Purpose/Objective(s) Stereotactic body radiotherapy (SBRT) for spine metastases affords excellent local control but may be associated with a dose-dependent risk of treatment induced vertebral compression fracture (VCF). Yet, dosimetric objectives to minimize VCF risk remain uncertain. We present a detailed dose-volume analysis of VCF risk following spine SBRT in a multi-institutional cohort. Materials/Methods Patients with spine metastases treated with SBRT at 10 international institutions between 2009 – 2024 were retrospectively reviewed. Those with prior surgical stabilization or radiation to the same site as SBRT were excluded. VCF was defined as new development or worsening of existing vertebral body height loss not attributable to tumor growth. Analyses were completed at the lesion level. Individual dose volume histograms (DVH) were tabulated for subset analysis. DVH availability varied by treatment era and institution. All fractionation schemes were converted to 3-fraction equivalent dose (3fxED) using the linear quadratic model with an α/β of 3 Gy. Dose-time normal tissue complication probability (NTCP) models were constructed by stratifying data into four dose groups using a k-medians clustering algorithm then performing time-to-event analysis to estimate overall 3-year VCF risk. In 3fxED, bin edges were 10.9 Gy (min), 20.8 Gy, 25.7 Gy, 28.8 Gy, and 37.1 Gy (max). Results 1,659 lesions were analyzed representing 2,203 spinal segments among 1,290 patients. Most were treated in 3 fractions (n = 531, 32.0%). The median prescription doses for 1-, 2-, 3-, and 5-fraction regimens were 18 Gy, 24 Gy, 27 Gy, and 30 Gy, respectively. Common histologies were prostate (16.4%), non-small cell lung cancer (15.3%), and kidney/renal cell carcinoma (15.3%). With median follow up of 38.9 mo (95% CI 35.8 – 40.5), local failure was 17.7% at 3 years (95% CI 15.7 – 19.7%). A total of 236 post-SBRT VCF events were reported, with a 3-year cumulative incidence of 11.7% (95% CI 10.1 – 13.4%). Multivariate logistic regression identified total SINS score (0-6 [reference]; 7-12 [OR 1.89, 95% CI 1.33 – 2.70, p<0.001]; 13-18 [OR 5.03, 95% CI 1.81 – 13.94, p = 0.002) and inclusion of the vertebral body in the planning target volume (PTV) (OR 2.58, 95% CI 1.27 – 5.23, p = 0.009) to be predictive for post-SBRT VCF. DVH data were available for 915 lesions. NTCP modeling suggested a steep response of VCF risk to the minimum dose that 90% of the PTV received (PTV D90%) (p = 0.053). In 3 fractions, PTV D90% of 21.3 Gy, 29.1 Gy, and 31.8 Gy corresponded with a 10%, 25%, and 33% risk of VCF at 3 years, respectively. Conclusion SBRT for de novo spine metastases offers durable local control, and the rate of post-SBRT VCF was low in this large international cohort. NTCP modeling suggests that doses to large subvolumes (PTV D90%) of the spine contributes more to VCF risk than high doses to smaller regions.
