Skip to main content

Table 2 TAP scoring precision for determining PD-L1 expression level in the internal study (including borderline cases)

From: Tumor Area Positivity (TAP) score of programmed death-ligand 1 (PD-L1): a novel visual estimation method for combined tumor cell and immune cell scoring

Study

Statistics

Agreement

Resultsa

Between-reader

APA %

99.3

Pass

n/N

296/298

95% CI

98.0–100.0

ANA %

99.3

Pass

n/N

300/302

95% CI

98.0–100.0

OPA %

99.3

Pass

n/N

298/300

95% CI

98.0–100.0

Within-reader

APA %

99.0

Pass

n/N

296/299

95% CI

98.0–100.0

ANA %

99.0

Pass

n/N

298/301

95% CI

98.0–100.0

OPA %

99.0

Pass

n/N

297/300

95% CI

98.0–100.0

Background

OPA %

100.0

Pass

n/N

600/600

95% CI

99.4–100.0

  1. Counts indicate the number of pairwise comparisons and do not represent the number of unique cases
  2. ANA Average negative agreement, APA Average positive agreement, CI Confidence interval, OPA Overall percent agreement, PD-L1 Programmed death-ligand 1, TAP Tumor Area Positivity
  3. aThe acceptance criteria for between-reader precision were ≥ 85% ANA and APA; the acceptance criteria for within-reader precision were ≥ 90% OPA, and ≥ 85% ANA and APA. The acceptance criterion for background staining was that the assay must produce acceptable levels of non-specific staining on ULTRA instruments in at least 90% of samples