Table 3.
Results of meta-analyses on diagnostic accuracy
| Sensitivity (%) (95% CI)a | Specificity (%)(95% CI)a | ||
|---|---|---|---|
|
N-staging | |||
| Our review |
PET (analyzed by patients) |
0 to 90e |
18.2 to 100e |
| PET/CT (analyzed by patients) |
14 to 87e |
95 to 100e |
|
| Xing and colleagues [17] |
PET (analyzed by patients and lesions; mixed)b |
30 (12 to 55) |
96 (87 to 99) |
| PET/CT (analyzed by patients and lesions; mixed)b |
11 (1 to 50) |
97 (97 to 100) |
|
| Jimenéz-Requena and colleagues [14] |
PET (analyzed by lymph nodes)b |
–c |
99 (97 to 99) |
| PET (analyzed by patients) |
–c |
–c |
|
| PET (analyzed by areas)b |
–c |
–c |
|
|
M-staging | |||
| Our review |
PET (analyzed by patients) |
33 to 97e |
56 to 98e |
| PET/CT (analyzed by patients) |
42 (15 to 72)d |
93 (81 to 99)d |
|
| Xing and colleagues [17] |
PET (analyzed by patients and lesions; mixed)b |
74 (51 to 88) |
75 (45 to 91) |
| PET/CT (analyzed by patients and lesions; mixed)b |
80 (53 to 93) |
87 (54 to 97) |
|
| Jimenéz-Requena and colleagues [14] |
PET (analyzed by patients) |
–c |
–c |
| PET (analyzed by lesions)b |
–c |
–c |
|
| PET (analyzed by areas)b |
86 (82 to 89) |
–c |
|
| PET (analyzed by Scans)b |
–c |
86 (77 to 92) |
|
|
N-staging and M-staging | |||
| Our review |
PET(analyzed by patients) |
0 to 100e |
18 to 100e |
| PET/CT (analyzed by patients) |
17 to 85e |
74 to 96e |
|
| Krug and colleagues [15] |
PET (analyzed by patients, lesions and scans; mixed)b |
83 (81 to 84) |
85 (83 to 87) |
| Mijnhout and colleagues [16] | PET (analyzed by patients, lesions and scans; mixed)b | 78 (70 to 84) | 88 (82 to 92) |
CT, computed tomography; PET, positron emission tomography. aPooled estimate with 95% confidence interval (CI), except for: Xing and colleagues, specification of median with credible intervals; Jiménez-Requena and colleagues, specification of mean values with CI; our review, specification of ranges. bThe results of these nonpatient-based analyses are not relevant for our review. cAuthors reported that quantitative synthesis was not possible due to heterogeneity. dData based on only one study. eSince the data were too heterogeneous, a quantitative synthesis of the results was not possible. Therefore a range of data is provided. Data of our systematic review is highlighted in bold font face.