Palliative treatment with carboplatin as late line therapy to patients with metastatic breast cancer

Introduction

Among patients with breast cancer, 20–25% will develop metastatic disease, and the treatment will be palliative [1]. In this situation, the purpose of treatment will be to ease symptoms in order to maintain or improve quality of life and if possible, to prolong life [2,3].

When the disease no longer respond to standard treatment regimes, the patient is either treated symptomatic or with experimental chemotherapy. The optimal cytostatic treatment of heavily pretreated breast cancer patients is not standardized, often of low efficacy, and limited by comorbidity and performance status. As the treatment is palliative, the benefits of tumor response and improvement in disease-related symptoms as a result of chemotherapy must be weighed against treatment-induced toxicity and its impact on quality of life [4]. Still more patients are receiving chemotherapy near end of life [5]. This is due to constant improvement in breast cancer treatment and an increasing patient wish for further treatment [6]. For several patients, even the smallest possibility of benefit appears to be worth fighting for and they prefer to do something actively [6,7].

Carboplatin has been proved effective in the treatment of early-stage breast cancer, both as monotherapy [8,9] and as combination therapy [10–12]. The efficacy of carboplatin in the treatment of pretreated metastatic breast cancer is more equivocal [13–19]. Furthermore, the efficacy in treating heavily pretreated metastatic breast cancer is still debatable and only few studies exist [20,21]. Recent studies have indicated effect of platinum-containing regimes in treatment of triple negative breast cancer [14,20,22,23], and international guidelines have included platinum-containing regimes in the treatment of BRCA1/2 associated triple-negative metastatic breast cancer [2]. Most of the studies involving carboplatin, show an advantageous side effect profile [8,9,15–17], making the drug suitable for palliative care. The aim of the current study was to evaluate whether a carboplatin regime could be suitable for late line treatment of heavily pretreated patients with disseminated breast cancer, and to evaluate whether a carboplatin regime could be an option for those patients, who desire additional treatment, but with a reasonable balance between chance of effect and amount of side effects.

Patients and treatments

The study was performed at two oncological departments in Denmark, Aarhus (cohort 1) and Odense (cohort 2). The databases were searched to find all patients ever treated with a carboplatin regime. All files were accessible from October 1990, where the first patient, later treated with carboplatin, was diagnosed with breast cancer.

The patients included women with advanced breast cancer, treated with a carboplatin regime between July 2004 and February 2012. Inclusion criteria for this retrospective study were histological verified breast cancer, advanced disease proven by biopsy or radiological investigations, and a minimum of one treatment course with a carboplatin-regime. In addition, the patients should not have been exposed to carboplatin (cohorts 1 and 2) or gemcitabine (cohort 2) as part of prior treatment. As carboplatin was not a part of the standard treatment, the use of carboplatin was different in the two hospitals.

Treatments

In cohort 1, the treatment consisted of carboplatin AUC 5 monotherapy or carboplatin and trastuzumab 6 mg/kg (loading dose 8 mg/kg) in case of HER2-positive disease. Both treatments were given on day 1 in a 3 week cycle. In cohort 2, the treatment regime was carboplatin AUC 5 on day 1 plus gemcitabine 1000 mg/m² on day 1 + 8 in a 3-week cycle. Before every treatment, side effects were evaluated. The treatment was continued until disease progression or unacceptable side effects. Dose reduction or postponement of treatment followed local instructions. Evaluation of treatment efficiency and side effects was performed at every third treatment cycle and included CT- and/or MRI scans, biochemistry, and physical examination.

Data collection and statistical analysis

Data were collected through patient records, treatment charts, laboratory rapports, imaging examination, and pathology reports. Baseline was first treatment day with carboplatin.

Performance status (PS) was evaluated according to WHO score, based on medical records. The side effect profile of carboplatin was evaluated by biochemical markers and the patient's subjective perception according to CTCAE 3.0 (GFR and creatinine from MedDRA Version 12.0, SPSS Inc., Chicago, IL, USA), need for hospital admissions and dose reduction, as well as postponement of treatment due to side effects. Tumor response was evaluated according to RECIST criteria (version 1.0/1.1) based on CT and MRI. In addition, the response was based on the clinical examinations described in the medical record. Stable disease was defined as no change for a minimum of 3 months.

The primary endpoint was time to progression (TTP). TTP takes stable disease (SD) into account, which is a desirable and achievable goal in palliation. Secondary end-points included side effect profile, objective response rate (ORR), clinical benefit rate, time to response, and time to death (TTD). Hypothesis generating analysis was performed to identify possible subgroups (hormone receptor (HR)-positive, HER2-positive and triple negative status) with different responses to treatment. All descriptive statistics and analytical statistics were performed in Stata/IC 12.0 (StataCorp LLC, College Station, TX, USA). Estimates for time-to-event data with censoring, were determined from Kaplan–Meier curves. Due to the number of patients in the study, hypothesis generating analysis of time-to-event data were performed as differences in median TTP.

