Initial dosimetric experience: Contura™ Multilumen Balloon Registry Trial

Oral Presentations / Brachytherapy 8 (2009) 105e180 Results: Dosimetry was outstanding with median V90, V150 and V200 of 96.2%, 25.7, and 12.9 cc. Thirty percent received antibiotics, and the infection rate was 5%. Keloid formation was seen in 1 patient. There were no symptomatic seromas and one report of asymptomatic fat necrosis seen on mammogram at 1 year. Fourteen patients (35.0%) were treated with a skin-cavity spacing of <7 mm. The median skin dose remained less than 100% in patients with skin bridges <7 mm. Even in patients with the device in close proximity to both skin and lung, dose to both structures remained less than 100%. There have been no local recurrences to date. Conclusions: Early clinical experience with the SAVI demonstrates the ease of placement of single entry brachytherapy devices combined with the increased dose modulation of interstitial brachytherapy. Dose to normal structures has remained exceedingly low. Almost half of evaluated patients were not candidates for other single entry brachytherapy devices because of skin spacing or breast size, demonstrating an expansion of candidates for single entry partial breast brachytherapy. OR26 Presentation Time: 10:15 AM A dosimetric comparison of AccuBoostÒ noninvasive partial breast brachytherapy to electron beam tumor bed boost and 3-D conformal accelerated partial breast irradiation Shirin Sioshansi, M.D.,1,3 Jessica R. Hiatt, M.S.,3 Mark J. Rivard, Ph.D.,1 Amanda A. Hurley, C.M.D.,3 Yoojin Lee, M.S., M.P.H.,2 Jaroslaw T. Hepel, M.D.,1,3 Gene A. Cardarelli, Ph.D., M.P.H.,3 Sean O’Leary, R.T.T.,1 David E. Wazer, M.D.1,3 1Radiation Oncology, Tufts Medical Center, Tufts University School of Medicine, Boston, MA; 2 Institute for Clinical Research and Health Policy Studies, Tufts Medical Center, Boston, MA; 3Radiation Oncology, Rhode Island Hospital, Brown Medical School, Providence, RI. Purpose: To perform 3D dose modeling and dose-volume analysis of the AccuBoostÒ system and compare it to electron boost and 3D-CRT APBI techniques. Methods and Materials: AccuBoostÒ is an image-guided non-invasive breast brachytherapy approach consisting of breast immobilization via moderate compression followed by 2D mammographic target localization. Tungsten applicators mounted on mammography paddles are centered on the target and direct HDR 192Ir emissions along orthogonal axes. Dosimetric characterization of single applicators has been performed. However, 3D dose modeling of all 4 applicator positions has not been examined due to limitations of conventional brachytherapy treatment planning systems to model the tungsten applicator collimation and variable tissue deformation as a consequence of breast compression in both axes. Consequently, within PinnacleÒ AccuBoost applicator dose distributions were modeled as single point-sources with appropriate simulated anisotropy using the 2D brachytherapy dosimetry formalism. This source data was applied to compressed breast CT data from 8 patients imaged in the prone position on a table such that the breast would fall forward and be compressed between parallel plates in the

Median AccuBoostÒ boost Median eboost p-value Median AccuboostÒ APBI Median 3DCRT APBI p-value

PTV Volume (cc)

PTV Dmax (Gy)

PTV Dmin (Gy)

PTV Dmean (Gy)

Chest Wall Max (cGy)

Skin Max (cGy)

