S524 I. J. Radiation Oncology d Biology d Physics Volume 72, Number 1, Supplement, 2008

FB, mDIBH, and DEBH. In the Pinnacle3 treatment planning system, the forward IMRT planning was completed on the first set ofmDIBH CT images from the primary CT simulation and the planning was copied by the special system order to the second set of CTimages from the primary CT simulation and first set of CT images from the repeat CT simulation with the primary of angle, di-rection, size and shape of the MLC field and prescribed dose, the volume covered by selected high dose area in the selected segmentwere compared.

Results: In the planning based on the first set of mDIBH CT images from the primary CT simulation, the volume irradiated byequal and more than 103% of prescribed dose in the segment was 1.155 ± 0.392 cm3, and the volume were 3.877 ± 1.066 cm3

and 51.659 ± 8.676 cm3 in the plannings copied from the first set of mDIBH CT images from the primary CT simulation respec-tively to the second set of CT images from the primary CT simulation and first set of CT images from the repeat CT simulation, thedifference of the volume in the two plannings based the two set mDIBH CT image from the primary CT simulation was not sta-tistically significant (t = -1.672,p = 0.103)and the difference of the volume respectively in the first sets of mDIBH CT images re-spectively from the primary and repeat CT simulations (t = -5.728, p \ 0.01).

Conclusions: If the same threshold of mDIBH is remained during IMRT after breast-conservative surgery, the influence of theintrafraction target displacement on the dose distribution is not significant; if set-up error is not corrected, the interfraction changeof position of the segment given to cove the high dose area in the IMRT planning will be significant to result in significant change ofdose distribution in the breast.

Author Disclosure: J. Li, None; Z. Ma, None; J. Wang, None; J. Lu, None; T. Fan, None; S. Guo, None; Y. Wang, None.

2794 Reduction in Cardiac and Lung Doses with Deep Inspiration Breath-hold for Left-sided Breast Cancers

Treated with Accelerated Partial Breast Radiotherapy

P. J. Kim, H. Gee, S. Anooshfar, L. X. Yang, M. C. Rounsaville

California Pacific Medical Center, San Francisco, CA

Purpose/Objective(s): Reducing cardiopulmonary toxicity from breast irradiation continues to be an active area of study. Strat-egies toward this endeavor have included the reduction of irradiated volume and treatment during deep inspiration breath-hold.This study investigates whether deep inspiration breath-hold (BH) can reduce cardiopulmonary irradiation compared with freebreathing (FB) in accelerated partial breast irradiation (APBI).

Materials/Methods: Between December 2005 through February 2008, seven patients with left-sided breast cancers were treatedwith APBI during deep inspiratory breath-hold. For each patient, both a free breathing scan and a deep inspiratory breath-hold scanwere obtained with a GE Lightspeed CT simulator. The Varian Real-time Position Management system was used to obtain breath-hold scans. Typically, four non-coplanar 6MV photon beams were utilized, except for one patient who was treated with a mixed-beam plan of 6MV photons and 9 MeV electrons. All cardiac and left lungs were contoured by a single physician according toNSABP B-39 guidelines for consistency. Retrospectively, each treatment plan was overlaid on the corresponding free-breathingscan obtained at simulation for each patient. Cardiac and left lung doses were compared and analyzed.

Results: Patient characteristics include: age (range, 31-75-years-old), histology (DCIS (n = 2), infiltrating ductal carcinoma (n = 4),invasive mucinous carcinoma (n = 1)), tumor size (0.6-2.9 cm), and location of tumor (superior lateral (n = 5), superior central (n =1), superior medial (n = 1)). On average, mean and maximum cardiac doses with BH were reduced compared with FB 43.1% (50.6cGy vs. 95.2 cGy, p = 0.038) and 45.4% (266.2 cGy vs. 518.0 cGy, p = 0.038), respectively (range, reduction in mean cardiac dose:16.8-68.9%, reduction in maximum cardiac dose: 8.9-74.2%). Cardiac volume decreased with BH versus FB, on average, 11.0%(range, -3.5 to 39.0%). Average reduction in mean left lung dose was 11.3% (235.0 cGy BH vs. 273.5 cGy FB, p = 0.25), whilethere was minimum change in maximum left lung dose (3,684.1 cGy BH vs. 3476.8 cGy FB). Due to small patient numbers, nodefinite conclusions could be made regarding the impact from the location of the tumor on cardiopulmonary irradiation.

