Page 6 - Canadian Urological Association/Genitourinary Medical Oncologists of Canada consensus statement: Management of unresectable locally advanced and metastatic urothelial carcinoma

CUAJ – Consensus Statement Warren et al
Unresectable locally advanced and metastatic urothelial carcinoma

rates, their greater toxicity combined with failure to impact meaningfully on OS has limited their
33
use. Despite low response rates of approximately 10%, paclitaxel and docetaxel are well
tolerated and have emerged as preferred regimens. 34–36 NAB-paclitaxel has demonstrated
37
efficacy in the second line setting and is a reasonable alternative.
2. Immunotherapy: Checkpoint inhibitors
The evidence supporting the use of checkpoint inhibitors in the second line setting is more robust
than in the first line setting. Five checkpoint inhibitors have been tested in this setting including
2 in randomized phase III studies.
Pembrolizumab was compared to investigators’ choice of single agent paclitaxel, docetaxel or
vinflunine in 542 patients with progressive disease during or after platinum based chemotherapy
in the randomized phase III Keynote 045 study. Overall response rate was 21% in the
pembrolizumab-treated patients compared with 11% in chemotherapy-treated patients. Median
PFS was only 2.1 months on pembrolizumab compared with 3.3 months on chemotherapy (HR
0.98, 95% CI 0.81 - 1.19, P=0.42) reflecting the small number of patients who responded to
pembrolizumab. Nonetheless, responses were durable with 68% ongoing at 12 months compared
with only 35% of chemotherapy responses ongoing at 12 months. Median OS was 10.3 months
with pembrolizumab compared to 7.4 months on chemotherapy (HR 0.73, 95% CI 0.59 – 0.91,
p=0.002). Grade 3 – 5 toxicity was seen in 15% of patients on pembrolizumab compared to 49%
on chemotherapy. Benefit was seen irrespective of the choice of chemotherapy given and
38
irrespective of PD-L1 expression by CPS%.
Atezolizumab was compared with investigators’ choice of paclitaxel, docetaxel or
vinflunine in the phase III IMvigor 211 study, sharing a similar overall design with the Keynote
045 study. One important study design difference was the use of a hierarchical fixed-sequence
procedure for endpoint analysis, according to PD-L1 expression. The primary endpoint of OS
was tested in successive subgroups defined by PD-L1 expression: IC2/3 (> 5% expression on
tumour-infiltrating immune cells) followed by IC 1/2/3 (> 1% expression on tumour-infiltrating
immune cells) followed by the overall ITT group. Demonstration of superior OS was required in
each group before the next group could be tested. No OS benefit was seen for atezolizumab
compared with chemotherapy in the first group of 234 patients with IC2/3 positive disease. In
this population, median OS was 11.1 months vs 10.6 months on chemotherapy (HR 0.87,
p=0.41). A survival benefit favoring atezolizumab was seen in the exploratory analysis of the
ITT population of 931 patients (HR 0.85, 95% CI 0.73 – 0.99). In this study, PD-L1 positivity
seemed to predict benefit from chemotherapy as well as immunotherapy. The theme of durable
immunotherapy responses was seen in this study with median duration of response of 21.7
months in the ITT population compared with 7.4 months for chemotherapy in the ITT
39
population.
Several single arm phase I and II studies have evaluated outcomes of checkpoint
inhibitors in the second line setting as outline in Table 3. 40–47 Response rates have ranged