Inherent change in MammoSite applicator three-dimensional geometry over time

Accelerated partial breast irradiation (APBI) with the MammoSite catheter is a new brachytherapy concept in breast conserving therapy for a subset of patients with early stage breast cancer [1, 2]. The catheter consists of an inflatable balloon and a central channel for HDR brachytherapy. The initial experience [3] describes the ideal technique for the initial placement of the catheter, either at the time of lumpectomy or percutaneously under ultrasound guidance. As the initial Phase I trial describes, in order to deliver a homogenous dose to the tumor cavity, the balloon on the catheter should be inflated with saline to achieve a uniform spherical shape. Asymmetry of the applicator, poor placement, and intrinsic applicator abnormalities are all grounds for removal of the applicator. In this trial, a number of applicators were removed, with poor balloon conformance the most common reason for removal. We describe three separate cases where asymmetric applicators corrected themselves over time without any intervention, allowing for subsequent treatment.

The MammoSite Catheter for APBI has shown to be well tolerated with acceptable cosmesis for the treatment of both invasive breast cancers and DCIS [4]. However, even in experienced hands, the initial MammoSite experience showed a 10% removal of implant due to asymmetry [3]. All patients described were treated using a single isotope dwell position. With a single dwell position, asymmetrical central channels would deliver an inappropriately asymmetrical dose [5]. Also, with a single dwell position, any non-spherical balloon placement would deliver an inhomogeneous dose. With a newer dose delivery technique, using dose optimization [6, 7] and multiple dwell positions, some applicators forming "imperfect" spheres can be correctly treated. However, with only one channel for isotope delivery, dose optimization cannot correct for channel asymmetry within the applicator. All optimized dwell positions still deliver dose around the channel symmetrically.

After a patient has a MammoSite applicator placed percutaneously, a CT scan for planning is done very shortly thereafter. The majority of clinics can facilitate placement of the applicator, a CT scan, and planning within 4–24 hours. In case 1, our patient initially followed the typical sequence of events. In her situation, the applicator would have normally been removed, but she wished to wait and retry applicator manipulation the next day. After only 20 hours, she had intra-balloon geometry change, placing her isotope channel in the center of the balloon. Our second case was placed using the closed technique, with her initial scan following placement on the same day. Her initial scan showed an oblong applicator, which would normally be characterized as an unsuccessful placement. She was re-scanned after 4 days and without manipulation, showed a successful placement. The final patient also underwent percutaneous placement, and her initial CT was performed approximately 1–2 hours later. Again, the balloon appeared asymmetric, with marked differences in the radius of the balloon on cross-sectional imaging, which would result in inhomogeneous surface doses on a brachytherapy plan. None of these patients would have been APBI candidates as per the MammoSite study guidelines. With time, however, these patients were converted to appropriate candidates and were treated successfully.

These three patients show that asymmetry of the MammoSite applicator on an initial planning CT might not be absolute contraindications for eligibility for MammoSite-based therapy. Our institutional practice, in line with current industry standards, would consider asymmetry of 2-mm or more to be unacceptable for MammoSite treatment. These cases would have been determined to have unsuccessful placements, necessitating applicator removal as defined by the initial study guidelines. However, these patients' later scans indicated adequate symmetry without further intervention. Once the MammoSite applicator became symmetrical and spherical, the patients were treated without any difficulty. Also the applicator did not change geometry again, as evidenced by daily CT scans. Moreover, there have not, to date been instances in which the MammoSite balloon symmetry did not improve on repeat imaging. Although our report is admittedly limited by the small number of cases, it is nonetheless encouraging that in all instances in which balloon asymmetry was discovered, this finding soon corrected itself and remained constant thereafter.

In our experience, of approximately 75 MammoSite treatments, these three patients represent the only patients who would have been deemed poor placement due to asymmetry only. All three of these patients converted from inadequate to adequate placement over the course of 1–4 days. No factors seem to indicate this would happen, since we had all 3 patients (100%) with asymmetrical applicators convert to symmetrical applicators. Our institutional policy was to wait as long as a week, before removing the applicators. This additional week makes the total time of an indwelling MammoSite catheter to be two weeks, which is approximately the time the catheter is indwelling in the patient if placed at the time of surgery and found to be tolerable [3].

The MammoSite is still a novel technology for partial breast irradiation. As the use of MammoSite catheters increases around the country, the learning curve will continue to increase. As seen in these cases, in some unsuccessful placements of applicators due to balloon geometry, the passage of time in days has seemed to correct the geometry. This suggests that some patients previously not considered candidates for treatment could still be treated with APBI using the MammoSite, warranting further study in a prospective fashion.