Publications by authors named "Alexandra Zlate"
Purpose: To investigate the performance of 4 atlas-based (multi-ABAS) and 2 deep learning (DL) solutions for head-and-neck (HN) elective nodes (CTVn) automatic segmentation (AS) on CT images.
Material And Methods: Bilateral CTVn levels of 69 HN cancer patients were delineated on contrast-enhanced planning CT. Ten and 49 patients were used for atlas library and for training a mono-centric DL model, respectively.
Salivary gland carcinomas (SGCs) are the most heterogeneous subgroup of head and neck malignant tumors, accounting for more than 20 subtypes. The median age of SGC diagnosis is expected to rise in the following decades, leading to crucial clinical challenges in geriatric oncology. Elderly patients, in comparison with patients aged below 65 years, are generally considered less amenable to receiving state-of-the-art curative treatments for localized disease, such as surgery and radiation/particle therapy.
View Article and Find Full Text PDFArticle Synopsis
- - The study aimed to evaluate the effectiveness of both four atlas-based and two deep learning methods for automatically segmenting organs-at-risk in head-and-neck treatments.
- - Results showed that deep learning algorithms generally outperformed atlas-based methods in accuracy, although one hybrid atlas-based approach sometimes matched the best performance of the deep learning algorithms.
- - While deep learning methods were faster and more efficient for manual corrections, significant dose differences were mostly noticed when organs were close to treatment targets, indicating careful consideration needed in contour placement.