Summary

METHODS
Fifty consecutive right-sided breast cancer patients treated with free-breathing VMAT were retrospectively analyzed. Doses were converted to the equivalent dose in 2-Gy fractions (EQD2; α/β=3 Gy). Pulmonary NTCP (Lyman-Kutcher-Burman model) and major adverse cardiac event risk (Darby model) were calculated. Multivariate analysis identified independent OAR-dose predictors.

RESULTS
IMNI (13 patients, 26%) significantly increased lung Dmean, pulmonary NTCP, and cardiac Dmean on univariate analysis (all p<0.001). After adjustment, IMNI independently predicted lung Dmean (β=+0.97 Gy; p=0.029), lung V5 (β=+8.28%; p=0.024), cardiac Dmean (β=+0.43 Gy; p=0.027), and NTCP (β=+0.60%; p=0.020). Thoracic lateral diameter independently correlated with lung Dmean (ρ=0.368; p=0.009) and cardiac Dmean (ρ=0.311; p=0.028). The Haller Index showed no significant correlation with OAR doses.

CONCLUSION
IMNI independently increases lung and cardiac doses and thoracic lateral diameter independently predicts OAR dose in free-breathing right-breast VMAT. These markers may identify higher-risk patients who could be prioritized for deep inspiration breath-hold (DIBH) evaluation, pending confirmation in paired planning studies.

Introduction

The landmark study by Darby et al.[3] demonstrated that the risk of major adverse cardiac events (MACE) increases linearly with the mean heart dose by 7.4% per gray, commencing within five years of irradiation and persisting for decades. Radiation-induced pneumonitis, a clinically relevant dose-limiting toxicity, has been quantitatively characterized through dose-volume metrics, with lung mean dose and V20 serving as principal predictors within the Lyman-Kutcher-Burman (LKB) modelling framework.[4] Internal mammary node irradiation (IMNI) has been increasingly incorporated into locoregional RT protocols for node-positive and high-risk breast cancer following randomized evidence of survival benefits from comprehensive regional nodal irradiation.[5] However, incorporating internal mammary targets substantially expands the irradiated volume, raising concerns about elevated OAR dose exposure.[6] In right-sided breast cancer, the cardiac anatomy results in comparatively lower baseline mean heart doses than in left-sided irradiation; nevertheless, the incremental OAR dose burden attributable to IMNI remains clinically relevant, particularly given the proximity of the right coronary artery, lung base, and liver to the internal mammary chain.

Deep-inspiration breath-hold (DIBH) is an established respiratory gating technique that reduces cardiac and pulmonary doses by increasing thoracic volume and creating greater geometric separation between the irradiated target and critical structures.[7] Recent dosimetric studies have demonstrated meaningful OAR sparing with DIBH in right-sided breast cancer, particularly when IMNI is included.[8] Despite these findings, DIBH requires additional institutional < resources, extended simulation times, and patient cooperation. Identifying pretreatment CT-based morphological features associated with higher free-breathing OAR dose could, in principle, help flag patients for whom more advanced dose-sparing approaches merit prospective evaluation.

This study aimed to assess the independent impact of IMNI and thoracic morphological parameters, including lateral diameter, AP diameter, longitudinal diameter, the Vertical-Transverse Lung Ratio, and the Haller Index, on OAR dose-volume indices and NTCP in a homogeneous free-breathing right breast VMAT cohort and to identify determinants of free-breathing OAR dose that may serve as candidate criteria for prospective DIBH evaluation.

Methods

Treatment Planning
Treatment plans were generated using the Monaco 6.2 planning system (Elekta AB, Stockholm, Sweden) and delivered using an Elekta Versa HD linear accelerator. Contouring was performed by a single radiation oncologist according to the ESTRO consensus guidelines.[9] Hypofractionated schedule: 40.05 Gy in 15 fractions (simultaneous integrated boost 48 Gy/15 fr for breast-conserving surgery patients). Conventional schedule: 50 Gy in 25 fractions with a sequential boost of 10 Gy/5 fractions. IMNI was selectively applied based on the axillary nodal status and medial tumor location.

