One-week external beam partial breast irradiation: survival and toxicity outcomes

According to ASTRO and ESTRO guidelines, external beam Partial Breast Irradiation (PBI) is a valid option for early-stage breast cancer patients. Nevertheless, there is lack of consensus about the best treatment schedule. We retrospectively analysed data of female patients treated at our institution from 2013 to 2022 with adjuvant “one-week” partial breast irradiation. Clinical Target Volume (CTV) was an isotropic expansion of 15 mm from the tumour bed (identified as the breast tissue between surgical clips). The treatment schedule was 30 Gy delivered with Volumetric Modulated Arc Therapy in 5 daily fractions. The primary endpoint was Local Control (LC). Disease-Free Survival (DFS), Overall Survival (OS) and safety were secondary endpoints. Three hundred and forty-four patients with a median age of 69 (33–87) years were included in the study. After a median follow-up of 34 (7–105) months, 7 patients (2.0%) developed a local recurrence. Three-year LC, DFS and OS actuarial rates were 97.5% (95% CI 96.2%–98.8%), 95.7% (95% CI 94.2%–97.2%), and 96.9% (95% CI 95.7%–98.1%), respectively. Ten (2.9%) patients experienced grade 2 late toxicities. Five (1.5%) patients reported late cardiac major events. Three (0.9%) late pulmonary toxicities were detected. One hundred and five (30.5%) patients reported fat necrosis. Good or excellent cosmetic evaluation following the Harvard Scale was reported in 252 (96.9%) cases by the physicians, while in 241 (89.2%) cases by the patients. “One-week” PBI is effective and safe, and this schedule is a valid option for highly selected early breast cancer patients.


Introduction
Since Halsted described his surgical technique for breast cancer (Halsted 1894), physicians attempted to shift the clinical paradigm from radical approaches to more conservative treatments. The "National Surgical Adjuvant Breast and Bowel Project" and Veronesi et al. demonstrated that lumpectomy + 50 Gy in 25 Fractions (Fx) (i.e. conventional schedule) of adjuvant Whole Breast Irradiation (WBI) resulted in comparable survival as mastectomy (Fisher et al. 2002;Veronesi et al. 2002).
Then, START Trialists' Group (Bentzen et al. 2008;Haviland et al. 2013) and Whelan et al. (Whelan et al. 2010) conducted three distinct randomised phase III trials. They showed that lumpectomy + hypofractionated WBI (42.40 Gy or 40.05 Gy in 16Fx or 15Fx) achieved similar disease control and toxicity rates compared to the conventional schedule. Hypofractionation is now considered the standard for treating the whole breast in the adjuvant setting (Meattini et al. 2022;Smith et al. 2018).
Several phase III randomised trials recently demonstrated that Partial Breast Irradiation (PBI) and WBI shared comparable outcomes (Coles et al. 2017;Livi et al. 2015;Meattini et al. 2020;Orecchia et al. 2021;Polgár et al. 2013Polgár et al. , 2017Shah et al. 2013;Strnad et al. 2016;Vaidya et al. 2014Vaidya et al. , 2020Veronesi et al. 2013;Vicini et al. 2019;Whelan et al. 2019;White et al. 2019). According to ASTRO (Correa et al. 2017) and ESTRO-ACROP recommendations (Meattini et al. 2022), PBI is a valid option for strictly selected earlystage breast cancer patients. There are many dose schedules, and no one has been demonstrated to be the most beneficial. RAPID trial-like therapy (38.5 Gy/10Fx twice daily) is not recommended, while 40 Gy in 15Fx or 30-26 Gy in 5Fx is considered the standard treatment (Meattini et al. 2022). Glynn et al. (2022) compared the cost-effectiveness of WBI and PBI delivered in either 15 or 5 fractions. One-week PBI had the least cost and best expected quality-adjusted life years. However, there are still concerns about its efficacy and safety (Glynn et al. 2022).
The current study collected and analysed data from consecutive patients who underwent external beam PBI at Fondazione IRCCS Istituto Nazionale dei Tumori di Milano between 2013 and 2022. The aim is to report long-term survival and toxicity outcomes of 30 Gy in 5Fx delivered in one week.

