Whole brain radiotherapy (WBRT) is the cornerstone in management of advanced brain metastases as it improves survival and neurological function [1,2]. Recently, indications for stereotactic radiosurgery (SRS) have rapidly expanded, and will continue to do so, but despite these advances many patients still require WBRT [3]. Classically, WBRT was performed using 2-dimensional (2D), parallel opposed lateral (OL) fields (2D-RT) aligned to the patient’s bony anatomy. While providing good coverage to the target volume, this rudimentary technique did not allow for the delineation of specific treatment volumes for targeting and avoidance leading to significant toxicity. The onset of computed tomography (CT)–based, forward-planned 3-dimensional conformal radiation therapy (3D-CRT) in the 1980s made it possible to calculate dosage to internal anatomy, allowing treatment planning to minimize organ at risk (OAR) dose, sparring patients toxicity [4]. The advent of inverse-planning intensity modulated radiation therapy (IMRT) in the 1990s, and further iterations of this technology more recently, resulted in volumetric modulated arc therapy (VMAT) allowing increasingly conformal doses and target delineation to minimize toxicity to OARs [5]. Classically, the median survival for patients receiving WBRT is generally 4–8 months, although this is likely higher in the age of immunotherapy, highlighting the importance of sparing acute toxicity, while being conscientious of end-of-life costs. Additionally, there are a small, but non-insignificant, number of patients who will live long enough to experience late toxicities.

It is well established that toxicities of WBRT include fatigue, alopecia, neurocognitive changes, xerostomia, and dry eye syndrome (DES) [6-14]. RTOG 0933 was a phase II trial that first showed neurocognitive changes could be reduced through the use of hippocampal-avoidant whole brain radiotherapy (HA-WBRT) [15], which has become the standard treatment at our institution and elsewhere for patients with good performance status with multiple brain metastases who are not candidates for SRS [3,16,17]. The ability of VMAT to spare additional OARs including the scalp, ear canals, cochleae, and parotid glands, without greatly extending the duration of treatment has been demonstrated [14]. Additionally, the risk of severe DES after head & neck radiotherapy has been well documented [18,19].

Ideally, every patient with brain metastases requiring WBRT would receive HA-WBRT delivered with IMRT or VMAT. Unfortunately, HA-WBRT is not always possible due to tumor location within 5 mm of the hippocampus, leptomeningeal dissemination, logistical constraints including time to generate a deliverable plan, costs related to treatment, and patient prognostic considerations i.e., median survival less than time to receive benefit from HA-WBRT. Therefore, not all patients are approved for their specific indication. In this subset of patients, treatment with OL 3D-CRT, or 3D optimized dynamic conformal arcs (optDCA) is reasonable, acceptable, and more affordable. The purpose of this paper is to dosimetrically compare four WBRT delivery types: OL, a novel opposed lateral sparing (OLS) 3D-CRT technique involving multi-leaf collimator (MLC) modifications to protect the lacrimal and parotid glands, optDCA to deliver a highly conformal dose and largely spare OARs without the cost and resource requirements of IMRT, and the well described HA-WBRT. Additionally, relative costs and dosimetry time required will be analyzed. We hypothesize that in addition to the parotid glands, the lacrimal glands can also be spared with OLS technique, without compromising planning target volume (PTV) coverage. We further believe that optDCA planning will provide a middle ground between standard OL and HA-WBRT in terms of reducing OAR dosage and costs, without compromising PTV coverage in the delivery of RT for the treatment of brain metastasis. We also believe optDCA plans will be more conformal than OL and OLS 3D-CRT plans.

The primary objective of this study was to dosimetrically compare three different 3D radiotherapy modalities for performing WBRT—static 3D OL planning, OLS, which is a modified static 3D opposed lateral ap­proach in which the lacrimal and parotid glands are spared using minor MLC modifications, and 3D optDCA—and to compare these to HA-WBRT planned with IMRT. Adequate target coverage was achieved with all 3D planning techniques. Lacrimal and parotid dosages can be greatly reduced with the implementation of minor MLC adjustments, without compromising whole brain PTV coverage. OptDCA therapy represented a further improvement over these MLC modifications in terms of OAR dosage without compromising PTV coverage. Here, we demonstrate the superiority of optDCA therapy over other 3D techniques, and show its utility in reducing OAR dose to similar levels of HA-WBRT with greatly reduced cost, but without the ability to spare the hippocampus.