44.2

2.31

1.77

2.07

30.8

91.2

68.8

2.27

1.11

2.14

214.2

229.3

0.023 77.9

NS 45.5

0.016 33.9

0.008 39.5

0.008 32.4

0.008 94.8

221.6

40.0

31.4

38.6

99.9

104

0.008

0.055

NS

NS

0.008

0.039

117

cranial-caudal and medial-lateral axes. Patient dose distributions were calculated using a 1 mm3 grid. PTV coverage and max dose to skin, lung, and chest wall were assessed and compared to conventional electron boost (using CT-based cavity + 0.5 cm, 2 Gy to 90% isodose line) and 3DCRT APBI (per NSABP B-39 protocol). Results: AccuBoostÒ target volumes are 35% and 65% smaller than the respective eboost and 3DCRT APBI volumes (p 5 0.023 for boost; p 5 0.008 for APBI). AccuBoostÒ max skin dose is 60% lower than eboost and 10% lower than 3DCRT. Compared to eboost and 3DCRT APBI, AccuBoostÒ delivers 70-90% less dose to the chest wall and lung. There is NSS difference between eboost and AccuBoostÒ boost for the PTV V110, V90, V80, V50, V20, Dmax, or D100. Eboost plans have a lower Dmin than AccuBoostÒ boost (1.11 Gy vs. 1.77 Gy, p 5 0.016), but higher V100, D90, and D50. With regards to PTV coverage, the only SS differences between the APBI techniques are: slightly higher D90 and V90 with 3DCRT and higher Dmax with AccuBoostÒ (45.5 Gy vs. 40 Gy p 5 0.055). Conclusions: By virtue of image-guidance and immobilization decreasing the need for extensive margin expansions, the AccuBoostÒ boost and APBI target volumes are substantially smaller than the respective eboost and 3DCRT volumes sparing more normal tissue. Max dose to skin, chest wall, and lung are also significantly lower with the AccuBoostÒ technique while PTV coverage remains comparable. Further efforts to model the tissue deformation as a result of serial compression are underway.

OR28 Presentation Time: 10:25 AM Initial dosimetric experience: ConturaÔ Multilumen Balloon Registry Trial Douglas W. Arthur, M.D.,1 Dorin Todor, Ph.D.,1 Frank A. Vicini, M.D.,2 Thomas B. Julian, M.D.3 1Radiation Oncology, Virginia Commonwealth University, Richmond, VA; 2Radiation Oncology, William Beaumont Hospital, Royal Oak, MI; 3Human Oncology, Allegheny General Hospital, Pittsburgh, PA. Purpose: The ConturaÔ Multilumen Balloon (MLB) registry trial is a multiple site accelerated partial breast irradiation (APBI), prospective, non-randomized study designed to compare the dosimetric efficacy of the ConturaÔ MLB with a single central lumen balloon device and to quantify the dosimetric improvement of multi-lumen use over single lumen use. The initial dosimetric comparison is presented. Methods and Materials: All patients are enrolled prior to ConturaÔ MLB placement. Eligibility criteria include early stage breast carcinoma confirmed to be <3 cm, unifocal, invasive or non-invasive histology that is estrogen receptor positive. Tumors must be resected with a negative pathologic margin, axillary nodes must be negative and the patient >50 years old. After successful post lumpectomy placement of the ConturaÔ MLB is confirmed by CT scan, three dosimetric plans are generated; multilumen (ML), centrallumen/multi-dwell (CL-M.D.) and central-lumen/single-dwell (CL-SD). All three plans are created with the goal of satisfying all five of the following dose criteria: >95% of the prescribed dose received by >95% of the target volume (with a 5% relaxation allowed on this parameter), max skin dose <125% (42.5 Gy), max rib dose <145% (50 Gy), V150 <50 cc and V100 <10 cc. All patients received 34 Gy delivered to the treatment target as 3.4 Gy bid over 5 days and continue with close followup. Secondary endpoints for this registry trial include disease control (local ipsilateral, contralateral, regional, and distant recurrences), cosmetic results and toxicity rates. Total accrual will be 342 patients. Results: Since opening in March 2008, the trial has accrued 52 patients. Twenty-five patients have completed data sets available for review. Seventy three percent of patients are >60 yo and overall 19% presented with disease stage Tis, 73.1% with T1 and 7.7% with T2. Median tumor size was 15 mm. When comparing the three dosimetric plans for each patient it was found that the ML approach succeeded in meeting all dosimetric criteria in 80% of the cases whereas the CL-M.D. and CL-SD plans only succeeded in 61.5% and 40%, respectively. There were 5 patients where CL-M.D. plans (20%) and 9 patients where CL-SD plans (36%) were unsuccessful in meeting all 5 dosimetric criteria but, with the use of the ML catheter, all criteria were successfully met. The ML plan delivered a max skin dose of <125% in 92% of cases as compared to 73.1% and 60% for the CL-M.D. and CL-SD plans generated for the same patients, respectively. When the balloon-skin

