 |
| About Bioline | All Journals | Testimonials | Membership | News |
|
||||||
|
||||||
Journal of Cancer Research and Therapeutics, Vol. 6, No. 2, April-June, 2010, pp. 172-178 Original Article Daily online localization using implanted fiducial markers and its impact on planning target volume for carcinoma prostate Robin Khosa, Sapna Nangia1 , Kundan S. Chufal, D. Ghosh, Rakesh Kaul 2 , Lalit Sharma Batra Cancer Centre, Batra Hospital and Medical Research Centre, Correspondence Address: Dr. Robin Khosa, Batra Cancer Centre, Batra Hospital and Medical Research Centre, 1, Tughalakabad Instituional Area, Mehrauli Badarpur Road, New Delhi - 110 062, India. robindr_25k@yahoo.com Code Number: cr10039 PMID: 20622364 DOI: 10.4103/0973-1482.65244 Abstract Background : Aim of the study was to assess prostate motion on daily basis with respect to setup and to compare the shifts based on bony anatomy and gold fiducial markers.Materials and Methods : Gold fiducial markers were inserted in prostate under U/S guidance and daily portal images were taken and compared with digitally reconstructed images, both using bony landmarks and fiducial markers as reference. A dose of 2 MU was given for two orthogonal images daily. The mean and standard deviation of displacement using gold seeds and bone were calculated. Systematic and random errors were generated. The planning target volume (PTV) was calculated using the Van Herk formula. Results : A total of 180 portal images from 10 patients were studied. The mean displacement along x, y and z axes was 1.67 mm, 3.58 mm, and 1.76 mm using fiducial markers and 2.12 mm, 3.47 mm, and 2.09 mm using bony landmarks, respectively. The mean internal organ motion was 1.23 mm (+1.45), 3.11 mm (+2.69 mm); and 1.87 mm (+1.67 mm) along x, y and z axes, respectively. The PTV to account for prostate motion if daily matching was not done was 4.64 mm, 10.41 mm and 4.40 mm along lateral, superoinferior, and anteroposterior directions, respectively. If bony landmarks were used for daily matching, margins of 3.61 mm, 7.31 mm, and 4.72 mm in lateral, superoinferior, and anteroposterior directions should be added to the clinical target volume. Conclusion : Daily alignment using gold fiducial markers is an effective method of localizing prostate displacement. It provides the option of reducing margins, thus limiting normal tissue toxicity and allowing the possibility of dose escalation for better long-term control. Keywords: Daily online localization, prostate internal organ motion, prostate planning target volume Introduction In developed countries prostate cancer is among the most common male cancers. Treatment for local disease by external beam radiotherapy is an effective alternative to radical prostatectomy. Preventing normal tissue toxicity is as important as effective local control. Daily positional changes may however compromise both local control and normal tissue, and need to be considered. The planning target volume (PTV) that is added to the clinical target volume, includes both external setup errors and internal organ motions. [1] The prostate is located between the bladder, placed anterosuperiorly, and the rectum, placed posteriorly. Both these structures are associated with variable filling not only during the day but also over the entire treatment. Since the rectum tends to become progressively less distended during the course of radiation treatment (mean decrease in diameter of 1.5 cm), prostate tends to move more posteriorly and inferiorly in the latter part of the treatment. Crook et al, reported more than 1 cm posterior shift in 30% of patients and similar shift inferiorly in about 11%. [2] Similarly, cystitis is seen in a significant number of patients, especially in the latter part of treatment. This results in increased urgency and decreased bladder urine retention. Thus, an adequate safety margin needs to be added to overcome the risk of daily variations. With newer techniques like 3DCRT and IMRT, highly conformal plans can be generated, providing an option for decreasing normal tissue toxicities. [3] However, caution needs to be exercised while limiting margins, so as to avoid geographical misses. Newer linear accelerators have facilities for portal imaging, allowing us the option of online correction of daily errors. Using bony landmarks for correction can only correct for setup errors. Direct localization of soft tissues like prostate is not possible using X-ray-based imaging. Internal fiducial markers act as a surrogate for prostate location. [4] Daily electronic portal image visualization and alignment with respect to planned fields help limit interfractional prostate movements, thus limiting the PTVs and normal tissue doses. [5] We conducted a study to assess daily shift in prostate location using implanted gold fiducial markers and compare shifts based on bony anatomy and gold fiducial markers. Materials and Methods This is a prospectively designed protocol for the department of radiotherapy at Batra hospital and Medical Research Centre, New Delhi. An Institutional ethics committee approval was taken prior to adopting this policy at our centre. Informed consent was taken from each and every patient before enrolling the patient in the protocol. Fiducial marker insertion Fiducial markers (Best Medical International Inc., Springfield, VA, USA) made of inert material gold, of thickness 1.2 mm and length 3 mm were used. Patients with localized prostate cancer, offered IMRT, were selected. Three sterile preloaded gold fiducial markers were inserted in prostate by under ultrasound guidance under antibiotic cover. The three seeds were inserted in such a pattern that they form an equilateral triangle, with one seed at the apex and two at the base in each of the lateral lobes. Seed position and arrangement were confirmed by a check X-ray. Treatment planning A gap of 5-6 days was given for prostatic inflammation to settle down and for the seeds to attain a stable position. A thermoplastic cast for the pelvis and head, to increase positioning accuracy and daily reproducibility, was made. No specific instructions for bladder filling and rectal emptying were given to the patient. A helical contrast-enhanced CT scan was performed on a flat couch and 2.5 mm axial cuts were generated, and transferred through the dicom system. Contouring of the scans was done on Somavision (Varian Medical Systems Inc., Palo Alto, CA, USA) by a radiation oncologist. The prostate and seminal vesicle constitute the clinical target volume. The PTV of 13 mm all around and 8 mm posteriorly was taken around the clinical target volume to account for setup errors and internal organ motion. Organs at risk including rectum, anterior rectal wall, bladder, femur heads, bone marrow, and intestine were marked. Prescription dose of 70 Gy in 28 fractions was advised and dose constraints to the organs at risk were given according to their tolerance limits. IMRT planning was done for each plan on an Eclipse planning system. Digitally reconstructed radiographs (DRR) were generated for daily comparison. Daily verification All patients were immobilized and treated in the supine position. Prior to each setup, patients were positioned using laser alignments to marks on their thermoplastic casts which acted as a surrogate for skin marks. AP and lateral films were taken using an electronic portal imager mounted on a dual energy Clinac 2100C (Varian Medical Systems). For each set of films, 2 MU were used. Portal images were compared with DRR's generated during planning, the latter being used as a reference image. Right and left shifts were represented by x coordinate with plus sign representing right shift and minus sign representing left shift. For the y axis, plus and minus signs denoted outward (away from gantry) and inward (toward gantry) patient movement, respectively. Lateral film was used to assess the z shift with plus and minus signs representing upward (anterior) and downward (posterior) shift. Additional marks were placed on the patient's body for positioning the pelvic cast. Online shifts along x, y, and z coordinates for both bony landmarks and gold fiducial markers (representing the prostate position) were noted by superimposing daily portal image over the DRR. To avoid any inter observer variations comparisons were done by a single radiation oncologist. Statistics Mean, median, standard deviation, minimum, and maximum range of variations along antero-posterior, supero-lateral, and in lateral dimensions for both bone and implanted gold seeds were calculated using SPSS version 16. A difference in the mean of seed and bone movements