Reader and case sample combinations needed for 0.80 power based on the data of:

Krupinski EA, Johnson J, Roehrig H, Engstrom M, Fan J, Nafziger J, Lubin J, Dallas WJ. Using a human visual system model to optimize soft-copy mammography display: influence of MTF compensation. Academic Radiology 2003;10:1030-1035.

Goal: measure observer performance on soft-copy display of mammographic images that were either unprocessed or processed to compensate for modulation transfer function (MTF) deficiencies in the CRT display.

Readers: 6 radiologists

Cases: 250 mammographic regions of interest (256 x 256) with micro-calcification clusters with different contrast levels on a CRT monitor -- 50 lesion free (0% contrast) and 200 of which had lesions (50 at each contrast level 25%, 50%, 75%, and 100%).

Rating: confidence in the presence of a microcalcification cluster on a 6-point scale.

Conditions: (1) without image processing, (2) with processing to compensate for MTF deficiencies in the CRT monitor.

ROC Model = propROC, Parameter = AUC,

AUCs = 0.833, 0.805, respectively.

Readers = 6, Cases = 250,

MS( Treatment*Reader ) = 0.01148499,

MS( Treatment*Case ) = 0.11132085,

MS( Treatment*Reader *Case ) = 0.10809497.

 

AUC Difference = 0.03

Readers	Readers and Cases Random	Cases Random	Readers Random
3	246	640	253
4	220	483	223
5	197	389	199
6	179	326	179
7	165	281	163
8	152	247	150
9	142	221	139
10	133	200	129
11	125	183	121
12	118	169	114
13	112	157	107
14	106	146	101
15	101	138	96

 

 

AUC Difference = 0.05

Readers	Readers and Cases Random	Cases Random	Readers Random
3	208	232	215
4	161	175	164
5	130	142	131
6	109	119	109
7	95	103	94
8	83	91	82
9	75	81	73
10	68	74	66
11	62	68	60
12	57	62	55
13	53	58	51
14	50	54	48
15	47	51	45

 

 

AUC Difference = 0.10

Readers	Readers and Cases Random	Cases Random	Readers Random
3	122	60	127
4	72	46	73
5	50	37	51
6	39	32	39
7	32	28	32
8	27	25	27
9	24	22	23
10	21	20	20
11	20	20	20
12	20	20	20
13	20	20	20
14	20	20	20
15	20	20	20

 

Cautionary note:  For this data set the estimated required number of cases when only cases are treated as random sometimes considerably exceeds the required number when both cases and readers are treated as random. If the variance components were all known, this would not be possible. It happens here because we are using unbiased estimates of the variance components rather than the true unknown parameter values; in particular, it happens because the treatment-by-reader variance component estimate is negative.  In this situation a conservative approach is to rerun the program with the treatment-by-reader variance component estimate set to zero.  Alternatively, in this situation you may want to consider pooling information from several similar studies, resulting in more precise variance component estimates and hence more precise sample size estimates.  The reader is referred to the Details section of the MRMC Sample Size User's Guide for further information.