Increasing demands for, and shortages of, trained staff already place a heavy burden on health care systems, which can lead to long delays for patients as radiotherapy is planned, and the continued rise in head and neck cancer incidence may make it impossible to maintain even current temporal reporting standards. The duration of resulting delays in treatment initiation ( Figure 1) is associated with an increased risk of both local recurrence and overall mortality. Segmentation is also very time consuming: an expert can spend 4 hours or more on a single case. However, the fact that this process is predominantly done manually means that results may be both inconsistent and imperfectly accurate, leading to large inter- and intrapractitioner variability even among experts and thus variation in care quality. Thus, the efficacy and safety of head and neck radiotherapy depends on the accurate delineation of organs at risk and tumors, a process known as segmentation or contouring. However, strategies are needed to mitigate the dose-dependent adverse effects that result from incidental irradiation of normal anatomical structures ( organs at risk). Where available, most patients will be treated with radiotherapy, which targets the tumor mass and areas at high risk of microscopic tumor spread. This incidence is rising and more than doubling in certain subgroups over the last 30 years. J Med Internet Res 2021 23(7):e26151 doi:10.2196/26151 KeywordsĮach year, 550,000 people worldwide are diagnosed with cancer of the head and neck. With appropriate validation studies and regulatory approvals, this system could improve the efficiency, consistency, and safety of radiotherapy pathways. The model’s generalizability was then demonstrated on 2 distinct open-source data sets, reflecting different centers and countries to model training.Ĭonclusions: Deep learning is an effective and clinically applicable technique for the segmentation of the head and neck anatomy for radiotherapy. We also introduced surface Dice similarity coefficient, a new metric for the comparison of organ delineation, to quantify the deviation between organ at risk surface contours rather than volumes, better reflecting the clinical task of correcting errors in automated organ segmentations. Results: We demonstrated the model’s clinical applicability by assessing its performance on a test set of 21 computed tomography scans from clinical practice, each with 21 organs at risk segmented by 2 independent experts. Methods: The model was trained on a data set of 663 deidentified computed tomography scans acquired in routine clinical practice and with both segmentations taken from clinical practice and segmentations created by experienced radiographers as part of this research, all in accordance with consensus organ at risk definitions. Objective: Adopting a deep learning approach, we aim to demonstrate a 3D U-Net architecture that achieves expert-level performance in delineating 21 distinct head and neck organs at risk commonly segmented in clinical practice. Although auto-segmentation algorithms offer a potentially time-saving solution, the challenges in defining, quantifying, and achieving expert performance remain. This planning process can delay treatment while also introducing interoperator variability, resulting in downstream radiation dose differences. Radiotherapy is an important curative treatment for this disease, but it requires manual time to delineate radiosensitive organs at risk. Online Journal of Public Health Informaticsīackground: Over half a million individuals are diagnosed with head and neck cancer each year globally.Asian/Pacific Island Nursing Journal 15 articles.JMIR Bioinformatics and Biotechnology 38 articles.JMIR Biomedical Engineering 75 articles.Journal of Participatory Medicine 84 articles.JMIR Perioperative Medicine 102 articles.JMIR Rehabilitation and Assistive Technologies 231 articles.JMIR Pediatrics and Parenting 313 articles.Interactive Journal of Medical Research 344 articles.JMIR Public Health and Surveillance 1266 articles.
0 Comments
Leave a Reply. |
Details
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |