This article is mentioned in:

Introduction

Bladder cancer is a significant cause of morbidity and mortality in the UK with ~10,000 new cases diagnosed and 5,000 deaths registered each year (1). The most common pathological subtype is transitional cell carcinoma (TCCU) (90% of cases) (2) with ~50% of all cases being stage II or beyond at the time of diagnosis (1). Standard of care for localized, muscle invasive TCCU is radical cystectomy, which historically, i.e., without modern neoadjuvant combination chemotherapy, leads to 5-year overall survival in the approximate range of 50–60% (3).

A meta-analysis including 11 trials and 3,005 patients confirmed the value of preoperative cisplatin based combination chemotherapy which was shown to reduce the risk of relapsing disease and conferring a 5-year absolute overall survival advantage of 5% (4). A recent meta-analysis comprising 945 patients and nine clinical trials similarly demonstrated the benefits of postoperative cisplatin based combination chemotherapy (5). There is thus significant evidence supporting the use of chemotherapy in localized disease. However, not all patients are eligible for curative-intent multimodal therapy, and about one in five patients present with stage IV disease (i.e., locally extending to other organs, to the pelvic or abdominal wall, and/or with evidence of distant metastases) at the time of diagnosis (1).

Patients with relapse following primary treatment, or with advanced disease at presentation, confer a significant challenge, and even among those fit for optimal platinum-based combination chemotherapy the median overall survival does not exceed the range of 12–15 months (6,7). The recommended first line chemotherapy for these patients are cisplatin based combinations and either MVAC (methotrexate, vinblastine, doxorubicin, and cisplatin) or GC (gemcitabine and cisplatin) (6,7) although the GC regimen is often preferred due to a milder toxicity profile (7). For patients with acceptable performance status and preserved organ functions, and where the relapse occurs later than 12 months following neoadjuvant/adjuvant cisplatin-based combination chemotherapy, change of platinum based regimen may be a feasible option (8). In selected cases the addition of paclitaxel to gemcitabine and cisplatin may be considered (9). For patients unfit for cisplatin combinations alternative although potentially less efficient combination regimens have been proposed, either with alternative platinum agents [oxaliplatin (10) or carboplatin (11)] or a platinum-free combination of paclitaxel and gemcitabine (12). In patients deemed ineligible for standard cisplatin based treatment, combination treatment with split dose cisplatin and gemcitabine has reported encouraging results (13).

Following failure of first line chemotherapy, be it early relapse following platinum based neoadjuvant/adjuvant chemotherapy, or progressive disease during palliative first-line chemotherapy, treatment options have so far been limited. Studies, mostly phase II and retrospective series, have reported activity with taxanes and pemetrexed (14). Until recently there were no randomized studies to confirm the benefit of second line chemotherapy for patients with metastatic TCCU. Vinflunine (Javlor®), a microtubule inhibitor of the vinca-alkaloid family of anticancer agents (15), is the first drug to obtain European Medicine Agency (EMA) approval for use in TCCU (2009) due to evidence of efficacy from phase II (16,17) and phase III trials (18,19). Considering the multiple challenges in the second-line setting, with declining performance status due to progressive disease, persistent side effects or complications from earlier treatments, and primary or acquired chemo resistance after primary chemotherapy, the safety profile and efficacy data from the recent vinflunine reports are encouraging. In the phase III trial (18,19) median overall survival was 6.9 months in the vinflunine plus best supportive care compared to 4.3 months in the best supportive care only population.

Further empirical studies have confirmed vinflunine to be a safe and effective second line approach in Spain [n=66 (20)], France [n=134 (21)] and Germany [n=77, Hegele (22)] with reported overall survival of 7.7–10.4 months. Based on the accumulating evidence, the ESMO guidelines now suggest vinflunine as the recommended second-line therapy in advanced bladder cancer (23).

Vinflunine as a second line therapy is not currently recommended for UK practice, nor is it available to the NHS patients through the approved list of drugs on cancer drug fund. It has however been made available through the Free of Charge Program (FOCP) sponsored by Pierre Fabre. Here we evaluate the outcome of patients treated with vinflunine as a second line therapy in this program.

Patients and methods

Data were collected retrospectively on patients with advanced metastatic TCCU diagnosed between 6th June 1999 and 20th June 2013. Data were collected via a pre-defined CRF adopted for the study and sent to local investigators for population. All patients who received vinflunine as a second line therapy following failure of first line therapy were eligible for inclusion.

All patients received vinflunine as a second-line therapy through the free of charge program (FOCP) and received at least one dose of vinflunine. All patients were included in the analysis, irrespective of any dose reductions or toxicities. The dose of vinflunine was 320/280/250 mg/m2 every 3 weeks as per the SPc. The aim was to document the toxicity, radiological RR and OS for patients treated with vinflunine in real life setting within the FOCP.

Demographic data were collected on patient gender, age, height, weight and performance status as well as the site of the disease recurrence and their hemoglobin level at the time that the decision to prescribe vinflunine was made. Details on the type of first-line therapy were requested as was whether this was in the neo-adjuvant or adjuvant setting.

