Risk-adapted SBRT () for recurrent lung-only metastases in patients with frailty or ultracentral anatomy Authors: Ismaell Massalha, Mhammad Abu-Juda, Reem Zabit, Nitzan Sagie, Kim Sheva, Leonid Bogolmoni, Olga Belochitski, Ofir Cohen, Amichay Meirovitz, Konstantin Lavrenkov Published: 2026 ABSTRACT AbstractPurposeStereotactic body radiation therapy (SBRT) is widely used for pulmonary oligometastatic disease. Guideline-based regimens target a biologically effective dose ; However, this threshold may not be achievable in patients with limited physiological reserve or ultracentral lesions. We examined outcomes after SBRT with versus in a consecutive cohort of patients with recurrent lung-only metastases.Methods and materialsConsecutive patients treated with SBRT for recurrent lung-only metastases between January 2019 and December 2022 were retrospectively reviewed. Primary endpoints were overall survival (OS), progression-free survival (PFS), and freedom from local–regional progression (FFLP). Kaplan–Meier estimates were compared using log-rank tests, and univariable Cox proportional hazards models estimated hazard ratios (HRs) with 95% confidence intervals (CIs). A propensity-score inverse probability of treatment weighting (IPTW) analysis adjusted for age, Eastern Cooperative Oncology Group (ECOG) performance status, Charlson Comorbidity Index (CCI), lung primary status, and number of metastases.ResultsFifty-three patients were included. Median OS was 47.3 months for versus 34.3 months for (HR 1.45; 95% CI 0.66–3.16; p=0.35). Median PFS was 22.8 months versus 14.7 months (HR 1.59; 95% CI 0.82–3.11; p=0.17). For FFLP, the HR was 2.67 (95% CI 0.60–11.93; p=0.20). In IPTW-weighted models, was not significantly associated with OS (HR 0.58; 95% CI 0.25–1.32; p=0.19) or PFS (HR 0.59; 95% CI 0.24–1.40; p=0.23).ConclusionIn this retrospective cohort, was delivered to selected higher-risk patients on the basis of frailty, ultracentral anatomy, or both. Differences in OS, PFS, and FFLP compared with did not reach statistical significance, and IPTW-weighted estimates were directionally consistent. These observations are consistent with the feasibility of sub-ablative SBRT in selected higher-risk patients when ablative dosing is not safely deliverable, and they define the patient population and dose parameters that warrant prospective, risk-stratified evaluation.Supplementary InformationThe online version contains supplementary material available at 10.1186/s13014-026-02864-3.Keywords: Stereotactic body radiotherapy, Oligometastatic disease, Biologically effective dose, Hypofractionation, Lung metastases, Dose de-escalation, Frailty, Ultracentral Introduction Oligometastatic disease denotes limited metastatic burden that may be amenable to metastasis-directed therapy. Randomized and prospective studies have established that local consolidative treatment improves PFS and, in some settings, OS [3, 5, 6]. SBRT is a standard non-invasive ablative approach for pulmonary metastases and early-stage lung tumors. Local control after SBRT for pulmonary metastases is dose-dependent. Consensus guidelines advocate when feasible [4], but this target is not universally attainable. Poor performance status and high comorbidity burden increase competing mortality risks and may limit tolerance of ablative doses. Geriatric screening tools, including the G8 score and CCI, have been associated with survival and toxicity following SBRT for early-stage lung cancer [1, 2, 7]. Ultracentral tumors present a separate constraint: proximity to proximal bronchial structures and major vessels elevates the risk of severe toxicity at ablative doses, sometimes precluding on anatomic grounds alone. The outcomes of SBRT delivered at in these situations remain poorly characterized. Two distinct, non-equivalent indications drive this prescription pattern in clinical practice: limited physiological reserve (frailty) and ultracentral anatomy. Each shifts the risk–benefit calculation through a different mechanism. We report a single-institution retrospective series describing how was used in a consecutive cohort with these indications, and we compare outcomes with , recognizing that dose selection reflected clinical and anatomic constraints rather than random assignment. Methods and materials Study design and patient selection Consecutive patients treated with SBRT for recurrent lung-only metastases between January 2019 and December 2022 at a single institution were included. Eligibility required lung-only recurrence confirmed by contemporaneous CT and/or PET-CT without extrapulmonary disease. Consistent with oligometastatic criteria, patients had up to four lung lesions. SBRT was delivered to all known active lesions at the time of treatment. The study was approved by the institutional review board (IRB-2021–86) with waiver of informed consent. Definitions: recurrent lung-only metastases, frailty, and ultracentral location “Recurrent lung-only metastases” denotes metachronous lung-limited recurrence after definitive treatment of the