A portal image alignment and patient setup verification procedure using moments and correlation techniques

Area of Science:

• Medical Physics
• Radiotherapy Imaging
• Image Analysis

Background:

• Accurate patient setup and treatment field verification are critical in radiotherapy.
• Portal imaging is a key tool for verifying daily treatment delivery.
• Existing methods may lack automation or precision in detecting subtle field shape or patient position errors.

Purpose of the Study:

• To develop an integrated, automated procedure for verifying treatment field shape and patient setup using portal images.
• To enhance the accuracy and efficiency of quality assurance in radiotherapy.

Main Methods:

• Utilized an approved portal image as a reference for automated comparison with subsequent images.
• Employed a moments method for aligning treatment field boundaries, compensating for detector repositioning errors.
• Applied a grey-scale image correlation method to assess anatomical misalignment (translations and in-plane rotations).

Main Results:

• The procedure demonstrated accuracy within 1 mm for translational shifts and 1 degree for in-plane rotations in phantom studies.
• The moments verification method achieved a sensitivity of 6 mm for detecting single leaf placement errors.
• The alignment procedure was computationally efficient, completing in under 12 seconds on a standard PC.

Conclusions:

• The developed integrated procedure offers a fast and accurate solution for automated field shape and patient setup verification in portal imaging.
• This method has the potential to improve radiotherapy quality assurance and patient safety.