint was OS, which was evaluated over a fixed three-year follow-up period. To adjust for baseline confounders between patients with and without mean heart dose (MHD) ≥5Gy, we performed 1:1 propensity score matching (PSM) using the following variables: performance status (PS); primary tumor type; number of bone metastases; number of distant metastases (visceral organs, non-regional lymph nodes, and serosal surfaces such as the pleura and peritoneum).
Results: A total of 108 of 303 patients were included after PSM, with 54 patients each in the groups with and without MHD ≥5 Gy. After matching, the baseline characteristics used as matching variables were well balanced, with no significant differences between the groups according to the chi-square test. Kaplan-Meier analysis demonstrated significantly poorer OS in patients with MHD ≥5 Gy than in those without (p=0.016); annualized event rates for OS were 30.2% and 21.6% in patients with and without MHD ≥5 Gy, respectively.
Conclusion: In propensity score-matched patients receiving palliative, but not curative, radiotherapy, higher MHD was significantly associated with poorer OS over a three-year follow-up period. Furthermore, in the palliative cohort, the MHD level associated with OS may be lower than that in the curative setting.
Introduction
Bone is a common site of metastasis in various malignancies, ranking third after the lung and liver in terms of frequency of secondary metastatic involvement (1-3). Metastatic lesions tend to develop in the vertebrae, pelvis, ribs, and metaphyseal regions of long bones, owing to the abundance of red bone marrow in these areas (4-6). In the spinal column, metastases are most commonly found in the lumbar region, followed by the thoracic, and least frequently in the cervical spine, with reported distributions of 52%, 36%, and 12%, respectively (7).
Radiation therapy (RT) is a commonly used treatment to relieve pain associated with bone metastases. RT is intended to deliver an adequate dose to the tumor while sparing adjacent normal tissues. However, in thoracic RT, such as for lesions in the thoracic spine, sternum, or ribs, some degree of heart radiation exposure may be unavoidable due to the close anatomical relationship with the heart.
Although previous studies (8-13) reported inconsistent results, there is growing evidence that cardiac radiation exposure may be associated with reduced overall survival (OS) in patients with locally advanced non-small cell lung cancer (LA-NSCLC) treated with curative-intent RT, suggesting that heart dose could serve as a prognostic indicator. Similarly, heart dose has also been shown to be an independent predictor of OS in patients with esophageal cancer undergoing curative radiotherapy (14). In breast cancer and lymphoma, several studies have reported elevated cardiac mortality among long-term survivors, with risk increasing over time (15-18).
By contrast, the effect of heart dose on OS in the palliative setting remains underexplored. Because palliative RT employs lower prescription doses than curative RT, doses to organs at risk (OARs) often receive less attention. However, in curative cohorts the heart-dose cut-off associated with reduced OS is frequently lower than the dose constraints routinely used during treatment planning (12, 19, 20). This finding raises the possibility that, even with the lower prescriptions typical of palliative RT, cardiac exposure could still affect survival.
Given these differences, it is necessary to evaluate the impact of heart dose on OS separately in the palliative setting. This study aimed to investigate the impact of heart dose on OS in patients with bone metastases treated with palliative radiotherapy over a three-year follow-up period.
Patients and Methods
Study design. This study retrospectively screened a cohort of patients treated with palliative radiotherapy for bone metastases at a single center from January 2013 to July 2022. The Institutional Review Board approved the study protocol (approval number: 2025-010) and granted a waiver of the requirement for individual patient consent. The research was performed in compliance with all relevant ethical guidelines. Relevant clinical and dosimetric data were sourced from the institutional radiation oncology database.
The patient selection process is summarized in Figure 1. Initially, 396 patients were identified as eligible for inclusion. Ninety-three patients were subsequently excluded for the following reasons: (i) absence of treatment planning data sufficient for heart contouring and dose assessment (n=1), and (ii) loss to follow-up within a three-year period (n=92). This resulted in a final study population of 303 patients.
