15 / 27
Eckstein, M et al. Eur J Cancer 2019
approved PD-L1 assays in
urothelial carcinoma
28-8 and the SP263 assays (
Fig. 1
E;
Table S4
).
Compared with the SP142 assay, pooled OPA across
prespecified algorithm cut-offs was 82.0% (range:
3.8. Agreement across different clinical important cut-offs
for first-line therapy selection with atezolizumab and
pembrolizumab
Table 1
Positive percentage agreement (PPA) and negative percentage agreement (NPA) in 251 muscle-invasive bladder cancer specimens across different
PD-L1 assay algorithms.
Reference assay
Comparator assay at matched expression cut-off
Dako 28-8
Dako 22c3
Ventana SP142
Ventana SP263
PPA
NPA
PPA
NPA
PPA
NPA
PPA
NPA
DAKO 28
-
8
<
/
!
5% TC
88.1% 96.4% 44.0% 99.4% 85.0% 95.2%
DAKO 22c3
<
/
!
10 CPS
95.5% 92.6%
46.6% 100% 92.0% 89.6%
Ventana SP142
<
/
!
5% IC
73.2% 86.7% 78.0% 90.0%
82.9% 81.0%
Ventana SP263
<
/
!
25% TC- or IC
90.0% 98.8% 85.7% 99.4% 30.0% 100%
TC, tumour cells; IC, immune cells; CPS, combined positive score.
PPA
!
85% indicates interchangeability of diagnostic assays.
M. Eckstein et al. / European Journal of Cancer 106 (2019) 234
e
243
238