Real Time Planning

Real Time Planning

The image shows a computer generated model of the prostate in the OR with the palladium seeds precisely positioned to maximize cure and minimize complications. Dr Doggett has an acknowledged expertise in intraoperative treatment planning. Dr. Doggett's published text (PDF Document) He was the first physician outside of Harvard Medical School to use the Burdette/CMS intraoperative treatment planning system.

Intraoperative treatment planning uses a computer in the operating room to measure the exact shape and position of the prostate, urethra, rectum, bladder and nerves controlling erection. A precise treatment plan sparing normal structures and delivering the proper dose of radiation to the cancererous portions of the prostate can be created with this technique in less than ten minutes. The older preplan technique is still commonly used at many institutions. In the preplan technique, an ultrasound is obtained several days or weeks prior to the implant. A plan is created with these images and this plan is used in the OR. Under anesthesia, the prostate gland commonly increases in size. This makes the preplan inaccurate and causes the urethra and rectum to receive higher doses of radiation than those received with intraoperative planning. Dr Doggett's research shows that the use of intraoperative treatment planning results in a 33.7% dose reduction to the rectum and a 12.1% dose reduction to the urethra compared to the old preplan technique. Additionally, patients treated with the intraoperative planning system had a 11.4% increase in the radiation dose to the prostate compared to the old preplan system. This research was presented at the International Brachytherapy Conference in Barcelona Spain on May 15 2004. The text of the presentation follows.

 

Improved Dosimetry by the Use of a Unique Interactive Image Registered Intra-Operative Treatment Planning System for Permanent Seed Prostate Brachytherapy

Stephen Doggett MD Tustin California USA
E. Clif Burdette PhD Champaigne Illinois USA

Poster presentation, Joint Brachytherapy Meeting GEC/ESTRO-ABS-GLAC, Barcelona, Spain,13-15 May 2004

 

Introduction

Ultrasound guided permanent seed placement for prostate cancer is a rapidly growing and technically demanding procedure with a significant learning curve. Until recently, standard treatment planning technique has had few changes over the past 15 years. There have been several modifications of seed positioning to reduce the central urethral dose and to increase target volume coverage. The creation of a preplan occurs several days or weeks in advance of the implant. Commonly, ultrasound images with the patient in the dorsal lithotomy position are obtained in the clinic by a technologist or physician without anesthesia. A treatment plan is then created by a medical radiation physicist or dosimetrist. Average physicist time to create the plan is 2.5 hours. (Prete JJ, et. Al. A Survey of Physics and Dosimetry Practice of Permanent Prostate Brachytherapy in the United States. Int J Radiat Oncol Biol Phys 1998 Mar1;40(4):1001-5).

The implant is then performed with the anesthetized patient in the OR with the physician attempting to replicate the anatomic position achieved during the preplan ultrasound study. Multiple adjustments of the ultrasound probe positioning assembly are made by the operator in an attempt to duplicate the same patient/probe position obtained in the clinic. The implant is commonly performed with the patient's hips hyperflexed, which can be quite difficult to duplicate in the clinic setting, especially with aged patients. Additionally, anesthesia induced muscle relaxation and prostatic vascular congestion may induce significant variations in prostate position. Uncertainties in translation of the preplan into the actual implant may be responsible for areas of overdosage or underdosage.

Image collection and treatment plan computation in the OR with an anesthetized patient has significant benefits. Fewer procedures and personnel are needed which lessens the potential for error. Ultrasound images of the prostate can be obtained and the plan computed with the patient in the implant position by the same operator. Changes in prostate size and shape due to hormonal influence or anesthetic induced pelvic muscle relaxation are reduced. Advances in high speed computing combined with anatomically based automatically generated loading patterns permit the creation of a plan in seconds. Thus, rapid comparison of multiple plans generated in the OR becomes practical. The implant is performed after creation and acceptance of a plan without moving the patient, ensuring greater plan/implant fidelity.