Study flow chart. A total of 396 patients with stage IV cancer treated with palliative radiotherapy for bone metastases were initially screened. Ninety-three patients were excluded according to the following criteria: missing treatment planning data (n=1) and loss to follow-up within three years (n=92). Therefore, 303 patients were included in the final analysis. Of these, 248 patients received mean heart dose (MHD) of <5 Gy, and 55 received MHD of ≥5 Gy. PSM was conducted for the two groups using Eastern Cooperative Oncology Group Performance Status (ECOG PS), primary tumor type, number of bone metastases, and number of distant metastases as matching variables. Following 1:1 propensity score matching (PSM), 108 patients were included in the matched cohort, with 54 patients in each group based on MHD (≥5 Gy vs. <5 Gy).
Treatments. Patients underwent computed tomography (CT) simulation in a supine position using an Aquilion LB CT system (Canon Medical Systems, Tochigi, Japan) with a 2-mm slice thickness. Radiotherapy planning was conducted using the Pinnacle system (version 9.10; Philips Medical Systems, Fitchburg, WI, USA). Based on these CT images, target volumes and OARs were delineated by experienced radiation therapists. The heart was contoured from the aortopulmonary window inferiorly to the cardiac apex to include the entire pericardium, following the guidelines of NRG-BR001 (21) and Alliance A082002 (22).
From the treatment plans, dose-volume histogram (DVH) data were extracted to determine the mean heart dose (MHD). To allow for comparison across various fractionation schedules, the MHD was converted to the equivalent dose in 2 Gy fractions (EQD2), assuming an α/β ratio of 2. In cases where patients did not complete the planned course of radiotherapy, the delivered MHD was estimated by scaling the planned MHD by the proportion of fractions completed. This adjusted value was subsequently converted to EQD2. The predominant radiotherapy regimen was 30 Gy in 10 fractions. All treatments were administered using three-dimensional conformal radiotherapy (3D-CRT) with 4-10 MV photon beams delivered by an Elekta Synergy linear accelerator (Elekta, Stockholm, Sweden).
Statistical analysis. Follow-up duration was measured from the date of radiotherapy initiation. The primary endpoint was OS, defined as the time from treatment initiation to death from any cause. Surviving patients were censored at three years post-treatment. Baseline characteristics were compared using the chi-square test for categorical variables (Table I). OS rates were estimated using the Kaplan-Meier method, and survival distributions were compared using the log-rank test. A two-sided p-value of <0.05 was considered statistically significant. All analyses were performed using R software (version 3.6.1; R Foundation for Statistical Computing, Vienna, Austria).
Patient characteristics.
To minimize the effect of confounding variables on the association between cardiac dose and OS, propensity score matching (PSM) was performed. Patients with a MHD ≥5 Gy were matched to those with an MHD <5 Gy. Based on previously published studies (23-26), primary tumors were classified into two groups: low-risk, including breast cancer, prostate cancer, and hematologic malignancies; and high-risk, comprising all other tumor types. The propensity score was calculated using the following baseline covariates: Eastern Cooperative Oncology Group Performance Status (ECOG PS; categorized as 0-1, 2, or 3-4), primary tumor type (low risk vs. high risk), number of bone metastases (single vs. multiple), and number of distant metastases (categorized as 0, 1, or ≥2). A 1:1 nearest-neighbor matching algorithm was applied without replacement, using a caliper width of 0.0005.