relative to the isocenter was done for internal prostate motion. For individual patient average displacement and standard deviation in each direction was calculated. Standard deviation of the average displacement for the whole population was calculated as systematic error. Root mean square of the individual standard deviation was considered as random error. [6] The PTV was generated using the Van Herk formula (2.5 ∑+0.7s) where ∑ represents the systematic error and s represents random error. [7] Results A total of 10 patients with prostate cancer attending Batra Hospital and Medical Research Centre were studied. Two hundred and eighty-five fractions were delivered and one hundred and eighty portal images generated. All shifts with respect to gold fiducial markers inserted inside the prostate gland and bony anatomy were checked online on a 4D treatment console of Varian Medical Systems. The mean and median shifts of gold fiducial markers and bony landmarks are shown in [Table - 1]. [Figure - 1], [Figure - 2] and [Figure - 3] show the mean shifts of both gold fiducial markers and bony landmarks of each patient in various directions. The positive direction in the [Figure - 1] shows right-sided displacement and negative values indicate left-sided displacement. Positive bars in the [Figure - 2] shows inferior displacement and negative values indicate superior displacement. The positive direction in the [Figure - 3] shows upward displacement and negative values indicate posterior displacement. The range of seed movement on the x axis ranged from +7 mm to -9 mm, for both fiducial markers and bony anatomy. For the y axis, the range was +13 to -15 mm and +11 mm to -16 mm for fiducial markers and bony anatomy; and for the z axis it was +8 mm to -8mm and +9 mm to -7 mm, respectively. Prostate displacement vis a vis bony displacement (calculated by subtracting bony shift from seed shift) is shown in [Table - 2]. [Figure - 4], [Figure - 5] and [Figure - 6] show relative displacement on all three axes for 10 patients. Seed movements which were representative of prostate motion was compared with bony anatomy by subtracting the two motions on all three axes are shown in [Table - 3]. Out of 10 patients, 6 received radical dose to prostate only and 4 received an additional 50.4 Gy to pelvic nodes, based on the risk of lymph node involvement as calculated by the Roach formula. One cm PTV was taken around the nodal clinical target volume. [Table - 4] shows systematic and random errors for all patients in lateral, superoinferior, and anteroposterior directions. All shifts more than 2 mm were corrected. Using fiducial markers, the PTV according to the Van Herk formula was 4.64 mm, 10.41 mm, and 4.4 mm in lateral, superoinferior, and anteroposterior directions, respectively. The same using the bony landmarks was 6.17 mm, 8.98 mm, and 5.43 mm in lateral, superoinferior, and anteroposterior directions, respectively. The PTV for internal organ motion was 3.61 mm, 7.31 mm, and 4.72 mm along x, y and z axes, respectively. Discussion IMRT is a state of art form of radiation treatment, delivering doses precisely to the areas of interest. The beauty of the technique is its ability to even achieve concave shapes and rapid dose fall off at the margins, not possible with conventional or 3D conformal radiotherapy. Rapid dose fall off can lead to geographical misses if the plan implementation is associated with improper patient setup and daily internal organ movements. External setup errors can be eliminated by use of better immobilization devices. Internal organ motion, including variable bladder filling and rectal emptying, can lead to daily changes in prostate localization starting from the treatment planning itself. [8] Systematic shift will occur if during initial CT scan, prostate position is different to the one present during daily treatment delivery. [9] Daily external setup errors occur because of movement of skin markers relative to the bony landmarks. To overcome this, daily online verification and correction using bony landmarks can be helpful. The mean and median movements in x, y and z axes were 2.12 mm (±2.08), 3.47 mm (±3.00), and 2.09 mm (±1.93); and 1.5 mm (±2.08), 3 mm (±3.00), and 2 mm (±1.93), respectively, highlighting the importance of online correction to limit the