118

Abstracts / Brachytherapy 8 (2009) 105e180

spacing measured a clinically relevant value of <7 mm, the ML plan delivered a reduced max skin dose in 36% as compared to plans reflecting CL-M.D. use and 44% reflecting CL-SD use. The mean skin dose reduction was 34% and 21% respectively. When the rib distance measured <5 mm, the ML plan delivered reduced dose to the closest rib in 28% of cases when compared to CL-M.D. plans and 36% when compared to CL-SD. The mean rib dose reduction was 16% and 4%, respectively. Conclusions: Location of the treatment balloon, skin and rib distance are all dependent on the specific location of the tumor within the breast and the resultant lumpectomy cavity. Without the dosimetric optimizing capabilities, the safe use of a single lumen balloon is limited to patients with ideal cavity geometry and balloon symmetry. This initial comparison shows that the use of multiple, offset lumens allows the optimization of target coverage while minimizing dose to nearby skin and chest wall, thus creating a new, improved dosimetric standard. Completion of this trial is needed for confirmation and to determine the clinical scenarios where only the ConturaÔ MLB could be safely used to deliver improved adjuvant APBI using balloon brachytherapy

OR28 Presentation Time: 10:35 AM A comparison of skin and chest wall dose delivered with multicatheter, ConturaÔ Multilumen Balloon, and MammoSiteÒ breast brachytherapy Laurie W. Cuttino, M.D., Dorin Todor, Ph.D., Lynn Gilbert, Douglas W. Arthur, M.D. Radiation Oncology, Virginia Commonwealth University, Richmond, VA. Purpose: Accelerated partial breast irradiation (APBI) continues to increase in popularity. Reports with increasing followup continue to support excellent in-breast control rates with acceptable toxicity, although late toxicities have been reported. Skin dose and toxicity has been correlated and a recent investigation using the MammoSiteÒ (MS) device has suggested that increased toxicity may occur when the dose to the chest wall exceeds 125% of the prescribed dose. The MS device is a single-lumen balloon with recognized dosimetric limitations. With the emergence of skin and chest wall dose constraints, the appropriateness of treatment with MS is dependent on location of the balloon within the breast and proximity of normal structures with limited ability for dosimetric correction. The multicatheter (MC) technique has a high degree of dosimetric flexibility and has not been associated with significant skin or chest wall toxicity. Several new intracavitary devices incorporate multiple lumens with the potential to address the dosimetric limitations experienced with a singlelumen device and to better approximate the doses delivered with the MC technique. This investigation compares the ability to control skin and chest wall doses between three cohorts of patients: those treated with MC, Multilumen balloon (MLB), and MS brachytherapy. Methods and Materials: The dosimetric data for 43 patients treated with the MC technique, 45 patients treated with the MLB, and 83 patients treated with the MS were reviewed. This cohort represents consecutively treated patients from our most recent experience to minimize any learning curve effect on dosimetry. Plans were generated using 3D software (Brachyvision, Varian Medical Systems, Inc., Palo Alto, CA). The maximum doses delivered to the skin and chest wall were calculated for all patients. The minimum distances from the balloon surface to the skin and chest wall were calculated for the MLB and MS. Results: The mean skin doses for the MC, MLB, and MS were 2.3 Gy (67% of prescription), 2.8 Gy (82% of prescription dose), and 3.2 Gy per fraction (94% of prescription dose), respectively. The mean skin distances for patients treated with the MLB and the MS were 1.1 and 1.3 cm, respectively. Although the skin distances were similar (p 5 0.23) for the two balloon techniques, the mean skin dose with the MLB was significantly lower than with the MS (p 5 0.05).The mean chest wall doses for the MC, MLB, and MS were 2.3 Gy (67% of prescription), 2.8 Gy (82% of prescription dose), and 3.6 Gy per fraction (105% of prescription dose), respectively. The mean chest wall distances for patients treated with the MLB and the MS were 1.4 and 1.2 cm, respectively. Again, although the chest wall distances were similar (p 5 0.36) for the two balloon techniques, the mean chest wall dose with the MLB was significantly lower than with the MS (p 5 0.002). The percentage of patients receiving

skin doses in excess of 125% for MC, the MLB, and the MS were 0%, 0%, and 9.6%, respectively. The percentage of patients receiving chest wall doses in excess of 125% for MC, the MLB, and the MS were 0%, 11.1%, and 38.6%, respectively. Conclusions: The MC and MLB techniques are associated with significantly lower mean skin and chest wall doses than the MS. Treatment with the MS was associated with significantly more patients receiving doses to the skin or chest wall in excess of 125% of the prescription. Doses delivered with the MLB were more similar to those obtained with multi-catheter interstitial brachytherapy. Given that the multi-catheter technique is associated with a very low incidence of skin and chest wall toxicity, treatment with the MLB may prove to be better tolerated than the MS.

