Liposome-encapsulated doxorubicin compared with conventional doxorubicin in a randomized multicenter trial as first-line therapy of metastatic breast carcinoma

Abstract

BACKGROUND

The objective of this study was to compare the efficacy and toxicity of the liposome-encapsulated doxorubicin, TLC D-99 (Myocet, Elan Pharmaceuticals, Princeton, NJ), and conventional doxorubicin in first-line treatment of metastatic breast carcinoma (MBC).

METHODS

Two hundred twenty-four patients with MBC and no prior therapy for metastatic disease were randomized to receive either TLC D-99 (75 mg/m²) or doxorubicin (75 mg/m²) every 3 weeks, in the absence of disease progression or unacceptable toxicity. The primary efficacy endpoint was response rate. Responses were assessed using World Health Organization criteria and were required to be of at least 6 weeks' duration. The primary safety endpoint was cardiotoxicity. Cardiac function was monitored by multiple-gated radionuclide cardioangiography scan, and the left ventricular ejection fraction (LVEF) was scored at a central laboratory. Patients were removed from study if LVEF declined 20 or more EF units from baseline to a final value of greater than or equal to 50%, or by 10 or more units to a final value of less than 50%, or onset of clinical congestive heart failure (CHF).

RESULTS

Median age was 54 years in both treatment groups. All relevant prog nostic factors were balanced, with the exception that there were significantly more progesterone receptor positive patients in the doxorubicin-treated group. Protocol-defined cardiotoxicity was observed in 13% of TLC D-99 patients (including 2 cases of CHF) compared to 29% of doxorubicin patients (including 9 cases of CHF). Median cumulative doxorubicin dose at onset of cardiotoxicity was 785 mg/m² for TLC D-99 versus 570 mg/m² for doxorubicin (P = 0.0001; hazard ratio, 3.56). The overall response rate was 26% in both treatment groups. The median TTP was 2.9 months on TLC D-99 versus 3.1 months on doxorubicin. Median survival was 16 versus 20 months with a nonsignificant trend in favor of doxorubicin (P = 0.09). Clinical toxicities, commonly associated with doxorubicin, appeared less common with TLC D-99, although the difference was not statistically significant. There was only one report of palmar-plantar erythrodysesthesia (Grade 2) with this liposomal formulation of doxorubicin.

CONCLUSIONS

Single-agent TLC D-99 produces less cardiotoxicity than doxorubicin, while providing comparable antitumor activity. Cancer 2002;94:25–36. © 2002 American Cancer Society.

Anthracyclines, especially doxorubicin, are among the most active agents in the treatment of breast carcinoma.1–5 Even with availability of taxanes, and other new agents, doxorubicin remains a mainstay of treatment for patients with metastatic disease.3, 4 Anthracycline-based regimens produce relatively high response6–9 and may be particularly valuable in certain patient populations, such as those women with HER-2 positive tumors.10, 11 Unfortunately, the clinical use of anthracyclines is limited by dose-related cardiomyopathy, which becomes more prevalent with increasing cumulative doses of anthracyclines.12–14

As myocardial injury is known to occur cumulatively, starting with the first dose of treatment,15 there has been an effort to develop alternative formulations of anthracyclines with lower inherent cardiotoxicity. The liposome encapsulation of anthracyclines, specifically doxorubicin, has the potential to decrease toxicity of the agents. TLC D-99 (Myocet, Elan Pharmaceuticals, Princeton, NJ), a liposomal formulation of doxorubicin, was designed to reduce the cardiotoxicity of doxorubicin while preserving its antitumor efficacy.16, 17

The purpose of this randomized, multicenter trial was to test the hypothesis that single-agent TLC D-99 would result in significantly less cardiac toxicity than the same dose and schedule of conventional doxorubicin, while providing comparable antitumor efficacy in first-line treatment of metastatic breast carcinoma.

PATIENTS AND METHODS

Patient Population

Eligible patients were 18 years or older, had histologically or cytologically proven breast carcinoma with measurable metastatic disease, an Eastern Cooperative Oncology Group (ECOG) performance status of 0–2, and a life expectancy of at least 3 months. Patients had to have adequate bone marrow, liver, and renal function as evidenced by a leukocyte count of greater than or equal to 3500 cells/μL, absolute neutrophil counts (ANC) greater than or equal to 2000 cells/μL, platelets greater than or equal to 100,000 cells/μL, serum bilirubin less than or equal to 1.2 times the upper limit of normal, aspartate and alanine transaminases up to 4 times upper limit of normal, and serum creatinine level of less than 1.5 mg/dL. Patients with resting left ventricular ejection fraction (LVEF) greater than or equal to 50% and no other active neoplasm other than carcinoma in situ of the cervix or nonmelanoma skin carcinoma were included. Women of childbearing potential had to be using reliable contraceptive methods.

