Radiation-related quality of life parameters after targeted intraoperative radiotherapy versus whole breast radiotherapy in patients with breast cancer: results from the randomized phase III trial TARGIT-A

This study was a single center cross-sectional analysis. To qualify for the analysis, patients had to be randomized in the TARGIT-A trial between 2002 and 2009 from the University Medical Centre Mannheim. The full protocol of TARGIT-A is available online[5].

All eligible TARGIT-A patients were treated as follows Arm A: IORT with 50 kV X-rays (INTRABEAM™ system, Carl Zeiss Surgical, Oberkochen, Germany) delivering 20 Gy at the applicator surface during BCS (IORT group). In the presence of risk factors this was followed by EBRT with 46 Gy in 23 fractions or 50 Gy in 25 fractions to the whole breast (IORT-EBRT group). Arm B: EBRT with 56 Gy in 28 fractions to the whole breast postoperatively without boost (EBRT group). The rationale of the EBRT treatment was based on the German breast cancer study group (GBSG) recommendations at the time of study initiation.

The patients of the control groups outside of TARGIT-A were treated between 2002 and 2006 at the same centre. Patients received BCS and either a combination of 20 Gy IORT and postoperative EBRT of 46 Gy in 23 fractions (IORT-boost group) or postoperative EBRT of 50 Gy in 25 fractions followed by an external-beam boost to the tumor bed of 16 Gy in 8 fractions (EBRT-boost group). EBRT was initiated after completion of wound healing and/or chemotherapy. All patients received systemic therapy according to the St. Gallen consensus recommendations. Chemotherapy was routinely given before EBRT. Endocrine therapy was started 8–14 days after surgery or after completion of chemotherapy.

From June 2002 to February 2009, 88 patients (72%) out of 123 patients accrued to the TARGIT-A trial at our center consented to participate in the QoL study. Forty-six and 42 patients were allocated in the targeted intraoperative radiotherapy (Arm A, IORT +/− EBRT) and external beam radiotherapy (Arm B, EBRT) groups, respectively. Sixteen patients allocated to intraoperative radiotherapy received intraoperative radiotherapy plus external beam radiotherapy postoperatively (IORT-EBRT group). Five patients did not receive IORT: Four patients were treated with 56 Gy external beam radiotherapy postoperatively, one patient refused external beam radiotherapy. All patients received treatment as mentioned above. The patients in Arm B were treated with 28 x 2 Gy to the whole breast, and none of them received an additional boost. The mean age at the TARGIT-A entry was 64.7 years (median 65 years, range, 47 to 84), the mean follow-up time was 32.1 months (median 25 months, range, 9 to 94).

In addition, this single-center subgroup of patients from the TARGIT-A trial was compared to patients of our center treated with IORT as a tumor bed boost followed by EBRT outside of TARGIT-A (IORT-boost group), and patients treated with EBRT followed by an external-beam boost to the tumor bed (EBRT-boost group). All patients treated between 2002 and 2006 with 20 Gy IORT as a tumor bed boost followed by 46/2 Gy EBRT to the whole breast outside of TARGIT-A were asked to participate in the IORT-boost control group. 90 patients (96%) out of 94 patients consented to participate. All patients treated with 50/2 Gy EBRT to the whole breast followed by 16/2 Gy EBRT-boost to the tumor bed between 2005 and 2006 were asked to participate in the EBRT-boost group, and 53 patients (85%) out of 62 patients participated. For these groups, the mean age at the time of surgery was 56.3 years (median 56 years, range, 31 to 81), and the follow-up time was 39.1 months (median 41 months, range, 8 to 64).

Most patients suffered from breast cancer stage I or stage II. One patient (2%) in the EBRT group, 4 patients (4%) in the IORT-boost group and 11 patients (21%) in the EBRT-boost group had breast cancer stage III. At the time of survey, all patients were disease free. Further patients’ characteristics are summarized in Table1. We compared the patients that participated in the QoL study with those who declined participation to assess for sample bias, and there were no differences regarding demographic and clinical variables (data not shown). Also compared with patients in the whole TARGIT A trial[1], TARGIT-A patients in our study had largely similar demographic and clinical characteristics.

The present study was performed according to the Declaration of Helsinki principles. All patients were informed about the study and given the option of participating or not. Patients who consented to participate received mailed questionnaires 8 to 94 months following treatment. It was the choice of the patient to answer the QoL questionnaires, and an answer by the patient was considered as informed consent. To maximize response rate, patients were reminded by telephone after non-response.

Quality of life was assessed using two validated questionnaires of the EORTC: the Quality of Life Questionnaire C30 (QLQ-C30, version 3) and the Breast Cancer Module (QLQ-BR23).

