Table 3.
Subgroup analysis of diagnostic accuracy of ultrasonographic markers to detect PCOS.
| Index test | N | Sensitivity % (95% CI) | Specificity % (95% CI) | DOR (95% CI) | Positive LR (95% CI) | Negative LR (95% CI) | |
|---|---|---|---|---|---|---|---|
| Follicle number per ovary (FNPO) | |||||||
| Cutoff value | ≥20 follicles | 3 | 87.64 (78.78–93.12) | 93.74 (82.40–97.96) | 106.18 (28.06–401.70) | 14.00 (7.51–26.11) | 0.13 (0.05–0.37) |
| <20 follicles | 13 | 83.54 (80.29–86.35) | 90.71 (85.08–94.35) | 49.54 (24.48–100.24) | 8.99 (6.88–11.74) | 0.18 (0.11–0.29) | |
| ≥12 follicles | 12 | 82.98 (79.85–85.71) | 90.20 (84.67–93.87) | 44.85 (23.72–84.81) | 8.46 (6.65–10.77) | 0.19 (0.12–0.29) | |
| <12 follicles | 4 | 87.18 (80.01–92.03) | 93.04 (83.34–97.28) | 90.89 (23.57–350.50) | 12.53 (7.07–22.20) | 0.14 (0.06–0.33) | |
| Age groups | ≥30 years old | 4 | 79.29 (74.84–83.13) | 89.93 (76.86–96.00) | 34.18 (12.32–94.85) | 7.87 (5.80–10.69) | 0.23 (0.11–0.50) |
| <30 years old | 14 | 85.80 (82.38–88.65) | 92.08 (88.03–94.84) | 70.29 (37.54–131.64) | 10.84 (8.23–14.26) | 0.15 (0.10–0.23) | |
| BMI groups | ≥30 kg/m2 | 2 | 83.33 (75.09–89.24) | 95.96 (89.73–98.48) | 118.75 (38.70–364.36) | 20.63 (12.34–34.46) | 0.17 (0.08–0.40) |
| <30 kg/m2 | 9 | 87.78 (83.27–91.21) | 92.92 (88.48–95.73) | 94.32 (46.75–190.30) | 12.40 (8.66–17.75) | 0.13 (0.08–0.21) | |
| Geography | North America | 3 | 78.63 (72.91–83.41)* | 87.24 (73.06–94.52) | 25.17 (9.48–66.83) | 6.16 (4.40–8.63) | 0.24 (0.12–0.51) |
| Europe | 8 | 83.14 (79.46–86.27) | 90.96 (84.73–94.80) | 49.61 (25.23–97.55) | 9.19 (7.12–11.88) | 0.19 (0.11–0.31) | |
| Asia | 5 | 89.33 (83.70–93.18)* | 93.05 (85.44–96.83) | 112.13 (42.96–292.67) | 12.86 (8.08–20.44) | 0.11 (0.05–0.24) | |
| Diagnostic criteria | 1990 NIH | 10 | 81.76 (79.23–84.05)* | 91.06 (86.40–94.24) | 45.69 (26.17–79.76) | 9.15 (7.51–11.16) | 0.20 (0.14–0.30) |
| Rotterdam | 6 | 89.81 (84.11–93.62)* | 89.79 (77.46–95.75) | 77.55 (26.62–225.86) | 8.80 (5.51–14.05) | 0.11 (0.05–0.26) | |
| Transducer frequency | ≥8 MHz | 9 | 84.27 (80.07–87.72) | 86.42 (79.62–91.20) | 34.09 (16.95–68.55) | 6.21 (4.59–8.38) | 0.18 (0.12–0.28) |
| <8 MHz | 5 | 86.94 (80.48–91.49) | 94.13 (87.07–97.45) | 106.83 (31.68–360.20) | 14.82 (8.75–25.08) | 0.14 (0.06–0.30) | |
