Benjamin Leporq
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Benjamin Leporq has a total of 159 co-authors. Here are the top co-authors by number of shared publications:
- Olivier Beuf
INSA‐Lyon Université Lyon 1, CNRS, Inserm, CREATIS, UMR 5220, U1294 Villeurbanne France
13 shared publications - Bernard Van Beers
Center for Research on Inflammation, Inserm U1149 Université Paris Cité Paris France
11 shared publications Frank Pilleul
6 shared publications- Simon Auguste Lambert
Université Lyon 1, Ecole Centrale de Lyon, INSA Lyon, CNRS, AMPERE Villeurbanne France
6 shared publications - Frank Pilleul
INSA‐Lyon, Université Lyon 1 CNRS Inserm CREATIS UMR5220 U1294 Villeurbanne France
5 shared publications Olivier Beuf
5 shared publications
Latest Publications
- Open Access
- 24 August 2026
- Published by Wiley in Journal of Applied Clinical Medical Physics
- Vol. 27 (9) , e70722
How bone mis‐segmentation affects MRI‐based sCT dosimetry in head and neck radiotherapy: A CT‐referenced study
Background: Magnetic resonance imaging (MRI) is commonly used for head‐and‐neck (H&N) delineation in radiotherapy (RT) owing to the lack of contrast in computed tomography (CT) images. MRI‐based dosimetry is not yet fully developed for the H&N area due to the inter‐patient variability and the difficulty of accurately quantifying cortical bone based on MRI measure. Purpose: Despite the difficulty of cortical bone segmentation with nearby air area, most papers in the literature show dosimetry results very close to the reference CT scan dosimetry. This study evaluated the impact of bone segmentation on dosimetry. Methods: A home‐made phantom was used to evaluate the Hounsfield unit (HU) values of different tissue at diverse scanner energies. Different dosimetry plans were considered and compared with the reference CT‐based RT plan used for the treatment: CT images with cortical bone voxels assigned to the density of water and three based on MR images using a density assignment (DA) method for synthetic CT (sCT) generation with various bone attribution. Images from 24 patients were compared using mean absolute error (MAE). The generated dosimetry parameters were compared on the basis of dose–volume histogram values and 3%/3 mm gamma pass rates. Results: Among the sCT, the best MAE were found for the DA sCT without bone. The mean of the median dose difference for the planning tumor volume (PTV) was less than 2% for all sCT. Mean gamma pass rate for the MR‐based sCTs were comparable. Conclusions: The DA of the bone can be of interest if correctly segmented, but in the absence of certainty, an overestimation will lead to significant errors while an underestimation gives results close to the reference. The relevance of very precise sCT can be questioned if it requires too many resources and overlooks patients with artifacts.
- 24 August 2026
- Published by Wiley in Journal of Applied Clinical Medical Physics
- Vol. 27 (9) , e70722
Open AccessHow bone mis‐segmentation affects MRI‐based sCT dosimetry in head and neck radiotherapy: A CT‐referenced study
Background: Magnetic resonance imaging (MRI) is commonly used for head‐and‐neck (H&N) delineation in radiotherapy (RT) owing to the lack of contrast in computed tomography (CT) images. MRI‐based dosimetry is not yet fully developed for the H&N area due to the inter‐patient variability and the difficulty of accurately quantifying cortical bone based on MRI measure. Purpose: Despite the difficulty of cortical bone segmentation with nearby air area, most papers in the literature show dosimetry results very close to the reference CT scan dosimetry. This study evaluated the impact of bone segmentation on dosimetry. Methods: A home‐made phantom was used to evaluate the Hounsfield unit (HU) values of different tissue at diverse scanner energies. Different dosimetry plans were considered and compared with the reference CT‐based RT plan used for the treatment: CT images with cortical bone voxels assigned to the density of water and three based on MR images using a density assignment (DA) method for synthetic CT (sCT) generation with various bone attribution. Images from 24 patients were compared using mean absolute error (MAE). The generated dosimetry parameters were compared on the basis of dose–volume histogram values and 3%/3 mm gamma pass rates. Results: Among the sCT, the best MAE were found for the DA sCT without bone. The mean of the median dose difference for the planning tumor volume (PTV) was less than 2% for all sCT. Mean gamma pass rate for the MR‐based sCTs were comparable. Conclusions: The DA of the bone can be of interest if correctly segmented, but in the absence of certainty, an overestimation will lead to significant errors while an underestimation gives results close to the reference. The relevance of very precise sCT can be questioned if it requires too many resources and overlooks patients with artifacts.
