Table 2.

Test accuracy of CEUS-guided SLN biopsy using intradermally injected microbubbles in females with normal (Centres 1 and 2) or indeterminate results from previous greyscale ultrasound (with or without previous biopsy)

Centre	Prevalence of LN metastases	TP	FP	TN	FN	Sensitivity	Specificity	PPV	NPV
1	22%	84	2	635	95	46.9% (39.4–55.5%)	99.7% (98.95–100%)	97.7% (91.9–99.7%)	87.0% (84.3–89.3%
2	27%	31	3	153	28	52.5% (39.1–65.7%)	98.1% (94.5–99.6%)	91.2% (76.3–98.1%)	84.5% (78.4–89.5%)
3	35%	13	0	52	15	46.4% (27.5–66.1%)	100% (93.2–100%)	100% (75.3–100%)	86.9% (82.4–90.3%)
4	29%	5	1	26	6	45.5% (16.7–76.6%)	96.3% (81.0–99.9%)	77.6% (31.3–96.3%)	86.2% (78.4–91.5%)

ALND, axillary lymphnode dissection; CEUS, contrast enhanced ultrasound; FP, falsepositive; LN, lymph nodes; NPV, negative predictive value; PPV, positivepredictive value; TN, true negative; TP, true positive; SLN, sentinel lymphnodes; SLNE, sentinel lymph node excision.

Reference standard is SLNE or ALND. PPV and NPV were calculated directly for Centres 1 and 2 thatemployed consecutive recruitment, and using Bayesian methods with 22% prevalenceat Centres 3 and 4 as these were not a consecutivelyor randomly selected group.