primary malignancy (surgery and/or radiotherapy), without concurrent extrapulmonary disease at SBRT. Synchronous presentations, metastatic disease identified within three months of primary diagnosis, were excluded. Frailty was operationalized a priori from routinely documented clinical surrogates at SBRT consultation: ECOG performance status ≥3 and/or CCI ≥8. The cross-tabulation of frailty status against group, and the indication breakdown for each prescription (frailty alone, ultracentral anatomy alone, both, or neither), are presented in Supplementary Table S1. Ultracentral location was defined as tumor proximity to critical central thoracic structures necessitating organ-at-risk constraint prioritization during planning. Radiotherapy planning and delivery Four-dimensional CT simulation was performed at 2-mm slice thickness. An internal gross tumor volume (iGTV) was expanded with institution-standard margins to generate the planning target volume (PTV); Treatment was delivered using volumetric modulated arc therapy (VMAT). Dose and fractionation were physician-selected based on anatomy, lesion location, and clinical patient factors. was calculated with ; patients were categorized as or (Table 1). Table 1Distribution of prescribed SBRT total doses, fractionation schedules, and corresponding Total Dose (Gy)Fractions (Gy)Patients, n (%)56895.218 (34%)60810524 (45%)60513211 (21%)Patients with multiple lung metastases received identical prescription doses across lesions. Variability in BED reflected individualized clinical decision-making that incorporated tumor location, performance status, primary tumor type, and comorbidity burden. Abbreviation: BED = biologically effective dose; Gy = Gray Distribution of prescribed SBRT total doses, fractionation schedules, and corresponding Patients with multiple lung metastases received identical prescription doses across lesions. Variability in BED reflected individualized clinical decision-making that incorporated tumor location, performance status, primary tumor type, and comorbidity burden. Abbreviation: BED = biologically effective dose; Gy = Gray was used when organ-at-risk (OAR) constraints and/or limited physiological reserve precluded ablative regimens. Limited physiological reserve was defined by frailty criteria (ECOG performance status ≥3 and/or CCI ≥8). High-risk anatomy was primarily defined by ultracentral location, in which OAR constraints governed the prescription. A representative plan is shown in Fig. 1. Representative SBRT plan demonstrating conformal 95% isodose coverage of the planning target volume (A–C) and peripheral dose fall-off at the 60% isodose level (D–F). This example illustrates standard institutional planning principles applied across BED regimens Follow-up and endpoints Cross-sectional imaging follow-up was performed at institution-standard intervals. OS was measured from SBRT start to death from any cause or last follow-up; PFS from SBRT start to progression (local, regional, or distant) or death; FFLP was defined as the absence progressive disease within or adjacent to the treated volume and/or regional nodal stations. Adverse events were graded per the Common Terminology Criteria for Adverse Events (CTCAE), version 5.0. Statistical analysis OS, PFS, and FFLP were estimated by Kaplan–Meier methodology and compared between BED groups by log-rank tests. Univariable Cox proportional hazards models provided HRs and 95% CIs. The proportional hazards assumption was not formally tested given the limited event counts and therefore, Cox model results should be interpreted accordingly. To partially address confounding by indication, a propensity score analysis was performed using IPTW. The propensity score was estimated by logistic regression with BED group as the outcome and age, ECOG performance status, CCI, lung primary status, and number of metastases as covariates. Weights were stabilized and truncated at the 1st and 99th percentiles. Weighted Cox models for OS and PFS are presented as exploratory sensitivity analyses; they do not eliminate residual confounding. Results Patient characteristics Fifty-three patients were included. Median age was 71 years (IQR 65–77); 34 patients (64%) were male. ECOG performance status was 1 in 19 (36%), 2 in 13 (25%), and 3 in 21 (40%). Characteristics stratified by frailty status are provided in Supplementary Table S1. Single-lesion disease was present in 40 patients (75%). All cases met the metachronous recurrence definition; no synchronous presentations were included. The median interval from primary diagnosis to SBRT was 11 months (IQR 6–26) in the group versus 24 months (IQR 13–37) in the group (p=0.084). Among patients with evaluable recurrence data (n=26), distant failure was most frequent (73%), followed by regional recurrence (15%) and local recurrence (12%). Median time from prior SBRT to recurrence was 9 months (IQR 6–16). Smoking history was present in 64%. The CCI distribution was concentrated, with an overall median of 6 (IQR 6–7), although values extended to 