2725
2
2723 Changes in the Distance Between an Internal Fiducial Marker and a Motion Tumor on Fluoroscopic Real-Time Tumor-Tracking System Evaluated With an In-Room CT System K. Taguchi, K. Ebe, A. Hiyama, K. Tsukamoto, H. Sasai, Y. Nakashima, S. Yamatogi, R. Kanzaki, T. Matsumoto, N. Matsunaga Department of Radiology, Ube, Japan Purpose/Objective(s): Fluoroscopic real-time tumor-tracking radiotherapy (RTRT) system with the aid of an internal fiducial marker for tracking has been shown to be useful for the precise irradiation for tumors in motion. However, there has been fear for the migration of the marker and change in the distance between the marker and target volume. We examined spatial relationship between a marker and a tumor withan in-room CT system before every RTRT. Reliability of an embedded marker for tracking is reported. Materials/Methods: Twenty four patients with pulmonary cancer (30 treatment sites, 77 markers) underwent CT examinations (n182). Treatment periods ranged from 3 to 22 days. Four patients with hepatic carcinoma (5 treatment sites, 10 markers) underwent CT examinations (n40). Treatment periods ranged from 12 to 18 days. Gold markers of 1.5 mm and 2.0 mm in diameter as an internal fiducial marker were embedded in adjacent area to pulmonary tumors and to hepatic tumors, respectively. An in-room CT scans were performed before every RTRT. CT helical imaging parameters were as follows: pitch, 1.5; FOV, 400 mm; slice thickness, 3 mm. CT data sets were transferred to TPS (Pinnacle-3, Hitachi Medico, Tokyo). The three-dimensional (3D) distance between each center of a marker and a tumor was measured after delineating each contour (n471 in the lung, n80 in the liver). Results: Mean discrepancy plus minus standard deviation(SD) was -0.40 plus minus 1.98 mm in the lung (n471), 0.35 plus minus 1.70 mm in the liver (n80). Conclusions: We evaluated changes in the distance between an internal fiducial marker and a motion tumor quantitatively with an in-room CT system. A value of SD was around 2 mm. An embedded marker either in the lung or the liver seemed to be reliable. Author Disclosure: K. Taguchi, None; K. Ebe, None; A. Hiyama, None; K. Tsukamoto, None; H. Sasai, None; Y. Nakashima, None; S. Yamatogi, None; R. Kanzaki, None; T. Matsumoto, None; N. Matsunaga, None. 2724 Verification of Average Tumor Position for Phase-Gated Lung Cancer Patients Using Integrated EPIDs J. Cuijpers, J. van So ¨rensen de Koste, B. Slotman, S. Senan VU University Medical Centre, Amsterdam, The Netherlands Purpose/Objective(s): The toxicity of concurrent chemo-radiotherapy for stage III lung cancer can be reduced if smaller margins can be used to account for respiration-induced mobility. One such example is the use of respiration-gated treatment, but the use of smaller margins implies that accurate patient setup becomes of major importance. Current off-line setup correction protocols like the SAL or NAL protocols use bony anatomy matching to reduce systematic setup errors. The tumor motion relative to the bony anatomy during gated treatment delivery can be derived from a 4D CT-scan. However, variations in breathing amplitude and residual tidal volume during gated delivery may cause a significant baseline shift in tumor position. We explored the use of integrated EPIDs for imaging of the tumor and carina-complex during treatment in order to assess systematic displacements. Materials/Methods: An Integrated EPID (IE) consists of the sum of all EPID frames that are acquired during the entire duration of a treatment beam. The technique of generating an IE using phase gating was tested using a moving phantom. An IE of one of the treatment fields was acquired during every treatment fraction in 7 patients with stage III lung cancer who were treated using audio coached phase-gating at the expiratory phase of the breathing cycle. A match of the bony anatomy between IE and DRR’s derived from the selected phases of a 4DCT used for gating was performed to assess the setup errors in the patient positioning. The impact of using a two-stage NAL protocol, with EPID imaging at the start and halfway during treatment, was investigated. In a recent analysis of 4DCT datasets (J. van So ¨ rnsen de Koste, unpublished data), we found a correlation between the carina