MENU

Click here to

×

Are you sure ?

Yes, do it No, cancel

Whether DOSIOMICS Could Benefit to IMRT Treated Patient Survival Prediction: A Feasible Study On Head-And-Neck Cancer Cases

A Wu1*, Y Li2 , M Qi1 , F Guo1 , Q Jia1 , L Zhou1 , T Song1 , (1) Department of Biomedical Engineering, Southern Medical University, Guangzhou, Guangdong, (2) Department of Radiation Oncology, Sun Yat-sen University Cancer Center, Guangzhou, Guangdong

Presentations

(Sunday, 7/14/2019) 1:00 PM - 2:00 PM

Room: Stars at Night Ballroom 2-3

Purpose: To investigate whether dosiomics can benefit to IMRT treated patient’s survival prediction or not by a comparative study on prediction accuracy inspection between traditional radiomics methods and that integrating dosiomics for head and neck cancer IMRT cases.

Methods: A cohort of 237 head-and-neck cancer patients from four different institutions was obtained by The Cancer Imaging Archive (TCIA) and utilized to train and test the traditional radomics algorithm and dosiomics algorithm respectively. For radiomics, initial features were firstly extracted from images including the CTs and PETs, to quantify tumor’s intensity, shape and texture; secondly, the predictive power for each of initial features was calculated naming the concordance index (CI) by a univariate Cox proportional hazards regression model; consequently, initial features were selected by their CI values and condensed by a principle component analysis (PCA) method; lastly, a multivariate Cox proportional hazards regression (MCPHR) model was constructed as overall survival prediction model by inputting the aforementioned condensed features. To the similar procedures, for integrating dosiomics, initial features were composed by not only images but also treatment plan 3-dimentional dose distributions. Head-and-neck cancer patient overall survival prediction models were built using radiomics and dosiomics respectively. Their prediction accuracy was measured by CI of MCPHR models and the differentiation of high and low risk population performed by Kaplan-Meier analysis.

Results: Observed from evaluation dataset, CI of MCPHR model for integrating dosiomics is 0.7475 , relative to that of 0.6956 for radiomics with significant increment (Wilcoxon test, p=3.84×10�³�) . The differentiation degree of high and low risk population also show the superiority of integrating dosiomics over traditional radiomics, with p value of 6.9×10�� and 0.021 respectively.

Conclusion: “Dosiomics� could benefit to IMRT treated patient’s survival prediction and hence should not be neglected for related investigations.

Funding Support, Disclosures, and Conflict of Interest: 1) National Key R&D Program of China (NO.2017YFC0113203); 2) National Natural Science Foundation of China (NO.81571771 and 81601577); 3) Post-doctoral Science Foundation of China (NO.2016M592510). 4) Public Welfare Research and Capacity Building Special Foundation of Guangdong, China (2015B020214002)

Keywords

Radiation Therapy

Taxonomy

Not Applicable / None Entered.

Contact Email