Dosimetric Parameters and EQD2 Conversion
DVH parameters were extracted for the ipsilateral lung (Dmean, V5, and V20) and heart (Dmean and V5). All doses were converted to EQD2 using a linear-quadratic model (α/β=3 Gy). For the hypofractionated plans, the V20 threshold was adjusted to V16 and V5 to V4 in the physical dose, corresponding to 20 Gy and 5 Gy EQD2, respectively.

Thoracic Morphological Measurements
Thoracic dimensions were measured on planning CT, including the AP, lateral, and longitudinal diameters. The Haller Index (HI = lateral² / (AP × sagittal)) was calculated. The right lung, cardiac, and breast/GTV volumes were derived from CT-based contours. All measurements were performed by a single observer using a digital caliper.

Table 1: Patient and thoracic morphological characteristics by internal mammary node irradiation status

NTCP Modelling
Pulmonary NTCP was calculated using the LKB model (TD50=30.5 Gy, m=0.37, n=0.99).[10] Cardiac MACE risk was estimated using the Darby linear model: MACE risk=3%×(1+0.074×cardiac Dmean).[3]

Statistical Analysis
Normality was assessed using the Shapiro-Wilk test. Normally distributed continuous variables are expressed as mean ± standard deviation (SD) and compared using the independent-samples t-test. Non-normally distributed variables are expressed as medians (25th–75th percentile) and were compared using the Mann-Whitney U test. Categorical variables were compared using Fisher’s exact test. Spearman’s rank correlation was used to assess the bivariate associations between the morphological parameters and the OAR dose metrics, given the non-normal distribution of several dose parameters. Multivariate linear regression, adjusted for fractionation schedule, surgical type, right lung volume, and peripheral nodal irradiation, identified independent predictors of OAR dose. Multicollinearity was assessed using variance inflation factors (VIFs), with values <5 considered acceptable. The robustness of the IMNI associations was further evaluated by restricting the analysis to patients who received regional nodal irradiation, by modelling nodal treatment as a three-level ordinal variable (breast/chest-wall only, regional nodal irradiation without IMNI, and regional nodal irradiation with IMNI), and by re-estimating the primary model using a non-parametric bootstrap (percentile confidence intervals over 2000 resamples), which makes no distributional assumption while targeting the same conditional-mean estimand. Linear regression was used descriptively to quantify explained variance and independent associations rather than for formal parametric inference. Cohen’s d quantified effect sizes. Analyses were performed using Python 3.11 (pandas 2.1, scipy 1.11, statsmodels 0.14) and SPSS 26.0 (IBM Corp., Armonk, NY, USA). Statistical significance was set at p<0.05 (two-tailed).

Results

Table 2: OAR dose parameters and NTCP estimates by internal mammary node irradiation status (EQD2, α/β=3 Gy)

No significant differences were observed between the IMNI and non-IMNI groups in terms of surgical type, fractionation, thoracic measurements, or anatomical volumes (all p>0.05); the only significant difference was a higher peripheral RT rate in the IMNI group (100% vs. 59.5%; p=0.003). Fifteen patients received breast or chest-wall irradiation alone, 22 received comprehensive regional nodal irradiation without IMNI, and all 13 IMNI patients also received regional nodal irradiation; the 22 patients who received regional nodal irradiation without IMNI served as the reference stratum enabling the effect of IMNI to be separated from that of comprehensive regional nodal irradiation.

OAR Dose Distribution and NTCP
Overall, the median ipsilateral lung Dmean was 9.95 (IQR: 8.43–11.00) Gy, median lung V20 was 15.5 (IQR: 12.0–19.0) %, lung V5 was 45.30±11.64%, cardiac Dmean was 2.15±0.64 Gy, and median cardiac V5 was 1.35 (IQR: 0.10–7.60) %. The median pulmonary NTCP was 2.97 (IQR: 1.64–4.20) % (range 1.39–5.54%), and the mean MACE risk was 3.48±0.14%.

Impact of Internal Mammary Node Irradiation
Univariate analysis demonstrated that IMNI was significantly associated with higher values for all OAR dose endpoints, with large effect sizes (Cohen’s d=1.27, 1.18, and 1.18 for lung Dmean, lung V5, and cardiac Dmean, respectively) (Table 2). The rate of lung V20 exceeding 15% was 92% in patients with IMNI versus 46% in patients without IMNI (p=0.004). Cardiac Dmean exceeded 2.5 Gy in 54% of IMNI patients versus 14% of non-IMNI patients (p=0.007).

