Table 5.

Concordance between a software-identified top match and whether two of three clinical geneticists would definitely have considered this diagnosis based on the image alone

Syndrome	Frequency	Correctly classified by neither	Correctly classified by both	Correctly classified by software, but not clinicians	Correctly classified by clinicians, but not software	McNemars Chi-square	p-value	Kappa
Williams	183	31	66	44	42	0.05	0.82925	0.02
Rubinstein-Taybi	155	28	61	31	35	0.24	0.62246	0.11
Floating Harbor	61	7	33	9	12	0.43	0.51269	0.16
Coffin Lowry	154	42	41	47	24	7.45	0.00634	0.10
Kabuki	195	23	88	18	66	27.43	<.00001	0.09
Smith Magenis	124	34	38	34	18	4.92	0.02650	0.17
PACS1	39	33	2	4	0	4.00	0.04550	0.46
Kleefstra	128	52	9	63	4	51.96	<.00001	0.05
Koolan-de Vries	120	30	37	29	24	0.47	0.49221	0.12
Cornelia de Lange	441	52	264	65	60	0.20	0.65472	0.26