Conclusions: Overall, cardiac and lung doses during APBI treatment appear to be relatively low, and can be reduced further withdeep inspiratory breath-hold. Further study and follow-up may clarify the impact of individual cardiac anatomy, tumor location,treatment plan characteristics, and delivered dose on cardiopulmonary toxicity.

Author Disclosure: P.J. Kim, None; H. Gee, None; S. Anooshfar, None; L.X. Yang, None; M.C. Rounsaville, None.

2795 In Vivo Dosimetry of Skin Dose during HDR Breast Brachytherapy using Balloon Catheter

S. Kim1, S. Oh2, A. Gale1, L. A. Vallow1

1Mayo Clinic, Jacksonville, FL, 2Catholic Medical Center, Seoul, Republic of Korea

Purpose/Objective(s): To estimate delivered skin dose and its variation during HDR brachytherapy for partial breast irradiationusing a balloon catheter.

Materials/Methods: In vivo dosimetry was performed to ensure delivered skin dose was not abnormally significant through thewhole treatment of partial breast in HDR brachytherapy with a balloon catheter. TN-502RD MOSFET in vivo dosimeters (BestMedical Canada, Canada) were calibrated for Ir-192 Nucletron HDR source in source to detector distances (SDDs) rangingfrom 21 to 61 mm. Calibration conditions were made using solid water slab phantoms. To assure full scatter condition, 10 cm thicksolid water slabs were placed both under the source and over the detector. On the other hand, a radio-opaque marker was placed onpatient skin close to the center of the balloon During CT simulation. Two calibrated MOSFET dosimeters were placed on the pointto measure and monitor variation of skin dose from fraction to fraction.

Results: Measured mean skin dose over 10 fractions was lower than calculated by planning system for the first 7 patients (range, -4.2% to -24.5%). Each patient also showed a significant amount of variation (standard deviation ranging from 4.5% to 7.9%),which indicates a possible geometric change due to both setup and anatomic variances.

Conclusions: It is found that delivered skin dose is likely to be less than calculation of planning system, which is mainly due to lackof backscatter during real patient treatment while calculation is performed with an assumption of full scatter. All of 7 patients also

Proceedings of the 50th Annual ASTRO Meeting S525

showed a large amount of dose variation from faction to fraction, which was considered to be because of geometric variation.Therefore, it may be useful to monitor skin dose through the whole treatment, especially for patients with a shallow thickness be-tween the balloon surface and the skin.

Author Disclosure: S. Kim, MT&T (Medical Tool and Technology, LLC), E. Ownership Interest; S. Oh, None; A. Gale, None;L.A. Vallow, None.

2796 Comparison of Displacement of the Silver Clips in the Cavity Determined by KV-plain Film and Cone-

beam Computed Tomography for the Breast Cancer Patients Treated by Three-dimensional ConformalExternal-beam Partial Breast Irradiation Assisted by Active Breathing Control

C. Liu, J. Li, J. Xing, C. Liang, S. Tian, T. Fan, M. Xu

Shan dong Tumor Hospital, Jinan, China

Purpose/Objective(s): To compare the margins of CTV extended to PTV depending on the displacement of the silver clips in thecavity determined by KV-plain film and cone-beam computed tomography (CBCT) for the breast cancer patients treated by three-dimensional conformal external-beam partial breast irradiation (EB-PBI) assisted by active breathing control (ABC).

Materials/Methods: Ten patients consistent with the condition of external-beam partial breast irradiation were enrolled the study.The patients received CT simulated positioning assisted by ABC to get the CT image sets based on the respiratory condition ofmoderate deep inspiration breath hold (mDIBH), and the clips located at the cephal, pedal, lateral border and bottom of the cavitywere delineated respectively. Before each irradiation, twice CBCT and KV-plain film were received assisted by ABC device in thesequence of CBCT/KV-plain film/CBCT/KV-plain film, and then 3D-3D registration based on pixel between the CBCTimage and the planning CT image and 2D-2D registration based on pixel between KV-plain film and the planning digitally recon-structed radiograph (DRR) were finished, and the shifts of the marked clips on LAT, LNG ,VRT directions were recorded. Themean of the shifts and the margins of CTV to PTV on LAT, LNG ,VRT direction were calculated , and then the margins determinedby KV-plain film and CBCT were compared to observe the significance of difference.