After multivariate adjustment, IMNI remained an independent predictor of lung Dmean (β=+0.97 Gy, 95% CI: +0.10 to +1.85; p=0.029; R²=0.616), lung V5 (β=+8.28%, 95% CI: +1.14 to +15.41; p=0.024; R²=0.338), cardiac Dmean (β=+0.43 Gy, 95% CI: +0.05 to +0.81; p=0.027; R²=0.372), and pulmonary NTCP (β=+0.60%, 95% CI: +0.10 to +1.11; p=0.020; R²=0.625). Lung V20 and cardiac V5 did not retain significance after adjustment (Table 3). No evidence of problematic multicollinearity was observed among the regression covariates (all VIF values <2; IMNI VIF=1.21, peripheral RT VIF=1.72).

Table 3: Multivariate linear regression: Independent effect of IMNI on OAR doses after adjustment for fractionation, surgical type, right lung volume, and peripheral nodal irradiation

Because all IMNI patients also received comprehensive regional nodal irradiation, additional analyses were performed to separate the two effects. When the analysis was restricted to the 35 patients receiving regional nodal irradiation, IMNI remained independently associated with cardiac Dmean (β=+0.48 Gy; p=0.047), while lung Dmean (β=+0.99 Gy; p=0.054) and lung V5 (β=+8.07%; p=0.052) showed similar effect sizes with borderline statistical significance. Modelling nodal treatment as a three-level ordinal variable demonstrated a stepwise dose gradient, with IMNI adding an incremental dose burden beyond regional irradiation alone (lung Dmean: +2.96 Gy vs. +1.93 Gy; cardiac Dmean: +0.96 Gy vs. +0.48 Gy, relative to breast/chest-wall irradiation). Re-estimation of the primary model using a non-parametric bootstrap (2000 resamples; Table 4) confirmed the robustness of the IMNI associations for the lung endpoints (lung Dmean p=0.013; lung V5 p=0.035), whereas the association with cardiac Dmean remained directionally consistent (β=+0.45 Gy) but did not reach conventional statistical significance (p=0.068), consistent with the modest sample size and exploratory nature of the analysis.

Table 4: Robustness of the independent effect of IMNI on organ-at-risk doses: Primary OLS model and non-parametric bootstrap

Thoracic Geometry as a Predictor of OAR Dose
Among the morphological parameters, the lateral thoracic diameter correlated significantly with lung Dmean (ρ=0.368, p=0.009), pulmonary NTCP (ρ=0.368, p=0.009), cardiac Dmean (ρ=0.311, p=0.028), and MACE risk (ρ=0.311, p=0.028). These associations were retained in the multivariate models after adjusting for IMNI and fractionation. The Haller Index showed no significant correlation with any OAR dose parameter (all p>0.05; |ρ| range: 0.016–0.076). The longitudinal diameter showed no significant correlation with lung V5 (ρ=-0.226, p=0.114).

Discussion

Our finding of elevated OAR doses with IMNI is consistent with the published dosimetric evidence. Refsgaard et al.[11] reported a mean ipsilateral lung dose increase of 2.7 Gy and an approximately 20% increase in irradiated volume with IMNI versus non-IMNI locoregional RT plans. Regional nodal irradiation, including IMNI, has consistently been shown to increase cardiac and lung dose-volume parameters compared with breast or chest wall irradiation alone.[12] Our data confirm these findings in a right-sided free-breathing VMAT-specific cohort with EQD2 normalization, a methodological refinement enabling valid cross-comparison across mixed fractionation schedules, which is a common limitation of prior dosimetric studies.

The independent predictive value of IMNI for cardiac Dmean (β=+0.43 Gy after multivariate adjustment) has a direct clinical significance. Applying the Darby model,[3] this increment corresponds to an approximate 3.2% relative increase in MACE risk attributable specifically to IMNI. While the absolute cardiac risks associated with right-sided breast RT are lower than those associated with left-sided irradiation, the cumulative lifetime risk — particularly in younger patients, those with pre-existing cardiovascular comorbidities, or those receiving anthracycline-based chemotherapy — may reach clinical relevance.[13] Given that the right coronary artery passes through the anterior aspect of the heart near the medial fields, minor increases in mean heart dose driven by IMNI cannot be dismissed as clinically negligible.

