An integrated deep learning model for the prediction of pathological complete response to neoadjuvant chemotherapy with serial ultrasonography in breast cancer patients: a multicentre, retrospective study

Table 2 Multivariate analysis of clinicopathological model, Model 2 (the early-treatment model) and Model 3 (the posttreatment model)

	Clinicopathological model		p-value	Model 2		p-value	Model 3		p-value
	OR	95%CI	p-value	OR	95%CI	p-value	OR	95%CI	p-value
Ki67
Negative	Reference
Positive	1.436	0.783–2.633	0.240	1.716	0.815–3.615	0.153	1.985	0.858–4.594	0.107
ER
Negative	Reference
Positive	0.742	0.333–1.657	0.465	0.559	0.212–1.470	0.236	0.700	0.242–2.024	0.509
PR
Negative	Reference
Positive	0.990	0.427–2.302	0.984	0.701	0.258–1.904	0.484	0.768	0.231–2.548	0.664
Her2
Negative	Reference
Positive	0.937	0.078–11.320	0.959	1.062	0.369–3.055	0.911	1.036	0.063–16.92	0.980
Subtype
Triple-negative	Reference
HER2(+)	2.047	0.161–25.97	0.578	1.375	0.083–22.90	0.823	1.271	0.075–21.44	0.867
HER2(−)&HR(+)	1.194	0.361–3.945	0.770	1.275	0.305–5.331	0.738	1.378	0.272–6.987	0.697
P0-Signature	–	–	–	2.357	1.531–3.629	< 0.001	1.881	1.222–2.894	0.004
P1-Signature	–	–	–	2.455	1.721–3.502	< 0.001	2.094	1.468–2.987	< 0.001
P2-Signature	–	–	–	–	–	–	2.097	1.543–2.849	< 0.001

Model 2 was built based on the P0-Signature plus the P1-Signature. And Model 3 was based on Model 2 plus the P2-Signature
ER estrogen receptor, PR progesterone receptor, HER2 human epidermal growth factor receptor, P0 Phase 0 (pretreatment), P1 Phase 1 (early-stage treatment, namely, during the first–second cycle of the neoadjuvant chemotherapy), P2 Phase 2 (posttreatment), OR odds ratio, 95%CI 95% confidence interval

ISSN: 1465-542X