Multimodality therapy, namely chemotherapy and radiotherapy (RT), remains the primary treatment of choice for most children and adolescents with Hodgkin lymphoma (HL). With this approach, over 95% of favorable subgroup patients can be long-term survivors (1). Consequently, this high cure rate warrants efforts to reduce treatment-related late morbidities. With regard to RT, late effects are related to exposure of normal tissues [organs at risk (OARs)]. For instance, mediastinal irradiation has been associated with late cardiovascular (2) and pulmonary toxicity (3), along with increased risks of secondary malignancies, including neoplasms of the thyroid, breast, and lung (4–6).

Despite major RT paradigm shifts to smaller treatment volumes and lower doses, long-term sequelae remain a primary concern when treating children and adolescents with RT, and any possible therapeutic adaptation to decrease the risk of such serious complications becomes paramount in such a curable disease. One such approach that has been relatively understudied involves patient positioning. Specifically, arm positioning for mediastinal RT typically follows one of two patterns: with the patient’s arms raised above the head, or at the patient’s sides, termed the akimbo position. Using these approaches, doses to cardiopulmonary OARs as well as breast tissue have, heretofore, not been characterized; furthermore, in some academic and community practices, arm positioning technique in these patients is simply based on physician preference. This study is the first to date to quantify the impact of arm position on spatial and dosimetric parameters for the heart, lungs, and breast tissue in pediatric HL patients.

There have been multiple high-quality reports published regarding the development of breast cancer after mediastinal RT for HL; a primary theme therein emphasizes the utmost importance of measures to decrease breast doses as low as reasonably achievable (8–12). This is the first known study that quantitates breast doses and spatial parameters based on arm positioning, noting that akimbo positioning decreases breast doses and increases breast spatial separation, particularly in cases without axillary disease. This may be of importance to academic and community practices that may position patients at the discretion of the individual physician, largely owing to a dearth of data such as that reported herein.

It is noteworthy that there were no significant differences in craniocaudal breast position between techniques; rather, by far, the dominant mechanism of movement is in the medial–lateral directions. Hence, though arms raised positioning could be theoretically appealing for patients with appropriately located mediastinal tumors that could pull breast tissue upwards, this is, in fact, not the case. It is possible that the use of a wingboard versus alpha cradle (which may necessitate differential amounts of arm abduction) could change this notion, but there is no evidence for this presently. That cardiopulmonary doses were different (albeit statistically insignificant) between groups warrants further study. Though differences could result from positioning, breathing technique, and anatomical causes, larger sample sizes are required to confirm whether these differences persist.

A notable strength of this work is the relatively homogeneous patient treatment per the contemporary COG AHOD0031 protocol; all patients in this study were treated in the akimbo position. It should be mentioned, however, that the protocol does not allow multiangle three-dimensional conformal radiation therapy, intensity-modulated RT, or proton beam therapy (PBT), contemporary modalities that could theoretically further reduce breast doses compared to the AP/PA technique specified by the COG protocol. Moreover, it is also important to note that these patients were all treated in the era of involved field RT, prior to the widespread institution and acceptance of more limited RT treatment volumes in HL, such as the currently recommended involved site radiation (ISRT) (13). Hence, though our dosimetric results may not be applicable to these modalities/techniques, our spatial analysis of breast/nipple separation distance as a function of arm positioning still applies. Similarly, though no statistical differences in breast dosimetric parameters were observed in patients with mediastinal and axillary diseases, akimbo positioning still offers spatial advantages that could translate into dosimetric numerical differences possibly deducible with larger sample sizes. Similar work in patients treated with other modalities and other subgroups is hence needed as an extension of this study. For instance, larger breast volumes may stand to benefit to a greater degree from akimbo positioning.

The recent rise in PBT to treat not only pediatric cancers but also many other neoplasms (14–18) has necessitated greater data to support its use. Pediatric malignancies are one of the most agreed-upon indications of PBT, and the lower integral dose provided may limit occurrence of secondary, radiation-induced neoplasms. Application to pediatric HL cases is certainly noteworthy, as decreased cardiopulmonary and breast doses with AP PBT beams may aid in achieving lower incidences of late effects to these areas, albeit with virtually no long-term data to date.

The major limitation to our study is the use of involved field planning; recently, involved site RT has become the standard of care in this patient population. However, the AHOD0031 protocol is a contemporary protocol that was published recently in 2014. Many community practices and other countries still use involved field RT, and the results of this study are thus most applicable to these scenarios. The lack of heterogeneity corrections in this protocol is also noteworthy and results in higher estimated lung doses. Next, the finding of non-statistical differences between left and right nipples/centroids to midline may be related to individual positioning within immobilization frames, excursion of the arm during either positioning technique, and/or unforeseen patient rotation. Other limitations include not only the retrospective nature but also the relatively small sample size. Further research is hence needed to corroborate these conclusions, but our statistically significant breast dosimetry findings in patients with mediastinal (±cervical) disease cannot be discounted. Though cervical disease could impact breast doses, the magnitude is likely low, compared with the presence of axillary lymphadenopathy. Our non-uniform use of DIBH is another limitation, but studies have demonstrated that its use results in negligible changes to breast dose (19, 20).

Taken together, this is the first report to quantitatively conclude that arm positioning in pediatric patients with HL undergoing mediastinal irradiation is substantially important to decrease breast doses. Further research using larger sample sizes and modern, conformal treatment modalities would also be of great interest in order to further expand this realm.

KD and CL conceived the project and performed data collection and analysis. VV, AB, and NB wrote and edited the manuscript. All authors approved the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.