Between June 1997 and August 1999, 32 women diagnosed with a total of 33 AJCC Stage I/II breast carcinomas underwent surgical breast excision and postoperative irradiation using HDR brachytherapy interstitial implantation as part of a multi-institutional clinical Phase I/II protocol. Eligible patients included those with T1, T2, N0, N1 (≤3 nodes positive), and M0 tumors of nonlobular histologic features with negative surgical margins, no extracapsular lymph node extension, and a negative postexcision mammogram. Brachytherapy catheters were placed at the initial excision, reexcision, or either sentinel or full-axillary sampling. Direct visualization, surgical clips, and ultrasound and/or CT scan assisted in the delineation of the target volume, defined as the excision cavity plus a 2-cm margin. High-activity ¹⁹²Ir (3–10 Ci) was used to deliver 340 cGy/fraction, 2 fractions/d, for 5 consecutive days, to a total dose of 34 Gy to the target volume. Source position and dwell times were calculated using standard volume optimization techniques.

Between June 1997 and August 2000, 32 patients with 33 Stage I and II breast cancers were enrolled in a multi-institutional institutional review board-approved protocol of HDR brachytherapy of the tumor bed plus margin after lumpectomy. Whole breast external beam RT was not given. Patients were eligible for the protocol if they met all the following criteria: (1) unicentric primary cancer with invasive ductal histologic features; (2) Stage T1, T2, N0, N1 (≤3 metastatic axillary nodes without no extracapsular extension, a minimum of 6 nodes in the specimen, or negative sentinel lymph node biopsy); (3) negative (≥1 mm) microscopically assessed surgical margins; (4) no collagen-vascular disease or concurrent pregnancy; (5) no known unresected residual carcinoma and no diffuse microcalcifications; (6) negative postlumpectomy mammogram if cancer presented with malignancy-associated microcalcifications; (7) no prior malignancy except nonmelanoma skin cancer <5 years before enrollment in study or continuously disease free if >5 years; and (8) signed consent form. Patients were considered ineligible for participation in the protocol if any of the following were present: (1) tumor histologic features with invasive or in situ lobular carcinoma or pure ductal carcinoma in situ; (2) skin involvement; (3) a breast unsatisfactory for brachytherapy (defined as having <1 cm thickness of breast tissue within the entire implant volume as measured from the skin to the pectoralis fascia or subareolar position of the lumpectomy cavity); or (4) last breast surgery >8 weeks before planned brachytherapy.

All tumor excisions were performed with the aim of complete gross tumor removal. All excision specimens were coated with India ink, and the margins at all surfaces were measured at a minimum of 10 levels. Specimens were classified as having an extensive intraductal component (EIC) if they met the following criteria: (1) intraductal carcinoma is prominently present within the tumor, and (2) intraductal carcinoma is present in sections of grossly normal adjacent breast tissue. To be enrolled in the protocol, patients with margins <1 mm underwent re-excision at a separate surgical procedure. Patients who did not undergo perioperative implantation had the lumpectomy cavity marked by the surgeon with 6 surgical clips. Axillary assessment was performed with a Level I/II dissection (10 patients) or through identification of 1–3 sentinel lymph nodes using technetium tracer and blue dye (22 patients).

All patients were treated with an HDR interstitial implant with the isotope ¹⁹²Ir having a source strength of 3–10 Ci. Implants were either placed perioperatively (e.g., at the time of reexcision) or within 8 postoperative weeks of the definitive breast surgery. The target volume was defined as that volume encompassed by the excision cavity plus a 2-cm margin of breast tissue as delineated by direct visualization or by surgical clips using orthogonal radiography and ultrasonography.

Although not a prerequisite to study participation, all implants in this study consisted of 2 planes. Implants were performed without a rigid template. Catheter entry and exit points were marked on the skin with the use of a customized spacing device that could be adjusted at 1-mm intervals. Intraplanar separation was fixed at 1.0 cm and interplanar separation was a variable dependent on tissue thickness in accordance with an optimization algorithm designed to achieve maximal dose homogeneity. In brief, the interplanar separation was determined for each catheter as target volume thickness (to a maximum of 4.0 cm) divided by the square root of 2 .