errors. Weight loss over the treatment, respiration, and relaxation of the patient can lead to skin movement. Movement of the patient within the orfit cast used for immobilization can add to this error. Inter observer variations while calculating the shifts can add to errors. In our study this was overcome with only one radiation oncologist doing daily checks. The prostate is not visible using a conventional megavoltage electronic portal imager. Image-guided radiotherapy of prostate can be achieved by either using internal fiducial markers or direct visualization of prostate using kilovoltage CT (kvCT) scanners. [9] Average and median prostate movements in our study was 1.67 mm (±1.92), 3.58 mm (±2.94) and 2.76 mm (±1.99); and 1 mm(±1.94), 3 mm (±2.97) and 1.5 mm (±2.0) along x, y and z axes. Comparison of our data with various studies is shown in [Table - 5]. In a retrospective analysis by Hanley et al, [14] 50 patients 1239 port films done using two oblique and one lateral film were studied. Set errors were assessed by comparing bony landmarks as a reference. The mean and standard deviation of systematic setup error in AP, lateral, and SI directions were 0.3±1.3 mm, -0.1±1.9 mm, and 0.4±1.4 mm, respectively. Random errors were generally evenly distributed in all directions. Average shift in AP, lateral, and SI directions was 1.9 mm, 2.0 mm, and 1.7 mm, respectively. Nederveen et al, [12] in an analysis of 23 patients reported displacements both using bony landmarks and prostate displacement. The mean systematic displacement using bony landmarks was 0+ 2.1 mm in the lateral direction, -1±4.4 mm in anteroposterior and 0.1±2.1 mm in superoinferior directions. The mean prostate displacement was 0+ 1 mm in the lateral direction, 0+ 2.3 mm in anteroposterior and 1±4.1 mm in superoinferior directions. A total of 2549 portal images from 33 patients were analyzed by Alasti et al. [15] Bony displacement characterized by standard deviations was 1.4 mm and 1.8 mm in anteroposterior and superoinferior directions in 23 patients. Ten patients were analyzed for prostate motion with displacement of 5.8 mm and 3.3 mm in anteroposterior and superoinferior directions. While a plan is made fulfilling all dosimetric criteria, routine implementation is done using bony landmarks as a reference. This is followed by routine reviews following the same parameters. However , this is based on the assumption that there is no internal organ motion. Daily changes in bladder and rectal volumes result in prostate displacements which are independent of pelvic bony anatomy. [2] In addition, prostate tissue is not attached to the pelvis. The difference in seed movement and setup error calculated by matching bony landmarks is indicative of internal organ motion. Little et al[16] studied 35 patients who received IMRT for prostate cancer. Weekly orthogonal portal films and daily B-mode acquisition and targeting (BAT) ultrasound localization were done. A total of 243 pairs of orthogonal portal images were taken. The mean prostate organ motion was -0.89±3.3 mm (R-L), -1.3±5.7 mm (AP), and -1.6±6.4 mm (SI). In our study mean internal organ motion was assessed in all directions, separately. The mean and standard deviation in our study were 1.23 mm (±1.45 mm) in lateral directions, 3.11 mm (±2.69 mm) in superoinferior directions; and 1.87 mm (±1.67 mm) in anteroposterior directions. Thus, most of the daily prostate movements were in SI and AP directions, with a maximum shift occurring in superior and posterior directions. These results were corresponding to ones seen in studies by Rudat et al, [10] and Altholf et al, [11] showing the maximum displacement in superior and inferior directions, while Osei et al, [6] Nederveen et al, [12] and Alasti et al, [15] showed maximum movements in anterior and posterior directions [Table - 6]. Our study is in concordance with all studies in determining the fact that lateral movements in prostate motion are the least. As radiation progresses, reactions develop resulting in increased symptoms related to bladder control along with frequent stools, leading to daily variations which can occur in an unexpected fashion. Also even prior to the appearance of radiation reactions, daily emptying habits of the people and different radiation timings can lead to variation in the rectal and