Table 2. CIN Grade 2 or 3 Detection Statistics for a Simulated Cohort of 1000 Women.
| Test type | PPV, % (95% CI) | NPV, % (95% CI) | False negatives per 1000 women, No. (95% CI) | False positives, ie, excess colposcopies, per 1000 women, No. (95% CI) |
|---|---|---|---|---|
| LBC test only | 13.7 (12.1-15.1) | 99.5 (99.3-99.6) | 4.9 (3.5-6.7) | 95.1 (93.1-97.0) |
| HPV test onlya | 15.4 (15.2-15.5) | 99.8 (99.5-99.9) | 2.0 (1.9-2.1) | 98.9 (98.0-101.0) |
| HPV with LBC triageb | 58.7 (55.3-60.1) | 99.4 (99.2-99.5) | 6.4 (5.1-8.0) | 9.6 (9.4-9.8) |
| LBC with HPV triageb | 58.7 (55.3-60.1) | 99.4 (99.2-99.5) | 6.4 (5.1-8.0) | 9.6 (9.4-9.8) |
| Cotesting, HPV followed by LBCc | 9.6 (9.5-9.7) | 99.9 (99.6-99.9) | 0.5 (0.3-0.7) | 184.4 (181.8-188.0) |
| Cotesting, LBC followed by HPVc | 9.6 (9.5-9.7) | 99.9 (99.6-99.9) | 0.5 (0.3-0.7) | 184.4 (181.8-188.0) |
| Screening rounds required to achieve ts <0.1% or NPV >99.99% | ||||
| Test type | Rounds required | Total tests required per 1000 women, No. (95% CI)d | False negatives per 1000 women, No. (95% CI) | False positives, ie excess colposcopies, per 1000 women, No. (95% CI) |
| LBC test only | 3 | 2690 (2688-2694) | 0.3 (0.1-0.7) | 258.4 (253.6-263.2) |
| HPV test onlya | 2 | 1883 (1881-1884) | 1.0 (0.9-1.0) | 137.7 (135.9-141.5) |
| Cotesting, HPV followed by LBC | 1 (combined) | 1883 (1881-1884) | 0.5 (0.3-0.7) | 184.4 (181.8-188.0) |
| Cotesting, LBC followed by HPV | 1 (combined) | 1890 (1890-1890) | 0.5 (0.3-0.7) | 184.4 (181.8-188.0) |
| Reflex testing, HPV followed by LBC | 4 | 4274 (4267-4290) | 1.0 (0.9-1.1) | 34.3 (34.0-35.3) |
Abbreviations: CIN, cervical intraepithelial neoplasia; HPV, human papillomavirus; LBC, liquid-based cytology; NPV, negative predictive value; PPV, positive predictive value; ts, confidence threshold.
HPV tests without reflex are not typically used but were shown for completeness.
Secondary test used as a triage test for initially positive results. These results are for the first round of screening plus triage and do not include subsequent management of triage negative women.
In cotesting simulation shown here, both tests were performed and a positive on either was referred to colposcopy. In more recent implementations, positive results can trigger a repeated cotest rather than immediate colposcopy, so this table shows worst-case scenario for excess colposcopy.
Total tests are the number of initial tests (1000) plus follow-up tests. For example, cotesting HPV followed by LBC requires 1000 initial HPV tests plus 883 (95% CI, 881-884) LBC tests.