Docetaxel versus docetaxel alternating with gemcitabine as treatments of advanced breast cancer: final analysis of a randomised trial.

BACKGROUND: Alternating administration of docetaxel and gemcitabine might result in improved time-to-treatment failure (TTF) and fewer adverse events compared with single-agent docetaxel as treatment of advanced breast cancer. PATIENTS AND METHODS: Women diagnosed with advanced breast cancer were randomly allocated to receive 3-weekly docetaxel (group D) or 3-weekly docetaxel alternating with 3-weekly gemcitabine (group D/G) until treatment failure as first-line chemotherapy. The primary end point was TTF. RESULTS: Two hundred and thirty-seven subjects were assigned to treatment (group D, 115; group D/G, 122). The median TTF was 5.6 and 6.2 months in groups D and D/G, respectively (hazard ratio 0.85, 95% confidence interval 0.63-1.16; P = 0.31). There was no significant difference in time-to-disease progression, survival, and response rate between the groups. When adverse events were evaluated for the worst toxicity encountered during treatment, there was little difference between the groups, but when they were assessed per cycle, alternating treatment was associated with fewer severe (grade 3 or 4) adverse effects (P = 0.013), and the difference was highly significant for cycles when gemcitabine was administered in group D/G (P < 0.001). CONCLUSION: The alternating regimen was associated with a similar TTF as single-agent docetaxel but with fewer adverse effects during gemcitabine cycles.

Bleomycin, vincristine, lomustine and dacarbazine (BOLD) in combination with recombinant interferon alpha-2b for metastatic uveal melanoma.

This EORTC multicentre study analysed the efficacy and tolerability in patients with metastatic uveal melanoma of BOLD chemotherapy in combination with recombinant interferon alpha-2b. The dose of bleomycin was 15 mg on days 2 and 5, of vincristine 1 mg/m(2) on days 1 and 4, of lomustine 80 mg on day 1, and of dacarbazine (DTIC) 200 mg/m(2) on days 1-5, given every 4 weeks for a minimum of two cycles. Subcutaneous (s.c.) interferon alpha-2b at a dose of 3 x 10(6) IU was initiated on day 8 of the first cycle, and continued at a dose of 6 x 10(6) IU three times per week after 6 weeks. A median of two cycles were administered to 24 patients (median age 60.5 years). None achieved an objective response (0%; 95% Confidence Interval (CI): 0-14), 2 (8.3%) remained stable, 20 showed progression, and 2 (8.3%) were invaluable. The median progression-free survival was 1.9 months (95% CI: 1.8-3.4) and overall survival 10.6 months (95% CI: 6.9-16.4). Overall survival improved with increasingly favourable pretreatment characteristics (median, 14.7 versus 6.9 versus 6.0 months for Helsinki University Central Hospital (HUCH) Working Formulation stages IVBa, IVBb and IVBc, respectively; P=0.018). Grade 3 alopecia and neurotoxicity occurred in 13% of the patients. This multicentre study did not confirm earlier reports that BOLD with human leucocyte or recombinant interferon would induce at least 15% objective responses in metastatic uveal melanoma.

Serum adhesion molecules and interleukin-2 receptor as markers of tumour load and prognosis in advanced cutaneous melanoma.

Cell adhesion molecules are cell surface glycoproteins that may act as mediators in the metastatic process. Soluble interleukin-2 receptor (sIL-2R) is an immunological marker that may also serve as an indicator of tumour progression. Normal and neoplastic cells are capable of releasing these molecules into circulation. We evaluated the association between pretreatment serum levels of soluble intercellular adhesion molecule 1 (sICAM-1), vascular cell adhesion molecule 1 (sVCAM-1) and sIL-2R and metastases and survival in 50 patients with advanced melanoma. The patients with liver and/or bone metastases had significantly higher sICAM-1 levels than those with soft tissue and/or lung involvement (P=0.002). In addition, there was a trend towards higher sIL-2R levels in patients with more metastatic sites (P=0.07). In univariate analysis, the number of metastatic sites (P=0.0001, odds ratio (OR) 3.0, 95% confidence interval (CI): 1.7-5.3), the metastatic site (P=0.01, OR 2.3, 95% CI: 1.2-4.4) and the levels of sICAM-1 (P=0.011, OR 2.5, 95% CI: 1.2-5.0), sVCAM-1 (P=0.036, OR 2.1, 95% CI: 1.0-4.3) and sIL-2R (P=0.0016, OR 3.0, 95% CI: 1.5-6.0) were found to be statistically significant prognostic factors for survival. In multivariate analysis, the number of metastatic sites was the dominant prognostic indicator. After it was excluded from the analysis, the sIL-2R level and the metastatic site were found to be significant. It can be concluded, that high sICAM-1 levels suggest liver metastases and sIL-2R seems to serve as a marker of tumour load in metastatic melanoma. Furthermore, the sIL-2R level appears to add to clinical data predicting the patient's outcome.

