An X-Ray C-Arm Guided Automatic Targeting System for Histotripsy.

Area of Science:

• Medical Physics
• Oncology
• Robotics

Background:

• Histotripsy is a precise, noninvasive cancer therapy using focused ultrasound.
• Current ultrasound-based targeting is limited by factors like bowel gas, bone, or deep tumor locations.
• A novel targeting method is needed to expand histotripsy's applicability.

Purpose of the Study:

• To develop and validate an automated X-ray C-arm based targeting system for histotripsy.
• To overcome limitations of ultrasound guidance for targeting deep or obstructed tumors.
• To enhance the precision and safety of histotripsy cancer treatment.

Main Methods:

• Utilized cone-beam CT (CBCT) for lesion localization and 2D fluoroscopy for transducer positioning.
• Employed a digital model and deep learning for precise pose estimation of the histotripsy transducer.
• Calibrated a robotic arm to the C-arm system for automated transducer alignment.

Main Results:

• Phantom studies demonstrated high accuracy, with a mean Euclidean distance of 1.4 ±0.5 mm between the target and ablation zone center.
• The mean absolute error for the predicted treatment radius was 0.5 ±0.5 mm.
• The system achieved accurate, automated targeting without reliance on ultrasound.

Conclusions:

• CBCT-based histotripsy targeting offers an accurate and automated alternative to ultrasound guidance.
• This approach reduces operator dependency and expands treatment possibilities for tumors previously inaccessible via ultrasound.
• The developed system shows significant potential for advancing noninvasive cancer therapies.