Abstract:WeiwenLai**, Chang Wen Chen**, Diana F. Nelson*, M. C. Schell*, •Rubin*Oncologic Imaging Research LaboratoryDepartments ofRadiation Oncology* and Electrical Engineering**University of Rochester, 601 Elmwood Ave., Box 647, Rochester, New York 14642-8647ABSTRACTA fully automated system is being developed for the portal verification of tangential breast fields inradiation therapy of breast cancer. The automated verification system involves image acquisition, imagefeature extraction, feature correlation between reference and portal images, and quantitative evaluation ofpatient setup. In this study, the portal images are acquired using a matrix liquid ion-chamber electronicportal imaging device (EPID), and have a matrix size of 256 x 256 pixels with 12-bit gray levels. Ahier&chical region processing technique is developed to extract poor contrast features in the portal imagegenerated by megavoltage photon beams at different levels sequentially. The treatment field is ithtiallyextracted from the portal image. The skin line is then extracted from the treatment field. Finally, thelung/soft tissue separation is extracted from the breast region. A Chamfer matching filter is used tocorrelate features in the portal image with those in the reference image. The resulting parameters forrotation, translation and scaling are used for the setup evaluation of the treatment field.Keywords: breast cancer, image processing, portal image, electronic portal imaging device2. INTRODUCTIONBreast cancer is one of the leading cancers in women in USA, causing 46,000 deaths in 1994.1Currently, radiation therapy is one of the most effective approaches for the management of cancers.2 Thegoal of radiation therapy is to deliver prescribed dose to the target volume while keeping the dose tonormal tissues within the tolerance. Accurate definition of radiation treatment field is the key for thesuccess of radiation therapy due to potential radiation damage to the normal tissue and the need for tumorlocal control. In order to ensure the treatment is delivered as planned, a portal film (or verification image)is taken for each set of treatment fields in the treatment machine and is compared to the simulation film (orreference image) which is selected as the gold standard for the treatment field. At present, thiscomparison is performed by radiation oncologists based on the manually-identified features, which is bothtime-consuming and potentially error-prone. With the introduction of various electronic portal imagingdevices,35 real-time patient positioning correction is becoming clinically feasible to replace time-delayedanalysis using films. However, this procedure requires present of radiation oncologists during patienttreatment which is not cost-effective and practically not realistic. Therefore, the quality of cancer patientcare and efficiency of radiation therapy could be substantially improved if this verification procedure canbe automated.During the past several years, some investigators had used computers to analyze portal images toimprove contrast and visibility and to verify the field size.69 There were significant efforts devoted todevelop methodology for the correlation of features between portal and simulation images for automatedcomparison.1044 However, most features used for the image correlation were identified manually. Therewere no fully automated verification systems yet for clinical use of portal verification, mainly due to lackof fine techniques for automated extraction and registration of anatomical features. Previous investigatorstried to use one method to extract anatomical features for all treatment sites, which would be desirable forvarious treatment ports but may not be suitable because the features selected and surrounding background

MR Image-Guided Portal Verification for Brain Treatment Field

Accuracy of UTE-MRI-based patient setup for brain cancer radiation therapy.

Reduction of artifacts in susceptibility-weighted MR venography of the brain.

Development of A Computerized Portal Verification System for Radiation Therapy of Breast Cancers

Registration of On-Board X-ray Images with 4DCT: A Proposed Method of Phase and Setup Verification for Gated Radiotherapy

Results of Images Acquired with Backscattered Mv Radiation Using A Pinhole Collimator

Distortion-Free Diffusion MRI using an MRI-Guided Tri-Cobalt 60 Radiotherapy System: Sequence Verification and Preliminary Clinical Experience.

High Quality Proton Portal Imaging Using Deep Learning for Proton Radiation Therapy: a Phantom Study

Image Registration to Compensate for EPI Distortion in Patients with Brain Tumors: an Evaluation of Tract‐Specific Effects

Technical note: Preprocessing of portal images to improve image quality of VMAT-CT

Clinical implementation of MRI simulation for IMRT of prostate cancer

[Impact of MRI-CT Image Registration on Target Delineation of Postoperative Radiotherapy for Gliomas].

MRI-based Treatment Planning for Brain Stereotactic Radiosurgery: Dosimetric Validation of a Learning-Based Pseudo-Ct Generation Method.

A method for in vivo treatment verification of IMRT and VMAT based on electronic portal imaging device

SU‐E‐J‐142: Performance Study of Automatic Image‐Segmentation Algorithms in Motion Tracking Via MR‐IGRT

Validation of a surface-based deformable MRI-3D ultrasound image registration algorithm toward clinical implementation for interstitial prostate brachytherapy

Automatic Detection of Brain Metastases Using 3D Mask R-CNN for Stereotactic Radiosurgery

Improvement of 2D cine image quality using 3D priors and cycle generative adversarial network for low field MRI‐guided radiation therapy

Tumor Classification and Segmentation of MR Brain Images

Evaluation of the ability of the Brainlab Elements Cranial Distortion Correction algorithm to correct clinically relevant MRI distortions for cranial SRT

Automated deep learning auto-segmentation of air volumes for MRI-guided online adaptive radiation therapy of abdominal tumors