Multimodal MRI radiomic models to predict genomic mutations in diffuse intrinsic pontine glioma with missing imaging modalities
PurposePredicting H3.1, TP53, and ACVR1 mutations in DIPG could aid in the selection of therapeutic options. The contribution of clinical data and multi-modal MRI were studied for these three predictive tasks. To keep the maximum number of subjects, which is essential for a rare disease, missing dat...
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2023-02-01
|
| Series: | Frontiers in Medicine |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fmed.2023.1071447/full |
| _version_ | 1797897166290681856 |
|---|---|
| author | Fahad Khalid Jessica Goya-Outi Thibault Escobar Thibault Escobar Volodia Dangouloff-Ros Volodia Dangouloff-Ros Antoine Grigis Cathy Philippe Nathalie Boddaert Nathalie Boddaert Jacques Grill Jacques Grill Vincent Frouin Frédérique Frouin |
| author_facet | Fahad Khalid Jessica Goya-Outi Thibault Escobar Thibault Escobar Volodia Dangouloff-Ros Volodia Dangouloff-Ros Antoine Grigis Cathy Philippe Nathalie Boddaert Nathalie Boddaert Jacques Grill Jacques Grill Vincent Frouin Frédérique Frouin |
| author_sort | Fahad Khalid |
| collection | DOAJ |
| description | PurposePredicting H3.1, TP53, and ACVR1 mutations in DIPG could aid in the selection of therapeutic options. The contribution of clinical data and multi-modal MRI were studied for these three predictive tasks. To keep the maximum number of subjects, which is essential for a rare disease, missing data were considered. A multi-modal model was proposed, collecting all available data for each patient, without performing any imputation.MethodsA retrospective cohort of 80 patients with confirmed DIPG and at least one of the four MR modalities (T1w, T1c, T2w, and FLAIR), acquired with two different MR scanners was built. A pipeline including standardization of MR data and extraction of radiomic features within the tumor was applied. The values of radiomic features between the two MR scanners were realigned using the ComBat method. For each prediction task, the most robust features were selected based on a recursive feature elimination with cross-validation. Five different models, one based on clinical data and one per MR modality, were developed using logistic regression classifiers. The prediction of the multi-modal model was defined as the average of all possible prediction results among five for each patient. The performances of the models were compared using a leave-one-out approach.ResultsThe percentage of missing modalities ranged from 6 to 11% across modalities and tasks. The performance of each individual model was dependent on each specific task, with an AUC of the ROC curve ranging from 0.63 to 0.80. The multi-modal model outperformed the clinical model for each prediction tasks, thus demonstrating the added value of MRI. Furthermore, regardless of performance criteria, the multi-modal model came in the first place or second place (very close to first). In the leave-one-out approach, the prediction of H3.1 (resp. ACVR1 and TP53) mutations achieved a balanced accuracy of 87.8% (resp. 82.1 and 78.3%).ConclusionCompared with a single modality approach, the multi-modal model combining multiple MRI modalities and clinical features was the most powerful to predict H3.1, ACVR1, and TP53 mutations and provided prediction, even in the case of missing modality. It could be proposed in the absence of a conclusive biopsy. |
| first_indexed | 2024-04-10T07:53:15Z |
| format | Article |
| id | doaj.art-8f6cd73c2fc94f7099f75bf6a92d7ef8 |
| institution | Directory Open Access Journal |
| issn | 2296-858X |
| language | English |
| last_indexed | 2024-04-10T07:53:15Z |
| publishDate | 2023-02-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Medicine |