- 4 April 2026
- Published by Agence Bibliographique de l'Enseignement Supérieur
Hépatopathies chroniques : méthodes de quantification en IRM à 1,5 T et 3,0 T pour le diagnostic et le suivi
La fibrose hépatique est la complication de toutes les hépatopathies chroniques et la cirrhose, correspondant à son stade évolué, est responsable d'une importante morbi-mortalité. En lien avec l'augmentation de l'incidence du diabète de type 2, des syndromes métaboliques et de l'obésité l'incidence de la stéatose hépatique non-alcoolique a considérablement augmentée au cours de ces 10 dernières années pour devenir la première cause d'hépatopathies chroniques dans les pays industrialisés. Face à une méthode de référence invasive pour le diagnostic et exposée à une importante variabilité inter et intra-observateur, aux erreurs d'échantillonnage ainsi qu'à un coût élevé, il existe un besoin clinique pour concevoir des méthodes non-invasives pour le diagnostic et le suivi des hépatopathies chroniques. Ainsi, l'objectif de ces travaux de thèse, était de proposer une méthodologie complète permettant la quantification de la graisse intra-hépatique et la quantification de la fibrose par IRM à 1,5 et 3,0 T. Pour la quantification de la fibrose, une méthode permettant la réalisation de cartographies des paramètres hémodynamiques hépatiques en imagerie de perfusion a été développée à 1,5 T. Ensuite, dans l'idée d'utiliser une approche multiparamétrique, nous avons optimisé un protocole permettant l'imagerie du mouvement incohérent intra-voxel (IVIM) sur le foie afin de l'associer à l'imagerie de perfusion à 3,0 T. Concernant la quantification de la graisse intra-hépatique, une méthode incluant une correction des effets des temps de relaxation par estimation séparée des valeurs de T1 et T2* des protons de l'eau et des lipides, une prise en compte des 5 composantes spectrales principales des lipides, ainsi qu'une procédure permettant la levée de l'ambigüité des composantes dominantes a été développée et évaluée à 1,5 et 3,0 T. Les paramètres hémodynamiques hépatiques quantifiés, en particulier le débit portal et l'index de perfusion hépatique, permettent d'établir la distinction entre l'absence de fibrose, une fibrose débutante, une fibrose avancée et une cirrhose. Pour la quantification de la graisse, la fraction volumique de graisse (FVG) donnée par notre méthode permet de quantifier de manière précise la quantité de graisse intra-hépatique en s'affranchissant des différents facteurs pouvant biaiser l’estimation. Par ailleurs, la FVG permet la distinction entre les différents stades histologiques de stéatose avec une excellente sensibilité/spécificité. L'association de l'imagerie de perfusion avec l'imagerie du mouvement incohérent intra-voxel est en cours d'évaluation à travers l'étude HEPATOMAP qui combine toute la méthodologie développée. Les résultats préliminaires de cette étude ont déjà permis de montrer que l'association entre l'IVIM et la méthode de quantification de la graisse permet de distinguer entre la stéatose pure et la stéato-hépatite
- 4 April 2026
- Published by Agence Bibliographique de l'Enseignement Supérieur
Hépatopathies chroniques : méthodes de quantification en IRM à 1,5 T et 3,0 T pour le diagnostic et le suivi
La fibrose hépatique est la complication de toutes les hépatopathies chroniques et la cirrhose, correspondant à son stade évolué, est responsable d'une importante morbi-mortalité. En lien avec l'augmentation de l'incidence du diabète de type 2, des syndromes métaboliques et de l'obésité l'incidence de la stéatose hépatique non-alcoolique a considérablement augmentée au cours de ces 10 dernières années pour devenir la première cause d'hépatopathies chroniques dans les pays industrialisés. Face à une méthode de référence invasive pour le diagnostic et exposée à une importante variabilité inter et intra-observateur, aux erreurs d'échantillonnage ainsi qu'à un coût élevé, il existe un besoin clinique pour concevoir des méthodes non-invasives pour le diagnostic et le suivi des hépatopathies chroniques. Ainsi, l'objectif de ces travaux de thèse, était de proposer une méthodologie complète permettant la quantification de la graisse intra-hépatique et la quantification