Journal of clinical neuroscience : official journal of the Neurosurgical Society of Australasia · 2025
Random survival forests-based survival prediction for spinal chordomas.
СтатьяНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Cai Ming, Sun Hailun, Zheng Jihang, Zhang Nu, Yu Lisheng
Аннотация
BACKGROUND: Spinal chordomas are rare, malignant tumors with poor prognosis due to their tendency for local recurrence and distant metastasis. Accurate survival prediction is critical but challenging due to the complexity of clinical and histological factors. This study aimed to develop and validate a Random Survival Forest (RSF) model for survival prediction in spinal chordoma patients and compare its performance to the traditional Cox proportional hazards model.
METHODS: Using data from the SEER database (2000-2018), we retrospectively analyzed 679 patients with histologically confirmed spinal chordomas. Predictive models were constructed using the RSF algorithm and Cox regression, incorporating clinical and demographic variables. Model performance was assessed using three key aspects: calibration (integrated Brier score [iBS]), discrimination (integrated area under the curve [iAUC] and C-index), and clinical utility (Decision Curve Analysis [DCA]). Prognostic stratification was performed to assess the RSF model's ability to categorize patients into risk groups.
RESULTS: The RSF model demonstrated superior predictive performance compared to the Cox model, with a higher C-index (0.790 vs. 0.770), greater iAUC (0.731 vs. 0.679), and lower iBS (0.130 vs. 0.148). The RSF model provided improved time-dependent survival predictions and exhibited greater clinical utility in DCA across 3-, 5-, and 10-year survival thresholds. Age and surgical intervention were identified as the most influential prognostic factors. Risk stratification using the RSF model effectively distinguished high- and low-risk patients, as validated by Kaplan-Meier survival analyses.
CONCLUSIONS: The RSF model represents a significant advancement in survival prediction for spinal chordoma patients, offering superior accuracy and clinical utility compared to the traditional approach. Its integration into clinical practice could enhance personalized treatment planning and improve patient outcomes. Further external validation is needed to confirm its generalizability.
METHODS: Using data from the SEER database (2000-2018), we retrospectively analyzed 679 patients with histologically confirmed spinal chordomas. Predictive models were constructed using the RSF algorithm and Cox regression, incorporating clinical and demographic variables. Model performance was assessed using three key aspects: calibration (integrated Brier score [iBS]), discrimination (integrated area under the curve [iAUC] and C-index), and clinical utility (Decision Curve Analysis [DCA]). Prognostic stratification was performed to assess the RSF model's ability to categorize patients into risk groups.
RESULTS: The RSF model demonstrated superior predictive performance compared to the Cox model, with a higher C-index (0.790 vs. 0.770), greater iAUC (0.731 vs. 0.679), and lower iBS (0.130 vs. 0.148). The RSF model provided improved time-dependent survival predictions and exhibited greater clinical utility in DCA across 3-, 5-, and 10-year survival thresholds. Age and surgical intervention were identified as the most influential prognostic factors. Risk stratification using the RSF model effectively distinguished high- and low-risk patients, as validated by Kaplan-Meier survival analyses.
CONCLUSIONS: The RSF model represents a significant advancement in survival prediction for spinal chordoma patients, offering superior accuracy and clinical utility compared to the traditional approach. Its integration into clinical practice could enhance personalized treatment planning and improve patient outcomes. Further external validation is needed to confirm its generalizability.