Publications by authors named "Masoomeh Rahimpour"
Purpose: The aim of this study was to develop a convolutional neural network (CNN) for the automatic detection and segmentation of gliomas using [F]fluoroethyl-L-tyrosine ([F]FET) PET.
Methods: Ninety-three patients (84 in-house/7 external) who underwent a 20-40-min static [F]FET PET scan were retrospectively included. Lesions and background regions were defined by two nuclear medicine physicians using the MIM software, such that delineations by one expert reader served as ground truth for training and testing the CNN model, while delineations by the second expert reader were used to evaluate inter-reader agreement.
Article Synopsis
- The study aimed to create a visual ensemble of deep CNNs to improve 3D segmentation of breast tumors using T1-DCE MRI scans from patients with aggressive breast cancer.
- The methodology involved acquiring multi-center MRI scans, segmenting them by radiologists for training and testing, and using different models to assess segmentation accuracy both quantitatively and qualitatively.
- The results indicated that using subtraction images alongside post-contrast images enhanced segmentation performance, achieving a level of accuracy comparable to inter-radiologist agreement and leading to a significant portion of segmentation regarded as excellent or useful.
Convolutional neural networks (CNNs) for brain tumor segmentation are generally developed using complete sets of magnetic resonance imaging (MRI) sequences for both training and inference. As such, these algorithms are not trained for realistic, clinical scenarios where parts of the MRI sequences which were used for training, are missing during inference. To increase clinical applicability, we proposed a cross-modal distillation approach to leverage the availability of multi-sequence MRI data for training and generate an enriched CNN model which uses only single-sequence MRI data for inference but outperforms a single-sequence CNN model.
View Article and Find Full Text PDF