position in superior-inferior axis and the position of the tumor. A method of adapting field borders using 4DCT information and the average position of the carina will be presented and applied to the clinical data sets. Results: The moving phantom test showed that the average position of an object, with a residual motion typical for an expiration phase gated treatment, can adequately be assessed within 1 mm accuracy using an IE. The standard deviation for systematic setup errors measured using bony anatomy was 3.9 and 4.1 mm in the lateral and longitudinal directions, respectively. This is larger than the respective standard deviations for random setup errors of 2.7 mm (lateral) and 3.3 mm (longitudinal). A significant time trend was observed in 4 patients, but use of a two-stage NAL protocol could compensate for the time trends and could be reduced to 1.9 mm and 1.3 mm, respectively. The standard deviation for systematic errors in average carina position at expiration phase was on average 0.2 mm (sd 1.2 mm), and the corresponding value for random errors was approximately 1.6 mm. Conclusions: Imaging of bony anatomy and carina complex using Integrated EPIDs provides an efficient manner for assessing both patient setup and average tumor position. Author Disclosure: J. Cuijpers, None; J. van So ¨rensen de Koste, None; B. Slotman, None; S. Senan, None. 2725 On Accounting for Patient-Specific Tumor Motion in Target Definition for Lung Cancer Treatment Planning: Comparison of a Multi-Phase CT Simulation Approach and MRI CINE Study L. Wang, S. Feigenberg, L. Chen, K. Paskalev, L. Jin, C. C. M. Ma Fox Chase Cancer Center, Philadelphia, PA Purpose/Objective(s): In order to determine patient-specific tumor motion, a multi-phase CT scanning and MRI cine imaging approaches were used. In this work, we compared tumor motion measured by the two approaches as well as the tumor coverage S612 I. J. Radiation Oncology ● Biology ● Physics Volume 66, Number 3, Supplement, 2006
Transcript of 2725
2723 Changes in the Distance Between an Internal Fiducial Marker and a Motion Tumor on FluoroscopicReal-Time Tumor-Tracking System Evaluated With an In-Room CT System
K. Taguchi, K. Ebe, A. Hiyama, K. Tsukamoto, H. Sasai, Y. Nakashima, S. Yamatogi, R. Kanzaki, T. Matsumoto,N. Matsunaga
Department of Radiology, Ube, Japan
Purpose/Objective(s): Fluoroscopic real-time tumor-tracking radiotherapy (RTRT) system with the aid of an internal fiducialmarker for tracking has been shown to be useful for the precise irradiation for tumors in motion. However, there has been fearfor the migration of the marker and change in the distance between the marker and target volume. We examined spatialrelationship between a marker and a tumor withan in-room CT system before every RTRT. Reliability of an embedded markerfor tracking is reported.
Materials/Methods: Twenty four patients with pulmonary cancer (30 treatment sites, 77 markers) underwent CT examinations(n�182). Treatment periods ranged from 3 to 22 days. Four patients with hepatic carcinoma (5 treatment sites, 10 markers)underwent CT examinations (n�40). Treatment periods ranged from 12 to 18 days. Gold markers of 1.5 mm and 2.0 mm indiameter as an internal fiducial marker were embedded in adjacent area to pulmonary tumors and to hepatic tumors,respectively. An in-room CT scans were performed before every RTRT. CT helical imaging parameters were as follows: pitch,1.5; FOV, 400 mm; slice thickness, 3 mm. CT data sets were transferred to TPS (Pinnacle-3, Hitachi Medico, Tokyo). Thethree-dimensional (3D) distance between each center of a marker and a tumor was measured after delineating each contour(n�471 in the lung, n�80 in the liver).
Results: Mean discrepancy plus minus standard deviation(SD) was -0.40 plus minus 1.98 mm in the lung (n�471), 0.35 plusminus 1.70 mm in the liver (n�80).
Conclusions: We evaluated changes in the distance between an internal fiducial marker and a motion tumor quantitatively withan in-room CT system. A value of SD was around 2 mm. An embedded marker either in the lung or the liver seemed to bereliable.
Author Disclosure: K. Taguchi, None; K. Ebe, None; A. Hiyama, None; K. Tsukamoto, None; H. Sasai, None; Y. Nakashima,None; S. Yamatogi, None; R. Kanzaki, None; T. Matsumoto, None; N. Matsunaga, None.
