Vito Chianca1, Renato Cuocolo2, Salvatore Gitto3, Domenico Albano4, Ilaria Merli5, Julietta Badalyan6, Maria Cristina Cortese7, Carmelo Messina8, Alessandro Luzzati9, Antonina Parafioriti10, Fabio Galbusera9, Arturo Brunetti11, Luca Maria Sconfienza8. 1. Clinica di Radiologia EOC, Istituto di Imaging della Svizzera Italiana (IIMSI), Lugano, Switzerland; Ospedale Evangelico Betania, Napoli, Italy. 2. Dipartimento di Scienze Biomediche Avanzate, Università degli Studi di Napoli (")Federico II", Napoli, Italy; Laboratory of Augmented Reality for Health Monitoring (ARHeMLab), Dipartimento di Ingegneria Elettrica e delle Tecnologie dell'Informazione, Università degli Studi di Napoli "Federico II", Naples, Italy. 3. Dipartimento di Scienze Biomediche per la Salute, Università degli Studi di Milano, Milano, Italy. Electronic address: sal.gitto@gmail.com. 4. IRCCS Istituto Ortopedico Galeazzi, Milano, Italy; Sezione di Scienze Radiologiche, Dipartimento di Biomedicina, Neuroscienze e Diagnostica Avanzata, Università degli Studi di Palermo, Italy. 5. UOC Radiodiagnostica, Presidio San Carlo Borromeo, ASST Santi Paolo e Carlo, Milano, Italy. 6. International Medical School, University of Milan and Russian National Research Medical University, Milano, Italy. 7. Istituto di Radiologia, Fondazione Policlinico A. Gemelli IRCCS - Università Cattolica Sacro Cuore, Roma, Italy. 8. IRCCS Istituto Ortopedico Galeazzi, Milano, Italy; Dipartimento di Scienze Biomediche per la Salute, Università degli Studi di Milano, Milano, Italy. 9. IRCCS Istituto Ortopedico Galeazzi, Milano, Italy. 10. Anatomia Patologica, ASST Pini-CTO, Milano, Italy. 11. Dipartimento di Scienze Biomediche Avanzate, Università degli Studi di Napoli (")Federico II", Napoli, Italy.
Abstract
PURPOSE: Spinal lesion differential diagnosis remains challenging even in MRI. Radiomics and machine learning (ML) have proven useful even in absence of a standardized data mining pipeline. We aimed to assess ML diagnostic performance in spinal lesion differential diagnosis, employing radiomic data extracted by different software. METHODS: Patients undergoing MRI for a vertebral lesion were retrospectively analyzed (n = 146, 67 males, 79 females; mean age 63 ± 16 years, range 8-89 years) and constituted the train (n = 100) and internal test cohorts (n = 46). Part of the latter had additional prior exams which constituted a multi-scanner, external test cohort (n = 35). Lesions were labeled as benign or malignant (2-label classification), and benign, primary malignant or metastases (3-label classification) for classification analyses. Features extracted via 3D Slicer heterogeneityCAD module (hCAD) and PyRadiomics were independently used to compare different combinations of feature selection methods and ML classifiers (n = 19). RESULTS: In total, 90 and 1548 features were extracted by hCAD and PyRadiomics, respectively. The best feature selection method-ML algorithm combination was selected by 10 iterations of 10-fold cross-validation in the training data. For the 2-label classification ML obtained 94% accuracy in the internal test cohort, using hCAD data, and 86% in the external one. For the 3-label classification, PyRadiomics data allowed for 80% and 69% accuracy in the internal and external test sets, respectively. CONCLUSIONS: MRI radiomics combined with ML may be useful in spinal lesion assessment. More robust pre-processing led to better consistency despite scanner and protocol heterogeneity.
PURPOSE:Spinal lesion differential diagnosis remains challenging even in MRI. Radiomics and machine learning (ML) have proven useful even in absence of a standardized data mining pipeline. We aimed to assess ML diagnostic performance in spinal lesion differential diagnosis, employing radiomic data extracted by different software. METHODS:Patients undergoing MRI for a vertebral lesion were retrospectively analyzed (n = 146, 67 males, 79 females; mean age 63 ± 16 years, range 8-89 years) and constituted the train (n = 100) and internal test cohorts (n = 46). Part of the latter had additional prior exams which constituted a multi-scanner, external test cohort (n = 35). Lesions were labeled as benign or malignant (2-label classification), and benign, primary malignant or metastases (3-label classification) for classification analyses. Features extracted via 3D Slicer heterogeneityCAD module (hCAD) and PyRadiomics were independently used to compare different combinations of feature selection methods and ML classifiers (n = 19). RESULTS: In total, 90 and 1548 features were extracted by hCAD and PyRadiomics, respectively. The best feature selection method-ML algorithm combination was selected by 10 iterations of 10-fold cross-validation in the training data. For the 2-label classification ML obtained 94% accuracy in the internal test cohort, using hCAD data, and 86% in the external one. For the 3-label classification, PyRadiomics data allowed for 80% and 69% accuracy in the internal and external test sets, respectively. CONCLUSIONS: MRI radiomics combined with ML may be useful in spinal lesion assessment. More robust pre-processing led to better consistency despite scanner and protocol heterogeneity.
Authors: Nicolas Serratrice; Joe Faddoul; Bilal Tarabay; Christian Attieh; Moussa A Chalah; Samar S Ayache; Georges N Abi Lahoud Journal: Front Oncol Date: 2022-01-27 Impact factor: 6.244