Results

The patient population initially consisted of 27 women in cohort 1 and 26 women in cohort 2. One patient in cohort 1 was excluded from the study because of possible primary lung cancer. Five patients in cohort 2 were excluded, because carboplatin was given as neoadjuvant/adjuvant therapy. For final analysis, in cohort 1, 12 women were treated with carboplatin monotherapy and 14 women were treated with carboplatin and trastuzumab. In cohort 2, 21 women were treated with carboplatin-gemcitabine combination therapy (Supplementary material). Trastuzumab was available in the entire period. Patient characterization and previous courses of treatment are summarized in Table 1.

Table 1. Patient characterization.

	Cohort 1	Cohort 2	All	p Value
Number of patients, n	26	21	47
Age (year)				.9
Median	49	48	47
Range	31–63	31–66	31–66
Histopathology				.1
Ductal	19 (73)^a	18 (86)	37 (79)
Lobular	3 (12)	1 (5)	4 (9)
Other	4 (15)	0 (0)	4 (9)
Unknown	0 (0)	2 (9)	2 (4)
ER status				.2
Pos	10 (38)	12 (57)	22 (47)
Neg	16 (62)	9 (43)	25 (53)
HER-2 status				.02
Normal	11 (42)	16 (76)	27 (58)
Pos	14 (54)	3 (14)	17 (36)
Unknown	1 (4)	2 (10)	3 (6)
Triple negative status				.6
Yes	7 (27)	7 (33)	14 (30)
No	19 (73)	14 (67)	33 (70)
Grade of anaplasia				.4
1	3 (12)	1 (5)	4 (9)
2	4 (15)	7 (33)	11 (23)
3	14 (54)	8 (38)	22 (47)
Unknown	5 (19)	5 (24)	10 (21)
Lines of palliative treatments				.2
0–1	2 (8)	5 (24)	7 (15)
2–4	13 (50)	12 (57)	25 (53)
5–7	7 (27)	2 (10)	9 (19)
8–10	4 (15)	2 (10)	6 (13)
Overall relapse duration (month)				.01
Median (CI)	28.3 (18.4;43.7)	9.5 (4.2;21.4)	17.4 (11.2;27.2)
Range	0.9–129.2	0.3–114.8	0.3–129.2
PS (WHO)				.1
0	12 (46)	6 (29)	18 (38)
1–2	12 (46)	15 (71)	27 (58)
3–4	2 (8)	0 (0)	2 (4)
Metastatic sites				.8
1	1(4)	2 (10)	3 (6)
2–3	15 (58)	12 (57)	27 (58)
4–5	10 (38)	7 (33)	17 (36)

Age (year): age at the time of primary diagnosis. Lines of previous treatments: the number of lines of systemic antineoplastic treatment (chemotherapy and/or anti-hormonal therapy) the patient has been exposed to before carboplatin treatment. Overall relapse duration (month): describes the patients’ overall duration of time of the disease from first relapse to the start of carboplatin treatment. PS: Performance status at the start of carboplatin treatment. Metastatic sites: metastatic sites at the start of carboplatin treatment.

^a Percent values in brackets.

The median number of treatments was 4 series with a range of 1–21 series. In 40 (85%) women, termination of treatment was due to disease progression and only in three (6%) women treatment terminated was due to side effects (Supplementary material).

The response to carboplatin containing-treatment is summarized in Table 2. A separate analysis was made for the most heavily pretreated women, i.e., those treated with 5–7 lines and 8–10 lines of systemic antineoplastic treatment respectively (Table 2).

Table 2. Response to therapy.

	Cohort 1	Cohort 2	All
Time to progression (month)
All
Median (CI)	2.3 (1.7;3.2)	4.1 (2.4;4.7)	2.8 (1.9;3.9)
Range	0.4–19.3	1.2–14.5	0.4–19.3
5–7 lines
Median (CI)	3.1 (1.9;9.6)	5.0 (5.0; .)	4.4 (1.9;9.6)
Range	1.9;19.3	5.0;9.1	1.9;19.3
8–10 lines
Median (CI)	7.0(1.8; .)	2.8(2.8; .)	4.4(1.8; .)
Range	1.8;8.5	2.8;4.4	1.8;8.5
Time to response (weeks)
Median (CI)	6.1 (3.0;7.6)	8.9 (6.1;11.6)	7.6 (4.1;8.9)
Range	3.0–8.4	6.1–12.9	3.0–12.9
Best tumor response, n (%)
CR	1 (4)^a	2 (10)	3 (6)
PR	6 (23)	6 (29)	12 (26)
SD	5 (19)	4 (19)	9 (19)
PD	11 (42)	7 (33)	18 (38)
Unknown	3 (12)	2 (10)	5 (11)
Objective response rate % (CI)
All	26.9 (11.6;47.8)	38.1 (18.1;61.6)	31.9 (19.1;47.1)
5–7 lines	28.6 (3.7;71.0)	50.0 (1.3;98.7)	33.3 (7.5;70.1)
8–10 lines	50.0(6.8;93.2)	0.0(0.0;84.2)	33.3 (4.3;77.7)
Clinical benefit rate % (CI)	46.2 (26.6;66.6)	57.1 (34.0;78.2)	51.1 (36.1;65.9)
Time to death (month)
Median (CI)	8.8 (5.3;13.7)	8.2 (6.2;13.4)	8.8 (6.8;11.0)
Range	0.4–54.2	1.7–26.0	0.4–54.2

