Table 1.

Performance comparison of major prostate CADx systems published.

Publication	Data size	Modality	Field strength	Ground truth	Candidate generation	Region	Classifier	Performance
Chan et al. 2003 [56]	15	T2WI, ADC, PD, and T2 map	1.5T	MSTR + biopsy	RS + cancer	PZ	FLD	AUC = 0.839 (±0.064)
Puech et al. 2009 [72]	100	DCE	1.5T	Biopsy + RS	Manual ROI	WP	Rule-based	AUC = 0.77
Vos et al. 2008 [54]	34	DCE	1.5T	Biopsy	Manual ROI	PZ	SVM	AUC = 0.83, CI [0.75, 0.92]
Vos et al. 2010 [66]	34	T2WI, DCE	1.5T	Biopsy	Manual ROI	PZ	SVM	AUC = 0.89, CI [0.81, 0.95]
Shah et al. 2012 [40]	31	T2WI, ADC, and DCE	3T	Biopsy	Manual ROI	PZ	SVM	F-measure = 0.89
Liu and Yetik 2011 [64]	20	T2WI, ADC, and DCE	1.5T	Biopsy	Voxel	WP	SVM	AUC = 0.89
Liu et al. 2013 [39]	54	T2WI, ADC, and DCE	3T	Biopsy	Biopsy spots	WP	SVM	AUC = 0.82, CI [0.71, 0.93]
Niaf et al. 2012 [53]	30	T2WI, ADC, and DCE	1.5T	MSTR + biopsy	Manual ROI	PZ	SVM	AUC = 0.89, CI [0.81, 0.94]
Moradi et al. 2012 [52]	29	DT, DCE	3T	Biopsy	Biopsy spots	WP	SVM	AUC = 0.96
Niaf et al. 2014 [73]	49	T2WI, ADC, and DCE	Not specified	MSTR + biopsy	Manual ROI	WP	P-SVM	AUC = 0.889
Peng et al. 2013 [51]	48	T2WI, ADC, and DCE	3T	MSTR + biopsy	Manual ROI	WP	LDA	AUC = 0.95 (±0.02)
Artan et al. 2010 [55]	21	T2WI, ADC, and DCE	1.5T	Biopsy	Voxel	PZ	CRF	AUC = 0.79 (±0.12)
Tiwari et al. 2013 [71]	29	T2WI, MRS	1.5T	MSTR + biopsy	Voxel	WP	SeSMiK-GE + random forest	AUC = 0.89 (±0.09)
Tiwari et al. 2012 [63]	36	T2WI, MRS	1.5T	MSTR + biopsy	Voxel	WP	Random forest	AUC = 0.89 (±0.02)
Litjens et al. 2014 [57]	347	T2WI, PDWI, DCE, and DWI	3T	Biopsy	Voxel	WP	Random forest	AUC = 0.889

PD: proton density; MSTR: manual segmented tumor by radiologist; RS: random sampling; AUC: area under the ROC curve; HMM: hidden Markov models; WP: whole prostate; CI: confidence interval; ROI: region of interest; RG: region growing; FSE: fast spin-echo; DT: diffusion tensor; P-SVM: probabilistic SVM; SeSMiK-GE: semisupervised multikernel graph-embedded, PDWI: proton density-weighted imaging.