The traditional preplanning technique cannot quantitatively match the treatment plan to the template and target volume anatomy. It is strongly operator dependent on manual estimated adjustments of the patient and probe position in an attempt to match the plan to the prostate in the OR. The Interplant system is the first interactive treatment planning system capable of optically encoding and spatially registering the prostate image and digitally transferring this information to the planning computer through the presence of sensors in the ultrasound probe holder. Such registration permits accurate quantitative alignment of the plan to the probe/template/prostate which greatly enhances operator confidence in the fidelity of the positioning of plan to target. Prior to the introduction of this technology, prostate implants have been hampered by the inability of the treatment planning system to register the treatment plan to the true 3D volume and to the probe/template assembly at any given time during the implant. Image registration is dependent on a continuous updating of spatial information to the treatment planning computer regarding the probe/template position. Introduction of image registration into the interactive treatment planning process permits the computer to register in three dimensions and in real time the positions of the seeds, prostate, probe and template. Dynamic dosimetry cannot be accomplished without spatial image registration. Other intraoperative implant technologies such as image fusion and real time virtual needle guidance are also dependent on spatial image registration data.

 

Intraoperative Treatment Planning System Description

The Interplant interactive treatment planning system (CMS, Inc., St. Louis, MO) was introduced into a community based permanent seed prostate brachytherapy practice. A unique optically-encoded probe positioning system permits precise localization of the ultrasound probe and images to the template and prostate and serves as a platform for planned technology advances, specifically dynamic dosimetry, i.e. seed location based dosimetry during the implant procedure. Interplant interactive treatment planning provides electronic importing of ultrasound images into the treatment planning computer without human interaction at the time of surgery. Prostate size and position are precisely documented and errors introduced by prostatic motion are mitigated. The treatment plan is generated in the operating room on a PC using automated tools. A dose optimization algorithm reduces operator interaction time. Interplant interactive treatment planning allows real time, pre implant, three dimensional superposition of translucent isodose surfaces over the prostatic anatomy for a simulated micro-dosimetric analysis of radiation exposure to any location in the prostate and surrounding structures. Alterations in plan for optimization can be efficiently carried out in the OR prior to implant.

 

Materials and Methods

A total of 54 patients consecutively implanted were evaluated. 35% had received pre-implant hormone blockade therapy. The lead author had performed more than 400 pre-planned implants by the time of the study and had created over 40 intraoperative treatment plans using this system at the initiation of the study. Dosimetric comparisons of office pre-plans and intraoperative plans registered to patient anatomy in the OR were performed.

Preimplant plans were generated in the office setting without anesthesia. Patient was placed supine on an exam table with his feet in stirrups. B&K 8558 ultrasound probe was inserted into the rectum and attached to the Interplant digital stepper. Transverse ultrasound images were obtained at 5mm increments and digitally entered into the treatment planning computer. Contours of the prostate and rectum were drawn digitally and entered into the planning computer. Treatment plan was generated and optimized electronically. The planning system includes volume-based tools for the auto generation of seedloading patterns based upon the actual anatomy structures. Several pattern generation schema are available for selection, or the user may create custom loading schemes. All use the anatomical information for loading boundary determination.

Within 21 days after the preplan was generated the implant was performed in the operating room under general or spinal anesthestic. Bladder was catheterized and patient was placed in the dorsal lithotomy position. B&K 8558 ultrasound probe with offset cover was inserted into the rectum and attached to the Interplant stepper. Image registration was calibrated by identification of the base plane as the vesico/prostatic junction by the operator and entering this as the Z axis. The template was attached and stabilizer needles were placed in the lateral prostate. Transverse ultrasound images were obtained at 5mm increments and digitally entered into the Interplant treatment planning computer. Contours of the prostate, urethra and rectum were drawn digitally and entered into the planning computer. Treatment plan was generated and optimized electronically. Generally, a peripheral loading pattern was used. Adjunct template guide holes were seldom used, with diagonal positions being preferentially loaded. More centrally located needles contained seeds loaded only at the base and apex. The Interplant system provides straightforward tools for the operator to set these parameters. Anatomical boundaries from the contoured structures are automatically integrated and applied to set the limits of the pattern. DVH-based dose optimization may be applied at the operator's discretion.