Population, treatment and endpoints
We retrospectively analysed data from breast cancer female patients who completed adjuvant PBI after conservative breast surgery with a minimum follow-up of 6 months, from July 2013 to April 2022.
Institutional inclusion criteria for PBI were age ≥ 50, unifocal, grade (G) 1-2, Luminal-like (Luminal A tumours are characterised by the presence of hormonal receptors and the absence of HER2 and have a less than 20% expression of cell proliferation marker Ki-67; Luminal B tumours are oestrogen receptor positive and can be progesterone receptornegative and have a greater than 20% expression of Ki-67 (Orrantia-Borunda et al. 2022), ≤ 20 mm, pN0/mi, non-lobular invasive carcinoma. Some inclusion criteria were not considered mandatory in the presence of other good prognostic factors or firm desires of the patient.
The dose schedule was 30 Gy in 5 daily fractions (1 week of treatment); the contouring procedure is shown in Supplementary Table 1. Photon plans employing an energy of 6 MV were elaborated on Eclipse treatment planning system version 15.6 (Varian Medical Systems, Palo Alto, California) and calculated using the Analytical Anisotropic Algorithm with a dose grid resolution of 2.5 mm. Delivery technique of choice was Volumetric Modulated Arc Therapy (VMAT) for all patients. Dose constraints are reported in Supplementary  Table 2. Radiation Therapy (RT) was delivered by a Varian TrueBeam ® Linac employing daily Image Guided Radiation Therapy with a Cone-Beam Computed Tomography (CBCT). All patients underwent a daily CBCT with online positioning corrections.
The primary endpoint of the study was Local Control (LC), defined as the evidence of no Ipsilateral Breast Tumour Recurrence (IBTR). Time to local recurrence was the time in months between the primary surgery and the first IBTR. Overall Survival (OS) and Disease-Free Survival (DFS) (the time between surgery and the first evidence of recurrence disease or a new contralateral breast cancer) were secondary endpoints.
Acute toxicity was assessed at the end of the treatment. Follow-up consisted of a surgical visit every six months for the first three years and then annually + an annual radiotherapy visit. Prescribed once per year imaging were bilateral mammography and echography. Side effects were assessed according to the toxicity criteria established by the RTOG/ EORTC (Cox et al. 1995). Fat necrosis was considered as present if explicitly described in imaging reports.
The Harvard breast cosmesis scale (Harris et al. 1979) was employed for cosmetic evaluation during the last clinical visit. Afterwards, every patient was asked to assess their cosmetic outcomes by giving a mark from 1 to 10 (1 being the worst and 10 being the best result), which was coded as follows: Poor: 1-4; Fair: 5-6; Good: 7; Excellent: 8-10.

Statistical analyses
Kaplan-Meier survival curves were used to calculate the 3-year actuarial rates of the events. If an event exceeded the incidence of 3%, the univariate Cox proportional-hazards regression survival analysis was performed. Factors with p ≤ 0.20 in the univariate analyses were included in the multivariate Cox regression. MedCalc ® v20.111 (MedCalc Software Ltd, Ostend, Belgium; https:// www. medca lc. org; 2022) was used for statistical analysis.
The median interval between surgery and radiotherapy was 90 (32-321) days. The median Overall Treatment Time (OTT) was 5 (2-12) days, and no treatment interruption occurred except in 2 patients (one due to traumatic injury and to a systemic progression of a pancreatic adenocarcinoma). Median Planning Target Volume (PTV) was 135.1 (30.1-545.3) cc and the median percentage of target volume receiving a dose equal to or higher than 95% of the prescribed dose (V 95% ) was 99.9 (96.0-100) %. The median percentage of target volume receiving a dose equal to or higher than 105% and 107% of the prescribed dose was 0.1 (0-8.8) % and 0.0% (0.0-2.5), respectively. All organs at risks dosimetry complied with the constraints.
As reported in Table 2, the median age at diagnosis of patients with local relapses was 80 (68-85) years compared to the 69 (33-91) of the study population. Six had a Luminal A-like, G1-2 ductal invasive carcinomas, while one had a G3, HER2-positive ductal invasive carcinoma with a significant component of ductal in situ carcinoma and another one did not complete the treatment course. In all patients the PTV was covered with a median V 95% = 99.68 (96.7-100) %. Endocrine therapy was prescribed in 1 (14.3%) case, compared to the 78.5% of the study population. All the relapses were treated: 5 with mastectomy, 1 with lumpectomy alone and 1 with lumpectomy + radiotherapy. All these patients with relapse are alive without any further disease progression.
Cox regression highlighted a detrimental association between DFS and tumour G3 (p = 0.03) and no use of ET (p < 0.0001) at univariate analysis. Protective factors were the use and treatment duration of ET (p < 0.0001 for each factor) and target coverage (p = 0.002). Nevertheless, none of these variables reached statistical significance in the multivariate analysis.
One patient died from progression of a contralateral breast cancer. Three patients died from other malignancies; 3-year actuarial rate of breast cancer-specific survival was 99.3% (95% CI 98.6-100%).
Six died of unknown reasons. Among these patients, the median age was 83 (73-91) years. They all had hypertension, and 5 used anticoagulant drugs. The median value for the Mean Heart Dose (MHD) was 0.9 (0.1-2.5) Gy in these patients, the median value for the percent volume of the ipsilateral lung receiving more than 9 Gy was 6.2 (1.6-9.0) %. No patients had late cardiac toxicity, and 1 developed lung fibrosis.