bladder volumes. In our study, the last mentioned source of error was minimized by treating them at same time every day. Online kvCT-based localization forms the basis of image-guided radiotherapy. It allows for direct localization of the soft tissues including prostate, bladder, and rectum. However, this modality is much more costly than use of fiducial markers, and also requires much more expertise and time. In a study by Moseley et al, [17] mean shift differences between MV markers and soft-tissue matching were −0.8 mm (SD 1.0), −0.2 mm (SD 3.0), −0.8 mm (SD 2.5) in the LR, AP, and SI directions, respectively. Less than 1 mm difference in the shift seems less cost-effective to change from normal megavoltage X-ray-based portal images to costly kilovoltage-based CT localization. PTVs include margins for daily setup errors and internal organ motion. As calculated in our study using fiducial markers as reference, the PTV was 4.64 mm, 10.41 mm, and 4.4 mm in lateral, superoinferior, and anteroposterior directions, respectively, if no correction is to be made. The PTV for relative prostate displacement with respect to bony anatomy, representative of internal organ motion was 3.61 mm, 7.31 mm, and 4.72 mm in lateral, superoinferior, and anteroposterior directions. In routine practice where bony landmarks are used for daily verification, the above margins around the clinical target volume need to be added, thus limiting our ability to reduce margins. This in turn results in large volumes of normal tissues being irradiated. Any displacement with respect to seed movement of more than 2 mm was corrected in our study. Daily online verification and correction using prostate fiducial markers can help localize internal organ motion. The seed localization allowed us to reduce our PTVs by 3 mm all around and 2 mm posteriorly. We now limit our PTV to 10 mm all around and 6 mm posteriorly to account for intrafraction motion. This can even further allow us the option of limiting our PTVs around the prostate even further and limit normal tissue toxicities. Conclusions Using gold fiducial markers for prostate localization is a useful tool for localizing prostate gland motion where facilities for kvCT are not available. Daily fiducial marker-based matching of DRR's and portal images can help reduce the uncertainties associated with prostate motion which are mainly in SI and AP directions. The PTV calculated using systematic and random errors was large, but daily verification using implanted fiducial markers can help limit these margins. Thus, larger PTV would however have to be added if bony landmarks were used for daily verification as the prostate moves independent of pelvic bony structures, but our very aim to limit radiation dose to organs at risk will fail. Higher doses for treatment have proven to be associated with better tumor control, but the same can only be safely achieved if we are able to deliver the dose to the organ it is intended for. Acknowledgement I sincerely thank all my colleagues at Batra Hospital and Medical Research Centre, Dr. Anshul Bhatnagar, Dr. Anchal Aggarwal, Dr. Manish Chomal, Maninder Mishra, C.P. Bhatt and Sandeep Rathore for their help and opinions for smooth conduct of my study. References
Copyright 2010 - Journal of Cancer Research and Therapeutics The following images related to this document are available:Photo images[cr10039f4.jpg] [cr10039f2.jpg] [cr10039t4.jpg] [cr10039t3.jpg] [cr10039t6.jpg] [cr10039f6.jpg] [cr10039f5.jpg] [cr10039f1.jpg] [cr10039t1.jpg] [cr10039f3.jpg] [cr10039t5.jpg] [cr10039t2.jpg] |
| |||||||||
![[Table - 1]](/showimage?cr/photo/cr10039t1.jpg){kind=link}
![[Figure - 1]](/showimage?cr/photo/cr10039f1.jpg){kind=link}
![[Figure - 2]](/showimage?cr/photo/cr10039f2.jpg){kind=link}
![[Figure - 3]](/showimage?cr/photo/cr10039f3.jpg){kind=link}
![[Table - 2]](/showimage?cr/photo/cr10039t2.jpg){kind=link}
![[Figure - 4]](/showimage?cr/photo/cr10039f4.jpg){kind=link}
![[Figure - 5]](/showimage?cr/photo/cr10039f5.jpg){kind=link}
![[Figure - 6]](/showimage?cr/photo/cr10039f6.jpg){kind=link}
![[Table - 3]](/showimage?cr/photo/cr10039t3.jpg){kind=link}
![[Table - 4]](/showimage?cr/photo/cr10039t4.jpg){kind=link}
![[Table - 5]](/showimage?cr/photo/cr10039t5.jpg){kind=link}
![[Table - 6]](/showimage?cr/photo/cr10039t6.jpg){kind=link}