基于常规影像征象和表观弥散系数的列线图预测鼻腔鼻窦内翻性乳头状瘤恶变的价值  

Value of nomogram based on conventional imaging characteristics and apparent diffusion coefficient value for predicting malignant transformation of sinonasal inverted papilloma

作　　者：齐萌 夏志鹏 任继亮[2] 张放 沙炎 QI Meng;XIA Zhipeng;REN Jiliang;ZHANG Fang;SHA Yan(Department of Radiology,Eye&ENT Hospital of Fudan University,Shanghai 200031,China;Department of Radiology,Shanghai Ninth People’s Hospital,Shanghai Jiao Tong University School of Medicine,Shanghai 200011,China)

机构地区：[1]复旦大学附属眼耳鼻喉科医院放射科,上海200031 [2]上海交通大学医学院附属第九人民医院放射科,上海200011

出　　处：《中国眼耳鼻喉科杂志》2023年第6期475-479,487,共6页Chinese Journal of Ophthalmology and Otorhinolaryngology

摘　　要：目的 探讨常规影像征象和表观弥散系数(ADC)对鼻腔鼻窦内翻性乳头状瘤(SIP)恶变的预测价值。方法 回顾收集我院2015年1月—2021年11月收治的209例SIP患者的临床资料,并分为训练集(n=140;良性90例、恶变50例)和验证集(n=69;良性45例、恶变24例)。由2名放射科医师在常规影像图像上评估骨质破坏、软组织侵犯、眼眶或颅内侵犯、脑回征中断及三维肿瘤最大径并测量病灶的平均ADC值。比较所有临床资料、常规影像特征、ADC值的组间差异。采用逐步逻辑回归分析筛选出独立的预测因子,并建立联合模型。采用受试者工作特征(ROC)曲线下面积(AUC)评估独立预测因子和联合模型的预测效能。基于联合模型制作列线图,并绘制校正曲线。最后采用决策分析曲线评价列线图的临床价值。结果 在训练集和验证集中,2组之间软组织侵犯、眼眶或颅内侵犯、脑回征中断、肿瘤最大径及ADC值差异均具有统计学意义(P值均<0.001)。逐步逻辑回归显示肿瘤最大径、脑回征中断和ADC值是预测SIP恶变的独立预测因子,三者于训练集的AUC分别为0.80、0.79、0.85,于验证集的AUC分别为0.79、 0.74、 0.89。三者的联合模型在训练集和验证集的AUC分别为0.93和0.92,其列线图的校正曲线在两数据集中均有较高的拟合度。决策曲线分析显示阈概率> 5%时该列线图的临床净收益较大。结论 基于常规影像征象和ADC的列线图对于预测SIP恶变具有良好的表现,能为临床决策提供一定帮助。Objective To investigate the value of conventional imaging characteristics and apparent diffusion coefficient(ADC)value in predicting the malignant transformation(MT)of sinonasal inverted papilloma(SIP).Methods This retrospective study included 209 SIPs with and without MT,which were assigned into a training cohort(n=140;90 SIPs and 50 MT-SIPs)and a validation cohort(n=69;45 SIPs and 24 MT-SIPs).Bony erosion,soft tissue invasion,orbital or intracranial invasion,loss of convoluted cerebriform pattern(CCP),three-dimensional maximum diameter and the ADC value were assessed or measured by two radiologists,respectively.The clinical characteristics,conventional imaging characteristics and ADC value were compared between the two groups.Stepwise logistic regression was used to identify independent predictors and to construct a combined model.The predictive performances of the predictors and combined model were assessed with area under the curve(AUC).The combined model was presented as a nomogram with its calibration assessed.Decision curve analysis was conducted to evaluate the net benefits of nomogram.Results All conventional imaging characteristics and ADC value showed significant differences between the two groups in both cohorts(all P<0.001).Maximum diameter,loss of CCP and ADC value were identified as independent variables,with AUCs of 0.80,0.79,0.85 in the training cohort and AUCs of 0.79,0.74,0.89 in the validation cohort,respectively.The combined model achieved AUCs of 0.93 and 0.92 in the two cohorts,respectively.The nomogram based on the combined model showed good calibration in both cohorts and displayed good clinical utility when the threshold probability was greater than 5%.Conclusions The nomogram based on conventional imaging characteristics and ADC value showed excellent performance in predicting malignant transformation of SIP preoperatively and can be used as a tool to assist clinical decision-making.

关 键 词：鼻腔鼻窦内翻性乳头状瘤 恶变 表观弥散系数 列线图 

分 类 号：R765[医药卫生—耳鼻咽喉科]