de la fibrose par IRM à 1,5 et 3,0 T. Pour la quantification de la fibrose, une méthode permettant la réalisation de cartographies des paramètres hémodynamiques hépatiques en imagerie de perfusion a été développée à 1,5 T. Ensuite, dans l'idée d'utiliser une approche multiparamétrique, nous avons optimisé un protocole permettant l'imagerie du mouvement incohérent intra-voxel (IVIM) sur le foie afin de l'associer à l'imagerie de perfusion à 3,0 T. Concernant la quantification de la graisse intra-hépatique, une méthode incluant une correction des effets des temps de relaxation par estimation séparée des valeurs de T1 et T2* des protons de l'eau et des lipides, une prise en compte des 5 composantes spectrales principales des lipides, ainsi qu'une procédure permettant la levée de l'ambigüité des composantes dominantes a été développée et évaluée à 1,5 et 3,0 T. Les paramètres hémodynamiques hépatiques quantifiés, en particulier le débit portal et l'index de perfusion hépatique, permettent d'établir la distinction entre l'absence de fibrose, une fibrose débutante, une fibrose avancée et une cirrhose. Pour la quantification de la graisse, la fraction volumique de graisse (FVG) donnée par notre méthode permet de quantifier de manière précise la quantité de graisse intra-hépatique en s'affranchissant des différents facteurs pouvant biaiser l’estimation. Par ailleurs, la FVG permet la distinction entre les différents stades histologiques de stéatose avec une excellente sensibilité/spécificité. L'association de l'imagerie de perfusion avec l'imagerie du mouvement incohérent intra-voxel est en cours d'évaluation à travers l'étude HEPATOMAP qui combine toute la méthodologie développée. Les résultats préliminaires de cette étude ont déjà permis de montrer que l'association entre l'IVIM et la méthode de quantification de la graisse permet de distinguer entre la stéatose pure et la stéato-hépatite
- 18 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-3
AWESOMME: A Web-Based Medical Imaging Processing Platform for Multiple Use Cases
Access to state-of-art algorithms in the field of medical imaging is often difficult in hospital environments with a limited adaptation capability. Addressing these challenges, the AWESOMME platform has been developed as a secure compre-hensive web-based toolbox for analyzing medical images, serving clinical, and research purposes while allowing action traceability. New developments have been made to offer improvements for administrators, through a new deployment system, and users with three additional application cases. Each case has a dedicated and independent viewer mode with tailored functionalities: the initial Awesomme Workflow (from segmentation to treatment prediction for osteosarcoma analysis), Prepanext for education on annotations, Score for post-stroke assessment, and HERESP for MRI reconstruction and heterogeneity assessment. It stands as a unique open-source web platform, providing several complete analysis pipelines and enabling a comparison feature, unprece-dented in existing solutions.
- 18 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-3
AWESOMME: A Web-Based Medical Imaging Processing Platform for Multiple Use Cases
Access to state-of-art algorithms in the field of medical imaging is often difficult in hospital environments with a limited adaptation capability. Addressing these challenges, the AWESOMME platform has been developed as a secure compre-hensive web-based toolbox for analyzing medical images, serving clinical, and research purposes while allowing action traceability. New developments have been made to offer improvements for administrators, through a new deployment system, and users with three additional application cases. Each case has a dedicated and independent viewer mode with tailored functionalities: the initial Awesomme Workflow (from segmentation to treatment prediction for osteosarcoma analysis), Prepanext for education on annotations, Score for post-stroke assessment, and HERESP for MRI reconstruction and heterogeneity assessment. It stands as a unique open-source web platform, providing several complete analysis pipelines and enabling a comparison feature, unprece-dented in existing solutions.