The IMNI-related lung dose elevation has important implications for pneumonitis-risk stratification. Consistent with QUANTEC recommendations,[14] an institutional lung V20 threshold of 15% was used as a practical risk boundary in the breast RT context, which is a lower threshold than the 30–35% recommended for lung cancer RT, reflecting the expectation of lower lung doses in adjuvant breast irradiation. The fact that 92% of IMNI patients exceeded this threshold versus only 46% of non-IMNI patients (p=0.004) identifies IMNI candidates as the highest-risk free-breathing subgroup, in whom prospective evaluation of DIBH benefit is warranted.

DIBH has been shown to meaningfully reduce OAR doses in right-sided breast cancer RT, particularly when IMNI is included. Yoo et al.[15] demonstrated reductions in ipsilateral lung Dmean from 15.9 to 12.6 Gy and cardiac Dmean from 2.5 to 1.6 Gy with DIBH in right-sided patients receiving regional nodal irradiation. A treatment planning study by Brown et al.[8] similarly confirmed considerable OAR sparing with DIBH in right-sided locoregional RT plans, including IMNI. Our data identified patients with IMNI as the highest-risk free-breathing subgroup, in whom the potential benefit of more advanced dose-sparing techniques would be worth evaluating prospectively.

To our knowledge, the identification of the lateral thoracic diameter as an independent predictor of lung and cardiac OAR dose is novel in the right-sided breast RT setting. The mechanistic basis is straightforward: A wider thorax necessitates wider tangential beam apertures that intercept a proportionally larger volume of the ipsilateral lung. Critically, the lateral diameter remained significant after adjustment for lung volume, indicating that thoracic width conveys distinct geometric information beyond volumetric parameters. The lateral diameter is directly measurable from any planning CT without specialized software, making it a readily available candidate parameter for pretreatment risk assessment, pending external validation. External validation is required before adopting specific lateral-diameter thresholds for routine clinical decision-making.

The Haller Index, a thoracic deformity index derived from the AP, sagittal, and lateral diameters, showed no significant correlation with OAR doses in our right-sided cohort (all p>0.05). This differs from the findings of left-sided breast cancer studies, in which thoracic shape may influence cardiac proximity to the irradiated chest wall. In right-sided RT, cardiac exposure is predominantly determined by the medial field extent (IMNI inclusion) rather than thoracic deformity, explaining the lack of HI predictive value. These findings argue against the use of thoracic shape indices as OAR dose predictors in right-sided breast RT.

The LKB-derived pulmonary NTCP estimates (median, 2.97%) and Darby-model MACE estimates (mean, 3.48%) reflect the low OAR doses characteristic of adjuvant breast RT.[16] Their principal utility in this study lies in their use as continuous quantitative outcome variables for multivariate analysis and as individualized risk communication tools, enabling explicit, patient-centered quantification of the incremental OAR risk attributable to IMNI, which may support the informed consent process and facilitate risk-adapted planning decisions.

Limitations of the Study
This study had several limitations. As this was a retrospective single-center analysis, selection bias cannot be excluded. The sample size of 50 patients limited the statistical power of the subgroup analyses. Accordingly, multivariable modelling should be considered exploratory and requires confirmation in larger external datasets. Demographic variables, including patient age and cardiovascular comorbidity profiles, were not available in the dosimetric dataset and could not be included as covariates. The absence of clinical toxicity data precludes the direct validation of NTCP estimates against observed outcomes, and a prospective cohort study with long-term toxicity endpoints is warranted. Because all patients who received IMNI also underwent peripheral lymphatic irradiation, these two variables were strongly associated, which may introduce collinearity into the multivariate models; the IMNI estimates should therefore be interpreted as reflecting comprehensive regional nodal irradiation that included the internal mammary chain. In addition, p-values were not adjusted for multiple comparisons, so borderline associations warrant cautious interpretation. DIBH planning was not performed in this cohort; a paired dosimetric comparison of free-breathing and DIBH plans for individual patients is required to directly quantify patient-specific DIBH benefits. Consequently, the proposed DIBH selection criteria remain inferential and should be validated in paired free-breathing and DIBH planning studies.

