EJNMMI Physics

Table 2 Statistical summary of metrics used in this study

From: Data-driven gated PET/CT: implications for lesion segmentation and quantitation

	PET/ACT			DDG-PET			DDG-PET/CT
	\(\tilde{x }\)± σ	Range	p	\(\tilde{x }\)± σ	Range	p	\(\tilde{x }\)± σ	Range	p
Metric ratio
Volume	1.00 ± 0.58	0.38–3.38	0.54	0.64 ± 0.26	0.20–1.61	***	0.74 ± 0.40	0.37–2.39	***
SUV_max	1.20 ± 0.43	0.90–3.30	***	1.26 ± 0.23	0.58–1.81	***	1.66 ± 0.66	1.07–4.48	***
SUV_mean	1.17 ± 0.40	0.93–3.28	***	1.19 ± 0.18	0.53–1.81	***	1.55 ± 0.52	1.05–3.50	***
LG	1.26 ± 1.23	0.46–6.21	*	0.78 ± 0.32	0.11–2.17	***	1.13 ± 1.12	0.46–7.23	0.19
TLG	1.12 ± 0.32	0.80–2.04	0.41	0.82 ± 0.15	0.54–1.13	**	1.10 ± 0.38	0.79–2.08	0.34
Metric^†
CCD	1.6 ± 2.5	0.2–15.1	‒	3.9 ± 3.6‡	0.3–18.2	***	4.6 ± 3.8‡	0.4–21.0	***
DICE	0.80 ± 0.15	0.26–0.94	‒	0.67 ± 0.18‡	0–0.94	***	0.64 ± 0.17‡	0–0.89	***
Lesion ΔHU	61.6 ± 311	− 65–906	***	− 8.7 ± 157	− 859–49	*	85.6 ± 352	− 64–953	***

\(\tilde{x }\) = median, σ = standard deviation, LG = lesion glycolysis, TLG = total lesion glycolysis, CCD = centroid-to-centroid distance (mm), DICE = DICE similarity coefficient, ΔHU = change in mean HU
All metric ratios and tests of significance (p) are relative to baseline PET/CT except for ^‡p relative to PET/ACT
^†CCD, DICE, and ΔHU values are calculated relative to baseline PET/CT lesion contours
*p < 0.01, **p < 0.001, ***p < 0.0001

Back to article page