Twenty implants were performed with an open cavity at the time of initial excision or reexcision. The extent of the cavity was marked on the skin surface and by six surgical clips. With the wound open, the implant needles were placed to encompass the entire target volume. After placement of all the afterloading needles, dead-end plastic implant catheters were threaded into place and secured with a locking button sutured to the skin. Before implant loading, the final pathologic margins were verified to be >1 mm. If inadequate margins were obtained, the patients were removed from the study, and the implant was either removed or used to deliver a boost to the tumor bed before conventional whole breast external beam RT.

Postoperative implantation with a closed cavity was performed in 13 patients. Before placement of the implant, an ultrasound examination was performed to delineate the extent of the excision cavity. In addition, orthogonal radiography was performed to identify the surgical clips and their position marked on the skin relative to superficial fiducial markers. The skin marks were then used to define the target volume and guide needle placement.

Treatment planning was initiated by obtaining orthogonal radiographs with dummy source loading of the implant catheters for computerized reconstruction. The Nucletron Planning System (Nucletron, B.V., Veenendaal, The Netherlands) was used for isodose calculations. Because an HDR afterloader was used, implants could be optimized to maximize homogeneity throughout the implant volume by adjusting the dwell times of the ¹⁹²Ir source. The dwell spacing in all cases was fixed at 0.5 cm. The dose homogeneity index (DHI) was used to assess implant quality and has been described previously; the higher the DHI, the more uniform the dose distribution within a treatment volume. The calculation of the DHI was achieved as follows: the mean central dose is the average of the local dose minima, or geometric center doses, in the central plane of the implant. In a double-plane implant, the geometric center doses occur at the intersection of the perpendicular bisectors of the sides of the triangles formed by the neighboring sources. The mean central dose is the arithmetic mean of the geometric center doses in the implant. The peripheral dose is the minimal dose at the periphery of the clinical target volume. The DHI was calculated as the ratio of the peripheral dose to the mean central dose. The DHI was prospectively assessed for each implant, and achieving a value of 0.85 was the primary planning goal in each case. Deviations from this goal were primarily reflective of suboptimal interplanar spacing of the implant catheters.

A dose of 3400 cGy in 10 fractions was prescribed as the minimal dose anywhere within the prescription volume. A minimum of 6 h elapsed between each implant dose fraction.

To minimize the skin dose, the source was kept a minimum of 5–7 mm from the skin surface at the ends of each catheter. Patients were examined twice daily to monitor for complications and to ensure that any displacement of the implant catheters was not significant. After the 10th treatment fraction, the implant catheters were removed in the outpatient clinic, and the patient was immediately discharged home.

Patients were seen in follow-up at 1–3-month intervals. Mammograms were obtained at 6 and 12 months after the implant and then yearly thereafter. A global cosmetic score was assigned in those patients with a minimum of 24 months’ follow-up in accordance with a previously published scale: EXCELLENT = perfect symmetry, no visible distortion or skin changes, and no visible catheter entry/exit sequelae; GOOD = slight skin distortion, retraction or edema, any visible telangiectasia, any visible catheter entry/exit scar, or mild hyperpigmentation; FAIR = moderate distortion of the nipple or breast symmetry, moderate hyperpigmentation, or prominent skin retraction, edema, or telangiectasia; POOR = marked distortion, edema, fibrosis, or severe hyperpigmentation. Cosmetic scoring was performed by a radiation oncologist, surgeon, and clinic nurse, and the lowest of the 3 scores was retained for analysis. Skin and s.c. toxicity were graded by a radiation oncologist according to the Radiation Therapy Oncology Group/Eastern Cooperative Oncology Group (RTOG/ECOG) system The cosmetic and toxicity scores recorded on the last follow-up visit were used for analysis.