12, reflecting a subset with substantial comorbidity burden. Baseline characteristics by group Eighteen patients (34%) received and 35 (66%) received . Among the 18 patients prescribed , 12 (67%) met frailty criteria alone, 4 (22%) met both frailty and ultracentral anatomy criteria, 1 (5.6%) had an ultracentral lesion without meeting frailty criteria, and 1 (5.6%) received reduced dose for other organ-at-risk constraints; the cross-tabulation of frailty status by group is presented in Supplementary Table S1. Baseline characteristics stratified by BED group are presented in Table 2. Table 2Patient demographics, baseline clinical characteristics, and tumor locations stratified by groupCharacteristicOverall (N = 53) (n = 35) (n = 18)p-valueAge (years), median (IQR)71 (65, 77)72 (66, 77)69 (63, 78)0.452Gender0.380 Female19 (36%)11 (31%)8 (44%) Male34 (64%)24 (69%)10 (56%)Smoking status0.143 Non-smoker19 (36%)10 (29%)9 (50%) Smoker34 (64%)25 (71%)9 (50%)ECOG performance status < 0.001 119 (36%)18 (51%)1 (5.6%) 213 (25%)10 (29%)3 (17%) 321 (40%)7 (20%)14 (78%)Charlson Comorbidity Index0.578 Median (IQR)6 (6, 7)6 (4, 9)7 (4, 12)Primary tumor site0.049 Lung33 (62%)25 (71%)8 (44%) Gastrointestinal tract11 (21%)6 (17%)5 (28%) Renal cell carcinoma9 (17%)4 (11%)5 (28%)Number of lung metastases0.760 140 (75%)25 (71%)15 (83%) 210 (19%)7 (20%)3 (17%) 31 (1.9%)1 (2.9%)0 42 (3.8%)2 (5.7%)0Tumor location (per lesion, n = 71) > 0.999 Peripheral43 (61%)36 (72%)7 (33%) Central20 (28%)11 (21%)9 (44%) Ultracentral8 (11%)3 (7%)5 (22%)Abbreviations: ECOG = Eastern Cooperative Oncology Group; BED = biologically effective dose Patient demographics, baseline clinical characteristics, and tumor locations stratified by group Abbreviations: ECOG = Eastern Cooperative Oncology Group; BED = biologically effective dose ECOG distribution differed markedly between groups (p < 0.001): 78% of patients had ECOG 3, compared with 20% in the higher-dose group. Primary tumor site also differed (p=0.049), with a higher proportion of non-lung primaries in the cohort. Age (p=0.452), smoking status (p=0.143), CCI (p=0.578), number of metastases (p=0.760), and tumor location (p > 0.999) showed no significant between-group differences. Ultracentral lesions were more frequent in the group (22% vs. 7% of lesions, respectively). The ECOG imbalance constitutes pronounced confounding by indication; survival comparisons between groups are strictly descriptive. Overall survival and progression-free survival Univariable Cox models found no statistically significant difference in OS, PFS, or FFLP between the dose groups, though point estimates favored throughout. Median OS was 47.3 months in the group versus 34.3 months in the group (HR 1.45; 95% CI 0.66–3.16; p=0.35; Figure 2). Median PFS was 22.8 versus 14.7 months, respectively (HR 1.59; 95% CI 0.82–3.11; p=0.17; Figure 3). For FFLP, the HR was 2.67 (95% CI 0.60–11.93; p=0.20), with the wide interval reflecting the small number of local–regional events (Fig. 4). Overall survival (OS) stratified by group ( vs. ) Progression-free survival (PFS) stratified by group ( vs. ) Freedom from local–regional progression (FFLP) stratified by group Patterns of recurrence Recurrence patterns by group are summarized in Fig. 5. Distant progression was the dominant mode of failure in both groups. Local–regional events were too infrequent to permit meaningful between-group comparisons. Recurrence patterns by group (absolute counts) IPTW sensitivity analysis In IPTW-weighted Cox models adjusting for age, ECOG performance status, CCI, lung primary status, and number of metastases, was not significantly associated with OS (HR 0.58; 95% CI 0.25–1.32; p=0.19) or PFS (HR 0.59; 95% CI 0.24–1.40; p=0.23). Directionality was consistent with the unadjusted estimates. Treatment-related toxicity No grade 3 or higher treatment-related events occurred in the group. Five patients in the group experienced grade 3 toxicity: two cases of grade 3 pneumonitis and one of grade 3 esophagitis occurred within 90 days of treatment; two additional cases of grade 3 esophagitis were identified beyond 90 days. All resolved with conservative medical management. No grade 4 or grade 5 toxicities were observed. Discussion The central finding is straightforward: in this retrospective cohort, characterized by marked performance status imbalance between dose groups, was not associated with statistically significantly worse OS or PFS compared with . Confidence intervals were wide and confounding by indication was substantial, as lower-dose patients had poorer baseline status. These data are descriptive and should not be read as establishing the safety profile of dose de-escalation; the absence of grade 4 or grade 5 events in either group is, however, reassuring. These findings reflect routine clinical decision-making. When ablative regimens were not feasible because of frailty, ultracentral anatomy, or both, was prescribed. Median survival differed numerically between groups; however, this difference cannot be attributed to dose given the substantial