ALTERNATIVE ANALYTICAL APPROACHES TO INVESTIGATING THE ASSOCIATION BETWEEN TIMELINESS OF LUNG CANCER TREATMENT AND SURVIVAL
Author(s)
Saba Dangpiaei, MBA, PharmD1, Nicole L. Stout, DPT, CLT-LANA, FAPTA2, Kimberly Kelly, MS, PhD, MS, FSBM3, Mohammed Almubarak, MD2, Sabina O. Nduaguba, PhD2;
1West Virginia University, Ph.D. Student, Morgantown, WV, USA, 2West Virginia University, Morgantown, WV, USA, 3West Virginia University, School of Pharmacy, Morgantown, WV, USA
1West Virginia University, Ph.D. Student, Morgantown, WV, USA, 2West Virginia University, Morgantown, WV, USA, 3West Virginia University, School of Pharmacy, Morgantown, WV, USA
OBJECTIVES: Non-small cell lung cancer (NSCLC) remains the leading cause of cancer mortality in the United States. Although timely treatment is important for cancer outcomes, its impact on overall survival (OS) remains unclear. In a prior analysis, we observed no significant association between treatment timeliness and OS. This study examined the moderating effect of cancer stage and applied a random survival forest to evaluate the association between treatment timeliness and OS in NSCLC.
METHODS: We conducted a retrospective analysis using West Virginia Cancer Registry data to identify patients with NSCLC who received treatment. Time to treatment was classified as early (<35 days) or delayed (≥35 days). Overall survival was evaluated using multivariable Cox proportional hazards models stratified by cancer stage and a machine learning-based random survival forest to assess the association between treatment timeliness and OS, adjusting for clinical and demographic covariates.
RESULTS: Delayed treatment was associated with worse survival in stage 1 disease (HR = 1.11 [1.02-1.20], p = 0.011) but improved survival in stage 2 (HR = 0.85 [0.74-0.96], p = 0.0095), stage 3 (HR = 0.85 [0.77-0.92], p = 0.0002), and stage 4 (HR = 0.64 [0.60-0.69], p < 0.0001). In the random survival forest, cancer stage was the strongest predictor of survival (relative importance = 1.00), followed by treatment type (0.32), comorbidity index (0.25), age at diagnosis (0.22), race (0.19), treatment timeliness (0.06), sex (0.03), rurality (0.02), and marital status (0.01).
CONCLUSIONS: After adjustment for clinical and demographic factors, delayed treatment showed stage-dependent effects, with worse survival in stage 1 but improved outcomes in stages 2-4. In the random survival forest, treatment timeliness ranked sixth among nine predictors, while cancer stage ranked first. These findings support cancer stage as a key moderator of the association between treatment timeliness and survival in NSCLC and emphasize stage-specific interpretation.
METHODS: We conducted a retrospective analysis using West Virginia Cancer Registry data to identify patients with NSCLC who received treatment. Time to treatment was classified as early (<35 days) or delayed (≥35 days). Overall survival was evaluated using multivariable Cox proportional hazards models stratified by cancer stage and a machine learning-based random survival forest to assess the association between treatment timeliness and OS, adjusting for clinical and demographic covariates.
RESULTS: Delayed treatment was associated with worse survival in stage 1 disease (HR = 1.11 [1.02-1.20], p = 0.011) but improved survival in stage 2 (HR = 0.85 [0.74-0.96], p = 0.0095), stage 3 (HR = 0.85 [0.77-0.92], p = 0.0002), and stage 4 (HR = 0.64 [0.60-0.69], p < 0.0001). In the random survival forest, cancer stage was the strongest predictor of survival (relative importance = 1.00), followed by treatment type (0.32), comorbidity index (0.25), age at diagnosis (0.22), race (0.19), treatment timeliness (0.06), sex (0.03), rurality (0.02), and marital status (0.01).
CONCLUSIONS: After adjustment for clinical and demographic factors, delayed treatment showed stage-dependent effects, with worse survival in stage 1 but improved outcomes in stages 2-4. In the random survival forest, treatment timeliness ranked sixth among nine predictors, while cancer stage ranked first. These findings support cancer stage as a key moderator of the association between treatment timeliness and survival in NSCLC and emphasize stage-specific interpretation.
Conference/Value in Health Info
2026-05, ISPOR 2026, Philadelphia, PA, USA
Value in Health, Volume 29, Issue S6
Code
CO148
Topic
Clinical Outcomes
Disease
SDC: Oncology