Neural-Network-Based Detection of Radiopharmaceutical Extravasation In PET/CT Data

The immediate identification of PET/CT radiopharmaceutical extravasation can eliminate many adverse effects such as misdiagnosis and improper therapy. Radiopharmaceutical extravasation is the leakage of an injected radiotracer from the patient’s intended vein into surrounding tissues. The detection of this phenomenon often requires the use of an external monitoring device; due to a lack of robust visual features that can provide indication that it has occurred. In this thesis, the feasibility of using neural networks trained on PET/CT data to identify extravasation is explored. This approach begins with a novel preprocessing methodology that automatically extracts body weight normalized standard uptake values (SUVbw) from specific regions of interest (ROIs). These measurements are then classified using fully connected artificial neural networks (ANNs) to determine if the patient was extravasated. An alternative approach is examined, where convolutional neural networks (CNNs) are leveraged to explore previously undiscovered features through automatic feature extraction. These two approaches were able to differentiate extravasated from non-extravasated patients with high performance. With additional data and architectural improvements, this work serves as a direction for future neural-network-based PET/CT extravasation detection.