Automation of immunohistochemical evaluation in breast cancer using image analysis

Table 1 Comparison of manual grading and image analysis of estrogen receptor

Manual grading	Image analysis			Total
	Grade 0	Grade 1	Grade 2
Grade 0	30	6	6	42
Grade 1	0	2	2	4
Grade 2	0	0	14	14
Total	30	8	22	60

Table 2 Comparison of manual grading and image analysis of progesterone receptor

Manual grading	Image analysis			Total
	Grade 0	Grade 1	Grade 2
Grade 0	35	6	2	43
Grade 1	0	2	5	7
Grade 2	0	0	10	10
Total	35	8	17	60

Table 3 Comparison of manual grading and image analysis of human epidermal growth factor receptor-2

Manual grading	Image analysis				Total
	Grade 0	Grade 1	Grade 2	Grade 3
Grade 0	6	0	1	0	7
Grade 1	0	0	0	0	0
Grade 2	0	0	2	5	7
Grade 3	0	0	0	46	46
Total	6	0	3	51	60

Table 4 Results of the classification of estrogen receptor, progesterone receptor and human epidermal growth factor receptor-2 expression images

	True positive	True negative	False positive	False negative	Sensitivity (%)	Specificity (%)	PPV1 (%)	NPV2 (%)
ER expression	14	38	8	0	100	82.6	63.6	100
PR expression	10	43	7	0	100	86.0	58.8	100
HER-2/neu expression	46	9	5	0	100	64.3	90.2	100

¹PPV: Positive predictive value;

²NPV: Negative predictive value. It can be seen that the sensitivity has been maintained at 100%. The specificity of the evaluations for ER and PR is good. The specificity of HER-2/neu expression is comparatively less than the ER and PR expression evaluations. ER: Estrogen receptor; PR: Progesterone receptor; HER-2/neu: Human epidermal growth factor receptor-2.

Table 5 Spearman correlation coefficient (95% confidence interval) of the different evaluations

	Expert vs automated analysis1	Expert vs pathologist2	Pathologist reading 1 vs pathologist reading 23
ER expression	0.73 (0.59 to 0.83)	0.53 (0.32 to 0.69)	0.46 (0.24 to 0.64)
PR expression	0.82 (0.73 to 0.90)	0.63 (0.45 to 0.76)	0.66 (0.49 to 0.78)
HER-2/neu expression	0.92 (0.88 to 0.96)	0.68 (0.52 to 0.80)	0.70 (0.54 to 0.81)

¹Represents the correlation of the automated analysis with the expert’s evaluation;

²Represents the correlation between two pathologists;

³Represents the correlation between readings by the same pathologists taken at different time periods. It can be seen that the Spearman correlation coefficients for evaluations by automated analysis are better for all receptor expressions.