High collagenase-1 expression correlates with a favourable chemoimmunotherapy response in human metastatic melanoma.

Matrix metalloproteinases (MMPs) are proteolytic enzymes that can degrade extracellular matrix and thus enhance metastasis. We have studied the expression of two collagenolytic MMPs in 37 samples obtained from 26 patients treated for metastatic melanoma. Interestingly, the samples showed a different expression pattern of collagenase-1 (MMP-1) and collagenase-3 (MMP-13). The samples with high expression levels of MMP-1 (n = 18) were more frequently MMP-13 negative (14 out of 18), whereas those with low expression levels of MMP-1 (n = 15) were predominantly positive for MMP-13 (nine out of 15) (P = 0.027). High expression levels of MMP-1 were associated with a favourable response to chemoimmunotherapy. Responders (n = 13) frequently had intensively MMP-1-expressing metastases (nine out of 13), especially those who achieved a complete response (five out of six). Response failures (n = 7) mainly had metastases with a low intensity of MMP-1 expression (six out of seven) (P = 0.019). There was a tendency towards longer survival among patients with intensively MMP-1-expressing tumours (median 14.3 versus 6.7 months, P = 0.068). The high expression levels of MMP-1 correlated with low MIB-1 (to nuclear antigen Ki-67) (P = 0.019) and positivity for MMP-13 was associated with high MIB-1 expression (P = 0.00048), suggesting that their different expression patterns may affect tumour growth and contribute to differences in patient survival.

Tumor doubling times in metastatic malignant melanoma of the uvea: tumor progression before and after treatment.

OBJECTIVE: To obtain estimates of growth rate of metastatic uveal melanoma to infer appropriate follow-up programs and to assess the impact of current chemoimmunotherapy regimens. DESIGN: Retrospective case series. PARTICIPANTS: Of 70 consecutive patients diagnosed with metastatic uveal melanoma from 1986 through 1998, 37 patients who attended regular follow-up and had measurable metastases were eligible for this study. METHODS: Tumor doubling time (DT) was calculated by the Schwartz formula using three presumed sizes of metastasis at last negative follow-up. DT was compared according to tumor characteristics, and time of micrometastasis was estimated. MAIN OUTCOME MEASURES: Doubling time of untreated and treated metastases. RESULTS: Doubling time of untreated metastases ranged from 34 to 220 days (median, 63 days). Regardless of the presumed size of metastasis at last screening, two thirds of the metastases had a DT between 30 and 80 days. No significant correlation between DT and the observed disease-free interval was detected. Assuming constant growth rate, most metastases had predictably initiated within 5 years before primary treatment. Mean DT during active treatment of metastases in 18 patients who did not show an objective response ranged from 25 to 2619 days (median, 255 days). CONCLUSIONS: Based on the estimated growth rates, a rational follow-up interval to detect metastatic uveal melanoma would be 4 to 6 months. Primary uveal melanomas that develop clinically detectable metastasis after conservative therapy may micrometastasize several years before treatment. These estimates are rough and must be confirmed by prospective studies. Current chemoimmunotherapy regimens slow down the growth rate of metastases even if objective response is not obtained.

The value of serum S-100beta and interleukins as tumour markers in advanced melanoma.