The QLQ-C30 measure consists of one global health status/quality of life scale, five functioning scales (physical, role, emotional, cognitive, and social), three symptom scales (fatigue, nausea/vomiting, and general pain), and six single item scales (dyspnea, insomnia, appetite loss, constipation, diarrhea, and financial impact). The QLQ-BR23 module consists of two functioning scales (body image and sexual), three symptom scales (systemic therapy side effects, breast, and arm symptoms), and three single item scales (sexual enjoyment, future perspective, and upset by hair loss) specific to breast cancer. The scoring of both questionnaires was performed according to the recommended EORTC procedures[6]. All scores can be linearly transformed to a 0–100 point scale, with higher scores of functioning indicating greater functioning, i.e. better QoL, and higher scores on symptoms indicating worse symptoms, i.e. worse QoL. The time frame for all questions is the situation in the last week, except for items related to sexual functioning where a 4-week time frame is taken. Five functioning and symptom scales of the QLQ-C30/QLQ-BR23 questionnaires were preselected during the design of this study, based on a pilot study by our group[7] and relevance for radiation-related QoL in breast cancer: Global health status, restrictions in daily activities (role functioning), and general pain, breast, and arm symptoms. Other subscales and items of the QLQ-C30/QLQ-BR23 questionnaires were outside the aim of this study and are not presented here. The item response rate was between 96% and 99%.

The Hospital Anxiety and Depression Scale (HADS)[8], the Functional Assessment of Cancer Therapy-Fatigue (FACT-F) subscale[9], the Rosenberg Self-Esteem Scale (RSES)[10], and the Body Image Scale (BIS)[11] were used to control for differences that may inherently exist between the treatment groups. The Hospital Anxiety and Depression Scale (HADS) is a questionnaire for the screening of anxiety and depression in patients with physical illness. It consists of two subscales: one for anxiety and one for depression. The Functional Assessment of Cancer Therapy-Fatigue (FACT-F) subscale assesses fatigue and its impact on daily activities. The Rosenberg Self-Esteem Scale (RSES) measures global feelings of self-worth and self-acceptance. The Body Image Scale (BIS) is a valid measure of the affective, cognitive, and behavioral components of body image in cancer patients. The BIS was constructed in collaboration with the EORTC Quality of Life Study Group, and includes the body image subscale of the QLQ-BR23 module.

The scores of the HADS, FACT-F, RSES and BIS questionnaires were summed for each scale. The HADS scores range from 0 (no anxiety/depression) to 21 (severe anxiety/depression). The range of the FACT-F score is from 0 (greatest fatigue) to 52 (lowest fatigue). The RSES score ranges between 10 (low self esteem) and 40 (high self esteem). The BIS score ranges from 0 (best body image) to 30 (worst body image).

Data analyses were performed with the IBM Statistical Package for Social Sciences software, version 19 (SPSS Inc., Chicago, IL). All analyses were performed on an intention-to-treat and as-treated basis. The level of statistical significance was set at 0.01 (0.05/5) to reduce type I errors in multiple comparisons. Chi-squared tests (or Fisher’s exact tests), Kruskal-Wallis one-way analyses of variance, and post-hoc Mann–Whitney U-tests (or univariate analyses of variance and post-hoc Scheffe tests) were used to compare the treatment groups. Independent effects of demographic and clinical factors on QoL were tested using univariate linear regression analysis. Variables with a p value < 0.05 were further analyzed with multiple linear regression analysis (stepwise forward method).

In the ITT analyses, patients allocated to IORT showed more professional and other daily activities and fewer general pain symptoms compared to patients allocated to EBRT, but none of the P values were below 0.01 (Table2).

The AT analyses demonstrated a significant benefit for patients treated with IORT alone. Patients receiving IORT alone reported more professional and other daily activities (mean 78.7, SD 35.2) compared to patients receiving EBRT (mean 60.5, SD 29.5; P = 0.007). Differences in the global health status subscale were not observed (IORT, mean 63.6, SD 24.2, EBRT, mean 52.4, SD 22.1, IORT-EBRT, mean 60.9, SD 19.9; P > 0.01). The mean scores for the general pain, breast, and arm symptom scales were significantly lower in IORT patients than EBRT patients (Figures1,2,3). The difference in breast symptoms between IORT and IORT-EBRT patients was also significant (Figure2). Between-group differences in the HADS, FACT-F, RSES and BIS scores were not observed (P > 0.01). Table3 summarizes the frequencies of moderate or severe breast and arm symptoms reported by patients. The most commonly reported symptoms were moderate or severe pain in the arm or shoulder, difficulty in raising or moving arm sideways, and pain in the area of the affected breast, with no significant differences between treatment groups (P > 0.01). We found only a few associations between demographic and clinical characteristics and QoL parameters. Figure4 shows the percentage of variance explained by multiple linear regression modeling. Having two or more medical comorbidities was associated with worse global health status, more restrictions in other daily activities, i.e. worse role functioning, and more general pain symptoms (P = 0.004 to 0.043). Breast and arm symptoms were independently predicted by tumor size > 2 cm (P = 0.003 and 0.002).

No differences between IORT-EBRT patients and IORT-/EBRT-boost patients were found for any of the QoL comparisons (P > 0.01; Figures1,2,3). Table4 shows the proportion of variance explained in the QoL parameters by each of the demographic and clinical variables. In the multiple linear regression analyses, only one factor was related to 4 out of the 5 QoL parameters: having two or more medical comorbidities was associated with worse global health status, restrictions in daily activities, i.e. worse role functioning, more general pain symptoms, and arm symptoms.