Prediction of Tumor Grade and Nodal Status in Oropharyngeal and Oral Cavity Squamous-cell Carcinoma Using a Radiomic Approach

• DOI: 10.21873/anticanres.13949
• Publication Date: 2019-12-31T20:55:12Z
• Authors: VALERIA ROMEO; RENATO CUOCOLO; CARLO RICCIARDI; LORENZO UGGA; SIRIO COCOZZA; FRANCESCO VERDE; ARNALDO STANZIONE; VIRGINIA NAPOLITANO; DANIELA RUSSO; GIOVANNI IMPROTA; ANDREA ELEFANTE; STEFANIA STAIBANO; ARTURO BRUNETTI
• Keywords: Head and neck squamous cell carcinoma; computed tomography; machine learning; texture analysis; Carcinoma; Head and neck squamous cell carcinoma; X-Ray Computed; Machine Learning; Oropharyngeal Neoplasms; 03 medical and health sciences; 0302 clinical medicine; Texture analysis; Squamous Cell; Machine learning; Carcinoma, Squamous Cell; Humans; Computed tomography; Head and neck squamous cell carcinoma; Machine learning; Texture analysis; Aged; Algorithms; Carcinoma, Squamous Cell; Humans; Lymph Nodes; Machine Learning; Mouth Neoplasms; Neoplasm Grading; Oropharyngeal Neoplasms; Tomography, X-Ray Computed; Mouth Neoplasms; Lymph Nodes; Neoplasm Grading; Tomography, X-Ray Computed; Computed tomography; Tomography; Algorithms; Aged

To investigate whether a radiomic machine learning (ML) approach employing texture-analysis (TA) features extracted from primary tumor lesions (PTLs) is able to predict grade (TG) and nodal status (NS) in patients with oropharyngeal (OP) oral cavity (OC) squamous-cell carcinoma (SCC).Contrast-enhanced CT images of 40 OP OC SCC were post-processed extract TA PTLs. A feature selection method different ML algorithms applied find the most accurate subset TG NS.For prediction TG, best accuracy (92.9%) was achieved by Naïve Bayes (NB), bagging NB K Nearest Neighbor (KNN). For NS, J48, NB, boosting J48 overcame 90%.A PTLs NS SCC.