Improved chest wall stability during deep-inspiration breath-hold in breast cancer radiotherapy following quality improvement initiatives Authors: Kristine Wiborg Høgsbjerg, Anders W. Mølby Nielsen, Esben Yates, Christina Maria Lutz, Per Rugaard Poulsen, Birgitte Vrou Offersen, Harald Spejlborg Published: 2026 ABSTRACT Intrafractional chest wall motion during radiotherapy in deep-inspiration breath-hold may compromise dose delivery to the internal mammary nodes in patients with breast cancer. The present study evaluated chest wall motion after implementation of four treatment quality improvement initiatives, including repositioning of the gating marker block from the upper abdomen to the chest wall. Geometrical errors were quantified using cine MV images. The occurrence of geometrical errors >3 mm was 4.7%, which was significantly lower than before the improvement initiatives (21.2%, p < 0.0001). However, predominantly mild skin-related side effects were reported by 61% of patients caused by the gating marker block overlapping with the treatment field. Introduction Internal mammary node irradiation improves survival in node-positive breast cancer patients [1]. With tangential field 3D conformal radiotherapy (3D-CRT) in deep-inspiration breath-hold (DIBH), the internal mammary node (IMN) target is located at the border of steep dose gradients. This makes the IMN dose coverage vulnerable to intrafractional chest wall motion. Accurate chest wall positioning throughout treatment delivery is therefore an essential component of radiotherapy quality assurance. Intrafractional chest wall motion has been investigated in several studies [2], [3], [4], [5]. This has previously been demonstrated to cause geometrical errors above 3 mm in 21.2% of all delivered fractions and above 5 mm in 3.0% of fractions in a cohort of left-sided breast cancer patients treated with loco-regional radiotherapy using DIBH and tangential field 3D-CRT in 2021 [6]. This patient group is hereafter referred to as the 2021 cohort. The geometrical errors translated into a reduced delivered dose to IMN such that 8% of the patients had more than a 9% reduction in the volume receiving the recommended dose [6]. Although no tumour control probability model exists for IMN irradiation, such impaired dose delivery may imply reduced treatment efficacy. Following the identification of these clinically significant intrafractional chest wall motions in the 2021 cohort, four quality improvement initiatives were implemented in the clinic to reduces motion. The initiatives included repositioning of the gating marker block to the sternal bone, improved breath-hold reproducibility, IMN-prioritised matching for daily image-guided radiotherapy (IGRT), and systematic evaluation of continuous portal megavoltage (cine MV) images. The primary endpoint of the present study was to assess the geometric impact of these changes on intrafractional chest wall motion. The secondary endpoint was to quantify the potential skin-related side effects associated with the repositioned gating marker block. Material and methods Patient cohort This single-centre quality assurance study consecutively included 39 patients with left-sided breast cancer scheduled to receive loco-regional radiotherapy in DIBH between January and May 2024. The cohort size was matched to that of the 2021 cohort (n = 39) [6]. The study was approved by the institutional review board. Treatment planning and dose delivery The DIBH was controlled with external respiratory gating monitoring using a gating marker block (RGSC, Varian Medical Systems, Palo Alto, Ca, USA) with a standard gating window of 2 mm. The clinical target volumes (CTV) included the residual breast or chest wall, the axillary levels 2–4 (level 1 if indicated), and the IMN according to the ESTRO consensus delineation guidelines [7], [8]. According to Danish Breast Cancer Group (DBCG) recommendations at least 98% of the IMN CTV was required to receive 90% or more of the prescribed dose to the breast, corresponding to a V90% ≥ 98% [9]. The prescribed dose was 40 Gy in 15 fractions delivered using 3D-CRT with two tangential fields supplemented with one or two periclavicular fields using beam energies of 6-15 MV. Doses to organs of interest were presented as mean heart dose and mean lung dose. Treatment planning was performed in Eclipse (Varian Medical Systems) and treatment was delivered on a Varian TrueBeam linear accelerator. Further planning details have been described in detail previously [6]. Initial online IGRT matching and positional correction were performed daily using an orthogonal kV image and a tangential MV image [6]. During treatment delivery, cine MV images (7 frames/s) were acquired at all fractions of both the medial and lateral tangential fields. Treatment quality improvement initiatives Four treatment-delivery initiatives were implemented to decrease the intrafractional chest wall motion. First, the DIBH gating marker block was repositioned from the upper abdomen to the middle of the sternal bone to provide a more direct control of the chest wall and to minimise the influence of abdominal motion. The marker block position was the same for all fractions. As this modified marker block position was within the tangential treatment fields, a potential adverse skin bolus effect was evaluated as a secondary study endpoint. Second, our clinical observations have indicated that some patients exhibit an initial anterior overshoot of the chest wall position at the start of DIBH. Consequently, images acquired immediately after reaching the trigger level, i.e. the gating window, may not be representative of the average chest wall position during the entire DIBH. Therefore, the clinical workflow was modified such that acquisition of daily setup MV images was delayed by 5 s after the DIBH trigger level was reached. This was done to improve the representativeness of the chest wall position during the entire DIBH. Third, in case of a pitch rotation of the chest wall in the 2021 cohort, daily IGRT matching of the chest wall was a compromise considering the entire tangential field from the cranial part including the IMN to the caudal part dominated by heart and lung anatomy [6]. This was modified such that the IGRT matching instead prioritised the cranial part of the field, as IMN target coverage is essential [1]. Fourth, to evaluate DIBH performance during treatments, off-line review of the cine MV images was performed systematically for each patient after the second radiotherapy fraction, which was assumed to be more representative of the treatment course than the very first fraction. This enabled identification of intrafractional motion and set-up variation early in the treatment course and allowed corrective actions for subsequent fractions, for example, expansion of the gating window from 2 mm to 3 mm or additional patient coaching to improve DIBH stability. The present study evaluated the combined effect of the four treatment-delivery modifications, as all included patients were treated according to the modified workflow. Geometrical errors Intrafractional geometrical errors were determined by comparing the chest wall position on the final cine MV frame acquired during treatment delivery with the planned chest wall position in the Digitally Reconstructed Radiograph (DRR). The cine MV to DRR matching was performed by two authors (AWMN, MD and KWH, MD). To ensure match agreement, the first three patients were matched independently by both assessors, and the results were subsequently compared. The final cine MV frame was chosen for pragmatic reasons as the vendor software only allowed direct matching of this frame with the DRR. However, previous work has shown that the chest wall was stable within the DIBH and the average chest wall position was well represented by the final frame [2]. The geometrical errors were presented as the group mean error (M), the systematic error (∑), and the random error (σ) [10]. Furthermore, geometrical errors were reported as the proportion of delivered field with errors above 3 mm and 5 mm, respectively. To directly assess the benefit of these newly implemented initiatives, the present data were compared to the geometrical errors from the 2021 cohort and presented in supplementary Table S1. Differences between the two cohorts were tested with the chi-square test. Comparative statistics did not account for clustering effects. All statistical tests were two-sided, and a p-value <0.05 was considered statistically significant. Adverse events To assess a potential bolus effect from the repositioned DIBH gating marker block, we collected data on skin-related side effects from 20 patients with left-sided breast cancer treated with DIBH using the marker block. Unlike the main study cohort, which was evaluated retrospectively to minimise potential changes in clinical practice related to awareness of data collection, this cohort was included prospectively to enable collection of patient-reported skin-related side