Skip to main content

Table 7 Diagnostic performances of the clinical model, radiomic model and RNWCF in the validation and external test cohorts

From: Noninvasive prediction of node-positive breast cancer response to presurgical neoadjuvant chemotherapy therapy based on machine learning of axillary lymph node ultrasound

Data set

Models

Accuracy (95% CI)

Sensitivity (95% CI)

Specificity (95% CI)

Validation cohort

Clinical

0.809 (0.757ā€“0.860)

0.814 (0.750ā€“0.875)

0.795 (0.684ā€“0.896)

Radiomics

0.706 (0.640ā€“0.772)

0.711 (0.630ā€“0.781)

0.692 (0.564ā€“0.821)

RNWCF

0.846 (0.794ā€“0.897)

0.907 (0.853ā€“0.955)

0.692 (0.568ā€“0.812)

External test cohort

Clinical

0.802 (0.716ā€“0.877)

0.808 (0.708ā€“0.893)

0.793 (0.656ā€“0.920)

Radiomics

0.728 (0.642ā€“0.815)

0.865 (0.787ā€“0.941)

0.483 (0.333ā€“0.647)

RNWCF

0.778 (0.691ā€“0.852)

0.731 (0.623ā€“0.830)

0.862 (0.750ā€“0.963)