Continuous data are presented as medians (IQR) and categorical data are expressed as frequencies of counts. The primary outcome measure of interest is overall survival (OS) which is measured as the time from intention to treat with vinflunine until death by any cause. Survival estimates obtained via method of Kaplan and Meier. Progression-free survival (PFS) is measured as the time from first vinflunine administration until progression or death by any cause. Objective response rate is defined using RECIST criteria (version 1.1) and determined at each local site. Univariate analyses are carried out to assess patient demographics as prognostic factors for overall survival using Cox proportional hazards models. Hazard ratios (HR) are presented with associated 95% confidence intervals (CI). No multivariate analyses were carried out. All P-values are considered significant at the 5% level. All analysis were carried out using R (version 3.1.2).

Reported toxicities are defined using CTCAE (version 4) and are reported based on being either grade 1–2 or 3–4, respectively.

Results

Data are provided on 49 patients from nine contributing sites throughout the UK and Ireland (CCC 15; Christie 5; Cork 6; Glasgow 9; St. George's 1; Royal Free London 2; Royal Marsden 6; Northampton 1; Newcastle 4) diagnosed with advanced metastatic TCCU between 6th June 1999 and 20th June 2013. All patients have progression of disease after first-line platinum based therapy. Patient demographics are provided in Table I. The patient group consists of 67% (33/49) males and have a median (IQR) age of 64 (57–70). As per the summary of product characteristics (SPC) for vinflunine, patients with a 0/1 performance status (ECOG) or a hemoglobin (Hb) level >10 g/dl were targeted. Thirteen patients (27%) had an ECOG performance status of 2 or a hemoglobin level <10. Sixteen patients (33%) had either bone or liver recurrence. Two patients (4%) had chronic constipation at presentation and two patients (4%) had coronary artery disease.

Table I Patient characteristics.

Table I

Patient characteristics.

Category	Level	n=49
Gender	Female	16 (33%)
	Male	33 (67%)
Weight, kg	Median (IQR)	71 (63–83)
Height, cm	Median (IQR)	168 (160–175)
Age, years	Median (IQR)	64 (57–70)
Performance status (n=46)	0	12 (26%)
	1	30 (65%)
	2	4 (9%)
Site of disease recurrence (n=46)	Bone	8 (17%)
	Liver	11 (24%)
	Lung	15 (33%)
	Other visceral	8 (17%)
	Other non-visceral	3 (7%)
	None	1 (2%)
Hemoglobin (g/dl)	<10	11 (28%)
	10–12	20 (50%)
	>12	9 (22%)

Information on first-line therapy is available for 48/49 (98%) of all participating patients. First line neo-adjuvant chemotherapy was given to 60% (27/48) and consisted of gemcitabine/cisplatin (74%; 20/27), gemcitabine/carboplatin (22%, 6/27), or gemcitabine/cisplatin switched to carboplatin in view of worsening renal functions (4%, 1/27). In the metastatic setting patients received platinum regimens consisting of either gemcitabine/cisplatin (30%, 13/43), gemcitabine/carboplatin (37%, 16/43), gemcitabine/cisplatin switched to carboplatin (2%, 1/43), or some other combination (30%, 13/43).

Following first-line therapy, patient best response rates were complete response 15% (7/48), partial response 48% (23/48), stable disease 17% (8/48) and progressive disease 19% (9/48). Best response was not available for one patient. The overall response rate for first-line therapy was 63% (30/48) and 98% (43/44) of patients demonstrated disease progression with 77% (34/44) demonstrating bone, liver, or lung metastases.

Table II gives the details of vinflunine administrations. At the point of data collection 248 cycles of vinflunine had been administered. The starting dose was 250 mg/m2 (4%, 2/49), 280 mg/m2 (76%, 37/49) or 320 mg/m2 (20%, 10/49). Three patients (6%) had a dose escalation and 10 (21%) had dose reduction. The median (IQR) number of cycles given was 3.5 (2–6.25). Eight patients (17%) had only a single cycle, while the maximum number of cycles at the time of analysis was 18.

Table II Vinflunine administration.

Table II

Vinflunine administration.

Category	Level
Starting dose (mg/m2)	250	2 (4%)
	280	37 (75%)
	320	10 (21%)
Escalation (n=47)	No	44 (94%)
	yes	3 (6%)
Reduction (n=47)	No	37 (79%)
	Yes	10 (21%)
Cycles	Median (IQR)	3.5 (2–6.25)

Toxicity

Toxicities are reported by sites as being either grade 1/2 or grade 3/4 hematological or non-hematological. Results are given in Table III. Overall 31% (15/49) of patients reported at least one grade 3/4 adverse event. The most common grade 3/4 toxicities are constipation reported by 8% (4/49) and asthenia/fatigue reported by 6% (3/49) of patients.

Table III Reported hematological and non-hematological toxicities.

Table III

Reported hematological and non-hematological toxicities.

