Squamous Cell Carcinoma and Lymphoma of the Oropharynx: Differentiation Using a Radiomics Approach

Abstract

The purpose of this study was to evaluate the diagnostic performance of magnetic resonance (MR) radiomics-based machine learning algorithms in differentiating squamous cell carcinoma (SCC) from lymphoma in the oropharynx. MR images from 87 patients with oropharyngeal SCC (n=68) and lymphoma (n=19) were reviewed retrospectively. Tumors were semi-automatically segmented on contrast-enhanced T1-weighted images registered to T2-weighted images, and radiomic features (n=202) were extracted from contrast-enhanced T1-and T2-weighted images. The radiomics classifier was built using elastic-net regularized generalized linear model analyses with nested five-fold cross-validation. The diagnostic abilities of the radiomics classifier and visual assessment by two head and neck radiologists were evaluated using receiver operating characteristic (ROC) analyses for distinguishing SCC from lymphoma. Nineteen radiomics features were selected at least twice during the five-fold cross-validation. The mean area under the ROC curve (AUC) of the radiomics classifier was 0.750 [95% confidence interval (CI), 0.613-0.887], with a sensitivity of 84.2%, specificity of 60.3%, and an accuracy of 65.5%. Two human readers yielded AUCs of 0.613 (95% CI, 0.467-0.759) and 0.663 (95% CI, 0.531-0.795), respectively. The radiomics-based machine learning model can be useful for differentiating SCC from lymphoma of the oropharynx.

Role of Texture Analysis in Oropharyngeal Carcinoma: A Systematic Review of the Literature

“…Among the 26 selected papers (Table 2), the largest number were from the USA [41][42][43][44][45][46][47][48][49][50][51] and Korea [52][53][54][55], whereas the rest of them were from Switzerland [56,57], the Netherlands [58,59], Italy [19,60], Canada [61], Japan [62,63], Taiwan [64,65] and China [45]. Fourteen articles [19,41,[43][44][45][46][47]49,52,53,55,57,58,62] had the oropharynx as the anatomic area of interest and seven articles [42,48,54,56,[59][60][61] analysed the head and neck district with sections dedicated t...…”

“…Fourteen articles [19,41,[43][44][45][46][47]49,52,53,55,57,58,62] had the oropharynx as the anatomic area of interest and seven articles [42,48,54,56,[59][60][61] analysed the head and neck district with sections dedicated to the oropharynx. Six, thirteen, and two articles, respectively, used MRI [53][54][55]58,59,61], CT [19,[41][42][43][44][45][46][47]52,56,57,60,62], and PET-CT [48,49] as imaging techniques on which texture analysis was performed (Table 3). Among the 26 studies, eleven of them used free software to extract textural data [19,44,[50][51][52]…”

“…Six, thirteen, and two articles, respectively, used MRI [53][54][55]58,59,61], CT [19,[41][42][43][44][45][46][47]52,56,57,60,62], and PET-CT [48,49] as imaging techniques on which texture analysis was performed (Table 3). Among the 26 studies, eleven of them used free software to extract textural data [19,44,[50][51][52][53][54][55]58,60,62,63]; five studies used licensed software [43,49,52,61,64], whereas the remaining ten papers used in house developed software not available to the public. Furthermore, ten studies focused on the distinction between HPV+ and HPV-neoplasms [19,[41][42][43][44]46,52,57,…”

“…[51]. Finally, two studies [53,63] focused on the ways to distinguish between squamocellular cancer and lymphoma through differences in textural features. Both studies elaborated models including both texture analysis features and image analysis techniques.…”

“…Given these differences in their microscopic structure, in a study conducted by Bae et al [53] it has been postulated that radiomics may show potential in discriminating between lymphoma and OPSCC. They identified on T1 post-contrast and T2 MRI sequences 19 features, 10 first-order features and 9 texture features capable of differentiating these two entities.…”