Recently serum S-100beta has shown promise as a tumour marker in melanoma; however, its use as a prognostic marker in the advanced stage needs to be confirmed. Interleukins (ILs) may mediate regression or progression of cancer. In order to study their relation to the metastatic profile and survival, we evaluated the association between pretreatment serum levels of S-100beta, IL-6, IL-10 and IL-12 and metastatic site and survival in 50 patients with advanced melanoma who were to receive chemoimmunotherapy. Patients with liver and/or bone metastases had significantly higher median concentrations of S-100beta, IL-6 and IL-10 than those with only skin, nodal and/or lung involvement. The differences in IL-12 levels were unremarkable. Using univariate analysis, the S-100beta level and metastatic profile were found to be statistically significant prognostic factors for survival. Using multivariate analysis the S-100beta level was the most powerful prognostic indicator, while the metastatic profile was found to be significant after exclusion of S-100beta. The findings suggest that elevated serum levels of S-100beta, IL-6 and IL-10 reflect concurrent liver or bone metastases in melanoma. S-100beta is also an independent prognostic marker. Pretreatment IL levels were not associated with outcome.

Prognostic value of beta1 integrin expression in metastatic melanoma.

The expression of integrin-type cell adhesion receptors is frequently changed in malignant transformation. Despite their important role in cancer cell behaviour, the value of integrins as prognostic markers is mostly unknown. We have examined the expression of beta1 integrins in 38 metastatic melanomas obtained from 27 patients treated with combined chemoimmunotherapy. On the basis of beta1 integrin expression, the melanoma samples were divided into two groups: beta1-negative tumours (<10% beta1 integrin immunostained cells) and beta1-positive tumours (with > or = 10% positive cells). Patients with beta1-positive tumours (n = 15) had significantly longer disease-free survival (median 38 versus 7 months, P < 0.0001) and overall survival (median 70 versus 23 months, P = 0.0001) evaluated after the diagnosis of primary disease compared with patients with beta1-negative metastases (n = 11). Moreover, the survival of the patients with beta1-positive tumours after the initiation of chemoimmunotherapy was significantly prolonged (median 18 versus 9 months, P = 0.017). The independent nature of beta1 integrin expression as a significant prognostic factor for survival after therapy was confirmed using Cox's multivariate analysis (P = 0.014). Our results indicate that the expression of beta1 integrins might have some major tumour growth regulatory role and can be used as a predictor for prognosis in patients with metastatic melanoma.

Analysis of G(1)/S checkpoint regulators in metastatic melanoma.

We have analyzed the expression of the CDKN1A (p21(CIP1)), CDKN1B (p27(Kip1)), TP53, RB1 and MDM2 proteins and tumor cell proliferation by immunohistochemical staining in 59 cases of metastatic melanoma. The genomic status of the CDKN2A (INK4-ARF, p16/p14(ARF)), CDKN2B (p15) and CDKN2C (p18) genes was determined by PCR-SSCP (single-strand conformation polymorphism) in 46 of these cases. These results were correlated with various clinico-pathological parameters, including the outcome of combined chemoimmunotherapy. We found positive correlations between the expression of CDKN1A and MDM2 (r = 0.5063, P = 0.001), between the expression of CDKN1B and RB1 (r = 0.5026, P = 0.001), and between RB1 expression and tumor cell proliferation (0.5564, P<0.001). Two mutations in the CDKN2A (p16) gene were detected, including a novel base change AAC-->ATC (Asn to Ile) at codon 71, that also changes the codon 85 of the alternative reading frame gene p14(ARF) from CAA to CAT (Gln to His). Homozygous deletion at exon 2 of the CDKN2A (INK4-ARF) gene was detected in six cases. In seven cases, the 540C-->G polymorphism in the 3'UTR of the CDKN2A (p16) gene was found in linkage disequilibrium with the 74C-->A polymorphism in intron 1 of the CDKN2B gene (P < 0.0001). These cases had significantly lower expression of the TP53 protein (P = 0.0032). Both 540C-->G and 580C-->T polymorphisms in the 3'UTR of the CDKN2A (p16) gene were associated with significantly shorter progression time from primary to metastatic disease (P = 0.0071). We conclude, that although none of the analyzed cell cycle regulators could be singled out as a major prognostic factor, G(1)/S checkpoint abnormalities remain one of the most significant factors in the development of malignant melanoma.

Serum matrix metalloproteinase-2 as a prognostic marker in advanced cutaneous melanoma.

The association was studied between serum concentration of matrix metalloproteinase-2 (MMP-2) and metastatic site, survival and disease progression in patients with advanced cutaneous melanoma. The patient population consisted of 50 patients who were treated with chemoimmunotherapy. The median baseline serum concentration of MMP-2 was 724 ng/ml (range 500-2,297 ng/ml). There were no significant differences in MMP-2 levels according to metastatic site. Baseline MMP-2 concentration did not have a prognostic value. The patients with levels below 800 ng/ml survived for 8.8 months and those with higher levels for 9.7 months. On serial measurements, median serum MMP-2 concentration at disease progression in 25 patients was significantly higher than before treatment. Only five samples at response were available, and the levels were not significantly different from baseline levels. In conclusion, serum MMP-2 is not a prognostic marker in advanced melanoma. It also appears to be of limited clinical value in monitoring.

MIB-1 immunoreactivity correlates with blood vessel density and survival in disseminated malignant melanoma.

The purpose of our current work was to evaluate the prognostic significance of tumor cell proliferation in advanced metastatic melanomas and to investigate a possible correlation between the proliferation index and blood vessel density in metastatic tissue. Sixty patients with disseminated malignant melanoma treated with four-drug chemotherapy combined with interferon-alpha were included in this study. Proliferative activity and vascularity in metastatic tissues were examined by immunohistochemistry with anti-Ki-67 (MIB-1) and anti-CD31 antibody, respectively. A significant relationship between MIB-1 index and blood vessel number was detected (rho = 0.323, p = 0.013). In survival analysis, the overall survival and disease-free survival were significantly longer (58 and 38 vs. 38 and 17 months) for patients with low MIB-1 immunoreactivity (p = 0.012 and p = 0.023, respectively). Likewise, the low MIB-1 labeling index was associated with the prolonged survival calculated from the initiation of the chemoimmunotherapy (12 vs. 7 months, p = 0.032). In multivariate Cox's proportional hazard analysis, MIB-1 positivity was an independent prognostic factor both for overall survival and for survival after beginning of the chemoimmunotherapy (p = 0.016 and p = 0.029).

Prognostic value of tumour vascularity in metastatic melanoma and association of blood vessel density with vascular endothelial growth factor expression.

Tumour angiogenesis is essential for tumour growth and metastasis. Several lines of evidence indicate that vascular endothelial growth factor (VEGF) is a major regulator both of physiological and pathological angiogenesis. In this study we assessed the blood vessel density and VEGF expression of 94 melanoma metastases of 70 patients by immunohistochemistry, utilizing antibodies against human platelet endothelial cell adhesion molecule-1 (PECAM-1, CD31) and VEGF. The number of blood vessels ranged from 4 to 131 vessels/high power field (HPF), with a mean value of 32 vessels/HPF (+/-21) and a median of 29 vessels/ HPF. Survival since diagnosis of the primary disease and from the start of chemoimmunotherapy, as well as the disease-free survival period, was significantly shorter in the high vascularity group of patients compared with the low vascularity group (P< 0.05 and P< 0.01, respectively). A high overall expression of VEGF in the metastatic melanoma samples was observed. The degree of VEGF expression appeared to have a strong association with the blood vessel density (P= 0.017). This study demonstrates the clinical role of tumour vascularity in the prognosis of patients with metastatic melanoma. In addition, the strong association between vascularity and VEGF expression suggests a crucial role for this growth factor in the neovascularization of metastatic melanoma.

DNA flow cytometry and the outcome of chemoimmunotherapy in metastatic melanoma.

DNA ploidy and S-phase fraction (SPF) were measured by flow cytometry on tumours from 80 patients with disseminated malignant melanoma. All patients received a four-drug chemotherapy regimen (dacarbazine, vincristine, bleomycin and lomustine) plus interferon. Specimens were taken for analysis from the latest metastases biopsied before the start of the treatment. In 13 patients we also analysed sequential samples taken after the treatment at the time of progression. DNA aneuploidy was observed in 65% of the patients. Among the 40 responders there were 26 with aneuploid tumours (65%). Aneuploidy did not reach statistical significance as a prognostic sign in the unstratified study population, but when stratified by response groups, DNA aneuploidy was a significant prognostic factor for a better response (P = 0.04). SPF could be calculated in 76 tumours. Out of 40 responding patients (complete or partial response), 23 had tumours with an SPF higher than the median. Accordingly, patients with a high SPF survived longer than those with a low SPF, with median survival times of 9.8 months and 8.7 months, respectively (P = 0.18). We conclude that DNA aneuploidy and a high SPF are associated with longer survival in patients with disseminated melanoma treated with a chemoimmunotherapy regimen. Based on our findings we claim that, among patients receiving chemoimmunotherapy, high SPF and aneuploidy are not signs of unfavourable prognosis, which is in contrast to previous observations in melanoma patients receiving heterogeneous therapy.

