Table 3.
Example of a final rule set to determine indication for biopsy
Rule No. | Rule | Decision | Historical No. of patients that correctly satisfy specific rule | Historical No. of patients that correctly satisfy all rules |
---|---|---|---|---|
1 | Composition of nodule is entirely solid | Likely nonbenign—recommend biopsy | 12/12 | 12/12 |
LNs appear normal | ||||
Tumor is unifocal | ||||
2 | Nodule > 50% cystic | Likely benign-defer biopsy | 7/7 | 19/19 |
Margin is regular | ||||
LNs appear normal | ||||
3 | Composition of nodule is entirely solid | Likely nonbenign–recommend biopsy | 4/4 | 23/23 |
Hypoechoic halo is either absent OR complete but not partial | ||||
Margin is irregular/microlobulated/ spiculated | ||||
4 | Composition of nodule is entirely solid | Likely nonbenign–recommend biopsy | 9/10 | 32/33 |
Margin is indistinct | ||||
Tumor is unifocal OR multifocal (unilaterally) | ||||
5a | Composition of nodule is mixed solid/ cystic < 50% cyst | Likely benign–defer biopsy | 10/11 | 42/44 |
Hypoechoic halo is absent OR complete (ie, not absent) | ||||
Tumor is unifocal OR multifocal (bilaterally) | ||||
6 | Composition of nodule is entirely solid | Likely nonbenign—recommend biopsy | 6/7 | 48/51 |
LNs are enlarged but normal appearing OR are suspicious for metastasis | ||||
7a | LNs are not visualized or are visualized contralateral to primary tumor (but not ipsilaterally) | Likely benign—defer biopsy | 3/4 | 51/55 |
Tumor is multifocal (unilaterally) | ||||
8 | Composition of nodule is entirely solid | Likely nonbenign—recommend biopsy | 3/4 | 54/59 |
Margin is indistinct | ||||
Tumor is multifocal and bilateral | ||||
9 | Otherwise | Likely nonbenign—recommend biopsy | 3/8 | 57/67 |
Abbreviation: LN, lymph node.
aThese misclassification are the least acceptable type of error—patients classified as likely benign when they were not.