NULL

NULL

NULL

NULL

NULL

NULL

Background: MRI has become one of the most important clinical tools for longitudinal analysis of multiple sclerosis (MS). Newly appearing lesions are indicative of the disease progression. Several automatic approaches have been proposed for the detection of newly appearing lesions, which can be classified as either supervised approaches that use intensity-derived features from the subtraction images or unsupervised approaches that use also deformation fields information. Aim: We present here a supervised approach for detecting newly appearing MS lesions that combines both subtraction and deformation field features. Specifically, we use a logistic regression classifier trained with features from the baseline and follow-up intensities, subtraction values, and deformation field operators to provide a final segmentation. Materials and methods: One year apart multi-channel brain MRI were scanned for 60 patients with a 3T magnet, including transverse T2-FLAIR, PD-w, T2-w and T1-w images. 36 of these patients presented new T2-w lesions that were semi-automatically annotated by expert neuroradiologists. The rest had no new lesions in the follow-up scans. All images were pre-processed and co-registered by multi resolution-multi stage affine registration, and a deformation field was also obtained using the Demons non-rigid registration algorithm. Results: We performed a leave-one-out cross-validation strategy using the 36 patients with new T2-w lesions. In terms of detection, we obtained a 74.30% true positive fraction and 11.86% false positive fraction with a mean Dice similarity coefficient of 0.77. In terms of segmentation, we obtained a mean Dice coefficient of 0.56. We compared these results with those obtained with state-of-the-art methods such as Sweeney et al. (2013), Ganiler et al. (2014), and Cabezas et al. (2016), and our model had significantly better results (p 0.05). When testing the model with the 24 patients with no new T2-w lesions, only 5 false positives were found in 4 cases. Conclusions: The proposed model decreases the number of false positives while increasing the number of true positives. The study also proves the benefits of using deformation field operators as features to train a supervised learning model. Our approach is simple and fully automated and reduces user interaction and inter- and intra-observer variability. Disclosure: M. Salem: nothing to disclose. M. Cabezas: nothing to disclose. S. Valverde: nothing to disclose. D. Pareto: has received speaking honoraria fron Novartis and Biogen. A. Oliver: nothing to disclose. J. Salvi: nothing to disclose. A. Rovira serves on scientific advisory boards for Biogen Idec, Novartis, Sanofi-Genzyme, and OLEA Medical, has received speaker honoraria from Bayer, Sanofi-Genzyme, Bracco, Merck-Serono, Teva Pharmaceutical Industries Ltd, Novartis and Biogen Idec, and has research agreements with Siemens. X. Lladó: nothing to disclose.

Background: Convolutional Neural Network (CNN) methods are being proposed for automated white matter lesion segmentation increasing the performance of typical state-of-the-art methods. However, their accuracy decreases significantly when using them on other image domains that those used for training, showing lack of adaptability to unseen imaging data and limiting its applicability in non-specialized hospitals. Aim: To analyze the effect of domain adaptation on multiple sclerosis (MS) lesion segmentation, investigating how transferable a CNN model is when applied to other unseen image domains. Methods: An automated lesion segmentation method based on a 11-layer CNN classifier was firstly fully-trained using 35 T1-w and FLAIR scans from the MS lesion segmentation challenges (MICCAI 2008 and 2016). Then, domain adaptation was independently evaluated on two different datasets composed of 60 and 61 T1w and FLAIR images from a clinical hospital and from the public ISBI2015 challenge, respectively. For each unseen dataset, the same source model was fine-tuned re-training only the last layers but using a single image (we tested the use of images with different lesion load). The Dice overlap (DSC) coefficient between the resulting segmentations and manual lesion annotations was compared with respect to the same model when was fully trained on the target domain and with respect to other methods such as LST. Results: On the clinical dataset, the performance of the model fully-trained with data from the target domain was DSC=0.53. When using the source model without readaptation, the performance dropped to DSC=0.25, while when adapting the source model using a single image the performance ranged between DSC=[0.30-0.48] depending on the lesion load of the image used. In all cases, showed a significant increase in the accuracy with respect to LST (DSC=0.29). On the ISBI2015 challenge, our fully-trained CNN method was ranked 3rd among 59 methods, showing human like segmentation performance. Interestingly, adapted models trained with only one image still yielded a remarkably higher performance than other state-of-the-art methods like LST or lesionToads, showing also a very similar performance to other CNN models trained on larger number of images. Conclusions: Domain adaptation allows to use pre-trained CNNs on unseen clinical settings. A manual delineation of the lesions in only one image is sufficient to obtain accurate automated lesion segmentation performances. Disclosure: S. Valverde: nothing to disclose. M. Salem: nothing to disclose. M. Cabezas: nothing to disclose. D. Pareto: has received speaking honoraria fron Novartis and Biogen. J.C. Vilanova: nothing to disclose. Lluís Ramió-Torrentà: has received compensation for consulting services and speaking honoraria from from Biogen, Novartis, Bayer, Merck, Sanofi, Genzyme, Teva Pharmaceutical Industries Ltd, Almirall, Mylan. A. Rovira serves on scientific advisory boards for Biogen Idec, Novartis, Genzyme, and OLEA Medical, and has received speaker honoraria from Bayer, Genzyme, Sanofi-Aventis, Bracco, Merck-Serono, Teva Pharmaceutical Industries Ltd, OLEA Medical, Stendhal, Novartis and Biogen Idec. A. Oliver: nothing to disclose. J. Salvi: nothing to disclose. X. Lladó: nothing to disclose.

