Automated Neural Architecture Search for Cardiac Amyloidosis Classification from [18F]-Florbetaben PET Images.
Authors
Affiliations (9)
Affiliations (9)
- Department of Information Engineering, University of Pisa, Via G. Caruso 16, 56122, Pisa, Italy. [email protected].
- Bioengineering Unit, Fondazione Toscana G Monasterio, Via Giuseppe Moruzzi, 56124, Pisa, Italy. [email protected].
- Department of Information Engineering, University of Pisa, Via G. Caruso 16, 56122, Pisa, Italy.
- Bioengineering Unit, Fondazione Toscana G Monasterio, Via Giuseppe Moruzzi, 56124, Pisa, Italy.
- Nuclear Medicine Unit, Fondazione Toscana G Monasterio, Via Giuseppe Moruzzi, 56124, Pisa, Italy.
- Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital and Harvard Medical School, Boston, MA, USA.
- Division of Cardiology and Cardiovascular Medicine, Fondazione Toscana G Monasterio, Via Giuseppe Moruzzi, 56124, Pisa, Italy.
- Health Science Interdisciplinary Center, Scuola Universitaria Superiore 'S. Anna", Piazza Martiri della Libertà 33, 56127, Pisa, Italy.
- CNR Institute of Clinical Physiology, Via Giuseppe Moruzzi, 56124, Pisa, Italy.
Abstract
Medical image classification using convolutional neural networks (CNNs) is promising but often requires extensive manual tuning for optimal model definition. Neural architecture search (NAS) automates this process, reducing human intervention significantly. This study applies NAS to [18F]-Florbetaben PET cardiac images for classifying cardiac amyloidosis (CA) sub-types (amyloid light chain (AL) and transthyretin amyloid (ATTR)) and controls. Following data preprocessing and augmentation, an evolutionary cell-based NAS approach with a fixed network macro-structure is employed, automatically deriving cells' micro-structure. The algorithm is executed five times, evaluating 100 mutating architectures per run on an augmented dataset of 4048 images (originally 597), totaling 5000 architectures evaluated. The best network (NAS-Net) achieves 76.95% overall accuracy. K-fold analysis yields mean ± SD percentages of sensitivity, specificity, and accuracy on the test dataset: AL subjects (98.7 ± 2.9, 99.3 ± 1.1, 99.7 ± 0.7), ATTR-CA subjects (93.3 ± 7.8, 78.0 ± 2.9, 70.9 ± 3.7), and controls (35.8 ± 14.6, 77.1 ± 2.0, 96.7 ± 4.4). NAS-derived network performance rivals manually determined networks in the literature while using fewer parameters, validating its automatic approach's efficacy.