Vascularity and prognosis of metastatic melanoma.

The clinical role of vascularity was examined in metastatic melanoma, analyzing the correlation of the blood vessel density and prognosis. Our study included 51 specimens of metastatic melanoma tissue samples from 31 patients treated with combined chemo-immunotherapy. PECAM-1 (CD31) was used for assessing vascularity by immunohistochemical staining. On the basis of blood vessel counts, patients were classified into 2 main groups: low and high vascularity. A higher blood vessel density was found to be associated with shorter survival, estimated from the primary diagnosis of the disease (38 months), compared with patients with low blood vessel counts (68 months). A similar tendency was observed when vascularity was correlated to the survival period after the detection of the first metastases (13 vs. 30 months) and with survival since the initiation of chemo-immunotherapy (8 vs. 16 months). When vascularity and some common prognostic factors, such as age, sex, DNA ploidy and WHO tumor response, were used for a Cox multivariate analysis, vascularity turned out to be the most significant independent prognostic factor. Our results suggest that counting the blood vessels identified by immunohistochemical staining for the endothelial cell-specific CD31 is a powerful predictor for prognosis in patients with metastatic melanoma and should be considered when selecting patients for therapy.

Chemosensitivity of human melanoma metastases in mouse subrenal capsule assay--can it predict tumour response to combined cytostatic plus interferon therapy in metastatic melanoma?

We have analysed the chemosensitivity of 121 human melanoma metastases in the 6-day mouse subrenal capsule assay (SRCA). A total of 13 different chemotherapy regimens were analysed in four successive series. By the original criteria of the assay, 75% of tumours were sensitive to at least one regimen. The most effective combinations, giving sensitivity rates > 40%, were of dacarbazine, vincristine, and BCNU (DOB) with or without metronidazole, followed by cisplatin plus etoposide (Plat-VP16) combined with interferon. On the basis of the SRCA screening we developed a four-drug chemotherapy regimen combined with interferon for metastatic melanoma. Retrospectively analysed the assay correctly identified 77% of clinical responders and 54% of clinical non-responders. The overall predictive accuracy was 65%. Despite the limitations of the assay, it can be used for screening new drugs or combinations.

The CD4+/CD8+ ratio as a prognostic factor in patients with metastatic melanoma receiving chemoimmunotherapy.

PURPOSE: As reported earlier, a chemotherapy regimen that consisted of dacarbazine, vincristine, lomustine, and bleomycin (DOBC) combined with natural leukocyte interferon (IFN) has been administered with favorable results to patients with metastatic melanoma. In this study, lymphocyte subsets (CD4+ and CD8+) were analyzed before and during treatment to elucidate if alterations in the CD4+/CD8+ ratio had any prognostic value. MATERIALS AND METHODS: Blood samples were systematically obtained from 54 patients with metastatic melanoma who received this chemoimmunotherapy. The frequencies of peripheral-blood lymphocyte subsets were monitored by flow cytometry using the monoclonal antibodies OKT4 (CD4+, T-helper cells) and OKT8 (CD8+, T-suppressor cells). RESULTS: Twenty-seven patients had a constantly increasing ratio, while the remaining 27 patients had a fluctuating or constantly decreasing ratio. The former group had a median survival time of 11.8 months, as compared with 6.5 months for the latter (P = .008, log-rank test). This difference was generated among patients who had an objective response. Responding patients with a constantly increasing ratio had a median survival time of 21.7 months, as compared with 10.2 months for patients with no constant increase in the ratio (P = .038, log-rank test). In nonresponders, no difference in survival was observed between the two groups. CONCLUSION: The monitoring of early changes in the CD4+/CD8+ ratio can provide valuable information that predicts the prognosis of metastatic melanoma patients receiving chemoimmunotherapy.

Response of subcutaneous and cutaneous metastases of malignant melanoma to combined cytostatic plus interferon therapy.

