Development and Internal Validation of an RPA-Based Model Predictive of Pain Flare Incidence After Spine SBRT

Abstract

The definition of oligometastatic (OM) disease has largely focused on the number of metastatic lesions. Here, we investigate the total volumetric burden of metastases at initial OM presentation as an independent prognostic factor for distant progression free survival (DPFS), widespread progression (WSP, i.e., > 5 new lesions), and overall survival (OS) in patients treated with SBRT. Materials/Methods: Patients with oligometastatic cancer treated with SBRT at 6 international institutions from 2007 − 2016 were retrospectively reviewed. Patients were included if they had 1) five or fewer metastases, 2) no history of brain metastases, 3) curative treatment for primary disease, and 4) SBRT to all known metastases. Multivariable Cox regression models were used to determine the relationship between DPFS and OS with the total planning target volume (PTV) at initial OM presentation. The relationship between WSP and total PTV was determined using multivariable competing risks regression. All models were adjusted for the potential confounders of histology, pre-SBRT systemic therapy, osseousonly vs. visceral lesions, and number of metastases. Results: In total, 961 OM patients were included. The median follow-up was 24.4mo (IQR: 13.8-37.5). The most common primary histologies were NSCLC (25.9%), colorectal (22.0%), prostate (13.3%), and breast (7.9%). The majority of patients (59.4%) had a solitary oligometastasis, while 22.6%, 9.6%, 5.5%, and 2.9% had 2, 3, 4, or 5 lesions, respectively. Twothirds (65.8%) did not have pre-SBRT systemic therapy, and 28.3% had only osseous lesions. The median total PTV was 40.0cc (IQR: 19.7-85.0). The median DPFS, time to WSP, and OS were 15.1, 43.5, and 44.7mo, respectively. Modeling total PTV using restricted cubic splines revealed significant non-linear effects, and log-transformation was optimal to account for this non-linearity. Total PTV had a significant effect on DPFS in the first 18mo after SBRT and this was most profound in the first 6mo, when each doubling of total PTV conferred a 40.6% increased risk of distant progression (P < 0.001). Similarly, each total PTV doubling increased the risk of WSP by 45.4% in the first 6mo (P < 0.001). Total PTV had a significant effect on OS in the first 2yrs after SBRT, with each PTV doubling increasing the risk of death by 60.7% during the first 6mo (P < 0.001) and by 34.6 % thereafter (P < 0.001). Exploratory GTV analysis, where complete data was available for 76.1% of patients due to variability in contouring GTV vs. CTV, confirmed the PTV-based observations. Conclusion: This large multi-institutional series demonstrates the strong impact of total volumetric burden of metastases at initial OM presentation on the risk of distant and widespread progression and OS. We propose that total volume of metastatic lesions may have prognostic merit beyond number of metastases in identifying OM patients who would benefit most from systemic therapy in addition to local metastasis-directed therapy.

Development and validation of a unifying pre-treatment decision tool for intracranial and extracranial metastasis-directed radiotherapy

“…Recursive partitioning analysis (RPA) was conducted with open-source packages in Python (version 3.8.0), with similar methodology to that reported in the literature (25)(26)(27)(28)(29)(30). The overarching training set was used for all modeling efforts.…”