Main Session
Sep
28
PQA 04 - Breast Cancer, Patient Reported Outcomes/QoL/Survivorship, Functional Radiation Medicine, Hematologic Malignancies, Palliative Care, and International/Global Oncology
2892 - Glasgow Prognostic Score as a Strong Predictor of Early Mortality in Real-World Palliative Radiation Therapy for Stage IV Lung Cancer
Presenter(s)
Hidekazu Tanaka, MD, PhD - Yamaguchi University Graduate School of Medicine, Ube, Yamaguchi
H. Tanaka, N. Yasuda, K. Aga, K. Ueda, T. Ono, Y. Manabe, M. Kajima, K. Fujimoto, and T. Shiinoki; Yamaguchi University Graduate School of Medicine, Department of Radiation Oncology, Ube, Japan
Purpose/Objective(s):
In palliative radiation therapy (RT), performance status (PS) guides fractionation but remains subjective. The Glasgow Prognostic Score (GPS), reflecting systemic inflammation and nutritional status, is an established prognostic marker across malignancies. We hypothesized that combining GPS with PS would strongly predict early mortality and allow identification of patients at exceptionally high risk of 3-month mortality, for whom prolonged fractionation may be unsuitable.Materials/Methods:
We evaluated 172 patients with stage IV lung cancer who received palliative RT: 133 men (77.3%) and 39 women (22.7%), median age 71. Irradiated sites were bone (n=84, 48.8%), brain (n=56, 32.6%), and others (n=32, 18.6%). Post-RT systemic therapy was administered in 116 patients (67.4%). PS 0–4 was observed in 27 (15.7%), 76 (44.2%), 44 (25.6%), 21 (12.2%), and 4 (2.3%) patients. GPS was defined as: 0=CRP=1.0 mg/dL and albumin=3.5 mg/dL; 1=CRP>1.0 or albumin<3.5; 2=CRP>1.0 and albumin<3.5. GPS 0,1,2 were seen in 64 (37.2%), 44 (25.6%), and 64 (37.2%). Three-month mortality was compared across GPS groups. Predictive accuracy of GPS (0–1 vs.2) was assessed by ROC analysis. Multivariable logistic regression examined predictors of 3-month mortality, including BMI, GPS, PS (0–1 vs.2–4), irradiated site (bone vs.non-bone), histology (NSCLC vs.SCLC), age, and post-RT systemic therapy. Risk groups were created by PS (0–1 vs.2–4) and GPS (0–1 vs.2): Group1 (PS0–1&GPS0–1), Group2 (PS2–4&GPS0–1), Group3 (PS0–1&GPS2), Group4 (PS2–4&GPS2). Survival was analyzed using Kaplan–Meier with log-rank testing; overall survival (OS) predictors were assessed using Cox models.Results:
Median follow-up was 5 months (IQR 3–13). A total of 105 deaths (61.0%) and 35 three-month deaths (20.3%) occurred. Three-month mortality for GPS 0,1,2 was 11.3%, 13.2%, and 41.1% (p<0.001). GPS (0–1 vs.2) yielded an AUC of 0.692 (95% CI 0.603–0.781). GPS2 (OR4.90; p=0.005) and receipt of post-RT systemic therapy (OR0.115; p<0.001) independently predicted 3-month mortality. Compared with Group1, Group3 (OR5.150; p=0.003) and Group4 (OR6.740; p<0.001) exhibited markedly elevated risk. Three-month survival for Groups1–4 was 89.5%, 86.6%, 63.3%, and 57.8%, and 6-month survival was 71.6%, 65.6%, 44.7%, and 36.3%. In the Cox model, GPS2 and post-RT systemic therapy remained independent OS predictors.Conclusion:
Combining PS and GPS provides a strong tool for predicting early mortality in stage IV lung cancer patients receiving palliative RT. This simple stratification method may help identify patients at high short-term mortality risk and guide selection of appropriate fractionation schedules.