OR29

Presentation Time: 10:45 AM

Does treatment time or source strength affect late toxicity in accelerate partial breast irradiation brachytherapy? Jill Heffernan, M.D., Aditya A. Bondal, B.S., Peck-Sun Lin, Ph.D., Laurie W. Cuttino, M.D., Douglas W. Arthur, M.D., Dorin A. Todor, Ph.D. Radiation Oncology, Virginia Commonwealth University, Richmond, VA. Purpose: Interstitial multicatheter (MC) brachytherapy and MammoSiteÒ (MS) balloon brachytherapy are the accelerated partial breast irradiation (APBI) delivery techniques associated with the most mature outcome data. The biologically effective dose (BED) for continuous low dose-rate implants has been shown to be affected by the dose rate. Modern APBI employs high-dose-rate (HDR) delivery. The source strength can impact the treatment duration by a factor of 3 or more. Given that treatments are usually administered over 10 fractions bid, the typical treatment length can vary from 5 to more than 7 days. All these are legitimate concerns that prompted us to study using accepted radiobiological formalisms and parameters whether prolonged treatment times or the overall treatment time correlates with clinical outcomes. Methods and Materials: Dosimetric and clinical data for 79 early stage breast cancer patients treated with accelerated partial breast radiation via MC (33) and MS (46) were analyzed. The prescribed dose for all treatments was 34.0 Gy in 10 fractions bid. Treatment parameters included source strength (mean 6.589 Ci, range 4.331-9.613), delivery time (mean 470.4s, range 232.6-837.5), overall treatment time (mean 6.1d, range 5-9), Planning Target Volume (PTV) size, Equivalent Uniform BED (EUBED) and EUD. For each plan, a dose matrix was exported and the BED calculated using the linear-quadratic formula for each voxel of the planning target volume (PTV). These individual values were then aggregated into EUBED for the entire PTV, essentially removing any effect of non-uniformity. EUBED0 accounted for the total dose, fraction size, and a/b ratio. Two other models were employed. EUBED1 included repair and the actual delivery time sequence, while EUBED2 added the overall treatment time. The generalized EUD was computed with a 5 2, 4, 8 and 10. Results: The mean followup time was 35.5 mo (range 12-126 mo). Clinical information included pain: Grade 0-73.1%, 1-23.1%, 2-1.9% and 3-1.9%, fibrosis: Grade 0-37.2%, 1-53.8%, 2-7.7%, 1.3%, cosmesis: Grade 152.6%, 2-44.8%, 3-1.3%, 4-1.3%, scar formation, edema, and skin changes. Using a/b 5 4 Gy, the mean EUBED0 was 78.4 Gy for MC and 79.3 Gy for MS. When repair and the actual delivery time were considered, EUBED1 showed a decrease of 5% relative to EUBED0 and EUBED2 decreased by up to 10% for the longest treatment times. The hypothesis we tested was simply that this drop in EUBED would be significant to impact on the late toxicity. Given its ability to mimic doseresponse more closely than DVH parameters, EUD was added to the list of parameters to be correlated with outcome. A multivariate analysis and multiple correlation tests were ran on the variables alone as well as composite variables (e.g., EUBED*PTV_Volume/ Curie$seconds) and no statistically relevant correlations between dosimetric and delivery parameters and clinical outcome were detected. Conclusions: Late tissue effects (cosmesis, late skin and subcutaneous toxicity) of MC HDR APBI were previously reported and correlated with static (or planed) dosimetric quantities (V150, V200, DHI). Variable dose rate delivery (source strength), the delivery time and pattern and the overall treatment time were not previously studied but are of significant