Table_3_Development and Validation of a Prognostic Model for Post-Operative Recurrence of Pituitary Adenomas.doc

BackgroundWe aimed to assess clinical factors associated with tumor recurrence and build a nomogram based on identified risk factors to predict postoperative recurrence in patients with pituitary adenomas (PAs) who underwent gross-total resection (GTR). MethodsA total of 829 patients with PAs who achieved GTR at Tongji Hospital between January 2013 and December 2018 were included in this retrospective study. The median follow-up time was 66.7 months (range: 15.6–106.3 months). Patients were randomly divided into training (n = 553) or validation (n = 276) cohorts. A range of clinical characteristics, radiological findings, and laboratory data were collected. Uni- and multivariate Cox regression analyses were applied to determine the potential risk factors for PA recurrence. A nomogram model was built from the identified factors to predict recurrence. Concordance index (C-index), calibration curve, and receiver operating characteristic (ROC) were used to determine the predictive accuracy of the nomogram. Decision curve analysis (DCA) was performed to evaluate the clinical efficacy of the nomogram. ResultsPseudocapsule-based extracapsular resection (ER), cavernous sinus invasion (CSI), and tumor size were included in the nomogram. C-indices of the nomogram were 0.776 (95% confidence interval [CI]: 0.747–0.806) and 0.714 (95% CI: 0.681–0.747) for the training and validation cohorts, respectively. The area under the curve (AUC) of the nomogram was 0.770, 0.774, and 0.818 for 4-, 6-, 8-year progression-free survival (PFS) probabilities in the training cohort, respectively, and 0.739, 0.715 and 0.740 for 4-, 6-, 8-year PFS probabilities in the validation cohort, respectively. Calibration curves were well-fitted in both training and validation cohorts. DCA revealed that the nomogram model improved the prediction of PFS in both cohorts. ConclusionsPseudocapsule-based ER, CSI, and tumor size were identified as independent predictors of PA recurrence. In the present study, we developed a novel and valid nomogram with potential utility as a tool for predicting postoperative PA recurrence. The use of the nonogram model can facilitate the tailoring of counseling to meet the individual needs of patients.**背景** 本研究旨在评估与垂体腺瘤（Pituitary Adenomas, PAs）患者肿瘤复发相关的临床危险因素，并基于筛选出的风险因素构建列线图（Nomogram），以预测接受全切术（Gross-Total Resection, GTR）的垂体腺瘤患者术后复发风险。 **方法** 本回顾性研究纳入2013年1月至2018年12月于同济医院接受全切术并达到全切效果的829例垂体腺瘤患者。患者中位随访时间为66.7个月（范围：15.6~106.3个月），按随机分组原则分为训练队列（n=553）与验证队列（n=276）。研究收集了患者的多项临床特征、影像学表现及实验室检测数据，通过单因素及多因素Cox回归分析筛选垂体腺瘤复发的潜在危险因素，基于筛选得到的风险因素构建列线图预测模型以评估复发风险。采用一致性指数（Concordance Index, C-index）、校准曲线及受试者工作特征曲线（Receiver Operating Characteristic, ROC）评估列线图的预测效能，并通过决策曲线分析（Decision Curve Analysis, DCA）评价其临床应用价值。 **结果** 基于假包膜的囊外切除术（Extracapsular Resection, ER）、海绵窦侵袭（Cavernous Sinus Invasion, CSI）及肿瘤大小被纳入列线图模型。训练队列与验证队列的列线图一致性指数分别为0.776（95%置信区间[CI]：0.747~0.806）与0.714（95%CI：0.681~0.747）。训练队列中，列线图对应4年、6年、8年无进展生存（Progression-Free Survival, PFS）概率的受试者工作特征曲线下面积（Area Under the Curve, AUC）分别为0.770、0.774及0.818；验证队列中对应值分别为0.739、0.715及0.740。两个队列的校准曲线均拟合良好，决策曲线分析结果显示该列线图模型可改善两个队列的无进展生存预测效能。 **结论** 基于假包膜的囊外切除术、海绵窦侵袭及肿瘤大小被确定为垂体腺瘤复发的独立预测因素。本研究构建了一种新型且有效的列线图模型，有望作为预测垂体腺瘤患者术后复发风险的工具，该模型可助力实现个体化的患者咨询与诊疗方案定制。