| Methodology | 2D real time | 14 | 84.30 (80.97–87.13) | 90.24 (85.19–93.70) | 49.63 (26.30–93.67) | 8.64 (6.67–11.19)* | 0.17 (0.12 0.26) |
| Offline | 2 | 83.33 (75.09–89.24) | 95.96 (89.73–98.48) | 118.75 (38.70–364.36) | 20.63 (12.34–34.46)* | 0.17 (0.08–0.40) | |
| Risk of bias | High risk | 9 | 86.52 (82.03–90.03) | 91.71 (85.33–95.47) | 71.04 (30.07–167.83) | 10.44 (7.27–15.00) | 0.15 (0.08–0.26) |
| Low risk | 7 | 81.75 (77.80–85.14) | 89.74 (82.09–94.34) | 39.18 (18.20–84.36) | 7.97 (6.00–10.58) | 0.20 (0.12–0.35) | |
| Index test | N | Sensitivity % (95% CI) | Specificity % (95% CI) | DOR (95% CI) | Positive LR (95% CI) | Negative LR (95% CI) | |
|---|---|---|---|---|---|---|---|
| Ovarian volume (OV) | |||||||
| Cutoff value | ≥10 cm3 | 5 | 82.56 (68.83–91.03) | 78.95 (67.26–87.25) | 17.75 (8.65–36.44) | 3.92 (2.24–6.87) | 0.22 (0.14–0.34) |
| <10 cm3 | 12 | 80.74 (74.54–85.71) | 82.01 (74.01–87.95) | 19.11 (10.01–36.46) | 4.49 (3.24–6.21) | 0.23 (0.16–0.35) | |
| Age groups | ≥30 years old | 4 | 75.58 (64.62–83.98) | 73.31 (59.19–83.88) | 8.50 (4.26–16.96) | 2.83 (1.96–4.10) | 0.33 (0.21–0.53) |
| <30 years old | 14 | 83.63 (77.19–88.52) | 83.77 (77.40–88.61) | 26.36 (15.11–45.99) | 5.15 (3.65–7.28) | 0.20 (0.14–0.28) | |
| BMI groups | ≥30 kg/m2 | 3 | 84.78 (79.75–88.73) | 77.82 (54.55–91.11) | 19.53 (6.33–60.31) | 3.82 (2.79–5.24) | 0.20 (0.08–0.49) |
| <30 kg/m2 | 11 | 81.91 (74.32–87.63) | 85.78 (77.39–91.40) | 27.31 (12.33–60.52) | 5.76 (3.79–8.75) | 0.21 (0.13–0.34) | |
| Geography | North America | 4 | 75.84 (67.50–82.59) | 72.98 (65.03–79.69)* | 8.48 (5.82–12.35)* | 2.81 (2.16–3.65)* | 0.33 (0.26–0.42) |
| Europe | 5 | 83.30 (72.83–90.27) | 87.96 (80.58–92.79)* | 36.44 (19.59–67.77)* | 6.92 (4.16–11.52)* | 0.19 (0.12–0.30) | |
| Asia | 7 | 86.53 (74.74–93.31) | 82.68 (69.05–91.08) | 30.67 (9.95–94.59) | 5.00 (2.59–9.64) | 0.16 (0.08–0.32) | |
| Diagnostic criteria | 1990 NIH | 8 | 77.48 (71.67–82.39) | 78.41 (71.18–84.22) | 12.50 (8.33–18.74) | 3.59 (2.86–4.51) | 0.29 (0.22–0.38) |
| Rotterdam | 8 | 87.61 (77.62–93.52) | 85.45 (72.88–92.77) | 41.54 (14.59–118.26) | 6.02 (3.26–11.13) | 0.14 (0.07–0.29) | |
| Transducer frequency | ≥8 MHz | 9 | 79.17 (73.65–83.79) | 79.38 (72.08–85.16) | 14.63 (9.39–22.80) | 3.84 (3.02–4.87) | 0.26 (0.19–0.36) |