Results
Baseline patient characteristics. A total of 303 patients who underwent palliative radiotherapy for bone metastases met the eligibility criteria for this study. Patients were stratified into two groups based on the MHD: 248 patients in the MHD <5 Gy group and 55 in the MHD ≥5 Gy group. Detailed baseline characteristics for both groups are presented in Table I. The median age was 71 years, with no significant difference between the two groups (p=0.163), and the proportion of male patients was significantly higher in the MHD <5 Gy group compared to the MHD ≥5 Gy group (63.7% vs. 49.1%; p=0.045). Lung was the most frequent primary tumor site, followed by breast and prostate. The most commonly involved sites of distant metastasis included the lung, liver, and non-regional lymph nodes. The vertebra was the most commonly irradiated bone site for palliative RT (MHD <5 Gy, 47.2%; MHD ≥5 Gy, 89.1%). Regarding the prescribed dose, expressed as biologically effective dose with α/β = 10 (BED10), 39 Gy was the most commonly used regimen in both groups (MHD <5 Gy, 64.5%; MHD ≥5 Gy, 49.1%). ECOG PS (p=0.407), primary tumor risk group (p=0.618), number of bone metastases (p=0.070), and number of distant metastases (p=0.820) showed no significant differences between the two groups in the overall cohort.
Matched cohort analysis. After PSM, 108 of the 303 eligible patients were included in the final analysis, with 54 patients in each MHD group (≥5 Gy vs. <5 Gy), as detailed in Table I. There were no significant differences in median age (p=0.178), sex (p=0.248), or prescribed dose (BED10; p=0.157). The four covariates used for matching were successfully balanced between the groups: ECOG PS, primary tumor type, and number of bone metastases (all p=1.000), as well as the number of distant metastases (p=0.91).
In the matched cohort, lung cancer remained the most prevalent primary tumor type (MHD <5 Gy, 51.9%; MHD ≥5 Gy, 31.5%), followed by breast cancer (MHD <5 Gy, 14.8%; MHD ≥5 Gy, 14.8%) and prostate cancer (MHD <5 Gy, 13.0%; MHD ≥5 Gy, 11.1%), which was consistent with the overall cohort. The distribution of distant metastatic sites also mirrored the pre-matching pattern, with the lung the most frequently involved (MHD <5 Gy, 18.5%; MHD ≥5 Gy, 29.6%), followed by the liver (MHD <5 Gy, 13.0%; MHD ≥5 Gy, 27.8%) and non-regional lymph nodes (MHD <5 Gy, 24.1%; MHD ≥5 Gy, 16.7%). The vertebra was the most common site of bone irradiation for palliative RT in both groups (MHD <5 Gy, 42.6%; MHD ≥5 Gy, 88.9%).
In the propensity score-matched cohort, OS was significantly lower in patients who received MHD ≥5 Gy compared to those with MHD <5 Gy (log-rank, p=0.016). As shown in Figure 2, Kaplan-Meier analysis over a three-year follow-up period demonstrated a clear separation between the two survival curves, with consistently poorer outcomes in the higher MHD group. The 3-year OS was 35.2% in the MHD <5 Gy group and 9.3% in the MHD ≥5 Gy group.
Kaplan-Meier analysis for overall survival in patients stratified by mean heart dose (MHD). The figure shows Kaplan-Meier curves for overall survival over a three-year follow-up in the propensity score-matched cohort stratified by MHD. Patients who received an MHD of ≥5 Gy demonstrated significantly poorer overall survival compared to those who received an MHD of <5 Gy (3-year OS: 9.3% vs. 35.2%; log-rank p=0.016).
Discussion
This study evaluated the impact of radiation dose to the heart on survival outcomes in patients with bone metastases who received palliative radiotherapy. The main findings were as follows: (i) in patients receiving palliative, but not curative, radiotherapy, higher heart doses were associated with poorer OS over a three-year follow-up period; (ii) this association was evident at heart doses lower than those previously reported to affect OS in curative contexts; and (iii) the validity of these observations was strengthened by the use of PSM, which rigorously adjusted for key prognostic factors to ensure well-balanced baseline characteristics between the comparison groups.