- 30 April 2024
- Published by Springer Nature in Journal of Imaging Informatics in Medicine
- Vol. 37 (5) , 2612-2626
Development of a Secure Web-Based Medical Imaging Analysis Platform: The AWESOMME Project
Precision medicine research benefits from machine learning in the creation of robust models adapted to the processing of patient data. This applies both to pathology identification in images, i.e., annotation or segmentation, and to computer-aided diagnostic for classification or prediction. It comes with the strong need to exploit and visualize large volumes of images and associated medical data. The work carried out in this paper follows on from a main case study piloted in a cancer center. It proposes an analysis pipeline for patients with osteosarcoma through segmentation, feature extraction and application of a deep learning model to predict response to treatment. The main aim of the AWESOMME project is to leverage this work and implement the pipeline on an easy-to-access, secure web platform. The proposed WEB application is based on a three-component architecture: a data server, a heavy computation and authentication server and a medical imaging web-framework with a user interface. These existing components have been enhanced to meet the needs of security and traceability for the continuous production of expert data. It innovates by covering all steps of medical imaging processing (visualization and segmentation, feature extraction and aided diagnostic) and enables the test and use of machine learning models. The infrastructure is operational, deployed in internal production and is currently being installed in the hospital environment. The extension of the case study and user feedback enabled us to fine-tune functionalities and proved that AWESOMME is a modular solution capable to analyze medical data and share research algorithms with in-house clinicians.
- 30 April 2024
- Published by Springer Nature in Journal of Imaging Informatics in Medicine
- Vol. 37 (5) , 2612-2626
Development of a Secure Web-Based Medical Imaging Analysis Platform: The AWESOMME Project
Precision medicine research benefits from machine learning in the creation of robust models adapted to the processing of patient data. This applies both to pathology identification in images, i.e., annotation or segmentation, and to computer-aided diagnostic for classification or prediction. It comes with the strong need to exploit and visualize large volumes of images and associated medical data. The work carried out in this paper follows on from a main case study piloted in a cancer center. It proposes an analysis pipeline for patients with osteosarcoma through segmentation, feature extraction and application of a deep learning model to predict response to treatment. The main aim of the AWESOMME project is to leverage this work and implement the pipeline on an easy-to-access, secure web platform. The proposed WEB application is based on a three-component architecture: a data server, a heavy computation and authentication server and a medical imaging web-framework with a user interface. These existing components have been enhanced to meet the needs of security and traceability for the continuous production of expert data. It innovates by covering all steps of medical imaging processing (visualization and segmentation, feature extraction and aided diagnostic) and enables the test and use of machine learning models. The infrastructure is operational, deployed in internal production and is currently being installed in the hospital environment. The extension of the case study and user feedback enabled us to fine-tune functionalities and proved that AWESOMME is a modular solution capable to analyze medical data and share research algorithms with in-house clinicians.
- Open Access
- 28 June 2023
- Published by MDPI AG in Journal of Personalized Medicine
- Vol. 13 (7) , 1062
Characterization of Breast Tumors from MR Images Using Radiomics and Machine Learning Approaches
Determining histological subtypes, such as invasive ductal and invasive lobular carcinomas (IDCs and ILCs) and immunohistochemical markers, such as estrogen response (ER), progesterone response (PR), and the HER2 protein status is important in planning breast cancer treatment. MRI-based radiomic analysis is emerging as a non-invasive substitute for biopsy to determine these signatures. We explore the effectiveness of radiomics-based and CNN (convolutional neural network)-based classification models to this end. T1-weighted dynamic contrast-enhanced, contrast-subtracted T1, and T2-weighted MR images of 429 breast cancer tumors from 323 patients are used. Various combinations of input data and classification schemes are applied for ER+ vs. ER−, PR+ vs. PR−, HER2+ vs. HER2−, and IDC vs. ILC classification tasks. The best results were obtained for the ER+ vs. ER− and IDC vs. ILC classification tasks, with their respective AUCs reaching 0.78 and 0.73 on test data. The results with multi-contrast input data were generally better than the mono-contrast alone. The radiomics and CNN-based approaches generally exhibited comparable results. ER and IDC/ILC classification results were promising. PR and HER2 classifications need further investigation through a larger dataset. Better results by using multi-contrast data might indicate that multi-parametric quantitative MRI could be used to achieve more reliable classifiers.
