Abstract:Objective: To investigate the effect of the number of medical case for the automatic delineation on organ at risk based on template database of MIM software, and to guide clinical practice to select optimal amount of cases so as to constitute database of automatic delineation. Methods: 150 patients with nasopharyngeal carcinoma were divided into four groups that included of group a (10 cases), group b (50 cases), group c (100 cases) and group d (150 cases). And the database of four groups was constructed in the MIM software. And the database included of CT image and the outline information of organ at risk that delineated by manual operation of doctor. At the same time, the CT images of head of other 10 patients who weren't in the database were automatically delineated for organ at risk. And the automatic delineations were compared with gold standard that delineated by manual operation of doctor. The dice similarity coefficient (DICE), Hausdorff distance and Jaccard distance were used to evaluate the organ at risk of automatic outline. Results: The DICE values of four group were between 0.5 and 0.9. And all of the DICE values of eyeball, inner ear, oral cavity, parotid gland and other organs were higher than 0.8. And all of the means of Hausdorff distance (standard deviation) of eyeball, inner ear, temporomandibular joint, crystal and other organ or tissue at risk were less than 0.2. The Jaccard distance was minishing with the increasing of amount of cases in the template database, and all of the means of inner ear, temporomandibular joint, crystal and optic nerve were less than 0.4. Conclusion:With the increasing of amount of cases in the template database, the DICE values of eyeball, inner ear, spinal marrow, optic nerve and other organs will be higher, while the Hausdorff distances (standard deviation) of eyeball, inner ear, optic nerve and other organs will be smaller, and the Jaccard distances of each organ at risk will also be smaller. The organs at risk included of eyeball, inner ear, spinal marrow, optic nerve and others should adopted template database with more amount of cases to achieve automatic delineation. Besides, the amount of cases in the template database has no effect for the automatic delineation of brainstem, temporomandibular joint, crystal, oral cavity, parotid gland and other organ or tissue at risk.

Exploration for automatically delineation of organ at risk based on medical case number of template database in MIM software

Automatic Segmentation of Organs at Risk for Nasopharyngeal Carcinoma with Smart Segmentation and MIM Atlas

Comprehensive and Clinically Accurate Head and Neck Cancer Organs-at-risk Delineation on a Multi-Institutional Study

Comprehensive and Clinically Accurate Head and Neck Organs at Risk Delineation Via Stratified Deep Learning: A Large-scale Multi-Institutional Study

Segmentation of liver and spleen based on computational anatomy models

A dosimetric evaluation of treatment planning based on optimal auto-segmentation

Synthetic MRI-aided Head-and-Neck Organs-at-Risk Auto-Delineation for CBCT-guided Adaptive Radiotherapy

Deep-learning magnetic resonance imaging-based automatic segmentation for organs-at-risk in the brain: Accuracy and impact on dose distribution

Investigation on performance of multiple AI-based auto-contouring systems in organs at risks (OARs) delineation

Clinical evaluation of the convolutional neural network‑based automatic delineation tool in determining the clinical target volume and organs at risk in rectal cancer radiotherapy

Application of Autosegmentation Software in Esophageal Cancer

Deep learning-based auto-contouring of organs/structures-at-risk for pediatric upper abdominal radiotherapy

A Deep Learning Based Automatic Segmentation Approach for Anatomical Structures in Intensity Modulation Radiotherapy

Evaluation of Deep Learning‐based Auto‐segmentation Algorithms for Delineating Clinical Target Volume and Organs at Risk Involving Data for 125 Cervical Cancer Patients

Multi-organ auto-delineation in head-and-neck MRI for radiation therapy using regional convolutional neural network

Automated Identification of Failure Cases in Organ at Risk Segmentation Using Distance Metrics: A Study on CT Data

A Feasibility Study of Deep Learning-Based Auto-Segmentation Directly Used in VMAT Planning Design and Optimization for Cervical Cancer

Quality assurance of organs-at-risk delineation in radiotherapy

Clinically Applicable Deep Learning Framework for Organs at Risk Delineation in CT Images

Automatic delineation of rectal cancer target volume and organs at risk based on convolutional neural network

Delineation of clinical target volume and organs at risk in cervical cancer radiotherapy by deep learning networks