The novel prognostic nomograms for predicting cancer-specific survival and overall survival in mixed medullary and follicular cell carcinoma: A SEER-based study

Abstract

Background: The aim of this study was to evaluate independent predictors of prognosis in patients with mixed medullary and follicular cell carcinoma (MMFCC) and to establish the novel prognostic nomograms in the academic community for 3-, 5-, and 10 year CSS and OS in patients with MMFCC.

Methods: Demographic information, clinicopathological characteristics, treatment information, and survival status information of 200 patients with MMFCC and 6615 patients with medullary thyroid carcinoma (MTC) from 2000 to 2020 in the SEER database were retrospectively analyzed. Independent predictors of prognosis in MMFCC patients were derived using univariate and multivariate Cox regression analyses after relevant comparisons based on pathologic typing. On this basis, we developed and validated clinical prognostic nomograms and risk-stratified the patient population.

Results: In this study, the clinical information of 200 patients with MMFCC was compared with that of 5947 patients with MTC (NOS) and 668 patients with MTC with amyloid stroma, and there was a significant difference in the relevant variables among the three, with the CSS being 88.5%, 87.5%, and 90.9%, and the OS being 76.5%, 75.4%, and 83.8%. Univariate and multivariate Cox regression analyses yielded that age at diagnosis, presence of distant metastases, thyroidectomy scope, and lymph node dissection status were significantly correlated with the prognosis of patients (P < 0.05), and were independent predictors of CSS and OS for patients with MMFCC, and the Kaplan-Meier survival curves plotted by these factors demonstrated their predictive power for the prognosis of patients with MMFCC. The concordance index of the prognostic nomograms of CSS and OS established on this basis was 0.838 and 0.794, respectively, and the time-dependent area under curve, calibration curve, and decision curve analysis curve showed that the model had good discriminative ability, accuracy, and clinical applicability.

Conclusions: In this study, we concluded that there are large differences between MMFCC and MTC in terms of demographic information, clinicopathological characteristics, treatment information, and survival status information, and we constructed the novel prognostic nomograms for 3-, 5-, and 10 year CSS and OS for patients with MMFCC with risk stratification, which will help clinicians to develop individualized protocols for their postoperative treatments and follow-ups.

Keywords: Cancer-specific survival (CSS); Mixed medullary and follicular cell carcinoma (MMFCC); Overall survival (OS); Prognosis nomograms; Risk stratification.

Fig. 1

Flowchart of patients enrolled in this study

Fig. 2

Optimal cut-off values for age (a) and tumor size (b) were identified by X-tile software analysis. According to cancer-specific survival, the optimal cut-off values of age at diagnosis were 61 and 72 years old, and the optimal cut-off values of tumor diameter were 20 mm and 48 mm

Fig. 3

Nomograms to predict 3-, 5-, and 10 year cancer-specific survival (a) and overall survival (b) for MMFCC patients

Fig. 4

Each independent predictive factor with the established nomogram in the one time-dependent AUC curve: a nomogram (CSS); b nomogram (OS)

Fig. 5.

3 year (a), 5 year (b), and 10 year (c) cancer-specific survival nomogram internal verification curves

Fig. 6.

3 year (a), 5 year (b), and 10 year (c) overall survival nomogram internal verification curves

Fig. 7

Each independent predictive factor with the established nomogram in the one DCA curve for assessing the performance of predicting 3 year (a), 5 year (b), and 10 year (c) cancer-specific survival

Fig. 8

Each independent predictive factor with the established nomogram in the one DCA curve for assessing the performance of predicting 3 year (a), 5 year (b), and 10 year (c) overall survival

Fig. 9

Kaplan–Meier survival curves with respect to CSS were plotted for each significant independent predictor. a Age (years), b Distant metastasis, c Thyroidectomy scope, d Lymph node dissection status. Note: Time (days)

Fig. 10

Kaplan–Meier survival curves with respect to OS were plotted for each significant independent predictor. a Age (years), b Distant metastasis, c Thyroidectomy scope, d Lymph node dissection status. Note: Time (days)

Fig. 11

Risk factor maps of the prognostic prediction nomograms for CSS (a) and OS (b). Note: Survival time (days)