Automated algorithm for CBCT-based dose calculations of prostate radiotherapy with bilateral hip prostheses

Area of Science:

• Medical Physics
• Radiotherapy
• Image Processing

Background:

• Cone-beam CT (CBCT) images exhibit scatter, leading to inaccurate Hounsfield Units (HUs) and precluding direct use in radiotherapy dose calculations.
• Prostate cancer patients with bilateral metallic hip prostheses present unique challenges for CT imaging and radiotherapy planning.

Observation:

• A multilevel threshold (MLT) algorithm was developed to segment CBCT images, categorizing pixel values into homogeneous HU segments while accounting for positional variations.
• An automated MLT algorithm was created to minimize operator time for CBCT image segmentation.

Findings:

• Dose calculations using segmented CBCT (sCBCT) with MLT algorithms demonstrated clinically acceptable accuracy, with gamma evaluation (3%/3mm) exceeding 97.7% for the automated method.
• The automated MLT algorithm achieved a dose difference of -1.3% compared to the planning CT (pCT), significantly reducing operator time from 8 hours.

Implications:

• CBCT image segmentation using MLT algorithms enables accurate dose calculations, crucial for adaptive radiotherapy.
• The automated MLT algorithm offers a feasible and time-efficient solution for adaptive radiotherapy in complex cases, like those with hip prostheses, potentially eliminating the need for replanning.
• This approach enhances the clinical utility of CBCT in radiotherapy, improving treatment delivery and patient management.