Patients were not eligible if they had bone disease only. Adjuvant doxorubicin up to a maximum lifetime dose of 300 mg/m² was allowed, but patients should not have received adjuvant treatment with other anthracyclines or anthracenediones. Further exclusions were any cytotoxic chemotherapy for metastatic disease or adjuvant chemotherapy within 6 months of entering the study. Patients who had prior radiation greater than 3500 centigrays (cGy) to the mediastinal area or radiation to greater than 50% of the bone marrow also were excluded. Patients with a history of congestive heart failure (CHF), serious cardiac arrhythmia, or myocardial infarction within 6 months before enrollment were excluded, as were those with brain metastasis. Pregnant or lactating women were ineligible to enroll.

The study was approved by the institutional review board (IRB) at each participating center, with regular renewal, and all patients signed a written informed consent before randomization.

Study Treatment

Patients were stratified by institution and prior adjuvant doxorubicin and were randomized to one of the two treatment groups using a balanced block design. Patients received either TLC D-99 75 mg/m² or conventional doxorubicin 75 mg/m². (All doses of TLC D-99 refer to the doxorubicin content delivered via liposome encapsulation.) The recommended infusion duration of TLC D-99 is 1 hour; therefore, for the sake of uniformity, both TLC D-99 and conventional doxorubicin were administered as a 1-hour intravenous infusion. Granulocyte colony-stimulating factor use was to begin 48 hours after chemotherapy and to continue for at least 10 days or until ANC was greater than or equal to 10,000 cells/μL post-nadir in 2 consecutive counts 24–72 hours apart. Granuloycte colony- stimulating factor had to be discontinued at least 48 hours before the next cycle of chemotherapy. Treatment cycles were repeated every 3 weeks, in the absence of disease progression or significant toxicity requiring drug discontinuation.

To optimize treatment, individual dose titration based on toxicity was allowed. Dose of study drug was increased by 15 mg/m² for a platelet nadir grater than or equal to 75,000 cells/μL and ANC nadir less than 500 cells/μL on at most one count, and no other Grade 3 or 4 toxicity. Dose was reduced by 15 mg/m² for a platelet nadir less than 50,000 cells/μL, ANC nadir less than 500 cells/μL on 2 consecutive nadir counts, or any other Grade 3–4 toxicity. Patients had to recover from toxicities to receive a subsequent treatment cycle. On the day of dosing, the platelet count had to be greater than or equal to 100,000 cells/μL and ANC greater than or equal to 1200 cells/μL. In addition all other toxicities had to recover to Grade 1 or lower.

Study Evaluations

Each patient underwent a complete physical examination at baseline, including vital signs, ECOG performance status, and clinical tumor assessment. Hematology, serum biochemistry, urinalysis, and electrocardiogram tests were required. We estimated LVEF by multiple-gated radionuclide cardioangiography (MUGA) scan. And ejection fraction determined using the global LVEF value. A chest radiograph, computed tomography scan, magnetic resonance imaging scan, sonogram, bone scan, and brain scan were obtained, as clinically indicated.

Imaging studies, required for tumor measurement, were performed at baseline, every two cycles, at the time the patient came off study, and 3-month after the last dose of study treatment. After a response was achieved, response status was determined every two cycles. Hematology tests were performed before each cycle and 96 hours into each cycle. Repeat complete blood counts were performed 24–72 hours apart until recovery. Serum biochemistry tests were repeated before each cycle. A physical examination was performed before each cycle.

Electrocardiograms and MUGA scans were done at baseline, and after reaching a lifetime cumulative doxorubicin dose of 300, 400, 500 mg/m², and before each subsequent dose, at off-study, and at 3-month follow-up. All MUGA scan data were transferred electronically to a core laboratory for blinded interpretation and estimation of LVEF. Before the first interim analysis, endomyocardial biopsies were performed at a select number of institutions after lifetime cumulative doxorubicin dose of 425 mg/m². All patients whose LVEF declined by greater than 10% to a value of greater than or equal to 50%, or by greater than 6% to a value of less than 50% were to have cardiac biopsies regardless of lifetime cumulative doses. All biopsies were read by a core pathologist, and the results were scored according to the Billingham scale.15 The purpose of biopsies was to validate the results of MUGA scans. After the first interim analysis, it was determined that MUGA scans were adequate to monitor cardiac function and cardiac biopsies were discontinued.

