Test characteristics of one-sample and two-sample FIT used in the model
| Cut-off level (ng Hb/ml) | Specificity (per person, %) | Sensitivity (per lesion, %)* | ||||
| Adenoma | CRC early preclinical† | CRC late preclinical† | ||||
| ≤5 mm | 6–9 mm | ≥10 mm | ||||
| One-sample FIT | ||||||
| 50 | 95.79 | 0.0 | 9.6 | 16.1 | 65.0 | 90.0 |
| 75 | 97.05 | 0.0 | 5.7 | 14.4 | 58.5 | 87.0 |
| 100 | 97.76 | 0.0 | 4.4 | 13.1 | 52.0 | 83.5 |
| 150 | 98.34 | 0.0 | 2.9 | 12.3 | 50.5 | 83.0 |
| 200 | 98.70 | 0.0 | 2.5 | 10.3 | 50.0 | 82.5 |
| Two-sample FIT, at least one sample positive | ||||||
| 50 | 93.01 | 0.0 | 14.2 | 16.7 | 75.0 | 93.5 |
| 75 | 94.90 | 0.0 | 8.4 | 15.5 | 71.0 | 92.0 |
| 100 | 96.03 | 0.0 | 6.9 | 14.4 | 66.0 | 90.0 |
| 150 | 97.03 | 0.0 | 5.2 | 14.3 | 66.0 | 90.0 |
| 200 | 97.65 | 0.0 | 4.9 | 12.5 | 66.0 | 90.0 |
| Two-sample FIT, mean of both samples positive | ||||||
| 50 | 95.51 | 0.0 | 12.6 | 17.0 | 67.0 | 90.0 |
| 75 | 96.90 | 0.0 | 7.5 | 15.1 | 61.0 | 87.5 |
| 100 | 97.66 | 0.0 | 5.4 | 13.8 | 54.0 | 84.0 |
| 150 | 98.31 | 0.0 | 3.3 | 12.8 | 51.0 | 83.0 |
| 200 | 98.63 | 0.0 | 2.1 | 10.7 | 49.0 | 81.5 |
| Two-sample FIT, both samples positive | ||||||
| 50 | 98.40 | 0.0 | 3.8 | 12.0 | 34.0 | 70.0 |
| 75 | 98.94 | 0.0 | 1.8 | 10.0 | 29.0 | 65.0 |
| 100 | 99.21 | 0.0 | 0.9 | 8.8 | 24.0 | 59.0 |
| 150 | 99.43 | 0.0 | 0.1 | 7.1 | 20.0 | 53.0 |
| 200 | 99.49 | 0.0 | 0.0 | 5.2 | 16.0 | 47.5 |
The test characteristics used in the model were fitted to the positivity rates and detection rates of advanced neoplasia and CRC from two Dutch randomised controlled trials.9–12 Sensitivity for adenomas smaller than 5 mm was assumed to be 0% for all tests, at any cut-off level.
↵* Excluding the probability that an adenoma or cancer is found due to a lack of specificity.
↵† It was assumed that the probability a CRC bleeds and thus the sensitivity of FIT for CRC depends on the time until clinical diagnosis, in concordance with findings for FOBT, which were based on a previous calibration of the MISCAN–colon model to three FOBT trials.16 This result is to be expected when cancers that bleed do so increasingly over time, starting ‘occultly’ and ending as clinically visible. This interpretation also holds for FIT.
CRC, colorectal cancer; FIT, faecal immunochemical test; FOBT, faecal occult blood tests; Hb, haemoglobin.