Towards defining morphologic parameters of normal parous and nulliparous breast tissues by artificial intelligence

Ogony, Joshua; de Bel, Thomas; Radisky, Derek C.; Kachergus, Jennifer; Thompson, E. Aubrey; Degnim, Amy C.; Ruddy, Kathryn J.; Hilton, Tracy; Stallings-Mann, Melody; Vachon, Celine; Hoskin, Tanya L.; Heckman, Michael G.; Vierkant, Robert A.; White, Launia J.; Moore, Raymond M.; Carter, Jodi; Jensen, Matthew; Pacheco-Spann, Laura; Henry, Jill E.; Storniolo, Anna Maria; Winham, Stacey J.; van der Laak, Jeroen; Sherman, Mark E.

doi:10.1186/s13058-022-01541-z

Table 3 Comparison of AI data for parous women by time since last birth versus nulliparous women

From: Towards defining morphologic parameters of normal parous and nulliparous breast tissues by artificial intelligence

Outcome	Association measure	Comparison between N = 103 parous women with a time since last birth ≤ 5 years and N = 545 nulliparous women (reference group)				Comparison between N = 77 parous women with a time since last birth > 5 years and N = 545 nulliparous women (reference group)
		Age-adjusted analysis		Adjusting for age, race, BMI, percent fat, and relative with breast or ovarian cancer^c		Age-adjusted analysis		Adjusting for age, race, BMI, percent fat, and relative with breast or ovarian cancer^c
		Estimate (95% CI)	P value	Estimate (95% CI)	p value	Estimate (95% CI)	p value	Estimate (95% CI)	p value
TDLU count^a	Multiplicative effect on mean	1.40 (1.12, 1.76)	0.003	1.27 (1.02, 1.57)	0.031	1.90 (1.45, 2.48)	< 0.001	1.66 (1.28, 2.15)	< 0.001
Adipose tissue frac^b	Additive effect on mean	− 0.02 (− 0.07, 0.04)	0.52	0.02 (− 0.01, 0.04)	0.27	− 0.00 (− 0.07, 0.06)	0.94	0.01 (− 0.03, 0.04)	0.71
Mean acini count^a	Multiplicative effect on mean	1.98 (1.67, 2.36)	< 0.001	1.91 (1.60, 2.28)	< 0.001	1.16 (0.95, 1.42)	0.14	1.14 (0.93, 1.41)	0.21
Dilated acini	Multiplicative effect on mean	2.18 (1.70, 2.80)	< 0.001	2.07 (1.59, 2.70)	< 0.001	1.14 (0.87, 1.50)	0.34	1.11 (0.84, 1.48)	0.46
Mean avg acini^b	Additive effect on mean	− 0.44 (− 0.55, − 0.32)	< 0.001	− 0.46 (− 0.57, − 0.34)	< 0.001	− 0.33 (− 0.47, − 0.19)	< 0.001	− 0.30 (− 0.45, − 0.16)	< 0.001
Mean cap size^b	Additive effect on mean	0.16 (− 0.02, 0.34)	0.081	0.17 (− 0.02, 0.36)	0.078	− 0.05 (− 0.26, 0.16)	0.62	− 0.04 (− 0.26, 0.18)	0.70
Mean epi size^b	Additive effect on mean	0.12 (− 0.05, 0.29)	0.16	0.11 (− 0.06, 0.29)	0.21	− 0.12 (− 0.31, 0.07)	0.22	− 0.12 (− 0.32, 0.08)	0.24
Mean epi stroma ratio^b	Additive effect on mean	− 0.01 (− 0.03, 0.01)	0.41	− 0.01 (− 0.03, 0.01)	0.40	− 0.02 (− 0.05, 0.00)	0.097	− 0.02 (− 0.05, 0.01)	0.25
Mononuclear cell count^a	Multiplicative effect on mean	1.85 (1.58, 2.18)	< 0.001	1.82 (1.54, 2.14)	< 0.001	1.06 (0.89, 1.27)	0.51	1.05 (0.87, 1.27)	0.61
Mean nearby fat^b	Additive effect on mean	0.68 (− 3.80, 5.15)	0.77	1.70 (− 2.40, 5.80)	0.42	0.01 (− 5.18, 5.19)	1.00	0.30 (− 4.54, 5.15)	0.90
Mean TDLU size^b	Additive effect on mean	0.17 (0.01, 0.32)	0.038	0.16 (− 0.00, 0.32)	0.055	0.03 (− 0.15, 0.22)	0.70	0.02 (− 0.17, 0.21)	0.83

CI = confidence interval; adp = adipose; avg = average; cap = capillary; epi = epithelial; TDLU = terminal duct lobular unit
^aNegative binomial regression models were used for comparisons between parous and nulliparous women; multiplicative effects on the mean and 95% CIs were estimated and are interpreted as the multiplicative effect on the mean outcome value.
^bLinear regression models were used for comparisons between parous and nulliparous women; additive effects on the mean and 95% CIs were estimated and are interpreted as the additive effect on the mean outcome value (on the natural logarithm scale for mean avg acini, mean cap size, mean epi size, and mean TDLU size, and on the cube root scale for mean nearby fat).
^cMultivariable models were adjusted for age as well as any variable that differed between parous and nulliparous women with a P value < 0.15. P values < 0.0045 are considered as statistically significant after applying a Bonferroni correction for multiple testing

Back to article page

ISSN: 1465-542X

Contact us

Submission enquiries: journalsubmissions@springernature.com

Breast Cancer Research

Contact us