Moreover, deep learning is capable of learning high-level and task-adaptive image features. Deep learning has the potential to serve as a more powerful tool to overcome these issues as shown in several studies. For instance, the robustness of the conventional hand-crafted radiomic features is variable based on changing parameters, including pixel size, region-of-interest (ROI) delineation, and signal-to-noise ratio. However, conventional radiomics (CR) has several disadvantages. Nomograms and radiomic pipelines have been used to predict SLN status with promising results. One such task is the development of a predictive model for non-invasive staging of the axillary lymph nodes as an alternative to SLN biopsy. Accurate non-invasive assessment of nodal involvement therefore is valuable in cancer staging, surgical risk, and financial cost reduction.īreast cancer is an area of peaked interest for the combination of radiomics and artificial intelligence, with clinical impact possible as both a diagnostic and prognostic tool. Moreover, studies have reported > 70% of biopsied SLNs are negative, indicating that such procedure is unbeneficial and potentially harmful to a significant amount of breast cancer patients. Īlthough lymph node management has become less invasive with the use of sentinel lymph node (SLN) biopsy as opposed to full axillary lymph node dissection, significant side effects including shoulder dysfunction, lymphedema, and nerve damage were still observed in as much as one-fourth of patients. As a result, lymph node status is critical for diagnosis, prognosis, and monitoring of treatments. Lymph node involvement increases the risk of recurrence and acts as a prognostic indicator, with the survival rate of node-positive patients being up to 40% lower than node-negative patients. This could indicate that DLB features can ultimately result in a more generalizable model.īreast cancer increases in stage and severity as it metastasizes to axillary lymph nodes. More interestingly, these improvements were seen particularly in the independent testing set of dissimilar resolution. Furthermore, in the independent testing set of the dissimilar resolution, the DLB model performed markedly better than the CR model (accuracy 77% vs 71%). The predictive performance of the DLB model outperformed the CR model for this task. Same feature reduction, feature selection, model creation procedures were used for both approaches. In the validation set (same resolution as training), the DLB model outperformed the CR model (accuracy 83% vs 80%). The threshold determined using the training set was applied to the independent validation and testing dataset. Two methods were employed: (1) conventional radiomics (CR), and (2) DLB features which replaced hand-curated features with pre-trained VGG-16 features. The remaining 35 subjects with a different in-plane resolution (0.78 × 0.78 mm 2) were treated as independent testing set for generalizability. Of these subjects, 163 had an in-plane resolution of 0.7 × 0.7 mm 2, which were randomly divided into a training set (approximately 67%) and a validation set (approximately 33%). Specifically, this study aimed to compare the generalizability of radiomics vs DLB features in an independent test set with dissimilar resolution. Dynamic contrast-enhancement images from 198 patients (67 positive SLNs) were used in this study. The purpose of this study was to develop a prediction model for preoperative prediction of SLN metastasis using deep learning-based (DLB) features and compare its predictive performance to state-of-the-art radiomics. Radiomics has been proposed to pre-operatively predict sentinel lymph node (SLN) status however, radiomic models are known to be sensitive to acquisition parameters. An accurate non-invasive assessment of nodal involvement is valuable in cancer staging, surgical risk, and cost savings. Lymph node involvement increases the risk of breast cancer recurrence.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |