In this issue of ONCOLOGY, Ash and colleagues review the current state of intraoperative radiotherapy (IORT) for early-stage breast cancer. Fractionated accelerated partial-breast irradiation (APBI) has a significant accumulated literature, but the utility and appropriate use of single-fraction radiotherapy as a method of APBI remains contentious and undefined. There are three major causes for concern with current IORT methods: 1) patients are selected for a treatment strategy without full pathological review of data, 2) dose to the target volume is not consistent with known radiobiology, and 3) the treating physicians are unable to graphically evaluate and monitor doses delivered to the target volume or the normal tissues. It is worrisome that recent updates of the two seminal randomized trials of IORT completed in Europe are demonstrating results that are inferior to those seen with both whole-breast irradiation and fractionated APBI.
The TARGIT (“targeted intraoperative radiotherapy”) approach employs an intraoperative spherical applicator to deliver a single dose of 20 Gy at the applicator surface with 50-kV x-rays, which results in a dose of 5 Gy at 1 cm. Due to the physical constraints imposed by this technology, a higher dose at depth would create an unacceptably high dose at the surface of the applicator and the breast tissue with which it is in contact. Initial findings of the TARGIT-A trial were published somewhat prematurely in The Lancet. Women were randomized to either fractionated whole-breast radiotherapy (with or without a boost) or targeted IORT at a dose of 20 Gy at the applicator surface. Patients were grouped into two strata: the pre-pathology stratum included patients who had their first definitive lumpectomy and TARGIT at the same sitting, while the post-pathology stratum consisted of patients confirmed to be without adverse pathological features and who were then taken back to the operating room for the TARGIT treatment. Notably, 21% of patients in the pre-pathology stratum were triaged for additional, “remedial” whole-breast irradiation on account of unfavorable pathological features. Patients in the post-pathology stratum were selected for TARGIT after pathological review and did not receive whole-breast irradiation. In the initial report, at a median follow-up of approximately 24 months, there was no significant difference in the Kaplan-Meier estimates for local failure at 4 years (1.20% vs 0.95% P = .41). Toxicity events were infrequent and comparable in both arms.
The TARGIT trial results were recently updated at the San Antonio Breast Cancer Symposium 2012 annual meeting. The median follow-up was not reported in the print abstract; 3,451 patients were randomized, of whom 1,010 have had a minimum follow-up of 4 years and 611 patients have had a minimum follow-up of 5 years. The 5-year ipsilateral breast recurrence rate was higher in the TARGIT-treated patients (3.3% vs 1.3%; P = .042). Surprisingly, the 5-year rate of ipsilateral breast recurrence in the post-pathology stratum (ie, well-selected, favorable patients treated with TARGIT alone) was higher than in the pre-pathology stratum, in which about 20% of patients received remedial whole breast RT due to adverse pathology. The absolute excess in local failure was 3.7% in the former stratum and 1% in the latter, compared to whole breast RT. These results appear to confirm that the TARGIT dose is too modest to be efficacious without whole-breast RT, even in the lowest-risk patients.
The Milan group, led by Vernonesi and Orrechia, has studied a different technology for the same purpose: intraoperative radiotherapy with electrons (ELIOT). In the ELIOT technique, the breast margins are re-apposed intraoperatively and an intraoperative dose of 21 Gy is prescribed to the 90% isodose line with 3–12 MeV electrons. The Italian group has completed accrual to a randomized trial of whole-breast irradiation vs APBI using the ELIOT technique, and initial results from this randomized trial were recently reported at the Groupe Européenne de Curiethérapie–European Society for Radiotherapy and Oncology (GEC-ESTRO) meeting in Barcelona. Unfortunately, the trial results were not submitted for publication in the printed abstract book, but they are viewable online at the meeting web site. A total of 601 patients were randomized to the whole-breast arm and 585 to the ELIOT arm. With a median follow-up of approximately 6 years, the 5-year ipsilateral breast tumor recurrence rate was 5.3% in the ELIOT arm and 0.7% in the whole-breast arm (P < .0001). The corresponding 5-year “true” local recurrence rates were 3.2% vs 0.7% (P = .0002). The regional recurrence rates were 1.1% vs 0.4%. A careful review of the study reveals that these results could have been anticipated based on well-established principles of breast radiotherapy.
With the ELIOT approach, a homogenous dose is delivered to a target volume that is generally consistent with the target volumes treated in the accumulated APBI literature. However, the ELIOT approach is agnostic to final pathology findings; no remedial treatment is offered to women found to have positive margins or positive lymph nodes. The impact of positive margins on local failure is well documented. Tellingly, in the ELIOT randomized trial, only 23% of all patients (269 out of 1,184) were suitable for APBI based on American Society for Radiation Oncology (ASTRO) consensus statement guidelines, and there was no difference in local control in a post-hoc analysis of this subset. However, 387 out of 1,184 patients (33%) were unsuitable for APBI in the ELIOT trial, and presumably many of these patients had either positive margins or positive lymph nodes. (The available slides do not describe the population characteristics.) The ELIOT results amply highlight the perils of poor patient selection that may come with IORT. Furthermore, the clinical fat necrosis rate in the ELIOT patients in the randomized trial was 12%, much higher than the 2.3% reported with balloon device–based APBI. The ELIOT dose is almost double the 2-Gy equivalent-dose of fractionated courses of APBI. Thus, the ELIOT experience as reported in the interim results of the randomized trial have produced not one but two major reasons for pessimism—suboptimal tumor control and higher toxicity.
Neither the TARGIT nor the ELIOT technique allows image verification of target volume coverage or dose to organs at risk. Although unlikely, it is certainly not impossible that a deep-seated lumpectomy bed could bring an applicator in close proximity to a rib or the left ventricle without the treating physicians ever realizing it. Current reimbursement rates for IORT reflect the reality that careful evaluation of graphic treatment plans or consideration of normal tissue dosimetry is simply not a part of IORT approaches.
In summary, we too are excited by the prospect of even more compressed radiotherapy options for women requiring adjuvant breast radiotherapy. However, we urge readers to be cautious in their interpretation of the current intraoperative literature as they implement their programs. To study the safety and efficacy of shorter courses of APBI, we are conducting a prospective study that delivers the entire adjuvant treatment in 2 days using a multi-lumen brachytherapy device (Contura MLB, Bard, Irvine, Calif). This approach is grounded in currently accepted radiobiological parameters, and it allows for both post-pathology selection of patients and image-based treatment planning.
Financial Disclosure: The authors have no significant financial interest or other relationship with the manufacturers of any products or providers of any service mentioned in this article.