| spelling | doaj.art-8f6cd73c2fc94f7099f75bf6a92d7ef82023-02-23T07:16:48ZengFrontiers Media S.A.Frontiers in Medicine2296-858X2023-02-011010.3389/fmed.2023.10714471071447Multimodal MRI radiomic models to predict genomic mutations in diffuse intrinsic pontine glioma with missing imaging modalitiesFahad Khalid0Jessica Goya-Outi1Thibault Escobar2Thibault Escobar3Volodia Dangouloff-Ros4Volodia Dangouloff-Ros5Antoine Grigis6Cathy Philippe7Nathalie Boddaert8Nathalie Boddaert9Jacques Grill10Jacques Grill11Vincent Frouin12Frédérique Frouin13Laboratoire d'Imagerie Translationnelle en Oncologie (LITO)-U1288, Institut Curie, Inserm, Université Paris-Saclay, Orsay, FranceLaboratoire d'Imagerie Translationnelle en Oncologie (LITO)-U1288, Institut Curie, Inserm, Université Paris-Saclay, Orsay, FranceLaboratoire d'Imagerie Translationnelle en Oncologie (LITO)-U1288, Institut Curie, Inserm, Université Paris-Saclay, Orsay, FranceDOSIsoft SA, Cachan, FranceDepartment of Paediatric Radiology, Hôpital Universitaire Necker Enfants Malades, Paris, FranceInstitut Imagine, Inserm U1163 and U1299, Université Paris Cité, Paris, FranceNeurospin, Institut Joliot, CEA, Gif-sur-Yvette, FranceNeurospin, Institut Joliot, CEA, Gif-sur-Yvette, FranceDepartment of Paediatric Radiology, Hôpital Universitaire Necker Enfants Malades, Paris, FranceInstitut Imagine, Inserm U1163 and U1299, Université Paris Cité, Paris, FranceDépartement Cancérologie de l'enfant et de l'adolescent, Gustave-Roussy, Villejuif, FrancePrédicteurs moléculaires et nouvelles cibles en oncologie-U981, Inserm, Université Paris-Saclay, Villejuif, FranceNeurospin, Institut Joliot, CEA, Gif-sur-Yvette, FranceLaboratoire d'Imagerie Translationnelle en Oncologie (LITO)-U1288, Institut Curie, Inserm, Université Paris-Saclay, Orsay, FrancePurposePredicting H3.1, TP53, and ACVR1 mutations in DIPG could aid in the selection of therapeutic options. The contribution of clinical data and multi-modal MRI were studied for these three predictive tasks. To keep the maximum number of subjects, which is essential for a rare disease, missing data were considered. A multi-modal model was proposed, collecting all available data for each patient, without performing any imputation.MethodsA retrospective cohort of 80 patients with confirmed DIPG and at least one of the four MR modalities (T1w, T1c, T2w, and FLAIR), acquired with two different MR scanners was built. A pipeline including standardization of MR data and extraction of radiomic features within the tumor was applied. The values of radiomic features between the two MR scanners were realigned using the ComBat method. For each prediction task, the most robust features were selected based on a recursive feature elimination with cross-validation. Five different models, one based on clinical data and one per MR modality, were developed using logistic regression classifiers. The prediction of the multi-modal model was defined as the average of all possible prediction results among five for each patient. The performances of the models were compared using a leave-one-out approach.ResultsThe percentage of missing modalities ranged from 6 to 11% across modalities and tasks. The performance of each individual model was dependent on each specific task, with an AUC of the ROC curve ranging from 0.63 to 0.80. The multi-modal model outperformed the clinical model for each prediction tasks, thus demonstrating the added value of MRI. Furthermore, regardless of performance criteria, the multi-modal model came in the first place or second place (very close to first). In the leave-one-out approach, the prediction of H3.1 (resp. ACVR1 and TP53) mutations achieved a balanced accuracy of 87.8% (resp. 82.1 and 78.3%).ConclusionCompared with a single modality approach, the multi-modal model combining multiple MRI modalities and clinical features was the most powerful to predict H3.1, ACVR1, and TP53 mutations and provided prediction, even in the case of missing modality. It could be proposed in the absence of a conclusive biopsy.https://www.frontiersin.org/articles/10.3389/fmed.2023.1071447/fullMRIradiomicspredictionmissing datagenomic mutationdiffuse intrinsic pontine glioma |