Patients were withdrawn from the study for any of the following reasons: disease progression, unacceptable toxicity, noncompliance with the protocol, and patient or physician request. Patients were withdrawn for cardiac toxicity, defined as a decrease in resting LVEF of 20 or more points from baseline to a final value of greater than or equal to 50%, a decrease of greater than or equal to 10 points from baseline to a final value of less than 50%, a cardiac biopsy of Grade 2.5 or higher, or clinical evidence of CHF.

Response Criteria

All randomized patients were assessed for efficacy (intent-to-treat). Tumor measurements (as recorded at the treating centers) and other efficacy data were evaluated by an independent evaluation committee blinded to treatment assignment. Complete response (CR) was defined as the complete disappearance of all evidence of disease, including disease-related signs and symptoms, lasting at least 6 weeks. Partial response (PR) was defined as a greater than or equal to 50% decrease in the sum of the products of the two longest perpendicular dimensions of all measured lesions for at least 6 weeks, with no evidence of progressive disease (PD). Stable disease was defined as no significant change in measurable and nonmeasurable disease. Progressive disease was defined as a greater than or equal to 25% increase in the product of the two longest perpendicular dimensions of any measurable lesion or in the estimated size of nonmeasurable disease, or the unequivocal appearance of a new lesion, or reappearance of old lesions.

Time to progression (TTP) was defined as the time from Day 1 of treatment to the first evidence of PD or death within 6 months of the last dose of study treatment. Time to treatment failure (TTF) was defined as the time from Day 1 of treatment to the first evidence of PD, onset of cardiotoxicity, off-study for toxicity or death within 6 months of the last dose. Overall survival was defined as the time from Day 1 of treatment to death.

All treated patients were evaluated for safety, including cardiac toxicity. All clinical and laboratory toxicities, regardless of causality, were graded by the National Cancer Institute Common Toxicity Criteria (NCI-CTC).

Statistical Methodology

A 50% response rate was assumed for the conventional doxorubicin therapy in this population of patients with advanced breast carcinoma. TLC D-99 was not expected to be more active than the conventional doxorubicin. The study was powered to detect significance if the true response rate to TLC D-99 therapy in this patient population is greater than or equal to 15% lower than the true response rate to doxorubicin therapy. A sample size of 144 patients per treatment group was considered to be sufficient to achieve that objective with a 5% alpha and 80% power. Three interim analyses were planned, after enrollment of 72, 144, and 216 patients, using O'Brien-Fleming adjustments to P values.

Analyses of all efficacy parameters were stratified by prior use of doxorubicin. Pretreatment characteristics, efficacy, and safety parameters for the treatment groups were compared using Fisher exact test for binary variables and Wilcoxon rank-sum test for continues and ordered variables.

Distribution of time-to-event was estimated by Kaplan–Meier product-limit method, and curves were compared using log-rank chi-square test. The Cox proportional-hazards model was used to estimate the hazard ratio (HR), which indicates the overall risk of experiencing an event in one group relative to the other group. In this report, a HR greater than 1 favors the TLC D-99 group. Two-sided confidence intervals were applied to all time-to-event parameters. Kaplan–Meier curves also were used to plot the lifetime cumulative doxorubicin dose at the onset of cardiac events, and curves were compared by a log-rank test, and a HR was estimated by the Cox regression.

RESULTS

From November 1992 through May 1998, a total of 224 patients were randomized into this study by 42 investigators in North America. The study was closed to further accrual after the third interim analysis, at 216 patients, met the O'Brien-Fleming stopping criteria for both the response rate (P < 0.023) and cardiotoxicity (P < 0.018). This report is based on data collected on all 224 patients with follow-up through February 1999, when all patients were off-study.

Pretreatment characteristics were balanced between the two treatment groups (Table 1), except there were significantly more doxorubicin patients with positive progesterone receptor status (50% vs. 33%, Fisher exact P = 0.02). Overall the median age was 58 years and 91% were ECOG performance status 0 and 1. Greater than 70% had visceral disease, with 42% liver and 47% lung involvement. There was soft tissue disease in 74% of patients and bone lesions in 41%.