A chemotherapy regimen consisting of dacarbazine (DTIC), vincristine, bleomycin and lomustine (CCNU) was combined with natural leucocyte interferon (IFN) in the treatment of 37 patients with cutaneous and subcutaneous metastases of malignant melanoma. Twenty-five also had concomitant lymph node, visceral, bone or brain metastases. Fifteen patients (41%) experienced complete response (CR) and 10 (27%) partial response (PR) of the superficial lesions. In addition, three patients were rendered surgically tumour-free after PR or disease stabilization during drug therapy. The median overall duration of response was 10.2 months (range 1-53 months). In disseminated disease, the combined therapy produced favourable results, particularly with regard to superficial lesions. It is also possible that this therapy may retard the growth of aggressive subcutaneous metastases in patients who have previously had multiple surgical procedures in an attempt to stabilize their disease. In a small proportion of patients, the long-term complete remission with the drug therapy alone, or in combination with surgery, suggests that this regimen might have been curative.

Decreasing CD4/CD8 ratio during prolonged four-drug chemotherapy plus interferon treatment for metastatic melanoma.

A chemotherapy regimen consisting of dacarbazine, vincristine, bleomycin and lomustine in combination with natural leukocyte interferon (IFN)-alpha was administered to 48 metastatic melanoma patients with very favorable results. The frequencies of peripheral blood lymphocyte subsets during this treatment were monitored by flow cytometry in order to elucidate the underlying antitumor mechanism. Analysis of 467 peripheral blood samples revealed a significant decrease during treatment in absolute values of all lymphocyte subsets (CD4, CD8, CD56) as well as in total lymphocyte count. All the values returned to normal after discontinuation of treatment. Significant alterations in proportions of lymphocyte subsets were detected. Notably, the CD4+/CD8+ ratio showed changes related to response, treatment duration, and disease progression. Exceptionally high CD4+/CD8+ ratios (up to 14) were observed soon after the start of treatment in some of the patients later achieving a complete response. During the course of treatment, in 76% of the patients receiving at least a 2-month treatment, the CD4+/CD8+ ratio decreased to less than 1.0. Peculiarly, the nadir CD4+/CD8+ ratio was consistently observed 1-6 months before clinically detectable disease progression, suggesting a relation between CD4+/CD8+ ratio and tumor control.

A promising interferon plus four-drug chemotherapy regimen for metastatic melanoma.

PURPOSE: The goal of this study was to determine the therapeutic efficacy and toxicity of a four-drug chemotherapy regimen combined with natural leukocyte interferon alfa (IFN-alpha) for metastatic melanoma. PATIENTS AND METHODS: Between December 1988 and December 1991, 48 consecutive unselected patients with metastatic melanoma were entered onto this phase II study. Forty-five of these patients were assessable for response and 46 for toxicity. The four-drug chemotherapy regimen was as follows: dacarbazine (DTIC) 200 mg/m2 days 1 to 5, vincristine 1 mg/m2 days 1 and 4, bleomycin 15 mg days 2 and 5 intravenously (IV), and lomustine (CCNU) 80 mg day 1 orally. IFN-alpha, initiated day 8, was administered 3 x 10(6) IU/d, subcutaneously (SC) for 6 weeks, followed by 6 x 10(6) IU three times per week. A small protocol modification was adopted from the 21st patient onwards whenever there was more than 2 months' stabilization or progression with the original protocol: IFN therapy was split into 2-week treatment periods interrupted by a 2-week rest period. RESULTS: Among the 45 assessable patients, the objective response rate was 62% (95% confidence limit, 48 to 77); six patients (13%) achieved a complete response (CR) and 22 (49%) a partial response (PR). Metastases in such sites as liver also responded favorably (one CR, six PR, one stable disease [SD], two progressive disease [PD]). After splitting IFN therapy for nonresponders, in two patients PD and in another two patients SD changed into regression. Three of the six patients with a CR have suffered a relapse, but the other three have been off treatment for 7, 18, and 31 months without recurrence. Most of the symptomatic patients also experienced rapid relief of symptoms. Overall toxicity of this mainly outpatient regimen seemed to be acceptable. One patient died of a septic fever with grade 4 leukopenia and thrombocytopenia. The most frequent side effects were transient fever, nausea/vomiting, fatigue, and grade I/II hematologic toxicity. CONCLUSION: Results demonstrate a remarkably high response rate in combining IFN-alpha and four chemotherapeutic agents. The apparent schedule-dependency of responses must be further explored in a controlled phase III study.