There is increasing evidence to suggest that heart dose serves as a prognostic indicator in patients undergoing curative-intent thoracic RT. A meta-analysis by Pan et al. demonstrated that higher cardiac dose metrics, including heart volume receiving ≥5 Gy, ≥30 Gy, and MHD, were associated with worse OS in patients treated with curative thoracic RT for lung cancer (27). Similarly, Speirs et al. reported that heart dose, particularly heart volume receiving ≥50 Gy, was an independent predictor of OS in patients receiving definitive chemoradiotherapy for LA-NSCLC (28). A comparable association has also been observed in esophageal cancer, with a study by Xu et al. identifying heart V30 (percentage of heart volume receiving ≥30 Gy) as an independent OS predictor after curative-intent chemoradiotherapy (14). A recent review further synthesized the growing evidence that cardiac irradiation during thoracic radiotherapy is associated not only with radiation-related cardiac injury but also with survival outcomes, reinforcing the clinical relevance of heart dose metrics in this context (29).
However, few studies have analyzed the association between heart dose and OS specifically in patients treated with palliative RT. A recent study by Nieder et al. found that the maximum heart dose was not significantly associated with OS in patients receiving palliative thoracic radiotherapy for NSCLC (30). In contrast, our study included MHD in the analysis and identified a significant link between higher MHD and poorer OS in patients undergoing palliative RT. Furthermore, we identified an MHD cutoff of 5 Gy as being associated with OS in our cohort. Although derived from a different disease site, contemporary breast volumetric modulated arc therapy (VMAT) planning studies have explicitly used an MHD objective of <5 Gy as a practical planning target, indicating that this dose range is achievable with modern techniques (31). This threshold is lower than the values previously reported in curative settings. For instance, Shepherd et al. analyzed 284 patients with NSCLC who underwent postoperative radiotherapy and reported a median MHD of 11.2 Gy, which significantly stratified patient survival (median OS 31.7 vs. 57.5 months; p<0.001) (32). Another study of 140 patients receiving definitive radiotherapy for LA-NSCLC identified a similar MHD cutoff of 11.88 Gy as a significant prognosticator for 3-year OS (13). In contrast, cardio-oncology literature has emphasized that even relatively low-dose cardiac exposure may contribute to clinically relevant cardiovascular risk, providing plausibility for the hypothesis that modest incidental irradiation could influence relatively long-term outcomes (33).
Although past studies have often focused on heart volumes receiving moderate-to-high doses (e.g., V30, V50) (14, 28), other reports have indicated that low-dose parameters such as Heart V5 are also associated with OS in the curative thoracic RT setting (27). Beyond whole-heart metrics such as MHD, emerging data indicate that dose to specific cardiac substructures can be independently associated with overall survival and other clinically relevant outcomes (34). Consistent with this trend, a recent review highlighted the potential clinical importance of evaluating and constraining substructure doses, including coronary arteries, because such exposure may contribute to cardiac dysfunction and adverse clinical outcomes (35). These findings suggest that even low-dose cardiac irradiation, in the range of ≥5 Gy, may negatively impact survival. Moreover, MHDs in palliative settings are inherently lower than those in curative RT due to reduced prescription doses. Consequently, employing higher cutoff values would be infeasible for analysis, as most patients would fall below such thresholds. Based on these considerations, an MHD cutoff of 5 Gy was selected for our palliative cohort, and it was found to be significantly associated with poorer OS. Additionally, intensity-modulated radiation therapy (IMRT)-based planning approaches (including VMAT) can meaningfully alter the cardiac dose-volume profile, including the low-dose region, suggesting that additional reductions in incidental heart exposure may be achievable in routine practice (36).
This result may reflect the fundamental differences in patient populations between palliative and curative settings. Patients receiving palliative treatment typically present with a more advanced disease burden, poorer overall condition, reduced PS, and the presence of distant metastases. As a result, this vulnerable population may be more susceptible to the detrimental effects of cardiac irradiation on survival. These findings suggest that implementing stricter heart dose constraints, even during palliative RT, could contribute to improved patient outcomes.