Table 1. Pretreatment Characteristics

Characteristic	TLC D-99(n = 108)		Doxorubicin(n = 116)
	No. of patients	%	No. of patients	%
Age (yrs)
< 50	33	31	35	30
50–59	25	23	34	29
60–69	37	34	27	23
70 +	13	12	20	17
Median	58	58
Range	26–85	29–82
Performance status
0	53	49	50	43
1	48	44	52	45
2	7	7	14	12
Estrogen receptor status
Positive	46	43	57	49
Negative	37	34	34	29
Unknown	25	23	25	22
Progesterone receptor status
Positive	36	33	58	50
Negative	46	43	34	29
Unknown	26	24	24	21
Dominant sites involved
Soft tissue only	18	17	16	14
Bone and soft tissue	13	12	16	14
Visceral	77	71	84	72
No. of disease sites
1–2	63	58	60	52
≥ 3	45	42	56	48
Sites of disease
Abdomen	6	6	14	12
Bone	43	40	49	42
Breast	27	25	25	22
Liver	45	42	48	41
Lung	54	50	52	45
Lymph node	60	56	63	54
Skin/soft tissue	32	30	37	32

Exposure to prior therapy was comparable in both treatment groups (Table 2). Overall, 44% of patients had radiation therapy, and 54% received 1 or more systemic hormonal therapies for breast carcinoma, including greater than 30% of patients with metastatic disease. Adjuvant chemotherapy was administered to 40% of patients, including doxorubicin in 39 patients (17%). One TLC D-99–treated patient and two doxorubicin-treated patients had prior cytotoxic chemotherapy for metastatic disease, a protocol violation.

Table 2. Prior Anticancer Therapy

Therapy type	TLC D-99(n = 108)		Doxorubicin(n = 116)
	No. of patients	%	No. of patients	%
Radiation therapy	47	44	51	44
Hormonal therapy	57	53	65	56
Adjuvant only	27	25	27	23
Advanced only	15	14	21	18
Adjuvant and advanced	15	14	17	15
Chemotherapy	43	40	47	41
Adjuvant only	41	38	46	40
Advanced only	2	2	1	1
Adjuvant doxorubicin	18	17	21	18
Median dose (mg/m^2)	240	240
Range (mg/m^2)	167–300	70–360

Analysis Groups

A total of 108 patients were randomized to the TLC D-99 arm and 116 patients were randomized to the doxorubicin arm. One patient randomized to receive TLC D-99 was found to have had more than 3500 cGy radiation to mediastinum before first dose, a protocol violation, and never received study treatment. This patient was included in the TLC D-99 group for efficacy analyses but excluded from safety analyses, including cardiotoxicity. Two more patients who were randomized to the TLC D-99 group, erroneously received conventional doxorubicin. Thus TLC D-99 was administered to 105 patients and doxorubicin to 118. Those two patients were included in the TLC D-99 group for efficacy and in the doxorubicin group for safety analyses.

Cardioprotection

All MUGA scan data were sent to an independent central laboratory for estimation and interpretation of LVEF values without knowledge of treatment arm. Cardiac events, sufficient for removal of a patient from study, were more than twice as frequent in doxorubicin-treated patients than TLC D-99–treated patients (29% vs. 13%, log-rank P = 0.0001). With the increasing lifetime cumulative dose of doxorubicin and TLC D-99, there was a gradual increase in the median change from baseline LVEF to the first post-treatment LVEF among patients treated with either agent, but this was more pronounced in the doxorubicin group (Table 3). A Kaplan–Meier estimate of the probability of the first onset of a cardiac event as related to the lifetime cumulative dose of doxorubicin or TLC D-99 (Fig. 1) shows that risk of cardiotoxicity was much higher with doxorubicin treatment than TLC D-99 (HR = 3.56) (P = 0.0001).

Table 3. Changes in Cardiac Function by Total Lifetime Doxorubicin Dose

Total lifetime dose (mg/m^2)	No. of patients in cohort		Median LVEF change from baseline to posttreatment (%)		No. of patients with a protocol-defined cardiac event
	LVEFa		CHF
	D-99	Dox	D-99	Dox	D-99	Dox	D-99	Dox
LVEF: left ventricular ejection fraction; CHF: congestive heart failure; Dox: doxorubicin. a Asymptomatic LVEF decrease sufficient for removal of patients from study.
0–99	105	118	0	0	0	0	0	0
100–199	85	103	−2	−4	0	1	0	0
200–299	80	94	−3	−2	1	1	0	0
300–399	74	87	−3	−7	2	3	0	0
400–499	54	69	−5	−6	2	11	0	0
500–599	39	37	−2	−8	4	7	0	3
600–699	21	10	−9	−20	2	1	0	4
700–799	8	3	−10	−29	1	0	1	2
800–899	4	1	—	−15	0	1	0	0
≥ 900	3	0	−10	—	0	—	1	—
All doses	12	25	2	9
Log-rank P value	0.008		0.0001

Figure 1. Lifetime dose of doxorubicin to a cardiotoxicity endpoint. Dox: doxorubicin; HR: hazard ratio; C.I.: confidence interval.