effects. Patients were prospectively included from November 13 to December 11, 2024. Skin reactions in the region of marker block placement over the sternum were assessed by telephone interview. The interview was carried out 14 days after completion of radiotherapy, to coincide with the expected peak of acute radiation-induced skin reactions [11]. Skin reactions were thus recorded as patient-reported chest wall skin reaction and rated on an adapted, nonvalidated Common Terminology Criteria for Adverse Events (CTCAE) grading scale ranging from 0 to 3 (0 = no symptoms, 1 = mild symptoms with itching, 2 = moderate symptoms with blisters, and 3 = severe symptoms with ulceration). Results The planned cohort of 39 consecutive patients was successfully included. Baseline patient, treatment, and dose/volume characteristics are presented in Table 1. Thirty-eight patients (97.5%) underwent the planned cine MV evaluation after the second treatment fraction, resulting in treatment-delivery interventions in six patients (15.4%).Table 1Patient characteristics, radiotherapy treatment, and organs of interest dose metrics.No. of patients, n (%)39(100%)Age, median (range)67(33–83)Surgery Lumpectomy, n (%)21(54%) Mastectomy, n (%)18(46%)Radiotherapy Loco-regional RT, n (%)39(100%) Lymph node boost, n (%)5(13%)Target coverage and doses to organs of interest Mean heart dose, median Gy (IQR)1.2(1.0–1.5) Mean lung dose, median Gy (IQR)11.7(10.7–12.9) V90 CTV_IMN, median % (IQR)99.6(95.5–100) Table 1Patient characteristics, radiotherapy treatment, and organs of interest dose metrics.No. of patients, n (%)39(100%)Age, median (range)67(33–83)Surgery Lumpectomy, n (%)21(54%) Mastectomy, n (%)18(46%)Radiotherapy Loco-regional RT, n (%)39(100%) Lymph node boost, n (%)5(13%)Target coverage and doses to organs of interest Mean heart dose, median Gy (IQR)1.2(1.0–1.5) Mean lung dose, median Gy (IQR)11.7(10.7–12.9) V90 CTV_IMN, median % (IQR)99.6(95.5–100) Patient characteristics, radiotherapy treatment, and organs of interest dose metrics. In the present cohort, geometric errors were M = 0.1 mm, ∑ = 0.7 mm, and σ = 1.3 mm. Errors exceeding 3 mm were observed in 4.7% of all evaluated treatment fields (55/1176). Errors larger than 5 mm were observed in 0.6% of all evaluated treatment fields (7/1176). A direct comparison of these results to the 2021 cohort can be found in the Supplementary data (Text S1 and Table S1). Data on skin-related side effects were obtained from 18 of the 20 contacted patients, corresponding to a response rate of 90%. The median time of interview was 16 days (range 14–26 days) after completion of radiotherapy. In the region of the marker block, 11 patients (61%) reported skin reactions of any grade. Among these, most patients (n = 9, 82%) experienced grade 1 skin-related side effects, but one patient reported grade 2 side effects with blisters of the skin and one patient experienced grade 3 side effects with ulceration beneath the marker block (Table 2).Table 2Patient-reported skin-related side effects in the marker block region following radiotherapy (n = 18).n = 18%Grade 0739%Grade 1950%Grade 216%Grade 316% Table 2Patient-reported skin-related side effects in the marker block region following radiotherapy (n = 18).n = 18%Grade 0739%Grade 1950%Grade 216%Grade 316% Patient-reported skin-related side effects in the marker block region following radiotherapy (n = 18). Discussion The present study showed that the combined implementation of sternal repositioning of the DIBH marker block and other IGRT practices resulted in substantially reduced intrafractional chest wall motion. Compared with the 2021 cohort, the proportion of geometric errors exceeding 3 mm decreased from 21.2% to 4.7%, while errors above 5 mm were decreased from 3.0% to 0.6%. Formal comparison with the 2021 cohort revealed that these differences were statistically significant. In addition to the reduction in outlier errors above 3 mm and 5 mm, the quality improvement initiatives reduced the group mean error M from 0.7 mm to 0.1 mm, the systematic error Σ from 1.4 mm to 0.7 mm, and the random error σ from 1.7 mm to 1.3 mm. The present study highlights the importance of quality assurance in DIBH radiotherapy [12]. Quality assurance is an essential part of the radiotherapy treatment chain in order to ensure correct delivery of the prescribed and planned dose. Cine MV imaging may be a useful tool for monitoring geometric precision during treatment delivery. Skin-related side effects were observed at the position