{"id":6282,"date":"2022-05-31T17:25:28","date_gmt":"2022-05-31T15:25:28","guid":{"rendered":"https:\/\/www.convegnonazionaleaiic.it\/3d-mri-radiomics-based-machine-learning-for-differentiation-of-atypical-cartilaginous-tumor-from-grade-ii-chondrosarcoma-of-long-bones\/"},"modified":"2022-05-31T17:25:28","modified_gmt":"2022-05-31T15:25:28","slug":"3d-mri-radiomics-based-machine-learning-for-differentiation-of-atypical-cartilaginous-tumor-from-grade-ii-chondrosarcoma-of-long-bones","status":"publish","type":"post","link":"https:\/\/www.convegnonazionaleaiic.it\/3d-mri-radiomics-based-machine-learning-for-differentiation-of-atypical-cartilaginous-tumor-from-grade-ii-chondrosarcoma-of-long-bones\/","title":{"rendered":"3D MRI RADIOMICS-BASED MACHINE LEARNING FOR DIFFERENTIATION OF ATYPICAL CARTILAGINOUS TUMOR FROM GRADE II CHONDROSARCOMA OF LONG BONES"},"content":{"rendered":"
\n
\n

\"\"\"\" <\/p>\n

AFFILIAZIONE<\/span>
\n<\/strong>universit\u00e0 degli studi di milano<\/p>\n

AUTORE PRINCIPALE<\/span>
\n<\/strong> Gitto Salvatore<\/p>\n

\nVALUTA IL CHALLENGE<\/span><\/strong>
Vota<\/nobr><\/td>
<\/div><\/td><\/tr><\/table><\/p>\n<\/div>\n
\n

<\/a><\/p>\n

GRUPPO DI LAVORO<\/span><\/strong>
\n Gitto Salvatore<\/strong> – universit\u00e0 degli studi di milano, milano<\/i>
Castiglioni Isabella<\/strong> – universit\u00e0 degli studi di milano-bicocca, milano<\/i>
Interlenghi Matteo<\/strong> – deeptrace technologies, milano<\/i>
Salvatore Christian<\/strong> – deeptrace technologies, milano<\/i>
Tordi Roberto<\/strong> – irccs istituto ortopedico galeazzi, milano<\/i>
Sconfienza Luca Maria<\/strong> – universit\u00e0 degli studi di milano, milano<\/i> <\/p>\n

AREA TEMATICA<\/span>
\n<\/strong>Gestione delle tecnologie biomediche: dati, modelli, risultati<\/p>\n

ABSTRACT<\/span>
\n<\/strong>Objective
\nTo develop a machine learning model based on 3D MRI radiomics to differentiate atypical cartilaginous tumor (ACT) from grade II chondrosarcoma (CS2) of long bones.<\/p>\n

Methods
\nWe retrospectively collected MRI scans from 36 patients. Eighteen patients (50%) belonged to CS2 class and 18 patients (50%) belonged to ACT class, according to histological diagnosis from definite surgery. All lesions were manually segmented by drawing 3D regions of interests including the whole tumor volume on T1-weighted MRI sequences. After radiomic feature extraction and selection, this image set was used for the training and cross-validation of different machine-learning models. A robust radiomic approach was applied, under the hypothesis that radiomic features could be able to capture the disease heterogeneity among the two groups. Three models consisting of 3 ensembles of machine-learning classifiers (random forests, support vector machines and k-nearest neighbor classifiers) were developed for the binary classification task of interest (CS2 vs. ACT), based on supervised learning. <\/p>\n

Results
\nThe best model (random forest) showed: ROC-AUC (%) of 87 (majority vote), 87.6** (mean) [CI (confidence interval): 82.3-92.8]; accuracy (%) of 81 (majority vote), 78.7** (mean) [CI: 70.7-86.7], sensitivity (%) of 72 (majority vote), 72.2** (mean) [CI: 72.2-72.2], specificity (%) of 89 (majority vote), 85.2 (mean) [CI: 69.2-100], positive predictive value (%) of 87 (majority vote), 83.3* (mean) [CI: 68.6-97.9], and negative predictive value (%) of 76 (majority vote), 75.4** (mean) [CI: 71.8-78.9] (*p<0.05, **p<0.005). \n\nConclusions\nOur 3D MRI radiomics-based machine learning method may potentially aid in clinical decision making by accurately distinguishing between ACT and CS2.<\/p>\n

 <\/p>\n