- 28 June 2023
- Published by MDPI AG in Journal of Personalized Medicine
- Vol. 13 (7) , 1062
Open AccessCharacterization of Breast Tumors from MR Images Using Radiomics and Machine Learning Approaches
Determining histological subtypes, such as invasive ductal and invasive lobular carcinomas (IDCs and ILCs) and immunohistochemical markers, such as estrogen response (ER), progesterone response (PR), and the HER2 protein status is important in planning breast cancer treatment. MRI-based radiomic analysis is emerging as a non-invasive substitute for biopsy to determine these signatures. We explore the effectiveness of radiomics-based and CNN (convolutional neural network)-based classification models to this end. T1-weighted dynamic contrast-enhanced, contrast-subtracted T1, and T2-weighted MR images of 429 breast cancer tumors from 323 patients are used. Various combinations of input data and classification schemes are applied for ER+ vs. ER−, PR+ vs. PR−, HER2+ vs. HER2−, and IDC vs. ILC classification tasks. The best results were obtained for the ER+ vs. ER− and IDC vs. ILC classification tasks, with their respective AUCs reaching 0.78 and 0.73 on test data. The results with multi-contrast input data were generally better than the mono-contrast alone. The radiomics and CNN-based approaches generally exhibited comparable results. ER and IDC/ILC classification results were promising. PR and HER2 classifications need further investigation through a larger dataset. Better results by using multi-contrast data might indicate that multi-parametric quantitative MRI could be used to achieve more reliable classifiers.
- 27 May 2023
- Published by Wiley in Journal of Magnetic Resonance Imaging
- Vol. 59 (3) , 894-906
Diffusion‐Weighted MRI of the Liver in Patients With Chronic Liver Disease: A Comparative Study Between Different Fitting Approaches and Diffusion Models
Background: Diffusion‐weighted imaging (DWI) has been considered for chronic liver disease (CLD) characterization. Grading of liver fibrosis is important for disease management.Purpose: To investigate the relationship between DWI's parameters and CLD‐related features (particularly regarding fibrosis assessment).Study Type: Retrospective.Subjects: Eighty‐five patients with CLD (age: 47.9 ± 15.5, 42.4% females).Field Strength/Sequence: 3‐T, spin echo‐echo planar imaging (SE‐EPI) with 12 b‐values (0–800 s/mm2).Assessment: Several models statistical models, stretched exponential model, and intravoxel incoherent motion were simulated. The corresponding parameters (Ds, σ, DDC, α, f, D, D*) were estimated on simulation and in vivo data using the nonlinear least squares (NLS), segmented NLS, and Bayesian methods. The fitting accuracy was analyzed on simulated Rician noised DWI. In vivo, the parameters were averaged from five central slices entire liver to compare correlations with histological features (inflammation, fibrosis, and steatosis). Then, the differences between mild (F0–F2) or severe (F3–F6) groups were compared respecting to statistics and classification. A total of 75.3% of patients used to build various classifiers (stratified split strategy and 10‐folders cross‐validation) and the remaining for testing.Statistical Tests: Mean squared error, mean average percentage error, spearman correlation, Mann–Whitney U‐test, receiver operating characteristic (ROC) curve, area under ROC curve (AUC), sensitivity, specificity, accuracy, precision. A P‐value Ds, steatosis: r = −0.46, D*, fibrosis: r = −0.24) and significant differences (Ds, σ, D*, f) were observed for Bayesian fitted parameters. Fibrosis classification was performed with an AUC of 0.92 (0.91 sensitivity and 0.70 specificity) with the aforementioned diffusion parameters based on the decision tree method.Data Conclusion: These results indicate that Bayesian fitted parameters may provide a noninvasive evaluation of fibrosis with decision tree.Evidence Level: 1Technical Efficacy: Stage 1
- 27 May 2023
- Published by Wiley in Journal of Magnetic Resonance Imaging
- Vol. 59 (3) , 894-906
Diffusion‐Weighted MRI of the Liver in Patients With Chronic Liver Disease: A Comparative Study Between Different Fitting Approaches and Diffusion Models
