Construction and application of a mortality risk prediction model for patients with lung squamous cell carcinoma: A competing risk analysis.

Abstract

Background: Lung squamous cell carcinoma (LUSC) is the dominant histological subtype of lung cancer, accounting for 30% of all cases. Most patients develop distant metastases by the time they are diagnosed with the disease, owing to a delay in the appearance of symptoms. Therefore, accurate prognostic prediction is essential for personalized treatment. However, existing models tend to ignore competing risks, leading to an overestimation of the incidence. This study aimed to construct an accurate mortality risk prediction model for LUSC patients from the perspective of competing risks.

Methods: A total of 28,312 patients with LUSC from 2000 to 2019 were identified from the Surveillance, Epidemiology, and End Results (SEER) database. Potential predictors included age, sex, treatments, race, marital status, tumor site, differentiation, and stage. Statistical analyses were performed using cause-specific hazard and Fine-Gray risk models to assess competing risks. Model performance was evaluated using Harrell's concordance index and calibration plots.

Results: Age, sex, treatment, marital status, tumor site, differentiation, and stage significantly affected the prognosis of LUSC. Both risk models showed good predictive power. The Fine-Gray risk model was slightly better than the cause-specific hazard model in terms of the 3- and 5-year mortality. The key risk factors for LUSC outcomes included age, male gender, absence of surgery, chemotherapy or radiotherapy, being unmarried or divorced, primary tumors in the lower lobe or main bronchus, low differentiation, and high tumor stage.

Conclusion: The Fine-Gray model excels in predicting LUSC death risk and holds significant clinical value.