Evaluation of conditional survival outcomes in patients with redefined anaplastic thyroid carcinoma

TL;DR

This study evaluates how survival chances improve over time for patients with redefined anaplastic thyroid cancer and creates a tool to predict these chances.

Contribution

The study introduces a novel conditional survival nomogram for redefined anaplastic thyroid carcinoma patients.

Findings

The 5-year conditional survival probability increases from 8% to 95% as patients survive longer.

Age, tumor stage, surgery, radiotherapy, and chemotherapy are key prognostic factors.

A predictive nomogram was developed and validated for dynamic survival probability estimation.

Abstract

Conditional survival (CS) offers a more precise assessment of prognosis by continuously updating to reflect the current state. This study aimed to analyze the CS outcome of redefined anaplastic thyroid carcinoma (rATC).

A cohort comprising 1424 patients diagnosed with rATC between 2000 and 2018 was extracted from the Surveillance, Epidemiology, and End Results database and subsequently divided into training and validation sets at a ratio of 3:1. We analyzed the CS prognosis of these patients. A Lasso Cox regression model was employed to determine independent risk factors for long-term survival and to develop a predictive nomogram integrating CS analysis, aiming to provide CS estimates and stratify the risk among rATC patients.

The likelihood of achieving 5-year CS escalated from 8% after diagnosis to 44%, 68%, 82%, and 95% after 1, 2, 3, and 4 years of survival, respectively. As patients endured longer, their probability of further years of survival augmented. The Lasso Cox regression analysis identified age, tumor stage, surgery, radiotherapy and chemotherapy as the most influential prognostic factors associated with outcomes. Ultimately, we successfully established a novel CS-nomogram, aiming to provide dynamic survival updates for these patients. Our predictive model can quantify patients’ risk and stratify it accordingly. Furthermore, after evaluation, the performance of our predictive model was found to be satisfactory.

Utilizing extensive SEER datasets, we devised and validated a CS-nomogram, enabling the prediction of the conditional probability of survival for rATC patients. This facilitates the incorporation of survivorship duration into prognostic assessments.

Figures