| spellingShingle | Fahad Khalid Jessica Goya-Outi Thibault Escobar Thibault Escobar Volodia Dangouloff-Ros Volodia Dangouloff-Ros Antoine Grigis Cathy Philippe Nathalie Boddaert Nathalie Boddaert Jacques Grill Jacques Grill Vincent Frouin Frédérique Frouin Multimodal MRI radiomic models to predict genomic mutations in diffuse intrinsic pontine glioma with missing imaging modalities Frontiers in Medicine MRI radiomics prediction missing data genomic mutation diffuse intrinsic pontine glioma |
| title | Multimodal MRI radiomic models to predict genomic mutations in diffuse intrinsic pontine glioma with missing imaging modalities |
| title_full | Multimodal MRI radiomic models to predict genomic mutations in diffuse intrinsic pontine glioma with missing imaging modalities |
| title_fullStr | Multimodal MRI radiomic models to predict genomic mutations in diffuse intrinsic pontine glioma with missing imaging modalities |
| title_full_unstemmed | Multimodal MRI radiomic models to predict genomic mutations in diffuse intrinsic pontine glioma with missing imaging modalities |
| title_short | Multimodal MRI radiomic models to predict genomic mutations in diffuse intrinsic pontine glioma with missing imaging modalities |
| title_sort | multimodal mri radiomic models to predict genomic mutations in diffuse intrinsic pontine glioma with missing imaging modalities |
| topic | MRI radiomics prediction missing data genomic mutation diffuse intrinsic pontine glioma |
| url | https://www.frontiersin.org/articles/10.3389/fmed.2023.1071447/full |
| work_keys_str_mv | AT fahadkhalid multimodalmriradiomicmodelstopredictgenomicmutationsindiffuseintrinsicpontinegliomawithmissingimagingmodalities AT jessicagoyaouti multimodalmriradiomicmodelstopredictgenomicmutationsindiffuseintrinsicpontinegliomawithmissingimagingmodalities AT thibaultescobar multimodalmriradiomicmodelstopredictgenomicmutationsindiffuseintrinsicpontinegliomawithmissingimagingmodalities AT thibaultescobar multimodalmriradiomicmodelstopredictgenomicmutationsindiffuseintrinsicpontinegliomawithmissingimagingmodalities AT volodiadangouloffros multimodalmriradiomicmodelstopredictgenomicmutationsindiffuseintrinsicpontinegliomawithmissingimagingmodalities AT volodiadangouloffros multimodalmriradiomicmodelstopredictgenomicmutationsindiffuseintrinsicpontinegliomawithmissingimagingmodalities AT antoinegrigis multimodalmriradiomicmodelstopredictgenomicmutationsindiffuseintrinsicpontinegliomawithmissingimagingmodalities AT cathyphilippe multimodalmriradiomicmodelstopredictgenomicmutationsindiffuseintrinsicpontinegliomawithmissingimagingmodalities AT nathalieboddaert multimodalmriradiomicmodelstopredictgenomicmutationsindiffuseintrinsicpontinegliomawithmissingimagingmodalities AT nathalieboddaert multimodalmriradiomicmodelstopredictgenomicmutationsindiffuseintrinsicpontinegliomawithmissingimagingmodalities AT jacquesgrill multimodalmriradiomicmodelstopredictgenomicmutationsindiffuseintrinsicpontinegliomawithmissingimagingmodalities AT jacquesgrill multimodalmriradiomicmodelstopredictgenomicmutationsindiffuseintrinsicpontinegliomawithmissingimagingmodalities AT vincentfrouin multimodalmriradiomicmodelstopredictgenomicmutationsindiffuseintrinsicpontinegliomawithmissingimagingmodalities AT frederiquefrouin multimodalmriradiomicmodelstopredictgenomicmutationsindiffuseintrinsicpontinegliomawithmissingimagingmodalities |