Toxicity	Grade
	1–2	3–4
Hematological
Anemia	5	2
Neutropenia	13	1
Neutropenic infection	0	2
Thrombocytopenia	2	0
Febrile infection	1	0
Leukopenia	1	1
Other	0	2
Non-hematological
Constipation	9	4
Asthenia/fatigue	20	3
Vomiting	8	0
Abdominal pain	2	0
Other	0	2

Overall survival

Patients were followed up for a median of 9.1 months. At the time of analysis, 41 (84%) patients had died. Overall survival estimates obtained via the method of Kaplan-Meier (Fig. 1) give a median (95% confidence interval) of 9.1 months (6.0–12.7). ECOG performance measures were investigated further as a key prognostic indicator of interest (Fig. 2). Median survival estimates of 13.1 (6.0, NA) months and 7.6 (5.98–10.5) months were observed for ECOG performance measures of 0 and 1–2, respectively. This was not significant statistically [HR (95% CI): 1.62 (0.74, 3.55); P=0.23].

Figure 1 Kaplan-Meier plot of overall survival.

Figure 2 Kaplan-Meier plot for progression-free survival.

Disease progression

A total of 45 (92%) of patients had disease progression at the time of analysis. Median progression-free survival estimates (Fig. 2) are 5.1 months (4.3–8.7).

Patient response

Forty-eight patients were evaluable for overall response (Table IV). One patient did not provide any response data.

Table IV Overall best response for vinflunine administration.

Table IV

Overall best response for vinflunine administration.

Best response
CR	0
PR	14 (29%)
SD	10 (21%)
PD	19 (40%)
NE	4 (8%)

Sensitivity analyses report the result only on patients who meet the SPC (ECOG <2, HB >10). Here 35 patients gave an estimated OS of 9.9 months (7.6–16.4) with an estimated PFS of 6.0 months (4.8–12.1) and a partial response rate of 30%.

Discussion

The prognosis for patients with advanced metastatic TCCU who have progressed after first-line therapy remains dismal, and whilst previously published randomised trials and empirical evidence have favoured the use of vinflunine in this patient group, the results have not been clear enough to favour its wide spread use in the UK. As a consequence, many patients will end antitumour therapy or proceed onto palliative care once the first-line regimen has failed.

We performed a retrospective observational study in this clinical setting to assess the effects of vinflunine in terms of both efficacy and toxicity. Data on 49 patients from sites across the UK and Ireland were collected and analysed. The dose of and number of cycles were consistent with previous studies, with a median of 3.5 cycles being similar to the number recorded in the phase III trial (18,19). The most common starting dose was 280 g/m2 (75%) and the dose was relatively well tolerated with only 10 (29%) patients requiring a reduction. There was no evidence of any correlation between the starting dose and dose reduction.

Further to this, the present median overall survival of 9.1 months (6.0–12.7) is consistent with or even slightly better than the 6.9 months of the vinflunine treated arm of the Bellmunt phase III study (18,19). Further, the present numbers compare favourably with the empirical 'real life' results obtained from Spain (16) Germany (22) and France (21) which all report median survivals between 7.7–10.4 months in similar patient populations.

Toxicity rates observed in our study also show little cause for concern with 39% (19/49) patients reporting a grade 3+ adverse event, the most common being constipation (4 patients) and fatigue (3 patients). Toxicity can be further improved with the use of prophylactic laxative and oral antibiotics and should provide little concern in the prescribing of vinflunine.

Further inspection of study efficacy shows that patients with a baseline ECOG PS of 0 display a median OS of 13.1 (6.0, NA) months, while patients of PS 1–2 have a poorer prognosis of 7.6 (6.0–10.5) months. This is consistent with the data of ref. 20 reporting survival estimates of 13.2 and 6.7 months, respectively, for the same patient groups and emphasise baseline PS as a key prognostic predictor although not ruling out the potential benefits even in the PS 1–2 population.

This study is limited by the retrospective nature of the data collection and the lack of a randomised comparative arm against which to compare the efficacy, which ultimately open the study results to possible bias. In saying that, however, the study analysis includes patients who did not meet the SPC for vinflunine as well as patients who only received a single dose. Coupled with the high rate of ECOG >0 and the distribution of metastases this suggests a patient population which is representative of general clinical practice rather than a narrowly selected clinical trial population. Recently there has been much interest in the role of immune checkpoint inhibitors in urothelial cancer with durable responses have been noted with PD1 [e.g., pembrolizumab (24) and PDL1 atezolizumab (25)]. The results of recently completed trials comparing anti-PD1/PDL1 antibodies, with standard of care chemotherapy of institution's choice that includes vinflunine in the 2nd line setting (NCT 02256436) are awaited with interest.

In conclusion, the empirical results obtained from UK and Ireland practice reflect what has been observed in randomised trials of vinflunine as well as other retrospective studies throughout the European Union. The accumulating evidence puts forward vinflunine single drug therapy as an efficient, safe and reasonably tolerable second-line therapy in patients with advanced TCCU.

References