鼻咽癌放射治疗复发预测标准数据集

本数据集专为原发性鼻咽癌放射治疗后复发风险的智能预测任务构建，旨在利用机器学习技术、深度学习解决临床上复发风险评估不精准、个体化随访方案缺乏量化依据的问题。数据集针对放疗预测模型的训练需求，提供了多维度的结构化特征。
数据集特征架构
输入端（Features）： 结合人口信息、放射治疗信息、住院信息、诊断信息、病理信息、检验信息、随访信息等多维度结构化特征。输出端（Labels）： 严格基于患者长期随访结果判定的复发状态标签，涵盖复发的时间节点与临床结局。
数据源完全采用临床真实数据，源自高发地区权威省级三甲肿瘤专科医院的脱敏病例库，确保了样本在真实世界临床环境下的代表性。所有复发终点事件均经过病理学证实或连续影像学随访的双重校验。数据集经由“临床专家规则核准+多轮随访数据清洗”的严格质检流程，是训练高精度复发预测模型、辅助临床医师制定个体化随访策略与干预措施的优质核心资源。

This dataset is specifically constructed for the intelligent prediction of post-radiotherapy recurrence risk in primary nasopharyngeal carcinoma, aiming to address the clinical issues of inaccurate recurrence risk assessment and lack of quantitative basis for individualized follow-up plans via machine learning and deep learning technologies. It provides multi-dimensional structured features tailored to the training needs of radiotherapy prediction models.

Dataset Feature Architecture
Input Side (Features): Combines multi-dimensional structured clinical features including demographic information, radiotherapy information, hospitalization information, diagnostic information, pathological information, laboratory test information, follow-up information and other relevant data. Output Side (Labels): Strictly defined recurrence status labels based on long-term follow-up outcomes of patients, covering the time points of recurrence and clinical outcomes.

All data in this dataset are real-world clinical data, sourced from the de-identified case database of an authoritative provincial-level Class A tertiary cancer hospital in a high-incidence region, which ensures the representativeness of the samples in real-world clinical settings. All recurrence endpoint events have undergone dual verification via either pathological confirmation or consecutive imaging follow-up. The dataset has passed a strict quality inspection process including "approval by clinical expert guidelines and multiple rounds of follow-up data cleaning", making it a high-quality core resource for training high-precision recurrence prediction models and assisting clinicians in formulating individualized follow-up strategies and intervention measures.

搜集汇总

背景与挑战

背景概述

本数据集专为原发性鼻咽癌放射治疗后复发风险的智能预测而构建，提供了多维度结构化临床特征和基于长期随访的复发状态标签。数据源自福建省肿瘤医院的脱敏真实病例，并经过严格的质控流程，旨在支持高精度预测模型的训练，以辅助临床复发风险评估与个体化随访管理。