Table 3

Prospective comparator studies of computed tomography (CT) and ¹⁸f-fluoro-2-deoxy-d-glucose positron emission tomography (FDG-PET) in mediastinal lymph node staging in non-small cell lung cancer, reporting the results per patient

First author [Ref.]	Year	n	Method of analysis	Sensitivity %	Specificity %	Accuracy %	Prevalence %	p-value
Saunders 70	1999	84	Independent	71	97	92	18	NR
			CT	20	90	77
Valk 71	1995	76	Complementary	83	94	91	32	<0.01
			CT	63	73	70
Vansteenkiste 72	1998	68	Complementary	93	95	94	41	0.0004
			CT	75	63	68
Kerstine 73	1999	64	Independent	70	86	84	25	<0.001
			CT	65	79	76
Vansteenkiste 74	1998	56	Fusion images	89	82	86	50	0.04
			Complementary	89	82	86
			CT	86	43	64
Vansteenkiste 75	1997	50	Independent	67	97	88	30	0.004
			Complementary	93	97	96
			CT	67	59	64
Bury 76	1996	50	Independent	90	86	NR	32	<0.05
			CT	72	81	NR
Weng 77	2000	50	Independent	73	94	87	38	0.03
			Complementary	82	96	91
			CT	73	77	76
Steinert 78	1997	47	Independent	92	97	96	28	0.01
			CT	54	88	79
Patz 79	1995	42	Independent	92	100	95	NR	NR
			CT	58	80	68
Sazon 30	1996	32	Complementary	100	100	100	50	0.01
			CT	81	56	69
Chin 80	1995	30	Complementary	78	81	80	30	NR
			CT	56	86	77
Scott 81	1996	27	Complementary	100	100	100	33	0.03
			CT	67	83	78
Albes 82	1999	24	Independent	77	79	78	66	NR
			CT	77	79	74
Wahl 83	1994	19	Independent	82	81	81	NR	<0.05
			CT	64	44	52

Independent: interpretation of PET without CT; Complementary: interpretation with aid of CT; Fusion images: digital fusion of CT and PET images; prevalence: proportion of cases with diseased mediastinal lymph nodes in the total group of studied patients; p-value: significance of the difference in performance between CT and PET, usually examined with a McNemar test; NR: not reported.