This retrospective single-institution study analyzed our experience with more complex and resource consuming stereotactic approaches for radiotherapy of bone metastases. We were interested in patient selection, treatment completion, radiotherapy very close to the end of life (RT30), and overall survival. Other endpoints such as pain relief, local control, fracture rates, toxicity and quality of life were not evaluated. However, recent reviews and meta-analyses of prospective clinical studies have already provided information about these endpoints [2, 5, 6, 10]. When introducing stereotactic radiotherapy for bone metastases in 2017, an active research program was already running, focusing on improved patient selection for palliative radiotherapy and development of better decision support tools. More recently, these efforts led to recommendation of the LabBM score [15], followed by the LabPS score [16]. In addition, a model named palliative appropriateness criteria score (PACS) contributes by estimating the proportion of remaining lifetime spent on radiotherapy [20, 21]. During this work-in-progress phase, we did not feel sufficiently confident to implement one of the scores as obligatory requirement for decision making. We did, however, try to select patients with poor prognostic features for short-course non-stereotactic treatment, especially 8 Gy × 1, by assessing KPS, disease extent, remaining systemic therapy options and blood test results.

The comparison of baseline parameters (Table 1) confirmed that major differences in prognostic factors were present. Obviously, patients were assigned to single-fraction treatment if adverse factors were found and to multiple fractions if expected survival was longer, with some individual variation and sometimes imperfect match between survival and time on radiotherapy. Fewer stereotactically treated patients received opioid analgesics and inpatient radiotherapy. They were also less likely to harbor progressive disease sites that were not included in the actual course of radiotherapy. KPS was different too (best in patients with fractionated stereotactic radiotherapy). Comparable differences were seen for the LabBM score, i.e. best prognosis in the stereotactic fractionated group. Better patient selection resulted in lower rates of RT30. There is little, if any, potential for further improvement of RT30 rates in our stereotactic radiotherapy groups. However, adhering to thorough prognostic assessment is expected to improve quality of care in the non-stereotactic groups by lowering RT30 rates in the future. The policy in place during the time period of the study resulted in stable rates of RT30 over time. Given the group’s research focus on prognostic models, this lack of positive development is surprising. On the other hand, offering 8 Gy × 1 is neither very resource-consuming nor toxic, and is more polite and supportive than refusing radiotherapy at all. This does not automatically justify the decision, however, many clinicians may prefer treatment, especially if they have limited training in palliative care and the patient’s appointment is radiotherapy-specific.

Gillespie et al. studied almost 3000 patients who received radiotherapy for bone metastases at a different institution in the 3-year time period between 2016 and 2018 [7]. Their analysis included close to 6000 radiotherapy episodes. SBRT was frequently employed (n = 2790, 47%), while 8 Gy × 1 was infrequent (n = 368, 6%). On multivariate logistic regression, factors associated with receipt of SBRT were high KPS, a non-radiosensitive primary tumor histology, and location in spine. Death within 30 days occurred among 24% of all 8 Gy × 1 treatments (19% in our study), 9% of 3 Gy × 10 courses, and 4% of all SBRT treatments (p < 0.01). Outside of our center, the LabBM score has so far not been studied specifically in the bone metastases setting. Our initial experience in this large cohort supports further validation by other groups.

Kowalchuk et al. developed and validated a decision-making tool predictive of overall survival for patients receiving SBRT for spinal metastases (single or multiple fractions), i.e. a subgroup of all bone SBRT [22]. Three hundred sixty-one patients at one institution were used for the training set, and 182 at a second institution were used for external validation. The final model consisted of the following variables and scores: Spinal Instability Neoplastic Score (SINS) ≥ 6 (1), time from primary diagnosis < 21 months (1), Eastern Cooperative Oncology Group (ECOG) PS = 1 (1) or ECOG PS > 1 (2), and > 1 organ system involved (1). Each variable was an independent predictor of survival (p < 0.001), and each 1-point increase in the score was associated with a hazard ratio of 2.0. Three groups were defined: favorable (0–1 points), intermediate (2 points), and poor survival (3–5 points), with 2-year survival rates of 84, 46 and 21%, respectively (p < 0.0001 for each). In the external validation set, the score was also predictive of survival. Just like Gillespie et al. [7], Kowalchuk et al. did not include blood test results. The same is true for Zeng et al. [23], who also focused on spine SBRT and prognostic factors associated with surviving less than 3 months versus greater than 3 years, meaning that different endpoints were selected in all studies.

At our center, median overall survival was 6.4 months after stereotactic single-fraction irradiation. Largely identical results were reported by Nguyen et al. (median 6.7 months in both arms of their randomized study (comparator: conventional multi-fraction irradiation) [9]. In our study, the corresponding figure was 8.2 months for conventional multi-fraction treatment. Multivariately, neither fractionation nor stereotactic treatment was associated with improved survival. Propensity score matching also confirmed similar survival outcomes. Therefore, if the primary aim of treatment is symptom palliation, priority is given to simple short-course regimens, which avoid toxicity and may be applied in a fast-track setting. Single-fraction or 2–3 fraction stereotactic radiotherapy is a promising alternative, but less easy to plan and deliver. In addition, in a larger randomized clinical trial than the one reported by Nguyen et al. superiority of stereotactic single-fraction irradiation for the primary end point of patient-reported pain response at 3 months was not found [24]. These researchers studied single-fraction doses of 16 or 18 Gy versus a conventional 8-Gy regimen, limited to spine metastases.

The present study has both strengths, such as the routine clinical practice setting and availability of blood test results, and limitations, such as the moderate number of stereotactic radiotherapy episodes, lack of other outcome data, and limited follow-up if treatment was administered after 2021. There is still debate about complete avoidance of palliative radiotherapy in the last month of life. Recently, Christ et al. reported that radiotherapy achieved high completion and success rates until one week before death, and suggested that treatment within one week of death should be restricted to carefully selected patients or avoided altogether [25]. In a study by Rautakorpi et al., treatment was discontinued in 41% of the patients irradiated during the last two weeks of life, and worsening of the general condition was the prevailing reason for discontinuation (70%) [26]. Despite controversy, RT30 has remained the most common quality of care indicator [27,28,29]. In a limited number of bone metastases treatment episodes (< 300), we recently identified three significant predictors of 30-day mortality (KPS (≤ 50, 60–70, 80–100), weight loss of at least 10% within 6 months (yes/no), pleural effusion (present/absent)) and employed these to construct a predictive model with 5 strata and mortality rates of 0–75% [17]. Future validation in a larger study with additional potential predictors is planned.