baseline imbalance. Frailty and ultracentral anatomy operate through distinct mechanisms: frailty reflects limited physiological reserve and competing-risk mortality, whereas ultracentral location imposes anatomic dose-limiting constraints on organs at risk. The two indications were not biologically equivalent, and their relative contributions could not be analytically separated at this sample size. The cohort therefore represents a mixed-indication real-world population. The geriatric oncology literature provides relevant context. CCI and G8 score have been associated with survival and late toxicity following thoracic SBRT in older patients [1, 2, 7]. Formal geriatric assessment was not performed systematically in this cohort; ECOG performance status and CCI served as pragmatic surrogates. The concentration of grade 3 toxicity in the group, five events versus none, is consistent with higher-dose regimens carrying greater morbidity risk, though the groups were not comparable at baseline and this observation is descriptive. Several limitations warrant consideration. The cohort is modest in size, and the analyses are hypothesis-generating rather than confirmatory. Confounding by indication is intrinsic to this design: patients receiving were preferentially those with limited physiological reserve or anatomic dose-limiting constraints. The IPTW analysis, presented as a supportive sensitivity analysis, mitigates but does not eliminate this imbalance, and the directional consistency of IPTW-weighted estimates with the unadjusted results is reassuring. FFLP event counts are low and widen the confidence intervals for that endpoint. Heterogeneity in primary tumor type and in systemic therapy before and after SBRT (Supplementary Tables S2 and S3) reflects real-world practice and contributes additional variability. The single-institution design supports internal consistency in planning and follow-up at the cost of external generalizability. Within these constraints, the dataset captures a clinically meaningful real-world question that the published literature has scarcely addressed. Prospective work addressing this question would require pre-specified frailty and anatomic risk strata, standardized toxicity capture, and competing-risk–appropriate endpoints. The present data do not support specific clinical recommendations; rather, they define the population and dose parameters for such a trial. Conclusions In this retrospective cohort, was used as a risk-adapted prescription in patients with frailty, ultracentral anatomy, or both. Differences in OS, PFS, and FFLP between the two BED groups did not reach statistical significance, and IPTW-weighted estimates were directionally consistent. These observations are consistent with the feasibility of sub-ablative SBRT in selected higher-risk patients when ablative dosing is not safely deliverable, and they define the patient population and dose parameters that warrant prospective, risk-stratified evaluation. Electronic supplementary material Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements We thank Prof. David Geffen and Dr. Taha Massalha for their useful discussions. Abbreviations Biologically effective dose ( Gy) Charlson Comorbidity Index Eastern Cooperative Oncology Group Freedom from local–regional progression Internal gross tumor volume Overall survival Progression-free survival Planning target volume Stereotactic body radiotherapy Volumetric modulated arc therapy Author contributions I.M. conceived the study, designed the methodology, supervised data collection, interpreted the results, and drafted the manuscript. N.S., O.C. performed the statistical analyses, including propensity score and IPTW modeling, and contributed to interpretation of the data. M.A.J., R.Z., K.S., L.B., and O.B. contributed to patient identification, clinical data acquisition, and verification of treatment details. A.M. and K.L. provided senior clinical supervision and critically revised the manuscript for important intellectual content. All authors reviewed and approved the final manuscript and agree to be accountable for all aspects of the work. Funding No funding was received for this study. Data availability The datasets used during the current study are available from the corresponding author on reasonable request. Declarations Ethics approval and consent to participate The study protocol was approved by the Institutional Review Board of Soroka Medical Center (IRB-2021–86). Given the retrospective design, informed consent was waived. Consent for publication Not applicable. Declaration of generative AI and AI-assisted technologies in the writing process During the preparation of this work the authors used Grammarly to correct wording and grammar. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Konstantin Lavrenkov, Ofir Cohen and Amichay Meirovitz contributed equally to this work. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement The datasets used during the current study are available from the corresponding author on reasonable request.