MENU

Click here to

×

Are you sure ?

Yes, do it No, cancel

Accelerate the Treatment Planning Design of Lung and Esophagus Cancer Radiotherapy Using Kernel Density Estimation Based DVH Prediction and Pinnacle Auto-Planning

Y Chen1 , j wang1 , W Hu1 , J Lu1*, (1)Fudan University Shanghai Cancer Center, Shanghai, Shanghai

Presentations

(Sunday, 7/29/2018) 3:00 PM - 6:00 PM

Room: Exhibit Hall

Purpose: To improve the treatment planning design efficiency and the plan quality of lung and esophagus cancer radiotherapy.

Methods: The improvement of the treatment planning design and plan quality for lung and esophagus cancer radiotherapy includes initial dose volume histograms (DVHs) prediction for circumjacent organs and automatic generation of clinically acceptable treatment plans. A kernel density estimation algorithm was used to predict the initial DVHs of a new case from the prediction model trained with 20 clinically accepted plans well-designed by superior medical dosimetrists. Then, optimization objective functions extracted from the predicted DVHs were utilized as the basic initial setting to automatically generate an optimized or sub-optimized plan using the Pinnacle3 Auto-Planning (AP) module (version 9.10, Philips Medical Systems).

Results: The usage of the predicted DVHs avoids the occurrence of irreconcilable contradictions and inadequate restrain and consequently increases the reasonability of the objective functions and speeds-up the optimization. The knowledge-based Pinnacle3 AP module adds assistant dose-shaping structures automatically and emancipates planners from repeated trial-and-error. Treatment plans of 10 patients were designed in the proposed way by an inferior dosimetrist. Both the efficiency and the quality were comparative to the plans designed by the experienced ones.

Conclusion: The combination of DVH prediction and Auto-Planning makes the objective functions more reasonable and accelerates the convergence, reduces the tedious human intervention and improves the plan quality and consistency.

Keywords

Not Applicable / None Entered.

Taxonomy

Not Applicable / None Entered.

Contact Email