ALL subgroup prediction accuracies using top 50 chi-square selected genes from U133A and B and artificial neural network (ANN) in decision tree format
. | Training set, %* . | . | Test set, %† . | . |
|---|---|---|---|---|
| Subgroup . | Apparent accuracy‡ . | True accuracy§ . | Sensitivity∥ . | Specificity¶ . |
| T-ALL | 100 | 100 | 100 | 100 |
| E2A-PBX1 | 100 | 100 | 100 | 100 |
| TEL-AML1 | 98 | 100 | 100 | 100 |
| BCR-ABL | 100 | 95 | 75 | 100 |
| MLL rearrangement | 100 | 100 | 100 | 100 |
| Hyperdiploid with more than 50 chromosomes | 100 | 100 | 100 | 100 |
. | Training set, %* . | . | Test set, %† . | . |
|---|---|---|---|---|
| Subgroup . | Apparent accuracy‡ . | True accuracy§ . | Sensitivity∥ . | Specificity¶ . |
| T-ALL | 100 | 100 | 100 | 100 |
| E2A-PBX1 | 100 | 100 | 100 | 100 |
| TEL-AML1 | 98 | 100 | 100 | 100 |
| BCR-ABL | 100 | 95 | 75 | 100 |
| MLL rearrangement | 100 | 100 | 100 | 100 |
| Hyperdiploid with more than 50 chromosomes | 100 | 100 | 100 | 100 |
Training set consisted of 100 cases with distribution: 12 T-ALL, 13 E2A-PBX1, 15 TEL-AML1, 11 BCR-ABL, 15 MLL, 13 hyperdiploid with more than 50 chromosomes, and 21 other.
Blinded test set consisted of 32 cases (2 T-ALL, 5 E2A-PBX1, 5 TEL-AML1, 4 BCR-ABL, 5 MLL, 4 hyperdiploid more than 50, and 7 other).
Apparent accuracy determined by 3-fold cross-validation.
True accuracy determined by class prediction on the blinded test set.
Sensitivity indicates (the number of positive cases predicted)/(the number of true positives).
Specificity indicates (the number of negative cases predicted)/(the number of true negatives).