Two patients (2%) on TLC D-99 developed clinical CHF (Table 4). One patient, after 13 cycles of TLC D-99 and a cumulative dose of 1110 mg/m², had a decrease of 14 EF units in her LVEF to 46% and was taken off-study. Two months after the last dose she presented with shortness of breath and bilateral pleural effusions and was hospitalized for CHF. Another patient, with prior adjuvant doxorubicin dose of 290 mg/m² and prior chest wall irradiation, received five cycles of TLC D-99 for a total lifetime doxorubicin dose of 785 mg/m², and went off-study for PD. Four months after the last dose, a MUGA scan showed a LVEF of 46% (a 16-point decrease from baseline). Later, the patient received five cycles of a mitomycin plus mitoxantrone, and 11 months after the last study treatment she received a diagnosis of CHF.

Table 4. Characteristics of Patients with Congestive Heart Failure

Treatment group	Prior chest RT	Adjuvant doxorubicin (mg/m^2)	Lifetime doxorubicin (mg/m^2)	Associated left ventricularejection fraction (%)			Mos since last dose
				Base	Nadir	Change
RT: radiation therapy. a Patient received five cycles of mitoxantrone after study before onset of congestive heart failure.
TLC D-99	No	0	1100	60	25	−35	2
TLC D-99	Yes	290	785a	62	20	−42	11
Doxorubicin	No	0	525	65	—	—	1
Doxorubicin	No	0	525	62	20	−42	2
Doxorubicin	No	0	585	58	26	−32	<1
Doxorubicin	Yes	0	600	63	15	−48	5
Doxorubicin	Yes	0	600	54	22	−32	1
Doxorubicin	No	0	675	64	35	−29	<1
Doxorubicin	Yes	195	690	60	16	−44	2
Doxorubicin	No	300	750	66	40	−26	3
Doxorubicin	Yes	0	765	67	14	−53	2

Nine patients (8%) on doxorubicin developed clinical CHF at lifetime doses of 525–765 mg/m². Three patients had CHF within 30 days of the last dose of study treatment, including one who died of CHF after 585 mg/m². All nine cases were attributed to study drug treatment.

Before the first interim analysis, 51 patients were treated at 8 participating institutions performing endomyocardial biopsies. Of those, 36 patients qualified for the procedure, and all 36 patients had cardiac biopsies. All biopsies were read by a core pathologist, blinded to treatment assignment, and the results were scored according to the Billingham scale. There was a significant difference between the two treatment groups favoring TLC D-99 and the number of patients who had a score of greater than or equal to 2.5 (26% vs. 71%; P = 0.02; Table 5).

Table 5. Myocardial Biopsy Scores by Billingham Scale

Patients biopsied	TLC D-99 (n = 19)	Doxorubicin (n = 17)
Billingham score (grade)
0	2	0
1.0	2	3
1.5	4	1
2.0	6	1
2.5	5	5
3.0	0	7
Grades 2.5 or 3.0	5 (26%)	12 (71%)
Fisher exact P value	0.02

Efficacy

The objective response rates (CR plus PR) for all randomized patients were 26% in both treatment groups (Table 6). The response rate was lower in the few patients in the prior adjuvant doxorubicin stratum. The response rate at visceral sites was also lower compared with soft tissue sites. The distribution of responses by demographic and disease-related characteristics showed no subgroup of patients who were more likely to respond to one agent or the other (Table 7).

Table 6. Objective Response to Treatment

Response	TLC D-99 (n = 108)		Doxorubicin (n = 116)
	No. of patients	%	No. of patients	%
CR: complete response; PR: partial response.
Objective response
CR	0	2	2
PR	28	26	28	24
Stable disease	37	34	45	39
Progressive disease	35	32	31	27
Not evaluable	8	7	10	9
Response rate (CR + PR) (95% confidence interval)	28	26 (18–35)	30	26 (18–35)
No prior doxorubicin	25/90	28	29/95	31
Prior doxorubicin	3/18	17	1/21	5
Cochran–Mantel–Haenszel Stratified chi-square P value	0.97
Stratified difference in response rates (D99/Dox) (%)	2
95% confidence interval (%)	(−9–13)
Response rate at major metastatic sites (%)
Liver	22	13
Lung	22	17
Lymph node	38	49
Skin/soft tissue	36	45

