Algorithm Assignment to Stereotactic Body Radiation Therapy Is Associated with High Risk of Post-Surgical Serious Adverse Events in Stage I Non-small Cell Lung Cancer Patients

Abstract

For clinical Stage I non-small cell lung cancer (NSCLC) patients, identifying patients at increased risk for adverse events from surgical intervention, and thus more appropriate for stereotactic body radiation therapy (SBRT), remains a clinical challenge. Using a prospective treatment allocation model, we evaluated whether serious surgical events could identify high risk patients prior to treatment selection. From 2016 – 2018, clinical Stage I NSCLC patients receiving treatment in either the Department of Radiation Oncology or Division of Thoracic Surgery were invited to participate in a prospectively maintained database that included physiologic, pathologic, and functional status metrics. A multivariable logistic regression model was made to characterize the likelihood of SBRT receipt for all participating patients, and used to generate a predicted group value (allocation to surgery versus SBRT). From this, ≥ Grade 3 Common Terminology Criteria for Adverse Events (CTCAE) were compared among those surgical patients that the model allocated to surgery (concordant), versus those assigned to SBRT (discordant). In this prospective database, 104 (53%) received SBRT, 91 (47%) received surgery. Variables independently associated with allocation to SBRT included current smoking status (Odds Ratio 8.2, 95% CI 1.5 – 44.7), Global Initiative for Chronic Obstructive Lung Disease (GOLD) score ≥2, (OR 2.7, 95% CI 1.1 – 6.6), increasing 15-foot walk test time (per second increase OR 1.7, 95% CI 1.2 – 2.5), and ECOG ≥1 (OR 8.0, 95% CI 3.0 – 21.4). Of the 73 (80%) surgical patients with complete data for model allocation, 60 patients (82%) were predicted to receive surgery and 13 (18%) were predicted to receive SBRT. The rate of ≥Grade 3 CTCAE among discordantly assigned surgical patients was significantly higher than concordantly assigned surgical patients, 6/13 (46.2%) versus 10/60 (16.7%) respectively, p=0.03. Discordant allocation to SBRT was significantly associated with experiencing a ≥Grade 3 CTCAE following surgery (OR 4.3, 95% CI 1.2 – 15.5). Prospectively collected patient information that includes baseline functional and performance status metrics may assist in identifying surgical patients that are high-risk for ≥ Grade 3 CTCAE events. Identifying these patients at the time of consultation may assist in consideration for SBRT and shared-decision making.