A Methodological Evaluation of Volumetric Measurement Techniques including Three-Dimensional Imaging in Breast Surgery

Table 3

VIZ3D in situ versus VIZ3D sampling (“trained” operator).

(a)

Subject	Side	Volume_in_situ	Volume_resection	diff
P001	Right	384.5	424.7	−40.2
P001	Left	507.7	520.8	−13.1
P002	Right	362.3	436.8	−74.5
P002	Left	276.0	402.3	−126.3
P003	Right	171.5	234.0	−62.5
P003	Left	69.1	156.7	−87.6
P004	Right	285.8	350.1	−64.3
P004	Left	345.1	398.5	−53.4
P005	Right	153.2	133.2	20.0
P005	Left	108.3	173.0	−64.7
P006	Right	168.9	205.0	−36.1
P006	Left	123.8	165.4	−41.6
P007	Right	525.7	574.9	−49.2
P008	Right	244.7	248.2	−3.5
P008	Left	230.3	240.3	−10.0
	Mean ± SD	264±140	311±142	−47.1±36.5
	Median	245	248	−49.2
	P25–P75	153–362	173–425	−64.7–−13.1
	SD (robust)	155	186	38.2

(b)

Method 1	Method 2	𝑁	ICC	Difference	Paired Student’s 𝑡-test	Wilcoxon	CV (%)	Bland-Altman
			(ICC*)		𝑃 value	𝑃 value		𝑟 (𝑃 value)
In situ	Resection	15	0.92 (0.42)	−47.1±36.5	0.0002	0.0004	14.5	Pearson: 𝑟=−0.06  (𝑃=0.82)
								Spearman: 𝑟=0.04  (𝑃=0.89)

We observed a significant difference 47.1±36.5 cm³  (𝑃=0.0002) between the two methods. The concordance between the two methods is very good (0.92) but the lower confidence limit is not particularly high.