Results: The median follow-up of all patients was 33 months, and the mean patient age was 63 years. The mean tumor size was 1.3 cm, and 55% had an extensive intraductal component. Three patients had positive axillary nodes. Two patients experienced moderate perioperative pain that required narcotic analgesics. No peri- or postoperative infections occurred. No wound healing problems and no significant skin reactions related to the implant developed. The Radiation Therapy Oncology Group late radiation morbidity scoring scheme was applied to the entire 33-case cohort. In the assessment of the skin, 30 cases were Grade 0–1 and 3 cases were Grade 2. Subcutaneous toxicity was scored as 11 patients with Grade 0, 3 with Grade 1, 8 with Grade 2, 3 with Grade 3, and 8 with Grade 4. Clinically evident fat necrosis occurred in 8 patients at a median of 7.5 months after HDR brachytherapy completion. The only variables significantly associated with Grade 3–4 toxicity were the number of source dwell positions and the volume of tissue encompassed by the prescription isodose shell. The global cosmetic scores after a minimum of 18 months’ follow-up were 0 cases with poor, 4 with fair, 5 with good, and 24 with excellent scores. One case of ipsilateral breast tumor recurrence was diagnosed 23 months after HDR brachytherapy, for a 4-year actuarial recurrence rate of 3%. This failure appeared to be a new primary tumor, because it was histologically distinct from the initial tumor and was located 9 cm from the initial tumor bed and 3 cm from the edge of the implant volume.

Local recurrences after conservative surgery and whole breast RT are most likely to occur in the immediate vicinity of the lumpectomy site. This fact has prompted the investigation of an experimental approach to postlumpectomy RT, whereby treatment is confined to the breast tissue surrounding the excision cavity through the use of interstitial brachytherapy. In the current 33-case prospective study with a median follow-up of 33 months, 1 patient developed an apparently new primary cancer in the treated breast remote from the initial disease-bearing quadrant, for an overall breast failure rate of 3%. The cosmetic results appear favorable, with 88% of cases receiving a global cosmetic score of good/excellent. Acute toxicity was mild, and late skin toxicity was acceptable, with no Grade 3 or 4 events. Prominent (Grade 2 or 3) s.c. fibrosis and/or clinically evident fat necrosis was seen, however, in 58% of cases.

Several other investigators have also explored brachytherapy as the sole radiation modality after lumpectomy  Both LDR and HDR techniques have been performed. These studies generally had small patient numbers (17–90 enrollees) with limited follow-up (14–72 months) and varied with respect to clinical entry criteria, extent of tumor excision, margin evaluation, requirement for negative margins, dose rate, fraction size, total dose, and so forth. Nonetheless, most of these studies showed local recurrence rates of <10% and cosmetic ratings of good/excellent in two-thirds or better of their enrolled cases.

Two contemporary institutional studies of this treatment approach are particularly noteworthy for their adherence to strict enrollment criteria, defined dosimetric parameters, and prospective evaluation of complications, cosmesis, and local control. The first, from the Ochsner Clinic, most recently reported by King , was a prospective Phase I/II trial that enrolled 50 patients (51 breast cancers) with intraductal or invasive tumors ≤4 cm in size, negative inked surgical margins, and ≤3 positive axillary nodes. Brachytherapy was delivered by either a LDR (45 Gy during 4 days) or a HDR (32 Gy in 8 fractions during 4 days) technique. At a median follow-up of 75 months, these authors reported 1 recurrence (2%) in the treated breast and 3 (6%) regional nodal failures. Good or excellent cosmetic scores were recorded in 75% of cases. These results compared favorably with an historical control group treated contemporaneously at their institution with conventional whole breast RT. Grade 3 complications requiring surgical intervention were seen in 8% of cases, including 2 cases of fat necrosis. The relationship of these complications to the use of the LDR or HDR technique was not stated.

More recently, Vicini  reported on 174 patients with Stage I or II breast cancer treated at the William Beaumont Hospital treated using three different protocols of lumpectomy followed by RT restricted to the tumor bed by means of an interstitial implant. A total of 120 patients (69%) were treated with an LDR implant that delivered 50 Gy during 96 h. The remaining patients were treated with an HDR implant that delivered 32 Gy in 8 fractions (46 patients) or 34 Gy in 10 fractions (8 patients). Patient eligibility for these trials included tumor size ≤3 cm, tumor-free margins of excision measuring ≥2 mm, nonlobular histologic features/no extensive intraductal component, and ≤3 positive axillary lymph nodes. The implants were performed with a high degree of quality control, including detailed descriptions of implant dosimetry . The authors found no adverse acute or late sequelae, and the cosmetic results were judged as good or excellent in 90% of cases. One case of axillary nodal failure, but no cases of ipsilateral breast failure, were detected in any of the brachytherapy patients, resulting in a 5-year actuarial rate of locoregional failure of 1%. In a matched-pair analysis of brachytherapy patients with a comparable group of patients treated with conventional whole breast RT, these authors found no significant differences in the rates of locoregional failure, distant metastasis, disease-free survival, overall survival, or cause-specific survival.

