(g) Censored results at time of cross-over

Non-censored results at time of cross-over: March 31 2011: HR (95% CI) = 0.47 (0.35, 0.62); October 3 2011: HR (95% CI) = 0.67 (0.54, 0.84); February 1 2012: HR (95% CI) = 0.76 (0.63, 0.93); December 20 2012: HR (95% CI) = 0.79 (0.66, 0.95)

Figure 1: Kaplan-Meier curves of overall survival – previously untreated patients (December 20, 2012 cut-off)

Table 6 shows the treatment effect for all pre-specified stratification variables which are established as prognostic factors.

Table 6: Overall survival in previously untreated patients with BRAF V600 mutation-positive melanoma by LDH, tumour stage and ECOG status (post hoc analysis, December 20, 2012 cut-off, censored results at time of cross over)

Stratification variable
 N
 Hazard Ratio
 95% Confidence Interval
 
LDH normal
 391
 0.88
 0.67; 1.16
 
LDH >ULN
 284
 0.57
 0.44; 0.76
 
Stage IIIc/M1A/M1B
 234
 1.05
 0.73; 1.52
 
Stage MIC
 441
 0.64
 0.51; 0.81
 
ECOG PS=0
 459
 0.86
 0.67; 1.10
 
ECOG PS=1
 216
 0.58
 0.42; 0.9
 

LDH: Lactate Dehydrogenase, ECOG PS: Eastern Cooperative Oncology Group Performance Status

Table 7 shows the overall response rate and progression-free survival in previously untreated patients with BRAF V600 mutation-positive melanoma.

Table 7: Overall response rate and progression-free survival in previously untreated patients with BRAF V600 mutation-positive melanoma

  vemurafenib
 dacarbazine
 p-value (h)
 
December 30, 2010 data cut-off date (i)
 
Overall Response Rate

(95% CI)
 48.4%

(41.6%, 55.2%)
 5.5%

(2.8%, 9.3%)
 

 <0.0001
 
Progression-free survival

Hazard Ratio

(95% CI)
 

 0.26

(0.20, 0.33)
 

 <0.0001
 
Number of events (%)
 104 (38%)
 182 (66%)
  
Median PFS (months)

(95% CI)
 5.32

(4.86, 6.57)
 1.61

(1.58, 1.74)
  
February 01, 2012 data cut-off date (j)
 
Progression-free survival

Hazard Ratio

(95% CI)
 

 0.38

(0.32, 0.46)
 
 <0.0001
 
Number of events (%)
 277 (82%)
 273 (81%)
  
Median PFS (months)

(95% CI)
 6.87

(6.14, 6.97)
 1.64

(1.58, 2.07)

(h) Unstratified log-rank test for PFS and Chi-squared test for Overall Response Rate.

(i) As of December 30, 2010, a total of 549 patients were eva luable for PFS and 439 patients were eva luable for overall response rate.

(j) As of February 01, 2012, a total of 675 patients were eva luable for the post-hoc analysis update of PFS.

A total of 57 patients out of 673 whose tumours were analysed retrospectively by sequencing were reported to have BRAF V600K mutation-positive melanoma in NO25026. Although limited by the low number of patients, efficacy analyses among these patients with V600K-positive tumours suggested similar treatment benefit of vemurafenib in te