Acute
As shown in Table 3, acute fatigue, skin toxicity and oedema were all grade ≤ 1; at the end of the treatment, 7 (2.0%) patients had grade 2 breast induration. No acute cardiac, pulmonary, nor grade ≥ 3 breast toxicity occurred.

Late
A total of ten (2.9%) grade 2 breast side effects were scored, with an increasing trend: 1.3% at 2 years and 2.1% at 5 years; no > grade 2 events were reported. Three patients had grade 2 fatigue, 2 and 5 had grade 2 oedema and fibrosis, respectively. The median time for the onset of grade 2 late toxicity was 36 (8-75) months. Major cardiac events occurred in 5 (1.5%) patients. The 3-year actuarial incidence of late cardiac toxicity was of  2.2% (95% CI 1.2%-3.2%). These events consisted of 1 coronaropathy, 1 heart attack, 1 heart failure and 2 atrial fibrillations. Four out of the 5 patients with late major cardiac events were treated on the right breast. The median "mean heart dose" was 0.5 (0.2-2.2) Gy and the median percent of the heart volume receiving more than 1.5 Gy was 2.88 (1.6-30.0) %. The patient who was treated on the left breast developed a heart attack, but she had a clinical history of metabolic syndrome (obesity, hypertension, hypercholesterolemia and diabetes).
Cosmesis results are presented in Fig. 2. Only 1 (0.4%) patient had a Poor evaluation result by the clinician, and the other 7 (2.7%) were Fair, for a total of 3.1% of bad events. At the multivariate analysis, only oedema was confirmed to be statistically related (OR = 11.87, 95% CI 1.84-76.5, p = 0.01). Seventeen (6.4%) patients assessed their cosmesis as Poor and 13 (4.9%) as Fair. Twenty-six out of thirty were deemed as Excellent or Good by the physician, and no one was Poor.

Discussion
For highly selected breast cancer patients, external beam PBI is a treatment option in several national and international guidelines: ESTRO-ACROP (Meattini et al. 2022), ASTRO (Correa et al. 2017), Breast Cancer Working Group of the German Society for Radiation Oncology (Strnad et al. 2020) and AIRO (Ciabattoni et al. 2022). We chose the same dose prescription of the Florence Trial (Livi et al. 2015) (30 Gy in 5 Fx) with the difference that we used daily fractions. In terms of absolute Biological Effective Dose (BED) 3.7 (= 78.7 Gy), these two schedules are equivalent, but the OTT-corrected BED 3.7 of our study was higher: 75.6 Gy VS 71.4 Gy. The choice was based on the essential gain in convenience and sustainability for the patient and the hospital (Glynn et al. 2022). A pilot study (Lozza et al. 2018) demonstrated the safety and feasibility of such a schedule.
As shown in Table 4, our findings differ from the most important randomised trials on external beam PBI: we reported 7 local (of which 2 "In Field") recurrences and a 5-year IBTR actuarial rate of 7.2% (95% CI 3.9%-10.5%)  (Coles et al. 2017) and the 1.5% (95% CI 0.1%-3%) of the Florence trial (Livi et al. 2015). This difference is mainly ascribed to patient selection or incomplete therapy courses (both endocrine and radiation therapy): relapsed patients had worse prognostic factors, as shown in Table 2. Conversely, several retrospective studies reported local control rates similar to our cohort. Belkacemi et al. (Belkacemi et al. 2018 Apr) found a 1%-2% per year, Lazzari et al. (Lazzari et al. 2018) 1.5% per year and Bucholz et al. (Buchholz et al. 2020) 0.5%-1% every year. The 7.2% recurrence rate in 5 years reported in our series equals to 1.4% of relapses/year. These data are well included between the best and the worst reports of the three previous "real life" works: 0.5%-2% (Lozza et al. 2018;Belkacemi et al. 2018 Apr; Lazzari et al. 2018). Acute toxicity is aligned with our previous results (Rocca et al. 2020) (G2: 2.0% VS 1.3%) and they are fully comparable to the Florence Trial (Meattini et al. 2020) (G2: 2.0% VS 2.0%), and much better than the RAPID Trial (Whelan et al. 2019) (G ≥ 2: 2.0% VS 28%).
Breast G2 late toxicities were few but showed an increasing trend: 1.3% at 2 years and 2.1% at 5 years, respectively, from the end of PBI. No G3 events were recorded. Experimental trials about PBI reported a range from 0 to 13% for rates of severe side effects events (Meattini et al. 2020;Vinante et al. 2019;Whelan et al. 2019). Our 5-year ≥ G2 late toxicity rate was 2.1% VS 0% compared to the Florence trial (Meattini et al. 2020). Even the update by Marrazzo et al. (Marrazzo et al. 2023) confirmed a 0% of major late toxicity in the 50 patients treated with a daily PBI.
Evidence of fat necrosis was high (30.5%) but similar to data reported in the literature (1%-50%) (Genova et al. 2022). The rate of this side effect is comparable to the study of PBI using brachytherapy: 36.8% in Lövey et al. (Lövey et al. 2007), 22.2% in Polgar et al. (Polgár et al. 2004), 27% in Wazer et al. (Wazer et al. 2001. External Beam PBI has yielded fewer rates of fat necrosis ranging from 3 to 25% (Russo and Taghian 2013). Even with a more intense dose schedule (40 Gy in 10 fractions, twice daily) proposed by Bourgier et al. (Bourgier et al. 2011), fat necrosis occurred only in 12% of the cases. This considerable incidence of fat necrosis could be because we did value the radiological appearance and not its clinical incidence. Patients did not perform a radiologic evaluation before radiation therapy. Therefore, it was unlikely to discriminate the surgical effects on fat necrosis, which range from 1 to 9% (Genova et al. 2023). The only factor found to have a strong correlation with fat necrosis was diabetes (OR = 2.1, 95% CI 1.1-4.0; p = 0.02), unlikely to other variables such as breast or PTV volume, cc of surgical excision, endocrine therapy and infections as reported by Russo et al. (Russo and Taghian 2013). Due to this finding, we will start a pilot study of dose deintensification PBI in these patients, decreasing the dose from 30 to 26 Gy/5Fx "One-week" as per Fast-Forward trial (Murray Brunt et al. 2020), which showed excellent results in terms of local control.
Ten (3.1%) patients scored with Fair/Poor cosmesis after the clinician's assessment. The same clinician assessed all the evaluations, which is among the strengths of this study. Livi et al. trial (Livi et al. 2015) and Marrazzo et al. (Marrazzo et al. 2023) stated a 0% and a 2% of fair cosmetic assessment, respectively.

Limitations of the study
The most important limitation of this study is its retrospective nature which makes it impossible to build a robust evidence due to multiple biases (selection, performance, measurement).
The median follow-up was too short. Many late events (side effects or recurrence of disease) could not have been detected, as it has been demonstrated that they might occur even 10 years after treatment (Heins et al. 2022;Pedersen et al. 2022).
The cosmesis endpoint lacks baseline and annual evaluations, making it difficult to assess the trend of this outcome. Moreover, the patient's assessment was not conducted using an established scoring system, limiting the transmissibility of this evaluation.
Quality of Life questionnaires were not administered to the patients due to the absence of a "time 0" assessment.

Conclusions
Based on our experience and the results of this study, "Oneweek" PBI is also effective and safe, and this schedule is a valid option for highly selected early breast cancer patients. For diabetic women, a de-escalation trial will soon start its accrual, to evaluate the effect of 26 Gy in 5Fx on fat necrosis.
Author contributions All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by RRC. The first draft of the manuscript was written by RRC and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Funding The authors declare that no funds, grants, or other support was received during the preparation of this manuscript. The authors have no relevant financial or non-financial interests to disclose.

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Conflict of interest
The authors declare that they have no conflict of interest.
Ethical approval This study was not funded. The authors declare that they have no conflicts of interest. All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. The study protocol was approved by the local ethical committee, with the registration number: INT 110/22. Informed consent was obtained from all individual participants involved in the study.