| <8 MHz | 7 | 86.70 (72.76–94.09) | 84.54 (70.90–92.47) | 35.66 (10.34–122.97) | 5.61 (2.60–12.09330) | 0.16 (0.08–0.32) | |
| Methodology | 2D real time | 11 | 82.53 (75.27–87.99) | 80.11 (71.80–86.43) | 19.02 (11.03–32.82) | 4.15 (2.95–5.83) | 0.22 (0.15–0.31) |
| Other | 5 | 78.99 (69.64–86.04) | 83.12 (72.41–90.24) | 18.52 (6.53–52.51) | 4.68 (2.80–7.82) | 0.25 (0.14–0.44) | |
| Risk of bias | High risk | 11 | 83.16 (76.88–88.00) | 78.17 (70.70–84.16) | 17.68 (9.83–31.81) | 3.81 (2.76–5.26) | 0.22 (0.15–0.31) |
| Low risk | 5 | 76.88 (70.06–82.53) | 86.56 (82.70–89.66) | 21.41 (13.51–33.95) | 5.72 (4.19–7.81) | 0.27 (0.21–0.34) | |
| Adolescent studies | Total | 4 | 81.84 (75.74–86.68) | 83.54 (77.06–88.46) | 22.87 (13.16–39.74) | 4.97 (3.62–6.82) | 0.22 (0.15–0.31) |
| TAUS | 3 | 83.77 (77.90–88.31) | 81.88 (74.56–87.45) | 23.32 (13.10–41.54) | 4.62 (3.36–6.37) | 0.20 (0.14–0.29) | |
| Index test | N | Sensitivity % (95% CI) | Specificity % (95% CI) | DOR (95% CI) | Positive LR (95% CI) | Negative LR (95% CI) | |
|---|---|---|---|---|---|---|---|
| Follicle number per cross-section (FNPS) | |||||||
| Age groups | ≥30 years old | 2 | 89.54 (81.83–94.21)* | 79.98 (62.36–90.59) | 34.21 (11.50–101.71) | 4.47 (2.65–7.54) | 0.13 (0.06–0.29) |
| <30 years old | 3 | 70.14 (58.51–79.64)* | 86.58 (79.71–91.37) | 15.15 (8.87–25.87) | 5.22 (3.50–7.80) | 0.34 (0.25–0.48) | |
| BMI groups | ≥30 kg/m2 | 3 | 82.37 (65.78–91.91) | 88.00 (74.57–94.83) | 34.29 (9.57–122.88) | 6.87 (3.09–15.25) | 0.20 (0.09–0.43) |
| <30 kg/m2 | 2 | 77.80 (60.74–88.81) | 80.87 (76.37–84.68) | 14.81 (6.27–35.01) | 4.07 (2.09–7.90) | 0.27 (0.21–0.36) | |
| Methodology | 2D real time | 2 | 83.17 (70.52–91.08) | 79.27 (70.99–85.66) | 18.89 (11.33–31.49) | 4.01 (2.38–6.77) | 0.21 (0.16–0.28) |
| Offline | 2 | 71.30 (62.08–79.03) | 95.10 (65.16–99.51) | 48.17 (4.48–518.13) | 14.54 (6.67–31.70) | 0.30 (0.06–1.60) | |
| Risk of bias | Total | 4 | 82.70 (75.15–88.31) | 81.07 (70.10–88.67) | 20.47 (12.69–33.02) | 4.37 (2.80–6.81) | 0.21 (0.15–0.29) |
| Low risk | 3 | 80.55 (65.70–89.95) | 90.92 (78.95–96.40) | 41.49 (11.90–144.68) | 8.88 (4.29–18.36) | 0.21 (0.10–0.47) | |
DOR: diagnostic odds ratio; LR: likelihood ratio.
* Significant difference between stratified groups (bolded).