It is now well established that reducing parotid dosage in the palliative setting is of clinical importance. Wang et al. [6] first demonstrated that parotid dosage is clinically correlated to the risk of acute and late xerostomia; specifically, that parotid V20Gy >47% was correlated with worsening xerostomia score at 1 month in patients predominantly undergoing WBRT to 30 Gy in 10 fractions. Furthermore, they showed dosimetrically how simple modifications to MLC placement using OL 3D-CRT could spare parotid dosage, and in theory reduce the risk of xerostomia [6]. Our findings similarly demonstrate a large absolute reduction in parotid V20Gy, as well as parotid D_mean, and D_max with the use of an OLS technique. Furthermore, similar OAR doses were achieved in optDCA plans when compared to HA-WBRT plans, validating the optDCA technique as an excellent tool in clinical situations where IMRT may not be indicated or achievable. The prospective validation of quantitative analysis of normal tissue effects in the clinic criteria (at least one parotid D_mean <20 Gy, or both glands D_mean <25 Gy) by Beetz et al. [27] resulted in <20% risk of xerostomia at 12, 18, and 24 months, still a reasonably high incidence. Based on this data, it is likely that parotid doses as low as reasonably achievable (ALARA) could provide improved clinical benefit. Using OL technique 7 of 20 individual parotid glands were found to have D_mean >25 Gy; all were brought below that threshold using our OLS and optDCA techniques. Another parotid sparing strategy by Park et al. [4] explored the addition of a third anterior-posterior field to reduce dosage in a simple and effective manner; however, their technique showed only a modest reduction in parotid D_mean from 16.2 Gy to 13.7 Gy. Additionally, the plans described here extended inferiorly only to C1. Extension of plans to C2 would lead to an even greater relative benefit to the parotids with the optDCA plans, as the parotid is often directly lateral to C1 and C2, overlying these segments from beam’s eye view.

There is a relative paucity of data on the ability and clinical utility of lacrimal gland sparing in WBRT. A progressive decrease in lacrimal gland dosage was seen from OL to OLS plans, and OLS to optDCA plans that was statistically significant. Multiple authors have demonstrated and advocated for mean and maximum lacrimal gland dosage constraints of <30 Gy and <25 Gy, respectively, to prevent the complication of DES. These constraints were initially extrapolated from parotid recommendations. DES generally develops between 1 month and 3 years after radiotherapy depending on dose and fractionation [12,13]. It has been shown that almost one-third of patients receiving WBRT report DES at 1 month after radiation, and that V20Gy dose >79% is correlated with its development [7]. Another study reported incidence of DES <5% by maintaining lacrimal gland D_mean <30.6 Gy [29]. Therefore, we advocate for OLS, or preferably optDCA planning in all situations where OL planning would otherwise be employed, as it represents a step-function reduction in parotid and lacrimal dosage that may be clinically relevant.

A number of studies have examined the relationship between scalp dosage, treatment modalities, and the development of alopecia [30-32]. We showed the ability to inadvertently spare the scalp using optDCA and HA-WBRT techniques due to the conformity of dosage delivered to the brain, without specific inverse-planning optimization efforts. Witek et al. [30] performed a similar study in which seven patients previously undergoing WBRT with OL fields were re-planned using HA-WBRT with the addition of scalp sparing optimization. They showed the ability to reduce scalp V10Gy and V20Gy by 46% and 35%, and commented on the ability of optDCA therapy to achieve similar reductions in scalp dose without, of course, hippocampal sparing [30]. A study by Mahadevan et al. [32] noted that inadvertent scalp sparing using HA-WBRT was clinically useful for preventing acute alopecia. At doses of 43 Gy, 50% of patients will develop permanent alopecia, and point doses >25 Gy have a 20% of developing permanent alopecia [12]. Therefore, maintaining V25Gy ALARA is clinically relevant.

When sparing the hippocampus is limited by safety, feasibility, or clinical utility, and the patient is not a good candidate for SRS, 3D optDCA represents an excellent and cost-effective option for performing WBRT. An important limitation of this study is its dosimetric nature. While we were able to show improvements across a broad range of OARs and techniques, it is unclear which of these may be clinically relevant. We showed adequate and equivalent target coverage using OL, OLS, and optDCA techniques. Lacrimal and parotid dosages can be greatly reduced with the implementation of minor MLC adjustments. OptDCA therapy represented a further improvement of these modifications in the delivery of WBRT, and was comparable to HA-WBRT in terms of OAR dose, while being about two-thirds the cost, and six times faster for dosimetric planning. Further study is warranted to determine the clinical significance of these findings.