Ramkumar S, Ranjbar S, Ning S, Lal D, Zwart C M, Wood C P, Weindling S M, Wu T, Mitchell J R, Li J, Hoxworth J M
From the School of Computing, Informatics, and Decision Systems Engineering (S.Ramkumar, S.N., T.W., J.L.).
Department of Biomedical Informatics (S.Ranjbar), Arizona State University, Tempe, Arizona.
AJNR Am J Neuroradiol. 2017 May;38(5):1019-1025. doi: 10.3174/ajnr.A5106. Epub 2017 Mar 2.
Because sinonasal inverted papilloma can harbor squamous cell carcinoma, differentiating these tumors is relevant. The objectives of this study were to determine whether MR imaging-based texture analysis can accurately classify cases of noncoexistent squamous cell carcinoma and inverted papilloma and to compare this classification performance with neuroradiologists' review.
Adult patients who had inverted papilloma or squamous cell carcinoma resected were eligible (coexistent inverted papilloma and squamous cell carcinoma were excluded). Inclusion required tumor size of >1.5 cm and preoperative MR imaging with axial T1, axial T2, and axial T1 postcontrast sequences. Five well-established texture analysis algorithms were applied to an ROI from the largest tumor cross-section. For a training dataset, machine-learning algorithms were used to identify the most accurate model, and performance was also evaluated in a validation dataset. On the basis of 3 separate blinded reviews of the ROI, isolated tumor, and entire images, 2 neuroradiologists predicted tumor type in consensus.
The inverted papilloma ( = 24) and squamous cell carcinoma ( = 22) cohorts were matched for age and sex, while squamous cell carcinoma tumor volume was larger ( = .001). The best classification model achieved similar accuracies for training (17 squamous cell carcinomas, 16 inverted papillomas) and validation (7 squamous cell carcinomas, 6 inverted papillomas) datasets of 90.9% and 84.6%, respectively ( = .537). For the combined training and validation cohorts, the machine-learning accuracy (89.1%) was better than that of the neuroradiologists' ROI review (56.5%, = .0004) but not significantly different from the neuroradiologists' review of the tumors (73.9%, = .060) or entire images (87.0%, = .748).
MR imaging-based texture analysis has the potential to differentiate squamous cell carcinoma from inverted papilloma and may, in the future, provide incremental information to the neuroradiologist.
由于鼻窦内翻性乳头状瘤可能伴有鳞状细胞癌，鉴别这些肿瘤具有重要意义。本研究的目的是确定基于磁共振成像（MR）的纹理分析能否准确区分不存在并存鳞状细胞癌的内翻性乳头状瘤病例，并将这种分类性能与神经放射科医生的评估进行比较。
纳入标准为接受过内翻性乳头状瘤或鳞状细胞癌切除术的成年患者（排除并存内翻性乳头状瘤和鳞状细胞癌的患者）。纳入要求肿瘤大小>1.5 cm且术前行轴向T1、轴向T2和轴向T1增强序列的MR成像检查。将五种成熟的纹理分析算法应用于最大肿瘤横截面的感兴趣区（ROI）。对于训练数据集，使用机器学习算法识别最准确的模型，并在验证数据集中评估性能。基于对ROI、孤立肿瘤和完整图像的3次独立双盲评估，2名神经放射科医生达成共识预测肿瘤类型。
内翻性乳头状瘤组（n = 24）和鳞状细胞癌组（n = 22）在年龄和性别上相匹配，而鳞状细胞癌的肿瘤体积更大（P = .001）。最佳分类模型在训练数据集（17例鳞状细胞癌，16例内翻性乳头状瘤）和验证数据集（7例鳞状细胞癌，6例内翻性乳头状瘤）中的准确率分别为90.9%和84.6%，二者相似（P = .537）。对于联合训练和验证队列，机器学习的准确率（89.1%）优于神经放射科医生对ROI的评估（56.5%，P = .0004），但与神经放射科医生对肿瘤（73.9%，P = .060）或完整图像（87.0%，P = .748）的评估无显著差异。
基于MR成像的纹理分析有潜力区分鳞状细胞癌和内翻性乳头状瘤，未来可能为神经放射科医生提供更多信息。
