The identification of reliable prognostic tools is a major area of research in Amyotrophic Lateral Sclerosis (ALS), since they would improve the design of clinical trials and the clinical management of patients, allowing timely treatment of disease-related complications and providing to patients the opportunity to plan their life in advance. We aimed to evaluate the accuracy of brain 18Fluorodeoxyglucose-Positron-Emission Tomography (18F-FDG-PET) as an independent predictor of survival in ALS patients. Methods. A prospective cohort study enrolled 418 patients with definite, probable, and probable laboratory-supported ALS from May 2009 until December 2019, who underwent brain 18F-FDG-PET at diagnosis and whose survival time was available. The study was conducted at the ALS Expert Center in Turin, Italy. We discretized the survival time in a finite number of classes in a data-driven fashion by employing a k-means-like strategy. Then we aimed to identify 'hot regions' within the brain with maximal power in discriminating survival classes (k-classes). To identify significant features (i.e., voxels) for each of the k-classes, the Laplacian Scores were evaluated in a class-aware fashion. Then, it was possible to retain the top-m features for each class and use them to train the classification systems (i.e. a Support Vector Machine, SVM). We considered as significant only clusters including >100 contiguous voxels. Results. Data were discretized in three survival profiles: 0 - 2 years; 2 - 5 years; > 5 years. SVMs and K-NN classifiers resulted in an error rate less than 20% for two out of three classes separately. As for class one, the three discriminant clusters included bilateral caudate body and left anterior cingulate cortex. The most discriminant regions were bilateral cerebellar pyramid in class two, and right cerebellar dentate nucleus and left cerebellar nodule in class three. Conclusions. Brain 18F-FDG-PET along with Artificial Intelligence was able to predict with 100% accuracy the survival time range in our ALS cohort. Healthcare professionals can benefit from this prognostic tool for planning patients' management and follow up. 18F-FDG-PET represents a promising biomarker for individual patients' stratification in clinical trials.
Brain 18F-FDG-PET is an independent predictor of survival in Amyotrophic Lateral Sclerosis
Canosa A;Pagani M
2023
Abstract
The identification of reliable prognostic tools is a major area of research in Amyotrophic Lateral Sclerosis (ALS), since they would improve the design of clinical trials and the clinical management of patients, allowing timely treatment of disease-related complications and providing to patients the opportunity to plan their life in advance. We aimed to evaluate the accuracy of brain 18Fluorodeoxyglucose-Positron-Emission Tomography (18F-FDG-PET) as an independent predictor of survival in ALS patients. Methods. A prospective cohort study enrolled 418 patients with definite, probable, and probable laboratory-supported ALS from May 2009 until December 2019, who underwent brain 18F-FDG-PET at diagnosis and whose survival time was available. The study was conducted at the ALS Expert Center in Turin, Italy. We discretized the survival time in a finite number of classes in a data-driven fashion by employing a k-means-like strategy. Then we aimed to identify 'hot regions' within the brain with maximal power in discriminating survival classes (k-classes). To identify significant features (i.e., voxels) for each of the k-classes, the Laplacian Scores were evaluated in a class-aware fashion. Then, it was possible to retain the top-m features for each class and use them to train the classification systems (i.e. a Support Vector Machine, SVM). We considered as significant only clusters including >100 contiguous voxels. Results. Data were discretized in three survival profiles: 0 - 2 years; 2 - 5 years; > 5 years. SVMs and K-NN classifiers resulted in an error rate less than 20% for two out of three classes separately. As for class one, the three discriminant clusters included bilateral caudate body and left anterior cingulate cortex. The most discriminant regions were bilateral cerebellar pyramid in class two, and right cerebellar dentate nucleus and left cerebellar nodule in class three. Conclusions. Brain 18F-FDG-PET along with Artificial Intelligence was able to predict with 100% accuracy the survival time range in our ALS cohort. Healthcare professionals can benefit from this prognostic tool for planning patients' management and follow up. 18F-FDG-PET represents a promising biomarker for individual patients' stratification in clinical trials.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