Background: Diffusion‐weighted imaging (DWI) has been considered for chronic liver disease (CLD) characterization. Grading of liver fibrosis is important for disease management.Purpose: To investigate the relationship between DWI's parameters and CLD‐related features (particularly regarding fibrosis assessment).Study Type: Retrospective.Subjects: Eighty‐five patients with CLD (age: 47.9 ± 15.5, 42.4% females).Field Strength/Sequence: 3‐T, spin echo‐echo planar imaging (SE‐EPI) with 12 b‐values (0–800 s/mm2).Assessment: Several models statistical models, stretched exponential model, and intravoxel incoherent motion were simulated. The corresponding parameters (Ds, σ, DDC, α, f, D, D*) were estimated on simulation and in vivo data using the nonlinear least squares (NLS), segmented NLS, and Bayesian methods. The fitting accuracy was analyzed on simulated Rician noised DWI. In vivo, the parameters were averaged from five central slices entire liver to compare correlations with histological features (inflammation, fibrosis, and steatosis). Then, the differences between mild (F0–F2) or severe (F3–F6) groups were compared respecting to statistics and classification. A total of 75.3% of patients used to build various classifiers (stratified split strategy and 10‐folders cross‐validation) and the remaining for testing.Statistical Tests: Mean squared error, mean average percentage error, spearman correlation, Mann–Whitney U‐test, receiver operating characteristic (ROC) curve, area under ROC curve (AUC), sensitivity, specificity, accuracy, precision. A P‐value Ds, steatosis: r = −0.46, D*, fibrosis: r = −0.24) and significant differences (Ds, σ, D*, f) were observed for Bayesian fitted parameters. Fibrosis classification was performed with an AUC of 0.92 (0.91 sensitivity and 0.70 specificity) with the aforementioned diffusion parameters based on the decision tree method.Data Conclusion: These results indicate that Bayesian fitted parameters may provide a noninvasive evaluation of fibrosis with decision tree.Evidence Level: 1Technical Efficacy: Stage 1
- 30 September 2022
- Published by Radiological Society of North America (RSNA) in Radiology: Imaging Cancer
- Vol. 4 (5) , e210107
Prediction of Histologic Neoadjuvant Chemotherapy Response in Osteosarcoma Using Pretherapeutic MRI Radiomics
The use of radiomics combined with machine learning based on common MRI can be easily used to provide a new biomarker to predict poor response to chemotherapy and to modify the monitoring of neoadj... Histologic response to chemotherapy for osteosarcoma is one of the most important prognostic factors for survival, but assessment occurs after surgery. Although tumor imaging is used for surgical p...
- 30 September 2022
- Published by Radiological Society of North America (RSNA) in Radiology: Imaging Cancer
- Vol. 4 (5) , e210107
Prediction of Histologic Neoadjuvant Chemotherapy Response in Osteosarcoma Using Pretherapeutic MRI Radiomics
The use of radiomics combined with machine learning based on common MRI can be easily used to provide a new biomarker to predict poor response to chemotherapy and to modify the monitoring of neoadj... Histologic response to chemotherapy for osteosarcoma is one of the most important prognostic factors for survival, but assessment occurs after surgery. Although tumor imaging is used for surgical p...
- 13 July 2021
- Published by Springer Nature in Magnetic Resonance Materials in Physics, Biology and Medicine
- Vol. 34 (6) , 833-844
Multicontrast MRI-based radiomics for the prediction of pathological complete response to neoadjuvant chemotherapy in patients with early triple negative breast cancer
Introduction To assess pre-therapeutic MRI-based radiomic analysis to predict the pathological complete response to neoadjuvant chemotherapy (NAC) in women with early triple negative breast cancer (TN). Materials and methods This monocentric retrospective study included 75 TN female patients with MRI (T1-weighted, T2-weighted, diffusion-weighted and dynamic contrast enhancement images) performed before NAC. For each patient, the tumor(s) and the parenchyma were independently segmented and analyzed with radiomic analysis to extract shape, size, and texture features. Several sets of features were realized based on the 4 different sequence images. Performances of 4 classifiers (random forest, multilayer perceptron, support vector machine (SVM) with linear or quadratic kernel) were compared based on pathological complete response (defined on the excised tissues), on 100 draws with 75% as training set and 25% as test. Results The combination of features extracted from different MR images improved the classifier performance (more precisely, the features from T1W, T2W and DWI). The SVM with quadratic kernel showed the best performance with a mean AUC of 0.83, a sensitivity of 0.85 and a specificity of 0.75 in the test set. Conclusion MRI-based radiomics may be relevant to predict NAC response in TN cancer. Our results promote the use of multi-contrast MRI sources for radiomics, providing enrich source of information to enhance model generalization.
