TABLE 5.
The results of meta‐regression in 13 miRNA diagnostic models
| Model Name | Covariate (Consistency between training set * and validation set) | Category | Validation set number | Sensitivity (95% CIa) | P1 † | Specificity (95% CI) | P2 ‡ |
|---|---|---|---|---|---|---|---|
| Latorre 2015 | Sample type | Consistency | 9 | 37% (0.24‐0.51) | .18 | 88% (0.82‐0.94) | .18 |
| Inconsistency | 1 | 64% (0.35‐0.92) | 63% (0.29‐0.96) | ||||
| Ethnicity | Consistency | 6 | 47% (0.25‐0.68) | .45 | 87% (0.79‐0.95) | .25 | |
| Inconsistency | 4 | 37% (0.20‐0.54) | 85% (0.74‐0.96) | ||||
| Pan 2019 | Sample type | Consistency | 9 | 58% (0.33‐0.83) | .59 | 61% (0.51‐0.72) | .10 |
| Inconsistency | 2 | 43% (0.00‐0.98) | 28% (0.00‐0.57) | ||||
| Ethnicity | Consistency | 6 | 66% (0.37‐0.94) | .44 | 62% (0.44‐0.80) | .96 | |
| Inconsistency | 5 | 48% (0.22‐0.73) | 55% (0.42‐0.69) | ||||
| Wang 2011 | Sample type | Consistency | 7 | 27% (0.13‐0.41) | .21 | 76% (0.54‐0.98) | .68 |
| Inconsistency | 1 | 55% (0.24‐0.85) | 77% (0.25‐1.00) | ||||
| Ethnicity | Consistency | 4 | 52% (0.30‐0.75) | .04 | 73% (0.42‐1.00) | .92 | |
| Inconsistency | 4 | 22% (0.08‐0.36) | 78% (0.52‐1.00) | ||||
| Zhou 2016 | Sample type | Consistency | 8 | 54% (0.38‐0.69) | .88 | 52% (0.39‐0.65) | .22 |
| Inconsistency | 2 | 50% (0.22‐0.78) | 30% (0.02‐0.58) | ||||
| Ethnicity | Consistency | 8 | 49% (0.32‐0.65) | .38 | 42% (0.28‐0.56) | .10 | |
| Inconsistency | 2 | 62% (0.39‐0.86) | 69% (0.46‐0.91) | ||||
| Barry 2018 | Sample type | Consistency | 1 | 100% (1.00‐1.00) | NA | 0% (0.00‐0.00) | NA |
| Inconsistency | 10 | 93% (0.84‐1.00) | 15% (0.06‐0.23) | ||||
| Ethnicity | Consistency | 6 | 90% (0.71‐1.00) | .93 | 15% (0.00‐0.31) | .51 | |
| Inconsistency | 5 | 96% (0.88‐1.00) | 11% (0.00‐0.22) | ||||
| Cui 2017 | Sample type | Consistency | 1 | 0% (0.00‐0.00) | NA | 100% (1.00‐1.00) | NA |
| Inconsistency | 10 | 9% (0.01‐0.41) | 91% (0.84‐0.95) | ||||
| Ethnicity | Consistency | 6 | 18% (0.00‐0.54) | .82 | 97% (0.91‐1.00) | .11 | |
| Inconsistency | 5 | 5% (0.00‐0.16) | 89% (0.82‐0.97) | ||||
| Duffy 2018 | Sample type | Consistency | 0 | NAb | NA | NA | NA |
| Inconsistency | 0 | NA | NA | ||||
| Ethnicity | Consistency | 0 | NA | NA | NA | NA | |
| Inconsistency | 0 | NA | NA | ||||
| Miotto 2013‐RVM | Sample type | Consistency | 0 | NA | NA | NA | NA |
| Inconsistency | 9 | 98% (0.62‐1.00) | 0% (0.00‐0.97) | ||||
| Ethnicity | Consistency | 5 | 88% (0.62‐1.00) | .05 | 5% (0.00‐0.22) | .15 | |
| Inconsistency | 4 | 100% (1.00‐1.00) | 0% (0.00‐0.02) | ||||
| Miotto 2013‐AIC logistic regression | Sample type | Consistency | 0 | NA | NA | NA | NA |
| Inconsistency | 9 | 58% (0.27‐0.83) | 42% (0.31‐0.55) | ||||
| Ethnicity | Consistency | 5 | 87% (0.69‐1.00) | <.01 | 50% (0.32‐0.68) | .21 | |
| Inconsistency | 4 | 28% (0.13‐0.44) | 35% (0.20‐0.50) | ||||
| Qi 2012 | Sample type | Consistency | 1 | 0% (0.00‐0.00) | NA | 0% (0.00‐0.00) | 1.00 |
| Inconsistency | 9 | 91% (0.91‐0.91) | 0% (0.00‐0.00) | ||||
| Ethnicity | Consistency | 6 | 82% (0.59‐1.00） | .53 | 0% (0.00‐0.00) | 1.00 | |
| Inconsistency | 4 | 95% (0.84‐1.00） | 0% (0.00‐0.00) | ||||
| Zhang 2013 | Sample type | Consistency | 0 | NA | NA | NA | NA |
| Inconsistency | 5 | 10% (0.01‐0.55) | 84% (0.51‐0.96) | ||||
| Ethnicity | Consistency | 3 | 13% (0.00‐0.56) | .56 | 63% (0.38‐0.88) | .77 | |
| Inconsistency | 2 | 3% (0.00‐0.14) | 100% (1.00‐1.00) | ||||
| Alipoor 2019 | Sample type | Consistency | 1 | 0% (0.00‐0.00) | NA | 100% (1.00‐1.00) | NA |
| Inconsistency | 10 | 17% (0.05‐0.46) | 76% (0.65‐0.85) | ||||
| Ethnicity | Consistency | 6 | 15% (0.00‐0.42) | .51 | 74% (0.57‐0.91) | .24 | |
| Inconsistency | 5 | 18% (0.00‐0.43) | 78% (0.66‐0.90) | ||||
| Hu 2019 | Sample type | Consistency | NA | NA | NA | NA | NA |
| Inconsistency | NA | NA | NA | ||||
| Ethnicity | Consistency | NA | NA | NA | NA | NA | |
| Inconsistency | NA | NA | NA |
a: confidence interval; b: non‐available.
*
The training set referred to the dataset used to train the model in this work and the detailed information of training set of each model was provided in Table S1.
†
P1 referred to the P value when comparing the sensitivity of each model in consistent subgroup and inconsistent subgroup.
‡
P2 referred to the P value when comparing the specificity of each model in consistent subgroup and inconsistent subgroup.