Table 1.
Benchmark datasets for the evaluation in different cancer types and ICB therapy strategies.
| Data Name 1 | Data Name 2 | Tumor Type | Sample # | Objective Response Rate * | Therapeutic Agent | Data Type | RECIST | Reference |
|---|---|---|---|---|---|---|---|---|
| Gide et al., 2019 | Gide et al., 2019 (aPD1) | Melanoma | 41 | 46.3% | anti-PD-1 (pembrolizumab/nivolumab) | RNA-seq | RECIST 1.1 | [20] |
| Gide et al. 2019 (aCTLA4 + aPD1) | 32 | 65.6% | anti-PD-1 + anti-CTLA-4 (ipilimumab) | |||||
| Riaz et al., 2017 | Riaz et al., 2017 (aPD1) | Melanoma | 25 | 26.1% | anti-PD-1 (nivolumab) | RNA-seq | RECIST 1.1 | [19] |
| Riaz et al., 2017 (aCTLA4 prog aPD1) | 26 | 15.4% | anti-CTLA-4 progression + anti-PD-1 (nivolumab, ipilimumab) | |||||
| Van Allen et al., 2015 | Van Allen et al., 2015 | Melanoma | 42 | 17.1% | anti-CTLA-4 (ipilimumab) | RNA-seq | RECIST 1.1 | [22] |
| Chen et al., 2016 | Chen et al., 2016 (aCTLA4) | Melanoma | 16 | 26.7% | anti-CTLA-4 (ipilimumab) | NanoString nCounter | RECIST | [23] |
| Chen et al., 2016 (aCTLA4 prog aPD1) | 16 | 6.7% | anti-CTLA-4 progression + anti-PD-1 | |||||
| Hugo et al., 2016 | Hugo et al., 2016 | Melanoma | 27 | 55.6% | anti-PD-1 (pembrolizumab/nivolumab) | RNA-seq | irRECIST | [18] |
| TCGA | TCGA | Melanoma | 18 | 36.4% | anti-CTLA-4 (ipilimumab) | RNA-seq | RECIST | |
| Prat et al., 2017 | Prat et al., 2017 (melanoma) | Melanoma | 25 | 36% | anti-PD-1 (pembrolizumab/nivolumab) | NanoString nCounter | RECIST 1.1 | [24] |
| Prat et al., 2017 (NSCLC) | NSCLC | 35 | 25.37% | |||||
| HNSCC | 5 | |||||||
| Mariathasan et al., 2018 | Mariathasan et al., 2018 | UC | 298 | 22.8% | anti-PD-L1 (atezolizumab) | RNA-seq | RECIST | [10] |
| Snyder et al., 2017 | Snyder et al., 2017 | UC | 26 | 35% | anti-PD-L1 (atezolizumab) | RNA-seq | RECIST 1.1 | [25] |
| Miao et al., 2018 | Miao et al., 2018 (aPD1) | ccRCC | 16 | 18.8% | anti-PD-1 (nivolumab) | RNA-seq | RECIST 1.1 | [26] |
| Miao et al., 2018 (ICB) | 17 | 29.4% | anti-PD-L1 (atezolizumab)/anti-PD-1 + anti-CTLA-4 (nivolumab, ipilimumab) |
1 Ten benchmark datasets for overall analysis. 2 Fifteen datasets for cancer-specific and treatment-specific analysis. # Total number of patients included in the dataset. Among them, there are a few of patients missed response information that were excluded in analysis. * Objective response rate (ORR) = (CR + PR)/total patients.