An additional report from the William Beaumont Hospital group  presented the preliminary results of an in-house protocol using HDR brachytherapy as the sole radiation modality after lumpectomy in 38 cases of Stage I or II breast cancer. A minimum dose of 32 Gy was delivered in 8 fractions during 4 consecutive days. With a median follow-up of 31 months, 1 patient had an ipsilateral breast recurrence, for a crude failure rate of 2.6%; no regional or distant failures were found. Cosmetic results were judged as good or excellent in all patients. Late normal tissue reactions were mild with some residual firmness in the region of the implant in 9% but no instances of symptomatic fat necrosis.

The importance of the prescription volume and the DHI as a predictors of complication risk after HDR breast brachytherapy requires additional study. Although not statistically significant in our analysis, it is noteworthy that a DHI value of <0.85 was encountered in 75% of cases with Grade ≥3 s.c. toxicity. In the Beaumont trial although a higher dose per fraction (4.0 vs. 3.4 Gy) was used, the s.c. toxicity was reported as less than that encountered at Tufts/Brown. The Beaumont group did not report an analysis of the volume and dosimetric variables as they related to complication risk. Nonetheless, the median DHI reported from their trial was 0.878 compared with a median (and mean) value of 0.84 achieved at Tufts/Brown. Of particular note, however, the Beaumont group reported a narrower range of values (0.83–0.916) compared with those achieved in the current study (0.68–0.94). Furthermore, differences were present between the Beaumont and Tufts/Brown trials with respect to the volume of tissue that received 100% and 150% of the dose prescription. In the Beaumont trial, the median volume encompassed within the 100% isodose shell of 215.8 cm³ (range 98.5–407.2) was greater than that seen at Tufts/Brown (median value 160 cm³, mean 171, range 63–394). However, complication risk may be more closely related to the volume of tissue encompassed within the 150% isodose shell. This is suggested by the fact that the median value of 26.4 cm³ (range 15.5–38.8) observed at Beaumont was both lower and within a narrower range than that seen at Tufts/Brown (median 37 cm³, mean 45, range 14–144). These observations underscore the importance of obtaining a larger sample size to define accurately the relative importance of dose-volume relationships as they contribute to optimization algorithms for HDR breast brachytherapy.

The data from the current study and that of prior reports suggest that, with proper patient selection, brachytherapy as the sole radiation modality after lumpectomy is associated with mild acute toxicity, a low risk of local failure, and acceptable cosmetic outcomes. No definitive statement can as yet be made with respect to long-term results. A cautionary note must be made, however, with respect to intermediate- and late-s.c. toxicity. In the current series, 33% of patients treated with HDR brachytherapy developed clinically significant Grade 3 or 4 s.c. toxicity. Standardized assessment of implant quality, DHI and UI, was not associated statistically with the risk of s.c. toxicity. Instead, the risk of Grade 3 or 4 toxicity was associated exclusively with the volume of the implant as reflected in the number of source dwell positions and volume of tissue exposed to fractional doses of 340, 510, and 680 cGy. This is consistent with our prior analysis of clinically evident fat necrosis in this same patient cohort. Additional analyses of institutional and cooperative group experiences with this technique need to be performed to clarify the relative importance of dosimetric and volume effects so that they may be incorporated into refined optimization schemes.

Conclusion: Radiotherapy of the tumor bed alone with HDR interstitial brachytherapy is associated with a 33% incidence of Grade 3–4 s.c. toxicity, but with generally favorable overall cosmetic results. The risk of toxicity appears to be primarily related to the implant volume. With limited follow-up, the incidence of ipsilateral breast tumor recurrence was low.