- 13 July 2021
- Published by Springer Nature in Magnetic Resonance Materials in Physics, Biology and Medicine
- Vol. 34 (6) , 833-844
Multicontrast MRI-based radiomics for the prediction of pathological complete response to neoadjuvant chemotherapy in patients with early triple negative breast cancer
Introduction To assess pre-therapeutic MRI-based radiomic analysis to predict the pathological complete response to neoadjuvant chemotherapy (NAC) in women with early triple negative breast cancer (TN). Materials and methods This monocentric retrospective study included 75 TN female patients with MRI (T1-weighted, T2-weighted, diffusion-weighted and dynamic contrast enhancement images) performed before NAC. For each patient, the tumor(s) and the parenchyma were independently segmented and analyzed with radiomic analysis to extract shape, size, and texture features. Several sets of features were realized based on the 4 different sequence images. Performances of 4 classifiers (random forest, multilayer perceptron, support vector machine (SVM) with linear or quadratic kernel) were compared based on pathological complete response (defined on the excised tissues), on 100 draws with 75% as training set and 25% as test. Results The combination of features extracted from different MR images improved the classifier performance (more precisely, the features from T1W, T2W and DWI). The SVM with quadratic kernel showed the best performance with a mean AUC of 0.83, a sensitivity of 0.85 and a specificity of 0.75 in the test set. Conclusion MRI-based radiomics may be relevant to predict NAC response in TN cancer. Our results promote the use of multi-contrast MRI sources for radiomics, providing enrich source of information to enhance model generalization.
- Open Access
- 28 October 2020
- Published by Springer Nature in Cancer Imaging
- Vol. 20 (1) , 1-8
MRI-based radiomics to predict lipomatous soft tissue tumors malignancy: a pilot study
Objectives: To develop and validate a MRI-based radiomic method to predict malignancies in lipomatous soft tissue tumors. Methods: This retrospective study searched in the database of our pathology department, data from patients with lipomatous soft tissue tumors, with histology and gadolinium-contrast enhanced T1w MR images, obtained from 56 centers with non-uniform protocols. For each tumor, 87 radiomic features were extracted by two independent observers to evaluate the inter-observer reproducibility. A reduction of learning base dimension was performed from reproducibility and relevancy criteria. A model was subsequently prototyped using a linear support vector machine to predict malignant lesions. Results: Eighty-one subjects with lipomatous soft tissue tumors including 40 lipomas and 41 atypical lipomatous tumors or well-differentiated liposarcomas with fat-suppressed T1w contrast enhanced MR images available were retrospectively enrolled. Based on a Pearson’s correlation coefficient threshold at 0.8, 55 out of 87 (63.2%) radiomic features were considered reproducible. Further introduction of relevancy finally selected 35 radiomic features to be integrated in the model. To predict malignant tumors, model diagnostic performances were as follow: AUROC = 0.96; sensitivity = 100%; specificity = 90%; positive predictive value = 90.9%; negative predictive value = 100% and overall accuracy = 95.0%. Conclusion: This work demonstrates that radiomics allows to predict malignancy in soft tissue lipomatous tumors with routinely used MR acquisition in clinical oncology. These encouraging results need to be further confirmed in an external validation population.