Results: By comparison, the displacement of the marked clips determined by KV-plain film and CBCT on every directions had nosignificant difference (p . 0.05), the margins from CTV to PTV calculated by the marked clips displacement determined by KV-plain film on LAT, LNG,VRT directions were 5.0, 7.8, and 9.3 mm of on the cephal clip , 4.4, 6.4, 6.7 mm of the pedal clip, 5.0, 8.6,and 10.5 of the lateral clip and 5.4, 8.5, and 10.8 of the bottom clip. The margins from CTV to PTV calculated by the marked clipsdisplacement determined by CBCT were 5.7, 8.0, and 9.8 mm of the cephal clip, 5.0, 7.1, and 7.8 mm of the pedal clip, 5.3, 12.8,and 10.6 mm of the lateral clip and 5.3, 7.5, and 10.9 mm of the bottom clip . By comparison, the margins from CTV to PTV de-termined respectively by CBCT and KV-plain in the same clip and the same direction had no significant difference (p . 0.05).

Conclusions: Silver clips in the cavity can be clearly discovered on the KV-plain film and CBCT image ; CBCT and KV-plain areall the ideal image-guided methods to determine the displacement of silver clips for EB-PBI, comparably KV-plain film is a morefeasible method.

Author Disclosure: C. Liu, None; J. Li, None; J. Xing, None; C. Liang, None; S. Tian, None; T. Fan, None; M. Xu, None.

2797 Inverse IMRT Planning for Breast Cancer with ‘‘Limited Regional Metastatic Cancer’’ Involving the

Ipsilateral Anterior Mediastinal Nodes

S. Yoo, J. Boyd, S. Das, F. Yin, E. Jones, K. Light, L. Marks

Duke University Medical Center, Durham, NC

Purpose/Objective(s): With the increasing use of PET imaging, we are seeing more patients with ‘‘limited regional metastaticcancer’’ involving the ipsilateral anterior mediastinal nodes. The design of conformal RT plan is extremely difficult in these cases.We herein report the dosimetric results of IMRT planning for a series of such patients, and suggest a ‘‘class solution’’.

Materials/Methods: Five patients with regionally-extensive nodal disease were treated in the last year (1 right-, 4 left-sided). All 5patients underwent concurrent chemotherapy with Xeloda or Navelbine. All underwent CT scanning; 2 of the left-sided patientswith deep inspiration breath-hold to maximize the separation between the heart and chest wall. GTVs, CTVs, and PTVs were de-fined by the physician to include the breast/chest-wall, IMN, supraclavicular (Scv) and mediastinal nodes. The targets were large,with the average depth of the mediastinal nodes 10 cm from the anterior chest (range, 7-13 cm). Prescribed doses were typically 45-50 Gy to the CTV, 53-60 Gy to the GTV in Scv and 60-66 to the GTV in the mediastinum; usually in 30-33 fractions. IMRT plansinclude 8-9 6MV photon beams: 2 ‘‘AP/PA type’’ beams were set to cover Scv area only, 2-3 ‘‘lateral-tangential type’’ beams tocover breast CTV only, 2-3 ‘‘medial-tangential type’’ beams and 1-2 ‘‘enface type’’ beams to cover IMN, breast, and mediastinalCTVs. Beams had collimator rotation and their jaws fixed to minimize lungs and heart exposure. Beam intensity maps were op-timized to satisfy physician-defined dose constraints.

Results: The use of 8-9 IMRT beams provided reasonably good coverage of the targets: Average D95% (% of the prescribed dose toa 95% of target) were 99.1 ± 3.6% for the breast CTV, 96.3 ± 3.9% for the Scv CTV, 99.9 ± 4.5% for the mediastinum CTV, and101.0 ± 3.8% for the GTVs. The average dose homogeneity index (% of target volume getting between 95% and 110% of pre-scribed dose) was 90.3 ± 0.1% for breast CTV. Average V20 and mean dose to entire lung were 21.4. ± 8.3% and 13.4 ± 3.7Gy. Average V30 and mean dose to heart were 12.9 ± 16.5% and 12.5 ± 9.6 Gy. Average maximum cord dose was 26.5 ± 7.8Gy. One patient had acute symptomatic pneumonitis at 20 Gy requiring treatment interruption although the dose to the lung forthis patient was not higher than other patients. The remaining 4 patients tolerated the treatment without acute complications.We found no reported late complications, although follow-up is short.

Conclusions: Inverse IMRT planning using multiple beams with fixed jaws/collimator rotation is proposed as a ‘‘class solution’’for patients with breast cancer and metastases to the ipsilateral mediastinal nodes. Reasonably good target coverage was achievedand doses to the adjacent critical normal tissues generally appeared acceptable.

Author Disclosure: S. Yoo, None; J. Boyd, None; S. Das, None; F. Yin, None; E. Jones, None; K. Light, None; L. Marks, None.