Reader and case sample combinations needed for 0.80 power based on the data of:
Krupinski EA, Lubin J, Roehrig H, Johnson J, Nafziger J. Using a human visual system model to optimize soft-copy mammography display: Influence of veiling glare. Academic Radiology 2006;13:289-295.
Readers: 6 board certified family physicians
Cases: 160 mammographic images, half containing a single mass
Conditions: images were processed to simulate four levels of veiling glare: (1) none, (2) comparable to a medical grade monochrome curved-screen cathode ray tube (CRT) display, (3) double that of the CRT and (4) quadruple that of the CRT. The images were shown on a liquid crystal display (LCD) that had essentially no veiling glare.
Results: Observer performance declined with increasing veiling glare, with quadruple veiling glare yielding significantly lower performance than the lower veiling glare levels.
ROC Model = PropROC, Parameter = AUC,
AUCs = 0.882, 0.872, 0.867, 0.798, respectively.
Readers = 6, Cases = 160,
MS( Treatment*Reader ) = 0.20555930,
MS( Treatment*Case ) = 0.09767099,
MS( Treatment*Reader *Case ) = 0.09186388.
AUC Difference = 0.03
|
Readers |
Readers
and Cases
Random |
Cases
Random |
Readers
Random |
|
3 |
. |
553 |
. |
|
4 |
. |
420 |
. |
|
5 |
. |
340 |
. |
|
6 |
. |
286 |
. |
|
7 |
. |
248 |
. |
|
8 |
. |
220 |
. |
|
9 |
. |
197 |
. |
|
10 |
. |
180 |
. |
|
11 |
. |
165 |
. |
|
12 |
. |
153 |
. |
|
13 |
. |
143 |
. |
|
14 |
. |
134 |
. |
|
15 |
. |
126 |
. |
Note: Cells containing a period indicate that more than 2000 cases would be required.
AUC Difference = 0.05
|
Readers |
Readers
and Cases
Random |
Cases
Random |
Readers
Random |
|
3 |
. |
201 |
. |
|
4 |
. |
153 |
. |
|
5 |
. |
124 |
. |
|
6 |
. |
105 |
. |
|
7 |
1518 |
91 |
1500 |
|
8 |
420 |
81 |
409 |
|
9 |
247 |
73 |
238 |
|
10 |
177 |
66 |
168 |
|
11 |
138 |
61 |
130 |
|
12 |
114 |
57 |
106 |
|
13 |
97 |
53 |
89 |
|
14 |
85 |
50 |
77 |
|
15 |
76 |
47 |
68 |
Note: Cells containing a period indicate that more than 2000 cases would be required.
AUC Difference = 0.10
|
Readers |
Readers
and Cases
Random |
Cases
Random |
Readers
Random |
|
3 |
. |
52 |
. |
|
4 |
224 |
40 |
234 |
|
5 |
86 |
33 |
87 |
|
6 |
54 |
28 |
54 |
|
7 |
40 |
25 |
39 |
|
8 |
32 |
22 |
31 |
|
9 |
27 |
20 |
25 |
|
10 |
23 |
20 |
22 |
|
11 |
20 |
20 |
20 |
|
12 |
20 |
20 |
20 |
|
13 |
20 |
20 |
20 |
|
14 |
20 |
20 |
20 |
|
15 |
20 |
20 |
20 |
Note: Cells containing a period indicate that more than 2000 cases would be required.