The ultrasound probe was set at the same base position as was determined at the start of the transverse image collection for treatment planning. 18 gauge afterloading needles were inserted in a single row, moving left to right. X-Y positioning was optimized with operator manipulation of the needles. Z-depth was determined by a combination of transverse ultrasound, AP fluoroscopy and operator tactile judgement. The Mick applicator was utilized to implant seeds according to the computer generated treatment plan.

Implant analysis was carried out for both preplan and intraoperative planning approaches. We recreated and measured electronically the use of a preplan in the operating room, which is a re-creation of the usual treatment planning process in widespread use. Intraoperative anatomic contours of the prostate and rectum were electronically superimposed over the preplan dosimetry cloud. Positioning of the intraoperative anatomic contours were electronically shifted to center the intraoperative gland contour within the pre-plan dosimetry cloud.

The prostate and rectum dose coverage obtained by overlaying the pre-plan dose clouds onto the pre-plan measured prostate and rectum contours was calculated. The prostate and rectum dose coverage obtained by overlaying the pre-planned dose clouds onto the intra-operatively measured prostate and rectum contours was calculated.

An analysis of the differences in prostate and rectum dose coverage due to the use of the pre-plan dose clouds with prostate and rectum contours obtained intra-operatively was prepared. A significant decrease in prostate volume dose coverage was seen with the using of the pre-plan measured contours during the implant procedure instead of using intra-operatively obtained contours. A significant increase in rectal volume dose coverage was seen with the use of the pre-plan measured contours during the implant procedure instead of using intra-operatively obtained contours. A significant degradation in prostate and rectal dosimetry can be attributed then to the use of only a preplan in the operating room.

 

Results

Significant changes in the preplan dosimetry are identified with respect to the traditional preplanning method. Dosimetry comparisons were made between plans based on anatomical images taken days prior to the implant procedure and plans based on images taken with patients in OR and generated intra-operatively using the Interplant system. The 54 implant patients showed an increase of 11.47% of CTV coverage over preplanning. Rectal and urethral D100 doses are decreased by 33.70 % and 12.13%, respectively with the use of intra-operative treatment planning. D80 rectal dose is reduced by 37.5% and the D150 urethral dose is reduced 28% compared to the traditional pre-plan. Advantages of interactive image registered intra-operative treatment planning are therefore quantified.

 

Conclusions

Interactive image registered intra-operative treatment planning is a significant technological advance in prostate permanent seed brachytherapy. Interactive treatment planning can help substantially improve delivered dose distribution and reduce variability due to prostatic anatomical differences between pre-operative and intra-operative anatomy. The traditional method of preplanning the implant in the clinic is shown to have significant inaccuracies with respect to intraoperative dose delivered. These are induced by operator reliance on manual estimated adjustments of the patient /probe position in an attempt to match the plan to the prostate and also to prostatic motion and size changes in the time between planning and surgery. The planning inaccuracies caused by operator attempts to manually match the patient/probe position to the plan are shown to create significant areas of overdosage or underdosage. These operator induced planning inaccuracies can be mitigated by the use of image registered interactive treatment planning. Rectal and urethral doses are lowered through the use of image registered interactive treatment planning. Intraoperative feedback to the physician during the procedure increases accuracy of seed placement and should produce a shorter operator learning curve, lower complications and higher local control rates. Spatial registration of prostate images is an essential platform for emerging technologies such as dynamic dosimetry, fusion of imaging modalities, and dynamic computer directed needle guidance.

Improved Dosimetry by the Use of a Unique Interactive Image Registered Intra-Operative Treatment Planning System for Permanent Seed Prostate Brachytherapy

Stephen Doggett MD Tustin California USA
E. Clif Burdette PhD Champaigne Illinois USA

Poster presentation, Joint Brachytherapy Meeting GEC/ESTRO-ABS-GLAC, Barcelona, Spain,13-15 May 2004