DEVELOPMENT AND VALIDATION OF A NOMOGRAM TO PREDICT OVERALL SURVIVAL IN NON-SMALL CELL LUNG CANCER
Author(s)
Riya Arora, MA1, Kirti Mirchandani, M.Sc.1, Neha Agrawal, MA1, Manvi Sharma, MBA, MS, RPh, PhD1, Rajender Aparasu, PharmD, PhD2.
1Peritia, Morrisville, NC, USA, 2University of Houston College of Pharmacy, Houston, TX, USA.
1Peritia, Morrisville, NC, USA, 2University of Houston College of Pharmacy, Houston, TX, USA.
OBJECTIVES: Socioeconomic disparities significantly influence survival in Non-Small Cell Lung Cancer (NSCLC) but remain underrepresented in existing prognostic models. This study aimed to develop and validate a nomogram integrating socioeconomic and prognostic factors to predict 5- and 10-year overall survival (OS) among patients with NSCLC.
METHODS: This retrospective study used 2011-2022 Surveillance, Epidemiology, and End Results (SEER) data to identify adults diagnosed with NSCLC. Patients were randomly assigned to training (75%) and validation (25%) cohorts. Univariate and multivariable Cox regression analyses were performed in training cohort to identify independent predictors of OS for development of nomogram. Model performance was assessed using the concordance index (C-index), receiver operating characteristic (ROC) curve, calibration curve, and decision curve analysis (DCA).
RESULTS: The study included a total of 191,743 NSCLC patients; mean age was 68.2 years. Most resided in urban areas (84.5%), and most (57.2%) had a median household income of ≥ $75,000. The training and validation cohort numbers were 143,808 and 47,935 respectively. Age, sex, race and origin, marital status, histology, primary site, tumor (T) size, metastatic involvement, treatment modalities, T stage, lymph node (N) stage and metastasis (M) stage independently predicted OS. Among socioeconomic variables, lower household income was significantly associated with worse OS (HR: 1.29 for < $50,000; HR: 1.18 for $50,000-74,999; ref: ≥ $75,000; p<0.05). The nomogram comprising independent predictors demonstrated strong discrimination (C-index‑: 0.762 training; 0.763 validation) and achieved an area under the curve of 0.857 and 0.884 for 5 and 10‑year OS in training and 0.858 and 0.885 in the validation cohort. Calibration plots showed close agreement between predicted and observed survival. DCA further confirmed greater net clinical benefit of the nomogram over TNM staging.
CONCLUSIONS: The prognostic nomogram demonstrated enhanced predictive accuracy, offering clinicians an effective tool for prognosis assessment and guiding individualized treatment strategies in NSCLC.
METHODS: This retrospective study used 2011-2022 Surveillance, Epidemiology, and End Results (SEER) data to identify adults diagnosed with NSCLC. Patients were randomly assigned to training (75%) and validation (25%) cohorts. Univariate and multivariable Cox regression analyses were performed in training cohort to identify independent predictors of OS for development of nomogram. Model performance was assessed using the concordance index (C-index), receiver operating characteristic (ROC) curve, calibration curve, and decision curve analysis (DCA).
RESULTS: The study included a total of 191,743 NSCLC patients; mean age was 68.2 years. Most resided in urban areas (84.5%), and most (57.2%) had a median household income of ≥ $75,000. The training and validation cohort numbers were 143,808 and 47,935 respectively. Age, sex, race and origin, marital status, histology, primary site, tumor (T) size, metastatic involvement, treatment modalities, T stage, lymph node (N) stage and metastasis (M) stage independently predicted OS. Among socioeconomic variables, lower household income was significantly associated with worse OS (HR: 1.29 for < $50,000; HR: 1.18 for $50,000-74,999; ref: ≥ $75,000; p<0.05). The nomogram comprising independent predictors demonstrated strong discrimination (C-index‑: 0.762 training; 0.763 validation) and achieved an area under the curve of 0.857 and 0.884 for 5 and 10‑year OS in training and 0.858 and 0.885 in the validation cohort. Calibration plots showed close agreement between predicted and observed survival. DCA further confirmed greater net clinical benefit of the nomogram over TNM staging.
CONCLUSIONS: The prognostic nomogram demonstrated enhanced predictive accuracy, offering clinicians an effective tool for prognosis assessment and guiding individualized treatment strategies in NSCLC.
Conference/Value in Health Info
2026-05, ISPOR 2026, Philadelphia, PA, USA
Value in Health, Volume 29, Issue S6
Code
EPH207
Topic
Epidemiology & Public Health
Disease
No Additional Disease & Conditions/Specialized Treatment Areas, SDC: Oncology