In our study, the variables selected for PSM, namely PS, primary tumor type, number of bone metastases, and number of distant metastases, were based on well-established prognostic factors identified in previous studies (23, 24, 37). The classification of primary tumors into low-risk (breast cancer, prostate cancer, hematologic malignancies) and high-risk (all other types) groups was adopted from a previously published prognostic model (23-26). Similarly, the categorization strategies for the number of bone and distant metastases, as well as for ECOG PS, were consistent with those used in prior studies (23, 24, 37, 38). Given that the matching variables and their classifications were derived from established prognostic models (23, 24, 37, 38) and that these covariates were successfully balanced between the groups after matching, the observed association between higher MHD and poorer OS can be considered both robust and reliable.
Clinical implications. The findings of this study have important clinical implications for the practice of palliative radiotherapy. Our analysis demonstrated that an elevated MHD is a significant predictor of worse OS in the palliative treatment setting. Importantly, the MHD threshold associated with this survival detriment was lower than those previously reported in curative-intent studies. This suggests that the palliative patient population may be particularly vulnerable to cardiac radiation, and that even modest incidental doses can be clinically consequential. The follow-up period of three years in this study indicates the relevance of these findings for long-term outcomes. Therefore, an approach that actively spares the heart as a critical OAR is justified for these patients, even within the palliative framework.
The use of advanced radiation technologies presents a concrete solution for achieving this goal of cardiac dose reduction. Major clinical trials such as the Radiation Therapy Oncology Group 0617 trial (39) have demonstrated that IMRT can significantly reduce cardiac exposure compared to 3D-CRT. Notably, technique-oriented reports have demonstrated that coronary artery substructure doses (e.g., the left anterior descending artery) can be improved through VMAT-based planning approaches, supporting the feasibility of substructure-aware cardiac sparing (40). In addition, evidence from other disease sites suggests that highly conformal techniques such as VMAT/IMRT can reduce OAR doses and treatment-related toxicities compared with conventional approaches, further supporting the practicality of adopting advanced planning for toxicity reduction (41). Combining this established evidence with our findings leads to the clinical hypothesis that for patients with bone metastases and a favorable prognosis, applying IMRT to minimize unnecessary cardiac irradiation may contribute to extended survival. This reframes the role of IMRT in the palliative setting, elevating it from a tool for mere dosimetric optimization to a strategic intervention for improving long-term outcomes.
Limitations. First, as a retrospective analysis conducted at a single institution, the generalizability of our findings may be restricted. Second, the inclusion of a heterogeneous cohort with respect to systemic therapies, including patients receiving various agents or radiotherapy alone, may have introduced confounding variables. Third, although primary tumors were classified into two groups based on previously published studies (23-26), alternative classification methods have also been reported (38, 42, 43). Finally, this study did not assess the incidence of clinical cardiac events or the dosimetric impact on specific cardiac substructures. Further investigation is necessary to clarify how these factors affect survival outcomes and to validate our results in a prospective setting.
Conclusion
Among patients undergoing palliative RT for bone metastases, an increased MHD was associated with significantly worse OS during the three-year follow-up period. Importantly, the threshold of MHD linked to survival in this palliative setting appeared to be lower than those reported in studies of curative-intent radiotherapy. This finding supports the implementation of stricter cardiac-sparing constraints, even when the treatment intent is palliative.
Footnotes
Authors’ Contributions
Conceptualization: Y.W.; methodology: Y.W.; software: Y.W.; validation: Y.W.; formal analysis: Y.W.; investigation: Y.W.; resources: Y.W. and T.T.; data curation: Y.W.; writing – original draft: Y.W.; writing – review and editing: T.T., A.I., H.M., H.H., and T.N.; supervision: Y.W. All Authors have read and agreed to the published version of the manuscript.
Conflicts of Interest
The Authors declare that they have no competing interests in relation to this study.
Artificial Intelligence (AI) Disclosure
No artificial intelligence (AI) tools, including large language models or machine learning software, were used in the preparation, analysis, or presentation of this manuscript.
- Received February 3, 2026.
- Revision received February 28, 2026.
- Accepted March 5, 2026.
- Copyright © 2026 The Author(s). Published by the International Institute of Anticancer Research.
This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.
References
Jump to section
Related Articles
Cited By...
- No citing articles found.