Table 7. Response Rate by Pretreatment Characteristics

Characteristic	Percentage of patients with CR and PR
	TLC D-99 (n = 108)	Doxorubicin (n = 116)
Age (yrs)
< 50	27	23
50–59	24	38
60–69	27	22
70 +	23	15
Performance status
0	28	26
1	23	29
2	29	14
Estrogen receptor status
Positive	28	28
Negative/unknown	24	24
Progesterone receptor status
Positive	22	21
Negative/unknown	28	31
Visceral involvement
Yes	19	18
No	42	47
Prior radiation therapy
Yes	19	18
No	31	32
Prior hormonal therapy
Yes	23	18
No	29	35
Prior adjuvant chemotherapy
Yes	16	13
No	32	35

All randomized patients were included in TTF, TTP, and overall survival analyses (Table 8). Median time to onset of response was 42 days in both groups. Time to treatment failure was similar for patients randomized to either group (Fig. 2), and the difference was not significant (log-rank P = 0.21; HR = 1.21). There was also no difference in TTP between the two randomized groups (Fig. 3) (log-rank P = 0.59; HR = 0.92).

Table 8. Time-to-Event Parameters

Parameter	TLC D-99(n = 108)	Doxorubicin (n = 116)
CI: confidence interval.
Time to treatment failure
Events (%)	81 (75)	96 (83)
Median (mos)	3.7	3.4
95% confidence limits	2.6–4.8	2.7–4.3
Log-rank P value	0.21
Hazard ratio (95% CI)	1.21 (0.90–1.63)
Time to progression
Events (%)	76 (70)	76 (66)
Median (mos)	3.8	4.3
95% Confidence limits	2.6–5.3	3.1–6.0
Log-rank P value	0.59
Hazard ratio (95% CI)	0.92 (0.66–1.26)
Overall survival
Events (%)	82 (76)	79 (68)
Median (mos)	16	20
95% confidence limits	13–18	15–27
Log-rank P value	0.09
Hazard ratio (95% CI)	0.76 (0.56–1.04)
One-year survival (%)	64	69
Fisher exact P value	0.48

Figure 2. Time to treatment failure. Dox: doxorubicin; HR: hazard ratio; C.I.: confidence interval.

Figure 3. Time to disease progression. Dox: doxorubicin; HR: hazard ratio; C.I.: confidence interval.

Figure 4. Overall survival time. Dox: doxorubicin; HR: hazard ratio; C.I.: confidence interval.

There was a survival trend in favor of the doxorubicin group, but the difference was not statistically significant (Fig. 4) (log-rank P = 0.09; HR = 0.76). There was no difference in the survival curves during the first 12 months (1-year survival rate, 64% for TLC D-99 vs. 69% for doxorubicin; P = 0.48).

Dose Modification and Toxicity

A total of 159 patients (71%) received cycles at higher doses after the first cycle of treatment. Overall, doses for 37% and 19% of cycles were 90 mg/m² and greater than or equal to 105 mg/m², respectively (Table 9). Dose reductions to 60 and less than or equal to 45 mg/m² occurred in 11% and 9% of cycles, respectively. The most frequent reason for dose reduction was myelosuppression, mainly thrombocytopenia. Prophylactic use of G-CSF was required and was administered only in 58% of TLC D-99 cycles and 70% of doxorubicin cycles. Blood transfusion occurred in 7% of cycles. A delay in dosing was infrequent. The median cycle length was 21 days. Median dose intensity was 25.4 mg/m²/week in the TLC D-99 arm versus 26.3 mg/m²/week in the doxorubicin arm.

Table 9. Exposure to Study Drug

Parameter	TLC D-99 (n = 105)	Doxorubicin (n = 118)
G-CSF: granulocyte colony-stimulating factor.
Total no. of courses	509	529
Median	4.0	4.0
Range	1–14	1–11
Median cumulative dose (mg/m^2)	360	390
Range (mg/m^2)	75–1110	75–840
Median dose intensity	25.4	26.3
Range (mg/m^2/week)	11.6–34.7	15.6–32.7
As fraction of planned dose (%)	102	105
Cycles with dose escalation (%)	55	58
Cycles with dose reduction (%)	22	15
For myelosuppression	21	14
For other toxicity	1	1
Cycles delayed by ≥ 7 days (%)	15	13
Cycles with G-CSF use (%)	58	70
Cycles with blood transfusion (%)	6	9

There was only one drug-related death: one doxorubicin-treated patient died on study of complications of CHF.

