Skip to main content

Table 3 Clinical utility comparison of IHC TPS and RNA-seq rank results

From: Next generation sequencing of PD-L1 for predicting response to immune checkpoint inhibitors

Prediction Method Sensitivity Specificity PPV NPV
Melanoma IHC ≥1% 31.3% 81.8% 55.6% 62.1%
Melanoma RNA-seq > 75 25.0% 93.2% 72.7% 63.1%
Melanoma IHC ≥1% & RNA-seq > 75 20.8% 94.6% 71.4% 64.8%
NSCLC IHC ≥1% 58.8% 60.6% 26.3% 71.4%
NSCLC IHC ≥50% 47.1% 84.5% 42.1% 87.0%
NSCLC RNA-seq > 75 58.8% 64.8% 28.6% 86.8%
NSCLC IHC ≥1% & RNA-seq > 75 63.6% 68.0% 30.4% 89.5%
NSCLC IHC ≥50% & RNA-seq > 75 46.2% 83.9% 37.5% 88.1%
RCC IHC ≥1% 28.6% 92.1% 40.0% 87.5%
RCC RNA-seq > 75 42.9% 84.2% 33.3% 88.9%
RCC IHC ≥1% & RNA-seq > 75 33.3% 93.9% 50.0% 88.6%
  1. Sensitivity = TP/(TP + FN)
  2. Specificity = TN/(TN + FP)
  3. Positive predictive value (PPV) = TP/(TP + FP)
  4. Negative predictive value (NPV) = TN/(TN + FN)