Radiotherapy Versus Surgery-Which Is Better for Patients With T1-2N0M0 Glottic Laryngeal Squamous Cell Carcinoma? Individualized Survival Prediction Based on Web-Based Nomograms

Abstract

Background: Both radiotherapy and surgery are now recommended for early stage glottic laryngeal squamous cell carcinoma (LSCC), and both have their own advantages in patients with different characteristics. For each patient, it is hard to determine whether radiotherapy or surgery is more appropriate.

Methods: Patients with T1-2N0M0 glottic LSCC who received radiotherapy or surgery in the 2004-2016 SEER database were reviewed, then randomly divided into training and validation cohorts. Propensity score matching was used to eliminate the baseline variations, and competing risk analyses helped to exclude the effects of other causes of death. Based on univariate and multivariate analyses, we built two nomograms to visually predict the survival of each patient with different characteristics who received radiotherapy or surgery, then validated the accuracy in both training and validation cohorts. Using nomogramEx, we quantified the algorithms of the nomograms and put the nomograms on the websites.

Results: A total of 6538 patients in the SEER database were included. We found that therapy (p = 0.004), T stage (p < 0.001), age (p < 0.001), race (p < 0.044), grade (p = 0.001), and marital status (p < 0.001) were independent prognostic factors. Two nomograms were built to calculate the survival for each patient who received radiotherapy (C-index = 0.668 ± 0.050 in the training cohort and 0.578 ± 0.028 in the validation cohort) or underwent surgery (C-index = 0.772 ± 0.045 in the training cohort and 0.658 ± 0.090 in the validation cohort). Calibration plots showed the accuracy of the nomograms. Using the nomograms, we found that 3872 patients (59.22%) had no difference between the two therapies, 706 patients (10.80%) who received radiotherapy had better survival outcomes, and 1960 patients (29.98%) who underwent surgery had better survival outcome.

Conclusion: Nomograms were used to comprehensively calculate independent factors to determine which treatment (radiotherapy or surgery) is better for each patient. A website was used to offer guidance regarding surgery or radiation for patients and physicians.

Keywords: SEER; glottic laryngeal squamous cell carcinoma; nomogram; radiotherapy; surgery.

FIGURE 1

(A) Flow diagram of selecting process (B) Mirror histogram of propensity scores for patients with radiotherapy and with surgery. Matched patients are presented in green color. (C) Standardized differences of baseline variables between patients with radiation and with surgery before and after propensity score matching.

FIGURE 2

Survival analyses for patients with radiotherapy and with surgery (A) Kaplan-Meier method. (B) Competing risk analysis.

FIGURE 3

Nomogram analyses for patients with radiotherapy (A) A nomogram for prediction of 3- and 5-year CSS rates of patients (B) Calibration curve of the nomogram predicting 3-year CSS rates in training cohort. (C) Calibration curve of the nomogram predicting 5-year CSS rates in training cohort. (D) Calibration curve of the nomogram predicting 3-year CSS rates in validation cohort (E) Calibration curve of the nomogram predicting 3-year CSS rates in validation cohort.

FIGURE 4

Nomogram analyses for patients with surgery (A) A nomogram for prediction of 3- and 5-year CSS rates of patients (B) Calibration curve of the nomogram predicting 3-year CSS rates in training cohort. (C) Calibration curve of the nomogram predicting 5-year CSS rates in training cohort. (D) Calibration curve of the nomogram predicting 3-year CSS rates in validation cohort (E) Calibration curve of the nomogram predicting 3-year CSS rates in validation cohort.

FIGURE 5

The histogram of the difference between CSS rates (radiotherapy-surgery) (A) the difference between 3-year CSS rates (B) the difference between 5-year CSS rates.