All patients who received at least one dose of study treatment were analyzed for adverse events. With comparable drug exposure, there were no new or unexpected toxicities in the TLC D-99 group, and there was no increase in incidence or severity of known doxorubicin toxicities. Adverse events were analyzed regardless of causality (Table 10). Myelosuppression was the most frequent and severe toxicity. There was no significant difference between the two treatment groups in incidence of Grade 3 or 4 toxicities. There was a trend toward fewer Grade 3 or 4 infections (Fisher exact test, P = 0.09) and Grade 3 or 4 nausea/vomiting (P = 0.06) in the TLC D-99 patients. There was only one report of palmar-plantar erythrodysesthesia (Grade 2) which occurred in the TLC D-99 group.

Table 10. Adverse Events

Event	TLC D-99 (n = 105)	Doxorubicin (n = 118)	P valuea
ANC: absolute neutrophil count. a Fisher exact test.
Hematologic toxicity (%)
Anemia (hemoglobin < 8 g/dL)	22	26	0.53
Thrombocytopenia (platelets< 20,000/μL)	13	10	0.53
Neutropenia (ANC < 500/μL)	50	58	0.28
Infection (Grade ≥ 3)	5	12	0.09
Neutropenic fever (Fever ≥ 38 °C ANC < 500/μL with intravenous antibiotics and/or hospitalization)	11	9	0.66
Other toxicity (%)
Nausea/vomiting (Grade ≥ 3)	13	24	0.06
Stomatitis/mucositis (Grade ≥ 3)	9	14	0.21
Diarrhea (Grade ≥ 3)	1	4	0.22
Asthenia/fatigue (Grade 3)	14	19	0.47
Cutaneous (Grade 3)	1	1	1.00
Alopecia (Grade 2)	84	88	0.44

DISCUSSION

Doxorubicin is one of the most active agents against breast carcinoma, even in the context of newer agents such as the taxanes.4 Cardiotoxicity from doxorubicin continues to be an ongoing medical concern and ultimately is the cumulative toxicity that is dose-limiting. When doxorubicin is combined with trastuzumab, cardiotoxicity appears to be particularly severe.11 Thus, despite a wide range of new agents, there is a continuing need for formulations of doxorubicin that protect the heart.

Concomitant administration of dexrazoxane, an iron-chelating agent, has been shown to reduce the cardiac toxicity of doxorubicin. Speyer et al. randomized 150 patients to receive fluorouracil, doxorubicin (50 mg/m²), and cyclophosphamide with or without dexrazoxane.18 There were two cases of CHF in the dexrazoxane group versus 20 in the control group (including 17 in patients treated at cumulative doxorubicin doses of < 550 mg/m²). Although not available at the time of initiation of this study, dexrazoxane is now approved in the United States for patients with metastatic breast carcinoma who already have received a cumulative doxorubicin dose of 300 mg/m², and who show initial response to doxorubicin.19 There is slightly greater myelotoxicity when dexrazoxane is added to doxorubicin,9 and there is a suggestion, in at least one study, of reduced efficacy when doxorubicin is combined with dexrazoxane. If there is any reduction in efficacy with dexrazoxane, it may be related to the finding that both doxorubicin and dexrazoxane bind to topoisomerase II and are antagonistic to one another in preclinical models.20, 21

The current study was initiated at a time (1992) when there were fewer alternatives to doxorubicin than there are now for patients with metastatic breast carcinoma. In this study, patients were closely monitored with MUGA scans at multiple timepoints: at baseline, within 4 weeks of study entry and after lifetime cumulative doxorubicin doses of 300, 400, and 500 mg/m² and for patients who continued therapy at cumulative doses greater than 500 mg/m², before each subsequent dose. Despite the careful monitoring, nine patients on the doxorubicin arm, compared with two patients on TLC D-99, developed clinical CHF. Although many physicians might not continue doxorubicin in patients above a lifetime cumulative dose of 450–500 mg/m², it was thought that continuing therapy could be of benefit to patients if they were responding to treatment. Numerous studies have demonstrated that continuation of therapy in patients with stable or responding disease prolongs the time to disease progression.22, 23

There was clear evidence of a cardioprotective effect of liposomal encapsulation. Not only was cardiac toxicity less frequent with TLC D-99, but in all cases, the onset of CHF on doxorubicin was at lower cumulative doses than the cases of CHF in the patients who had received TLC D-99. The cardioprotective effect of TLC D-99 also has been observed in two other randomized combination trials in first-line treatment of metastatic breast carcinoma. Batist et al. treated 297 patients with either TLC D-99 (60 mg/m²) or doxorubicin (60 mg/m²), both in combination with cyclophosphamide (D-99/C vs. AC).24 There was no clinical cardiotoxicity in the D-99/C group, but 5 cases of CHF were observed in 154 AC-treated patients at lifetime doses of 350–480 mg/m². In the other trial by Chan et al., 160 anthracycline-naive patients receive a maximum 8 cycles of TLC D-99 (75 mg/m²) or epirubicin (75 mg/m²), both in combination with cyclophosphamide (D-99/C vs. EC).25 Again, there was no evidence of clinical cardiotoxicity.