of the gating marker block, due to a potential bolus effect when placing the marker block within the treatment field. Ideally, skin-related side effects would have been compared before and after repositioning of the marker block. However, skin-related side effects can be assumed to have been minimal in the 2021 cohort because the marker block was placed caudally, outside the treatment field. The present results indicate that the marker block should remain within the radiotherapy field to maintain a high DIBH reproducibility, as abdominal placement outside the target area reduced control of chest wall position [2], [9]. Consequently, strategies are needed to minimise skin dose while maintaining contact between the block and the sternal bone. This can be achieved by increasing the distance between the skin and the block—for example, using a low-density spacer to elevate the block above the skin while maintaining its sternal attachment. Ongoing work at our institution is investigating this approach, and further quality assurance is warranted to evaluate its effectiveness. Alternatively, skin reactions may be reduced by decreasing the amount of material in contact with the skin, and our institution has inquired with the vendor regarding this option. These findings regarding skin-related side effects are only relevant to DIBH monitoring systems that rely on a marker block. Alternative DIBH monitoring systems, such as surface-guided radiation therapy gating and Active Breathing Coordinator (Elekta), do not require a marker block. In breast cancer radiotherapy, highly conformal techniques such as volumetric modulated arc therapy (VMAT), intensity-modulated radiation therapy (IMRT), and the combination of 3D-CRT, IMRT, and VMAT (hybrid techniques) are increasingly used [13], [14]. Such highly conformal techniques require greater geometric accuracy and stable intrafractional chest wall positioning throughout treatment delivery. The improved chest wall stability demonstrated in the present study may therefore become increasingly important as highly conformal treatment techniques are more widely adopted. Future research and quality assurance studies should focus on quantifying intrafractional errors with arc techniques, where cine MV images provide less direct monitoring of geometrical errors. A strength of our study was that the present 2024 cohort was collected using the same methodology as the 2021 cohort, allowing direct comparison between the cohorts. Although not randomised, this gave us the opportunity to evaluate the four implemented improvement initiatives. Some limitations of the present work should however be mentioned. First, the sample size and follow-up are too small to evaluate late clinical outcomes. Second, no delivered IMN dose calculations were performed in this study. However, because of the substantial reduction in geometrical variation, this was deemed unnecessary for the present analysis. Since the four quality improvement initiatives were gradually introduced to clinic in the period between the 2021 cohort and the 2024 cohort, the individual contribution from each initiative could not be determined. In conclusion, sternal positioning of the DIBH marker block and improved IGRT practices resulted in clinically meaningful reductions of intrafractional chest wall motion in left-sided breast cancer patients treated in DIBH. While more than half of the patients reported skin-related side effects, severe skin-related side effects were rarely observed. Funding The project was funded by the Health Research Foundation of Central Denmark Region. CRediT authorship contribution statement Kristine Wiborg Høgsbjerg: Writing – review & editing, Writing – original draft, Validation, Project administration, Investigation, Data curation, Visualization. Anders W. Mølby Nielsen: Writing – review & editing, Writing – original draft, Visualization, Project administration, Investigation, Formal analysis, Data curation, Conceptualization, Validation. Esben Yates: Writing – review & editing, Investigation, Data curation. Christina Maria Lutz: Writing – review & editing, Investigation, Conceptualization. Per Rugaard Poulsen: Writing – review & editing, Supervision, Conceptualization. Birgitte Vrou Offersen: Writing – review & editing, Supervision, Conceptualization. Harald Spejlborg: Writing – review & editing, Writing – original draft, Supervision, Investigation, Formal analysis, Data curation, Conceptualization. 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.