2724 Verification of Average Tumor Position for Phase-Gated Lung Cancer Patients Using Integrated EPIDs
J. Cuijpers, J. van Sorensen de Koste, B. Slotman, S. Senan
VU University Medical Centre, Amsterdam, The Netherlands
Purpose/Objective(s): The toxicity of concurrent chemo-radiotherapy for stage III lung cancer can be reduced if smallermargins can be used to account for respiration-induced mobility. One such example is the use of respiration-gated treatment,but the use of smaller margins implies that accurate patient setup becomes of major importance. Current off-line setupcorrection protocols like the SAL or NAL protocols use bony anatomy matching to reduce systematic setup errors. The tumormotion relative to the bony anatomy during gated treatment delivery can be derived from a 4D CT-scan. However, variationsin breathing amplitude and residual tidal volume during gated delivery may cause a significant baseline shift in tumor position.We explored the use of integrated EPIDs for imaging of the tumor and carina-complex during treatment in order to assesssystematic displacements.
Materials/Methods: An Integrated EPID (IE) consists of the sum of all EPID frames that are acquired during the entire durationof a treatment beam. The technique of generating an IE using phase gating was tested using a moving phantom. An IE of oneof the treatment fields was acquired during every treatment fraction in 7 patients with stage III lung cancer who were treatedusing audio coached phase-gating at the expiratory phase of the breathing cycle. A match of the bony anatomy between IE andDRR’s derived from the selected phases of a 4DCT used for gating was performed to assess the setup errors in the patientpositioning. The impact of using a two-stage NAL protocol, with EPID imaging at the start and halfway during treatment, wasinvestigated. In a recent analysis of 4DCT datasets (J. van Sornsen de Koste, unpublished data), we found a correlation betweenthe carina position in superior-inferior axis and the position of the tumor. A method of adapting field borders using 4DCTinformation and the average position of the carina will be presented and applied to the clinical data sets.
Results: The moving phantom test showed that the average position of an object, with a residual motion typical for anexpiration phase gated treatment, can adequately be assessed within 1 mm accuracy using an IE. The standard deviation � forsystematic setup errors measured using bony anatomy was 3.9 and 4.1 mm in the lateral and longitudinal directions,respectively. This is larger than the respective standard deviations for random setup errors of 2.7 mm (lateral) and 3.3 mm(longitudinal). A significant time trend was observed in 4 patients, but use of a two-stage NAL protocol could compensate forthe time trends and � could be reduced to 1.9 mm and 1.3 mm, respectively. The standard deviation for systematic errors inaverage carina position at expiration phase was on average 0.2 mm (sd 1.2 mm), and the corresponding value for random errorswas approximately 1.6 mm.
Conclusions: Imaging of bony anatomy and carina complex using Integrated EPIDs provides an efficient manner for assessingboth patient setup and average tumor position.
Author Disclosure: J. Cuijpers, None; J. van Sorensen de Koste, None; B. Slotman, None; S. Senan, None.
2725 On Accounting for Patient-Specific Tumor Motion in Target Definition for Lung Cancer TreatmentPlanning: Comparison of a Multi-Phase CT Simulation Approach and MRI CINE Study
L. Wang, S. Feigenberg, L. Chen, K. Paskalev, L. Jin, C. C. M. Ma
Fox Chase Cancer Center, Philadelphia, PA
Purpose/Objective(s): In order to determine patient-specific tumor motion, a multi-phase CT scanning and MRI cine imagingapproaches were used. In this work, we compared tumor motion measured by the two approaches as well as the tumor coverage
S612 I. J. Radiation Oncology ● Biology ● Physics Volume 66, Number 3, Supplement, 2006
in the subsequent treatment delivery when the two sets of motion values are used for generating planning target volumes (PTVs)in treatment planning.
Materials/Methods: Consented patients underwent CT simulation consisting 2 short scans taken at maximum inspiration andexpiration breath holding conditions, in addition to a normal free-breathing scan. Consequently, these patients went throughMRI cine scanning. Targets were delineated on all 3 sets of CT scans and a composited gross tumor volume (GTV) wasconstructed after image registration which was based on the alignment of a stereotactic body localizer used for patientimmobilization during simulation. The maximum displacements from the respective geometrical center of the 3 targets weremeasured and considered as the target motion determined by multi-phase CT method. Such motion was compared with MRIcine study by measuring the variation of the tumor center along the three major directions (anterior-posterior (AP), right-leftlateral (Lat), superior-inferior (S-I)). For treatment planning and delivery, we constructed a PTV from the composited GTV byapplying 3 mm margin (termed PTV_3CT). For this study, another PTV, which was generated from the GTV outlined from thenormal-breathing scan by margins that were determined by MRI cine study of the target motion (termed as PTV_MRI), wasalso constructed and planned. Since a localization CT scan was obtained for each patient using a CT-on-rails system prior toeach treatment, we were able to obtain the actual GTV in treatment position (GTV_tx) and compared their coverage for theplans designed for either the PTV_3CT or PTV_MRI.