Doxil (Caelyx), a pegylated liposomal formulation of doxorubicin, is also believed to be less cardiotoxic than conventional doxorubicin,26 based on endomyocardial biopsies performed in patients with Kaposi sarcoma, who had received Doxil (Caelyx). However, no comparative study of this drug and doxorubicin has been reported.

Anthracycline cardiotoxicity is an issue not only when it is administered as a single-agent, but also when it is combined with other very promising agents, such as trastuzumab. In a randomized clinical trial, 281 anthracycline-naïve patients received doxorubicin (60 mg/m²) or epirubicin (75 mg/m²) plus cyclophosphamide, with or without trastuzumab, for a maximum of six cycles.11 Incidence of symptomatic or asymptomatic cardiac dysfunction was 27% (including 16% CHF) in the trastuzumab group versus 8% in the control group (including 3% CHF). One patient in each treatment group died of CHF.

The trend toward a shorter survival in the patients receiving TLC D-99 merits further discussion. It is unusual to see differences in survival in studies comparing two or more first-line treatments in women with metastatic breast carcinoma. Moreover, when survival differences have been observed, there are usually also significant differences in response rates and TTP.10, 27 An exception to this was a study conducted by Bishop and colleagues comparing single-agent paclitaxel and cyclophosphamide, methotrexate, 5-fluorouracil, and prednisone (CMFP) chemotherapy, in which there was a difference in survival favoring the paclitaxel arm without any differences in response rate or TTP. Note, however, that women randomized to CMFP never received paclitaxel, and thus the improvement in survival in that study was likely because of the finding that women in the paclitaxel arm received an additional active agent during the course of their illness.28 Although the survival trend in favor of doxorubicin in the current study cannot be ignored, it seems unlikely, given the other efficacy parameters, that this is a consequence of reduced efficacy of TLC D-99 compared with doxorubicin. It is possible that the excess of progesterone receptor positivity in the doxorubicin arm may denote a better prognostic group. Other unmeasured prognostic factors (e.g., HER-2 expression) also may have played a role in the natural history of the disease.

Although the protocol did not control for nor prospectively collect information on subsequent therapy, attempts were made to collect this information retrospectively. Unfortunately, available information was sparse. There were no records for greater than 20% of the patients, and many IRBs did not wish to share nonprotocol clinical data. For these reasons, it is not possible to analyze postprotocol treatment.

Other evidence also suggests that the use of TLC D-99 should be associated with an similar overall survival as doxorubicin. Two other randomized trials of TLC D-99 combination therapy versus doxorubicin or epirubicin containing combination therapy show similar survival times in both treatment groups. In Batist et al. the median survival time was 19 months on TLC D-99/cyclophosphamide versus 16 months on AC (log-rank P = 0.79; HR = 1.04),24 and Chan et al. show a median survival time of 18 months on TLC D-99/cyclophosphamide versus 16 months on EC (log-rank P = 0.51; HR = 1.15).25 AC and EC combinations are the most common way that these anthracyclines are administered and have proven an enduring clinical applicability.

Oncologists have learned to minimize the risk of anthracycline cardiotoxicity by keeping the total lifetime cumulative dose of doxorubicin below the recommended threshold. Trends in breast carcinoma treatment during the last decade have created a need for a less cardiotoxic doxorubicin for the treatment of metastatic breast carcinoma. For instance, with the widespread use of doxorubicin-containing regimens in the adjuvant setting, many patients who experience recurrence after adjuvant therapy with anthracycline-containing regimens may not be able to receive further doxorubicin treatment, even though they might benefit from it. In addition, whereas chest wall radiation has been shown to extend survival in some populations of patients, it may add to the risk of anthracycline cardiotoxicity. Finally, the availability of a less cardiotoxic anthracycline might allow for the concurrent administration of an anthracycline and trastuzumab. Given the role of anthracyclines in the treatment of HER-2 positive carcinoma, this approach has great promise.29, 30

TLC D-99 is an active agent in patients with metastatic breast carcinoma, showing similar activity to single-agent doxorubicin. Ongoing trials are evaluating potential roles for this agent in the management of women with breast carcinoma. Studies of TLC D-99 in combination with trastuzumab are ongoing.31