Background: Image synthesis is gaining attention in many domains, including medical imaging. For instance, the generation of synthetic lesions can be used as a solution to the lack of large datasets with multiple sclerosis (MS) lesions manually annotated, which is one of the main limitations to train robust and generalisable supervised machine learning algorithms. Objectives: To propose a fully convolutional neural network (CNN) model for MS lesion synthesis in magnetic resonance images. Materials and methods: T2-FLAIR and T1-w images from a dataset of 65 patients with a clinically isolated syndrome or early relapsing-remitting MS were used to train a CNN able to synthesise new lesions. The inputs of the CNN were processed images without lesions, while the original images with lesions were the outputs of the CNN. To obtain the input images, we computed a white matter hyperintensity (WMH) mask along with several intensity level masks that encoded the intensity profiles of the WMH voxels. Then, the WMH mask was filled with intensities resembling white matter. The CNN architecture to perform the image synthesis consisted of two encoders (one per each modality) that learned the latent representation for the input modalities, and two decoders, that allowed to generate new lesions in both modalities. To evaluate the synthesis, we tested a state-of-the-art MS lesion segmentation approach (Valverde et al. 2017) on an in-house dataset and the public ISBI2015 challenge dataset, showing the performance in different scenarios such as using synthetic images for data augmentation. Results: For the in-house dataset, when adding to a single original image several synthetic ones, the performance of the lesion segmentation increased the sensitivity from 41% to 50% and the positive predictive value (PPV) from 53% to 65%. Repeating the experiment on the ISBI2015 dataset, the sensitivity increased from 44% to 51% and the PPV from 76% to 78%. With the inclusion of few original images and the synthetic data, we were able to increase the detection performance to that of the segmentation algorithm fully trained using the entire available training set, yielding a comparable human expert rater performance. Conclusions: The proposed CNN was able to generate T1-w and T2-FLAIR images with synthetic MS lesions. The combination of original images with synthetic ones of the same domain increased the lesion segmentation accuracy, reducing also the number of manually annotated images of the database. Disclosure: M. Salem: nothing to disclose. S. Valverde: nothing to disclose. M. Cabezas: nothing to disclose. D. Pareto: has received speaking honoraria fron Novartis and Biogen. A. Oliver: nothing to disclose. J. Salvi: nothing to disclose. A. Rovira serves on scientific advisory boards for Novartis, Sanofi-Genzyme, Icometrix, SyntheticMR, Bayer, Biogen and OLEA Medical, and has received speaker honoraria from Bayer, Sanofi-Genzyme, Bracco, Merck-Serono, Teva Pharmaceutical Industries Ltd, Novartis, Roche and Biogen. X. Lladó: nothing to disclose.

Background: Magnetic Resonance Imaging (MRI) has become one of the most important clinical tools for diagnosing and monitoring multiple sclerosis (MS). In particular, new T2 lesions on brain MRI are considered a good biomarker for monitoring and predicting treatment response. Therefore, building automated and accurate methods for the detection of new T2 lesions is a need. Objectives: To propose a fully convolutional neural network (CNN) to detect new T2 lesions in longitudinal brain MRI images. Materials and methods: One year apart multi-channel 3T brain MRI were obtained in 60 MS patients, including transverse T2-FLAIR, PD-w, T2 and T1 images. 36 of those patients presented new T2 lesions that were visually and semi-automatically annotated by expert neuroradiologists. The remaining 24 cases had no new lesions. All Images were pre-processed and co-registered by affine registration. Afterwards, a fully CNN where the inputs were the basal and follow-up images was trained to detect new MS lesions. The first part of the network was a U-Net block that automatically learned the deformation fields (DFs) which nonlinearly registered the basal image to the follow-up space. The learnt DFs together with the basal and follow-up images were then feed to a second block, another U-Net that performed the final detection and segmentation of the new T2 lesions. Results: We performed a leave-one-out cross-validation strategy using the 36 patients with new T2 lesions. The model obtained 82.67% of true positive fraction (TPF), 15.06% of false positive fraction (FPF), and a mean detection and segmentation Dice similarity coefficient of 0.79 and 0.52, respectively. Our model had significantly better results (p 0.05) than those of other state-of-the-art approaches such as Sweeney et al. (2013), Cabezas et al. (2016) and Salem et al. (2018). Regarding the 24 cases with no new T2 lesions, a trained model with all the 36 cases provided only 2 false positive detections. The proposed CNN model was faster in testing time than other state-of-the-art methods since there is no need to perform a non-rigid registration. Conclusions: The proposed CNN approach provides better results than other state-of-the-art methods both in terms of sensitivity and specificity. In addition, the end-to-end learning framework avoids the use of complex processes such as the non-rigid registration and the definition of hand-crafted image features. Disclosure: M. Salem: nothing to disclose. S. Valverde: nothing to disclose. M. Cabezas: nothing to disclose. D. Pareto: has received speaking honoraria fron Novartis and Biogen. A. Oliver: nothing to disclose. J. Salvi: nothing to disclose. A. Rovira serves on scientific advisory boards for Novartis, Sanofi-Genzyme, Icometrix, SyntheticMR, Bayer, Biogen and OLEA Medical, and has received speaker honoraria from Bayer, Sanofi-Genzyme, Bracco, Merck-Serono, Teva Pharmaceutical Industries Ltd, Novartis, Roche and Biogen. X. Lladó: nothing to disclose.