MENU

Click here to

×

Are you sure ?

Yes, do it No, cancel

The Content-Based Standardizing Nomenclatures (CBSN) in Radiotherapy for Nasopharyngeal Carcinoma OARs

X MAI1,2, S HUANG1,2, Z Zhong1,2, W Zheng1,3, S Chen1,2, S Zhou2, S Huang1, Y XIA1, X HUANG1, X Yang1*, (1) Sun Yat-sen University Cancer Center; State Key Laboratory of Oncology in South China; Collaborative Innovation Center for Cancer Medicine; Guangdong Key Laboratory of Nasopharyngeal Carcinoma Diagnosis and Therapy, Guangzhou, Guangdong, 510060, China,(2) XinHua College of Sun Yat-sen University, Guangzhou, Guangdong, 510520, China, (3)Southern Theater Air Force Hospital of the People's Liberation Army, Guangzhou, Guangdong, 510050, China

Presentations

(Sunday, 7/12/2020)   [Eastern Time (GMT-4)]

Room: AAPM ePoster Library

Purpose: Based on AAPM TG-263, a Content-Based Standardizing Nomenclatures (CBSN) is proposed to explore the feasibility of its standardization verification for OARs of nasopharyngeal carcinoma (NPC) patients.

Methods: The radiotherapy structure files of 855 NPC patients who received IMRT from 2017 to 2019 are collected, of which 840 are retrospective and 15 are clinically abnormal structures. The self-developed MATLAB software is used to obtain the (a) image location/body posture, (b) the geometric features of the OARs outlined by the doctors, (c) the first-order gray histogram and (d) the texture features of the Gray-level Co-occurrence Matrix (GLCM), etc., for establishing the CBSN Knowledge Library and CBSN Location Verification Method. In addition, Fisher discriminant analysis is used to establish a CBSN OARs Classification Model, and the model is evaluated using internal-validation, cross-validation, and external validation.

Results: (1) There are different characteristic parameter values between CBSN Knowledge Library and abnormal structures, and between different OARs, such as volume and slice amount, with the significant statistical differences, P<0.001; (2) CBSN OARs Classification Model’s internal-validation, cross-validation and external-validation accuracy rates are 93.50%, 93.40%, and 91.38%, respectively; (3) In symmetrical structure, the accuracy of the left and right location verification is 100%; (4) 15 clinical abnormal structures are successfully detected by CBSN with an accuracy rate of 100%.

Conclusion: CBSN is applicable to NPC OARs verification and discrimination of abnormal structures. It could be developed as a contouring check tool for clinical usage. And this method could be extent to other diseases and to other cancer centers.

Keywords

Radiation Therapy, Structure Analysis

Taxonomy

IM/TH- Informatics: Data archiving - Therapy

Contact Email