Purpose/Objective(s) To evaluate predictors of toxicity following re-irradiation (reRT) for recurrent Grade 3-4 glioma (HGG), including the impact of cumulative dose and treatment related factors. Materials/Methods A multi-institutional retrospective cohort of patients treated with fractionated reRT for recurrent HGG was analyzed. Inclusion criteria were in-field recurrent HGG treated with ≥3 fractions of reRT, completion of reRT and ≥ 6 months between RT courses. Endpoints were grade ≥2 adverse radiation effects (ARE) and a composite grade ≥3 toxicity endpoint including ARE, new neurologic deficits and new or worsening seizures. RT doses were calculated as biological effective doses (BED) using an a/b ratio of 2. Cumulative dose was calculated using an assumption of 25%, 50%, 60% and 70% at 6, 12, 24 and 36 months respectively with no further recovery after 36 months and interpolation of recovery in between time points. Associations between toxicity and cumulative dose, tumor diameter, treatment factors and systemic therapy were evaluated using logistic regression. Results 215 patients in the dataset met inclusion criteria. Of those 215 patients, 61 (28%) developed grade ≥2 ARE and 58 (27%) developed the composite grade ≥3 toxicity measure. Median cumulative RT dose was 134.3 Gy BED2. 20% of patients received proton reRT (PRT), 46% received bevacizumab (BEV), 55% received TMZ and 56% had tumors >5cm in diameter. On multivariable analysis, cumulative dose was not associated with grade ≥2 ARE (OR 1.02 per 10 Gy BED2, p = 0.75); TMZ exposure (OR 1.20, p = 0.62) and a maximal tumor diameter > 5cm (OR 1.08, p = 0.83) were not associated with increased grade ≥2 ARE. PRT (OR 2.98, p = 0.01) and BEV exposure (OR 2.24, p=0.03) were associated with higher rates of ≥ grade 2 ARE. No factors were associated with the grade ≥3 composite endpoint on multivariate analysis. Patients receiving PRT had similar median survival (13 vs 10 months, p = 0.067), a longer time between RT courses (median 55 vs 22 months, p < 0.01), were less likely to have a tumor > 5 gt; 5 cm at reRT (37% vs 61%, p < 0.01), and receive TMZ with reRT (28% vs 49%, p = 0.02) and had similar rates of reRT regimens greater than 35 Gy / 10 fraction equivalent (12% vs 5%, p = 0.18). Patients receiving BEV with reRT had a shorter time between RT courses (median 11 vs 34 months, p < 0.01) and were less likely to receive TMZ with reRT (27% with TMZ vs 59% with BEV, p < 0.01). Conclusion Cumulative reRT dose was not associated with clinically significant toxicity with adjustments for tissue recovery after initial radiation. PRT and BEV were associated with increased grade ≥2 ARE. While these findings may reflect patient and treatment selection factors, PRT patients were more likely to have multiple clinical factors previously reported to be associated with decreased ARE. Prospective studies are needed to better define patient specific toxicity risk after reRT to guide reRT dose, use of PRT and BEV.
Purpose/Objective(s) Preoperative (preop) stereotactic radiosurgery (SRS) is a feasible alternative to postoperative (postop) SRS with potential benefits in adverse radiation effect and meningeal disease. The goal of this study was to determine if there is an association between the cancer primary site and oncologic outcomes after preop SRS for resected brain metastases (BM). Materials/Methods Patients with BM from solid cancers, of which at least 1 lesion was treated with preop SRS and underwent planned resection were included from 8 institutions. SRS to synchronous intact BM was allowed. Exclusion criteria included classically radiosensitive or non-solid cancers (including small cell lung cancer) and prior or planned whole brain radiotherapy (WBRT). Cancer primary site was categorized into 6 groups: non-small cell lung cancer (NSCLC), breast, melanoma, kidney, gastrointestinal (GI), and other. Intracranial outcomes were estimated using cumulative incidence with competing risk of death. Results The cohort consisted of 404 patients with 416 preop SRS treated index lesions. Primary site distribution was NSCLC for 192 patients (48%), breast 62 (15%), melanoma 50 (12%), kidney 29 (7%), GI 28 (7%), and other 43 (11%), respectively. The other category included predominantly gynecologic, sarcoma, and bladder/urothelial primary sites. Median interval between preop SRS and surgery was 2 days. Median cranial imaging follow-up interval was 9.5 months overall and 17.5 months for alive patients. There were baseline imbalances between primary site groups for age, graded prognostic assessment score, median index tumor volume, gender, and posterior fossa location. Univariate analyses results at 2-years by primary site are included in the table below. In multivariate analysis controlling for other prognostic factors, primary site was found to have significant association with risk of local recurrence (LR), distant brain failure (DBF), meningeal disease (MD), adverse radiation effect (ARE), and overall survival (OS). Relative to NSCLC, GI primary site was associated with significantly increased risk of LR; melanoma and kidney primaries with increased risk of DBF; breast, melanoma, and other primary sites with increased risk of MD; and melanoma with increased risk of ARE. Melanoma and kidney primary sites were associated with significantly improved OS compared to NSCLC. Conclusion Intracranial tumor control, ARE, and overall survival after preoperative SRS for resected brain metastases are highly influenced by cancer primary site in both univariate and multivariate analyses. Some of these findings may be related to inherent differences in primary site biology and/or differences in systemic therapy options and efficacy by cancer primary site. Additional confirmatory studies are needed.
Purpose/Objective(s) The three-arm randomized NRG/RTOG 0534 SPPORT trial (n=1792) examined treatment intensification on the outcome of men treated with salvage radiotherapy (SRT) for a detectable PSA. The arms were prostate bed RT (PBRT) alone (Arm 1), PBRT + short term androgen deprivation therapy (STADT; Arm 2), and PBRT + STADT + pelvic lymph node RT (PLNRT; Arm 3). With treatment intensification in Arms 2 and 3, there were significant incremental gains in the freedom from progression (FFP) endpoint, but not metastasis free-survival (MFS) with 8 yr median follow-up; although metastatic events were reduced with intensification. The PAM50 genomic classifier categorizes prostate cancers into luminal A (LumA), luminal B (LumB) and basal cell subtypes with LumB tumors found to benefit most from postoperative ADT. We hypothesized that PAM50 LumB tumors are a predictive biomarker of response to postoperative ADT. Materials/Methods Prospectively collected prostatectomy tissue was profiled for clinical-grade whole-transcriptomes from samples passing quality control to assign PAM50 LumB (vs not-LumB) subtype. This National Clinical Trials Network (NCTN)-approved integrated biomarker pre-specified statistical plan included the secondary objective of validating the predictive nature of PAM50 LumB for ADT use. The protocol FFP endpoint included biochemical (nadir+2 ng/mL) failure, clinical failure, or death from any cause. MFS included distant metastasis or death from any cause. Multivariable (MVA) Cox models adjusted for age, treatment arm, race, seminal vesicle involvement, Gleason grade group (1-3 vs 4-5), pre-SRT PSA and stage. Results PAM50 subtype scores were obtained for 709 patients (median follow-up 7.9 yr), with 215 in Arm 1, 247 in Arm 2, and 247 in Arm 3. The arms were balanced for key covariates. There were 226 FFP and 136 MFS events. On MVA, the PAM50 LumB subtype was prognostic for FFP (HR 1.33, 95% CI 1.02-1.74, p = 0.03) and MFS (HR 1.52, 95% CI 1.08-2.14, p = 0.02). Adding STADT ± PLNRT to PBRT resulted in a greater benefit in patients with PAM50 LumB subtype than non-LumB subtype, with a 5-yr FFP absolute benefit of 31% vs. 10% in PAM50 LumB vs non-LumB patients, respectively, with a significant treatment interaction (p-int = 0.01). In patients treated with STADT ± PLNRT + PBRT, the 10-yr MFS absolute benefit was 20% vs -3% in PAM50 LumB v non-LumB patients, respectively, with a significant treatment interaction (p-int = 0.05). Conclusion In this randomized phase 3 trial, PAM50 LumB subtype was both prognostic and predictive of benefit from postoperative ADT-based treatment intensification. LumB tumors have increased proliferative and androgen receptor signaling, providing a mechanistic rationale for their preferential response to ADT. These findings support molecular subtyping to personalize salvage radiotherapy strategies for localized prostate cancer.
: Interdisciplinary science and environmental management involve bringing together data and expertise at multiple spatial scales. The most challenging part of merging scales is aligning the scale of inquiry with the research application. Through the Louisiana case study relating wetland loss and commercial fishing, we examine how the nature and strength of the relationship changes depending on the scale of investigation. Resulting management implications also vary because of tradeoffs in choosing the scale of inquiry. State-level fisheries managers may miss effects of wetland loss in fishing communities because they are looking at aggregate data. Scientific information must directly address the constituent scale, where managers can enact policy. The case study demonstrates why scalar considerations should be an explicit part of the planning process for both science and management.