Results: For 10 patients, the motions determined by 3 CT scans and MRI cine measurements are very comparable, as shownby large p-values (0.162 in the AP, 0.733 in Lat, and 0.073 in the S-I directions). On the other hand, the GTV_tx (contouredfrom pre-treatment CT) for 5 patients and 20 fractions (20 sets of CT), were found all within the PTV_MRI geometrically andhave full dose coverage in terms of D95. However, for plans designed for PTV_3CT, 9 out of 20 of the GTV_tx were found,at a varying degree, to be partially out of the PTV_3CT and the coverage varies from 95.9% to 100% of the planned values.Although the difference in coverage is small, it is statistically significant, as the p-value for the t-test is 0.006.
Conclusions: The MRI-cine is a better motion study tool and the PTV design based on these motion data will provide bettercoverage for a moving target.
Author Disclosure: L. Wang, None; S. Feigenberg, None; L. Chen, None; K. Paskalev, None; L. Jin, None; C.C.M. Ma, None.
2726 On-Board Breath-Hold Digital Tomosynthesis (DTS) for Target Localization Prior to Breath-HoldTreatment
D. J. Godfrey, Z. Wang, S. Yoo, J. Wu, M. Oldham, C. Willett, F. Yin
Duke University Medical Center, Durham, NC
Purpose/Objective(s): Breath-hold treatment is useful for reducing target motion, thereby eliminating the need for expandedinternal target volumes (ITV), when treating sites prone to respiratory motion. Due to its long scan duration (�1 min.continuous scan), cone-beam CT (CBCT) localization of respiration-affected anatomy is cumbersome, requiring complicatedgating or segmented breath-hold acquisition schemes if an expanded ITV is to be avoided. DTS is a fast, low-dose alternativeto CBCT for 3-D image-guided radiation therapy (IGRT). This study examines on-board breath-hold digital tomosynthesis(DTS) as a simple alternative to breath-hold CBCT for image-guided target localization of abdominal and thoracic sites priorto breath-hold treatment.
Materials/Methods: On-board DTS images of 15 abdominal and thoracic subjects were reconstructed from limited-angle(�45o) subsets of kV CBCT projection data, acquired on a Varian 21EX Clinac equipped with an on-board imager (OBI), eitherduring a breath-hold or while subjects were freely breathing. Matching reference DTS (RDTS) images were reconstructed fromplanning CT data. Soft-tissue visibility and the size and location of reconstructed target volumes were compared amongstbreath-hold DTS, free-breathing DTS, and breath-hold and free-breathing CBCT, to assess the potential of breath-hold DTS fortarget localization prior to breath-hold treatment.
Results: Substantial volume savings were achieved using breath-hold localization and treatment, with typical PTV reductionsof more than 100 cm3, or 50% for moderately-sized CTVs. Free-breathing DTS and CBCT reconstructions often exhibitedartificially enlarged and displaced target volumes and poor visibility of soft-tissue anatomy (Fig. 1), and were thereforeinadequate for breath-hold treatment localization. Breath-hold CBCT yielded excellent soft-tissue visibility and target local-ization, but the scan must be parsed into multiple breath-hold acquisition segments, and usually required more than 5 min. tocomplete. Breath-hold DTS provided high-quality image information (Fig. 1), similar to breath-hold CBCT. Furthermore, DTSacquisition was completed in less than 10 sec. in every case, making rapid breath-hold DTS a much simpler method foracquiring motion-free 3-D localization images than breath-hold CBCT.
Conclusions: Rapid (� 10 sec.) breath-hold DTS is far more convenient than breath-hold CBCT (� 5 min.), yet still enablesdaily 3-D localization of soft-tissue targets for breath-hold treatment. The improvement in workflow provided by DTS willlikely allow for increased use of volume-sparing (�100 cm3 typical) breath-hold treatments.
S613Proceedings of the 48th Annual ASTRO Meeting