- 28 October 2020
- Published by Springer Nature in Cancer Imaging
- Vol. 20 (1) , 1-8
Open AccessMRI-based radiomics to predict lipomatous soft tissue tumors malignancy: a pilot study
Objectives: To develop and validate a MRI-based radiomic method to predict malignancies in lipomatous soft tissue tumors. Methods: This retrospective study searched in the database of our pathology department, data from patients with lipomatous soft tissue tumors, with histology and gadolinium-contrast enhanced T1w MR images, obtained from 56 centers with non-uniform protocols. For each tumor, 87 radiomic features were extracted by two independent observers to evaluate the inter-observer reproducibility. A reduction of learning base dimension was performed from reproducibility and relevancy criteria. A model was subsequently prototyped using a linear support vector machine to predict malignant lesions. Results: Eighty-one subjects with lipomatous soft tissue tumors including 40 lipomas and 41 atypical lipomatous tumors or well-differentiated liposarcomas with fat-suppressed T1w contrast enhanced MR images available were retrospectively enrolled. Based on a Pearson’s correlation coefficient threshold at 0.8, 55 out of 87 (63.2%) radiomic features were considered reproducible. Further introduction of relevancy finally selected 35 radiomic features to be integrated in the model. To predict malignant tumors, model diagnostic performances were as follow: AUROC = 0.96; sensitivity = 100%; specificity = 90%; positive predictive value = 90.9%; negative predictive value = 100% and overall accuracy = 95.0%. Conclusion: This work demonstrates that radiomics allows to predict malignancy in soft tissue lipomatous tumors with routinely used MR acquisition in clinical oncology. These encouraging results need to be further confirmed in an external validation population.
- 7 April 2020
- Published by Wolters Kluwer Health in Neurology
- Vol. 94 (14) , e1480-e1487
Fat fraction distribution in lower limb muscles of patients with CMT1A
Objective To quantitatively describe the MRI fat infiltration pattern of muscle degeneration in Charcot-Marie-Tooth (CMT) type 1A (CMT1A) disease and to look for correlations with clinical variables. Methods MRI fat fraction was assessed in lower-limb musculature of patients with CMT1A and healthy controls. More particularly, 14 muscle compartments were selected at leg and thigh levels and for proximal, distal, and medial slices. Muscle fat infiltration profile was determined quantitatively in each muscle compartment and along the entire volume of acquisition to determine a length-dependent gradient of fat infiltration. Clinical impairment was evaluated with muscle strength measurements and CMT Examination Scores (CMTESs). Results A total of 16 patients with CMT1A were enrolled and compared to 11 healthy controls. Patients with CMT1A showed a larger muscle fat fraction at leg and thigh levels with a proximal-to-distal gradient. At the leg level, the largest fat infiltration was quantified in the anterior and lateral compartments. CMTES was correlated with fat fraction, especially in the anterior compartment of leg muscles. Strength of plantar flexion was also correlated with fat fraction of the posterior compartments of leg muscles. Conclusion On the basis of quantitative MRI measurements combined with a dedicated segmentation method, muscle fat infiltration quantified in patients with CMT1A disclosed a length-dependent peroneal-type pattern of fat infiltration and was correlated to main clinical variables. Quantification of fat fraction at different levels of the leg anterior compartment might be of interest in future clinical trials.
- 7 April 2020
- Published by Wolters Kluwer Health in Neurology
- Vol. 94 (14) , e1480-e1487
Fat fraction distribution in lower limb muscles of patients with CMT1A
Objective To quantitatively describe the MRI fat infiltration pattern of muscle degeneration in Charcot-Marie-Tooth (CMT) type 1A (CMT1A) disease and to look for correlations with clinical variables. Methods MRI fat fraction was assessed in lower-limb musculature of patients with CMT1A and healthy controls. More particularly, 14 muscle compartments were selected at leg and thigh levels and for proximal, distal, and medial slices. Muscle fat infiltration profile was determined quantitatively in each muscle compartment and along the entire volume of acquisition to determine a length-dependent gradient of fat infiltration. Clinical impairment was evaluated with muscle strength measurements and CMT Examination Scores (CMTESs). Results A total of 16 patients with CMT1A were enrolled and compared to 11 healthy controls. Patients with CMT1A showed a larger muscle fat fraction at leg and thigh levels with a proximal-to-distal gradient. At the leg level, the largest fat infiltration was quantified in the anterior and lateral compartments. CMTES was correlated with fat fraction, especially in the anterior compartment of leg muscles. Strength of plantar flexion was also correlated with fat fraction of the posterior compartments of leg muscles. Conclusion On the basis of quantitative MRI measurements combined with a dedicated segmentation method, muscle fat infiltration quantified in patients with CMT1A disclosed a length-dependent peroneal-type pattern of fat infiltration and was correlated to main clinical variables. Quantification of fat fraction at different levels of the leg anterior compartment might be of interest in future clinical trials.