Conclusion

References

1) Bray F, Laversanne M, Sung H, Ferlay J, Siegel RL, Soerjomataram I, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2024;74(3):229–63.

2) Darby S, McGale P, Correa C, Taylor C, Arriagada R, Clarke M, et al; Early Breast Cancer Trialists’ Collaborative Group (EBCTCG). Effect of radiotherapy after breast-conserving surgery on 10-year recurrence and 15-year breast cancer death: Meta-analysis of individual patient data for 10,801 women in 17 randomised trials. Lancet 2011;378(9804):1707–16.

3) Darby SC, Ewertz M, McGale P, Bennet AM, Blom-Goldman U, Brønnum D, et al. Risk of ischemic heart disease in women after radiotherapy for breast cancer. N Engl J Med 2013;368(11):987–98.

4) Marks LB, Yorke ED, Jackson A, Ten Haken RK, Constine LS, Eisbruch A, et al. Use of normal tissue complication probability models in the clinic. Int J Radiat Oncol Biol Phys 2010;76(3 Suppl):S10–9.

5) Whelan TJ, Olivotto IA, Parulekar WR, Ackerman I, Chua BH, Nabid A, et al. Regional nodal irradiation in early-stage breast cancer. N Engl J Med 2015;373(4):307–16.

6) Poortmans PM, Collette S, Kirkove C, Van Limbergen E, Budach V, Struikmans H, et al. Internal mammary and medial supraclavicular irradiation in breast cancer. N Engl J Med 2015;373(4):317–27.

7) Haussmann J, Corradini S, Nestle-Kraemling C, Bölke E, Tamaskovics B, Djiepmo-Njanang FJ, et al. Recent advances in radiotherapy of breast cancer. Radiat Oncol 2020;15(1):71.

8) Brown S, Kügele M, Edvardsson A, Berg L, Alkner S, Andersson Ljus C, et al. Dose reduction to organs at risk with deep-inspiration breath-hold during right breast radiotherapy: A treatment planning study. Radiat Oncol 2019;14(1):223.

9) Offersen BV, Boersma LJ, Kirkove C, Hol S, Aznar MC, Biete Sola A, et al. ESTRO consensus guideline on target volume delineation for elective radiation therapy of early stage breast cancer. Radiother Oncol 2015;114(1):3–10.

10) Seppenwoolde Y, Lebesque JV, de Jaeger K, Belderbos JS, Boersma LJ, Schilstra C, et al. Comparing different NTCP models that predict the incidence of radiation pneumonitis. Int J Radiat Oncol Biol Phys 2003;55(3):724–35.

11) Refsgaard L, Buhl ES, Nielsen AWM, Thomsen MS, Andersen K, Jensen I, et al. Internal mammary node irradiation in early breast cancer — target coverage and implications on dose to organs at risk. Acta Oncol 2025;64:795–802.

12) Overgaard M, Hansen PS, Overgaard J, Rose C, Andersson M, Bach F, et al. Postoperative radiotherapy in high-risk premenopausal women with breast cancer who receive adjuvant chemotherapy. N Engl J Med 1997;337(14):949–55.

13) Taylor CW, Wang Z, Macaulay E, Jagsi R, Duane F, Darby SC. Exposure of the heart in breast cancer radiation therapy: A systematic review of heart doses published during 2003 to 2013. Int J Radiat Oncol Biol Phys 2015;93(4):845–53.

14) Kwa SL, Lebesque JV, Theuws JC, Marks LB, Munley MT, Bentel G, et al. Radiation pneumonitis as a function of mean lung dose: An analysis of pooled data of 540 patients. Int J Radiat Oncol Biol Phys 1998;42(1):1–9.

15) Yoo S, Jung H, Kim S, Jeong J, Lim YK, Kim JY, et al. Comprehensive analysis of the role of deep inspiration breath-hold in right-sided breast cancer radiotherapy: A focus on cardiac substructures and right coronary artery. J Appl Clin Med Phys 2025;26(7):e70067.

16) Gagliardi G, Constine LS, Moiseenko V, Correa C, Pierce LJ, Allen AM, et al. Radiation dose-volume effects in the heart. Int J Radiat Oncol Biol Phys 2010;76(3 Suppl):S77–85.