Assessing the Applicability of the TALK Score: A Modification for Concurrent Tobacco Use During Treatment

Objective The previously described TALK score (T-stage, Albumin, Liquor, Karnofsky Performance Status) has been proposed as a method to predict laryngectomy-free survival (LFS) in patients undergoing definitive chemoradiation (CRT). This study assesses its use as well as a modification to include continued tobacco use. Study Design Retrospective chart review. Setting Academic institution from 2004 to 2020. Methods Patients diagnosed with larynx or hypopharynx cancer undergoing CRT were reviewed. Clinically relevant variables were collected (TALK), which were dichotomized per previously set cutoffs. Concurrent tobacco use was evaluated and also dichotomized as 0 or 1. Multivariate analysis was conducted to determine which factors were most predictive of the key outcomes of survival and LFS. Results A total of 2514 patient charts were evaluated. Patients treated for larynx cancer with primary CRT with complete data were included, ultimately totaling 300. Of these, 78 patients required salvage total laryngectomy (TL). Multivariate analysis demonstrated that LFS was best predicted by tobacco use during treatment (odds ratio [OR] 0.3465, 95% confidence interval [CI] 0.1862-0.6300) and Karnofsky Performance Status (OR 0.1646, 95% CI 0.0673-0.3662). Tobacco use during treatment was also strongly predictive of survival. Excluding T4 tumors, the utilization of tobacco in place of T-stage improved the accuracy of the predictive model in this cohort. Conclusion Given that a T-stage of 4 is typically treated with total laryngectomy, modification of the TALK score to include tobacco use during treatment (tALK) for patients with T < 4 can be used to improve prediction of 3-year LFS and overall survival.

Automated summary

This study evaluated the effectiveness of the TALK score and its modification tALK (including tobacco use during treatment) in predicting laryngectomy-free survival in patients undergoing chemoradiation. The results showed that tobacco use during treatment was one of the most predictive factors for survival outcomes. Additionally, replacing T-stage with tobacco use in patients with T < 4 improved the accuracy of the predictive model.