Table 2.
Results of pooled estimates of studies
| Study number | Sensitivity | Specificity | PLR | NLR | DOR | AUC | I 2 sensitivity | I 2 specificity | |
|---|---|---|---|---|---|---|---|---|---|
| Diagnostic test accuracy analysis: high‐grade gliomas vs low‐grade gliomas | |||||||||
| All | 14 | 0.93 [0.87‐0.96] | 0.90 [0.82‐0.94] | 9.2 [5.1‐16.6] | 0.08 [0.05‐0.15] | 113 [42‐305] | 0.96 [0.94‐0.98] | 57.25 [31.82‐82.69] | 41.57 [4.74‐78.41] |
| Two‐compartment model | 12 | 0.91 [0.85‐0.95] | 0.89 [0.81‐0.94] | 8.4 [4.5‐15.6] | 0.10 [0.06‐0.17] | 82 [31‐216] | 0.96 [0.93‐0.97] | 47.97 [13.28‐82.67] | 40.25 [0.00‐80.85] |
| model‐independent | 2 | Study number is too small | |||||||
| Ktrans | 8 | 0.93 [0.84‐0.97] | 0.91 [0.82‐0.96] | 10.2 [4.9‐21.1] | 0.08 [0.04‐0.18] | 128 [44‐368] | 0.97 [0.95‐0.98] | 67.59 [43.47‐91.71] | 44.58 [0.00‐89.71] |
| ve | 5 | 0.87 [0.77‐0.93] | 0.95 [0.82‐0.99] | 18.9 [4.4‐80.1] | 0.13 [0.07‐0.25] | 141 [22‐883] | 0.96 [0.93‐0.97] | 35.46 [0.00‐98.77] | 43.83 [0.00‐100.00] |
| vp | 6 | 0.83 [0.67‐0.92] | 0.91 [0.77‐0.97] | 9.0 [3.2‐25.9] | 0.19 [0.09‐0.41] | 49 [9‐256] | 0.94 [0.92‐0.96] | 75.46 [55.49‐95.43] | 45.74 [0.00‐96.03] |
| hot‐spot ROI | 10 | 0.95 [0.89‐0.98] | 0.90 [0.82‐0.95] | 9.3 [5.0‐17.2] | 0.05 [0.02‐0.13] | 175 [52‐594] | 0.96 [0.94‐0.98] | 66.15 [43.47‐88.83] | 39.94 [0.00‐84.35] |
| whole volume ROI | 4 | 0.85 [0.73‐0.92] | 0.92 [0.68‐0.98] | 10.9 [2.2‐53.7] | 0.16 [0.08‐0.32] | 67 [9‐514] | 0.92 [0.90‐0.94] | 0.00 [0.00‐100.00] | 58.80 [13.48‐100.00] |
| Diagnostic test accuracy analysis: recurrence vs treatment‐related changes | |||||||||
| All | 9 | 0.88 [0.74‐0.95] | 0.86 [0.78‐0.91] | 6.4 [3.8‐10.5] | 0.13 [0.06‐0.32] | 47 [14‐156] | 0.89 [0.86‐0.91] | 72.77 [54.46‐91.08] | 0.00 [0.00‐100.00] |
| Two‐compartment model | 6 | 0.77 [0.65‐0.86] | 0.85 [0.75‐0.92] | 5.2 [2.9‐9.3] | 0.27 [0.17‐0.44] | 19 [8‐47] | 0.87 [0.84‐0.90] | 45.62 [0.00‐96.02] | 0.00 [0.00‐100.00] |
| model‐independent | 4 | 0.94 [0.86‐0.98] | 0.85 [0.74‐0.93] | 6.5 [3.4‐12.3] | 0.07 [0.03‐0.16] | 93 [29‐300] | 0.96 [0.94‐0.97] | 0.00 [0.00‐100.00] | 0.00 [0.00‐100.00] |
| Ktrans | 6 | 0.75 [0.63‐0.84] | 0.79 [0.68‐0.87] | 3.6 [2.3‐5.8] | 0.32 [0.21‐0.49] | 11 [5‐25] | 0.82 [0.78‐0.85] | 40.32 [0.00‐95.54] | 0.00 [0.00‐100.00] |
| ve | 2 | Study number is too small | |||||||
| vp | 2 | Study number is too small | |||||||
| hot‐spot ROI | 2 | Study number is too small | |||||||
| whole volume ROI | 7 | 0.91 [0.73‐0.97] | 0.88 [0.78‐0.93] | 7.3 [3.8‐13.8] | 0.11 [0.03‐0.34] | 68 [14‐328] | 0.91 [0.88‐0.93] | 76.12 [58.35‐93.89] | 0.00 [0.00‐100.00] |
| Diagnostic test accuracy analysis: primary central nervous system lymphomas vs high‐grade gliomas | |||||||||
| All | 5 | 0.78 [0.63‐0.89] | 0.81 [0.67‐0.90] | 4.1 [2.1‐7.7] | 0.27 [0.14‐0.51] | 15 [5‐50] | 0.86 [0.83‐0.89] | 51.10 [2.09‐100.00] | 69.63 [41.11‐98.16] |
| Two‐compartment model | 4 | 0.75 [0.53‐0.89] | 0.83 [0.69‐0.92] | 4.5 [2.0‐10.3] | 0.30 [0.14‐0.67] | 15 [3‐70] | 0.86 [0.83‐0.89] | 45.87 [0.00‐100.00] | 67.52 [32.79‐100.00] |
| model‐independent | 1 | Study number is too small | |||||||
| Ktrans | 2 | Study number is too small | |||||||
| ve | 3 | Study number is too small | |||||||
| vp | 1 | Study number is too small | |||||||
| hot‐spot ROI | 1 | Study number is too small | |||||||
| whole tumor ROI | 4 | 0.82 [0.67‐0.91] | 0.81 [0.64‐0.91] | 4.3 [2.0‐9.2] | 0.22 [0.11‐0.47] | 19 [5‐77] | 0.88 [0.85‐0.90] | 50.13 [0.00‐100.00] | 77.41 [54.83‐99.99] |
The numbers in the parentheses are 95% confidence intervals.
Abbreviations: AUC, area under the curve; DOR, diagnostic odds ratio; NLR, negative likelihood ratio; PLR, positive likelihood ratio; ROI, region of interest.