Smoking Duration and Log Cig-Years Outperform Pack-Years in Predicting Survival in HNSCC: Study Finds

"Their effectiveness remained consistent across different patient subgroups and tumor sites, reinforcing their potential as superior metrics for assessing the impact of smoking on survival outcomes," the researchers wrote.

Cigarette smoking plays a critical role in the development and prognosis of head and neck squamous cell carcinoma, significantly impacting patient survival. While smoking is a well-established risk factor, there is limited evidence on the most effective metric for accurately modeling its association with survival outcomes. Existing survival models often rely on smoking status or pack years to quantify exposure, but these measures may not fully capture the long-term effects of smoking.

To address this limitation, Andrew C. L. Lam, Department of Medicine, University of Toronto, Toronto, Ontario, Canada, and colleagues have explored alternative metrics that provide a more precise evaluation of smoking impact on survival in HNSCC patients.

For this purpose, the researchers conducted a retrospective multicenter cohort study across six clinical epidemiological studies, including five from the Human Papillomavirus, Oral and Oropharyngeal Cancer Genomic Research (VOYAGER) consortium. The study analyzed data from patients aged 18 and older with pathologically confirmed HNSCC, collected between January 2002 and December 2019 and evaluated between January 2022 and November 2024.

The primary outcome was overall survival (OS), assessed using eight smoking metrics, including pack-years, duration, and log cig-years. These metrics were compared based on their association strength in Cox proportional hazard models, linear trend significance, Akaike information criterion (AIC), and spline curve assessments. Secondary analyses explored OS across clinicodemographic subgroups and tumor subsites, while exploratory outcomes included cancer-specific and noncancer survival.

The following were the key findings:

• The study included 8,875 patients with HNSCC, with a median age of 61 years and 24% female participants.
• Among the eight smoking metrics evaluated, smoking duration (aHR: 1.11) and log cig-years (aHR: 1.11) had the highest adjusted hazard ratios (aHRs).
• Both smoking duration and log cig-years showed a statistically significant linear association with overall survival.
• Log cig-years demonstrated the best model fit, with the lowest Akaike information criterion (AIC) linear value and the most visually linear spline curve for OS modeling.
• Smoking duration and log cig-years outperformed pack-years in predicting OS across all age groups, smoking statuses, and cancer stages.
• Both metrics performed well in lip and oral cavity cancers, while duration alone showed significance in laryngeal cancer and human papillomavirus (HPV)-negative oropharyngeal cancers.
• In an exploratory analysis, smoking duration had the highest aHR (1.15) for noncancer survival, while log cig-years had the lowest AIC linear value in modeling noncancer survival.

The researchers found that smoking duration and log cig-years demonstrated the strongest linear association with overall survival in patients with HNSCC, maintaining consistent predictive performance across various clinicodemographic subgroups and tumor subsites.

"While conventional survival models often rely on smoking status or pack-years to quantify smoking exposure, these findings suggest that duration and log cig-years may serve as superior metrics for assessing the impact of smoking on survival. Their robustness in modeling survival outcomes highlights their potential for both clinical decision-making and future research applications in HNSCC," the researchers concluded.

Reference:

Lam ACL, Hueniken K, Pienkowski M, et al. Performance of 8 Smoking Metrics for Modeling Survival in Head and Neck Squamous Cell Carcinoma. JAMA Otolaryngol Head Neck Surg. Published online February 20, 2025. doi:10.1001/jamaoto.2024.5392