5–7 lines: women pretreated with 5–7 lines of systemic antineoplastic treatment.

8–10 lines: women pretreated with 8–10 lines of systemic antineoplastic treatment.

^a Percent values in brackets.

Exploratory analysis showed no correlation between TTP and ER status (difference in TTP −1.0; 95% CI −2.5 to 0.5; p = .2 for cohort 1 and 0.7; 95% CI −2.7 to 4.1; p = .7 for cohort 2), HER2 status (difference in TTP −0.4; 95% CI −1.7 to 1.0; p = .6 for cohort 1 and −0.1; 95% CI −4.0 to 3.8; p = 1.0 for cohort 2) or those described as triple negative (difference in TTP 0.5; 95% CI −2.0 to 3.0; p = .7 for cohort 1 and 0.4; 95% CI −3.6 to 4.4; p = .8 for cohort 2).

The main reason for termination of treatment was progressive disease. Side effects resulted only in few dose reductions, treatment postponements, hospitalizations, and/or cessations of treatments. The dominant side effects were hematological or nausea/vomiting and fatigue. Myelosuppression, in particular neutropenia and thrombocytopenia was of medium severity, with only few grades 3 and 4 events. Non-hematologic side effects were generally mild and mainly of grads 0–2 and with only one grade 3 event in fatigue and non-grade 4 events at all (Supplementary material).

Discussion

The study included women treated with several lines of antineoplastic treatments prior to the carboplatin containing-regimes. Despite this, the women in general had a good PS prior to initiation of treatment, meaning that their quality of life could make additional treatment relevant.

The carboplatin regimes in our study were administered considerably later in the disease course than in the previously mentioned studies. Nevertheless, the treatment effect was comparable with what has been observed in previous studies [8,9,12–21,23]. Our study demonstrates that carboplatin regimes were effective in treatment of heavily pretreated patients with up to ten lines of previous treatments, and both as monotherapy and combination therapy with trastuzumab or gemcitabine. Despite being heavily pretreated the women in our study showed a TTP of 2.3–4.1 months, and a relatively high ORR of 26.9–38.1%. The response to carboplatin containing treatment was independent of the number of previous lines of relapse treatments. Neither did our results indicate a difference in treatment response depending on tumor subgroups and triple negative status. However, the number of patients was small and the results of the subgroup analysis should, therefore, be taken with caution.

As the study is retrospective, there is a risk of selection bias, such as the fact that the women in our study in general responded well to previous treatment lines (data not showed), which could indicate a population with more chemo-sensitive disease compared to the average patient.

The treatment was generally well tolerated. Still the degree of myelosuppression was not insignificant. Especially, neutropenia was frequent and of higher grades, but only few cases of hospitalization due to febrile neutropenia were recorded. The women treated with carboplatin-gemcitabine combination therapy showed a tendency towards more severe myelosuppression, indicating a more burdensome treatment.

The extend of side effects was consistent with what has been observed in previous studies of carboplatin monotherapy [8,9,15,16] and in studies of carboplatin in combination with gemcitabine [13,14,19,21], and less than in some studies where more frequent and treatment limiting side effects have been observed [12,17,18].

Cytostatic treatment of heavily pretreated patients with metastatic breast cancer near the end of life should be considered carefully [24]. Often patients and their families want to continue anti-neoplastic therapy and value even small benefits greatly, and consider side effects as less important [5]. Thus, a number of studies have shown that chemotherapy can improve the quality of life of cancer patients and result in better performance status, despite low objective responses [3,4,24].

In conclusion, treatments with carboplatin regimes showed an encouraging effect and were well tolerated and may be candidates for late line treatment of heavily pretreated patients with metastatic breast cancer, where the primary treatment goal is to minimize the negative impact from the disease upon the quality of life of the patient. However, it is also clear that more studies concerning treatment in heavily pretreated patients are needed.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by Kræftens Bekæmpelse, [10.13039/100008363].