Skip to main content

Table 4 Summary of total error, false negative and false positive rates for the three scanner brands, and sub-classification of false positives and false negatives by possible causative factors (with % of the total errors of each type—false positives or false negatives)

From: Algorithm-assisted diagnosis of Hirschsprung’s disease – evaluation of robustness and comparative image analysis on data from various labs and slide scanners

Scanner (by manufacturer)

Image set total

Total errors (False positive and false negative)

Error type

Total (% of image set total)

Technical

Missing Specificity/Sensitivity

New/abnormal findings

Unknown

Hamamatsu Photonics

348

11 (3.2%)

False positive

3 (0.9%)

2 (66.7%)

1 (33.3%)

0 (0%)

0 (0%)

False negative

8 (2.3%)

0 (0%)

7 (87.5%)

1 (12.5%)

0 (0%)

3D Histech

447

44 (9.8%)

False positive

40 (8.9%)

7 (17.5%)

10 (25%)

0 (0%)

23 (57.5%)

False negative

4 (0.9%)

0 (0%)

4 (100%)

0 (0%)

0 (0%)

Philips

667

26 (3.9%)

False positive

23 (3.45%)

0 (0%)

20 (87%)

3 (13%)

0 (0%)

False negative

3 (0.45%)

0 (0%)

3 (100%)

0 (0%)

0 (0%)