Skip to main content
. 2017 May 22;70(6):500–507. doi: 10.1136/jclinpath-2016-204089

Table 4.

Performance of the assay for different histological types

Training*
Validation, entire set
Validation, agreement set
Histological type #Samples† % Correct‡ #Samples† % Correct‡ #Samples†§ % Correct‡
Medullary 5 100 [48–100] 3 100 [29–100] 1 (33.3%) 100 [3–100]
PTC classic 48 94 [83–100] 17 88 [64–99] 15 (88.2%) 100 [78–100]
FVPTC 40 81 [65–92] 37 84 [68–94] 23 (62.2%) 96 [78–100]
FC 16 56 [30–80] 3 67 [9–99] 1 (33.3%) 100 [3–100]
PDC 5 100 [48–100] 1 100 [3–100] 0 (0%) NA
Papillary, other 6 100 [54–100] 0 NA 0 NA
FA 90 76 [66–84] 95 76 [66–84] 82 (86.3%) 82 [72–89]
Nodular hyperplasia 48 75 [60–86] 28 64 [44–81] 23 (82.1%) 74 [52–90]
CLT 9 82 [44–99] 5 40 [5–85] 5 (100.0%) 40 [5–85]
Total 267 189 150

*Only indeterminate training samples are listed in the table. Estimates for the training performance are based on the mean of 10 10-fold cross-validation runs. Samples with very low expression in any of the classification steps, as well as medullary samples, are not included.

†Number of samples includes only those that passed quality control steps.

‡95% CIs are in square brackets.

§Numbers in parentheses signify the percentage of test samples in the agreement set.

CLT, chronic lymphocytic thyroiditis; FA, follicular adenoma; FC, follicular carcinoma; FVPTC, follicular variant of papillary carcinoma; NA, not applicable; PDC, poorly differentiated carcinoma.