Improving the dosimetric coverage of interstitial high-dose-rate breast implants

Purpose/Objective: We performed a retrospective computed tomography (CT)-ba sed three-dimensional (3D) dose-volume analysis of high-dose-rate (HDR) int erstitial breast implants to evaluate the adequacy of lumpectomy cavity cov erage, and then designed a simple, reproducible algorithm for dwell-time ad justment to correct for underdosage of the lumpectomy cavity. Methods and Materials: Since March 1993, brachytherapy has been used as the sole radiation modality after lumpectomy in selected protocol patients wit h early-stage breast cancer treated with breast-conserving therapy. In this protocol, all patients received 32 Gy in 8 fractions of 4 Gy over 4 days. Eleven patients treated with HDR brachytherapy who underwent CT scanning af ter implant placement were included in this analysis. For each patient, the postimplant CT dataset was transferred to a 3D treatment planning system, and the relevant tissue volumes were outlined on each axial slice. The impl ant dataset, including the dwell positions and dwell times, were imported i nto the 3D planning system and then registered to the visible implant templ ate in the CT dataset, The calculated dose distribution was analyzed with r espect to defined volumes via dose-volume histograms. Due to the variabilit y of lumpectomy cavity coverage discovered in this 3D quality assurance ana lysis, dwell times at selected positions were adjusted in an attempt to imp rove dosimetric coverage of the lumpectomy cavity. Using implant data from 5 cases, a dwell-time adjustment algorithm was designed and was then tested on 11 cases. Tn this algorithm, a point P was identified using axial CT im ages, which was representative of the underdosed region within the cavity, The distance (d) from point P to the nearest dwell position was measured. A number of dwell positions (N) nearest to point P were selected for dwell t ime adjustment. The algorithm was tested by increasing the dwell times of a variable number of positions (N = 1, 3, 5, 7, 10, and 20) by a weighting f actor (alpha), where alpha = f(d) and alpha > 1, and subsequently performin g 3D dose-volume analysis to evaluate the improvement in lumpectomy cavity coverage. Results: Before adjustment in the 11 implants, the median proportion of the lumpectomy cavity and target volume that received at least the prescriptio n dose was 85% and 68%, respectively. After dwell-time adjustment, lumpecto my cavity coverage was significantly improved in all 11 cases. The median d istance from point P to the nearest dwell position (d) was 1.4 cm (range 0. 9-1.9). The median volume of the lumpectomy cavity receiving 32 Gy increase d from 85.3% in the actual implant to 97.0% (range 74-100%) by increasing t he dwell time of a single dwell position by a median factor (alpha) of 12.2 according to the above algorithm. With N = 3, the median proportion of the cavity volume receiving 32 Gy was improved to 97.5% (range 77-100%), with a median alpha of 5.7, Further improvement in lumpectomy cavity coverage wa s relatively small by increasing additional dwell times, To addition, with N = 20, the median absolute volume of breast tissue receiving 150% of the p rescription dose was 70.3 cm(3) compared to 26.3 cm(3) in the actual implan t; whereas with N = 1 or N = 3, this median volume was only 35.9 and 42.0 c m(3), respectively. Conclusion: Lumpectomy cavity coverage sometimes appears suboptimal with in terstitial HDR breast brachytherapy using our current technique. A simple d well-time increase at only 1-3 dwell positions can compensate for some unde rdosage without creating significant regions of overdosage. Using simple me thodology, a single reference point representing the underdosed region can be utilized for initial selection of the dwell positions to be increased. ( C) 2000 Elsevier Science Inc.