背景：BCI-HOXB13/IL17BR 比值（BCI-H/I）已被證明可以作為預測內分泌 tx（ET）和延長內分泌獲益指標。在 NSABPB-42 研究中，我們評估了 BCI-H/I EET 在已完成 5 年內分泌治療的HR+乳腺癌患者中作為延長來曲唑內分泌治療（ELT）獲益的預指標的作用。

Background: The BCI HOXB13/IL17BR ratio (BCI-H/I) has been shown to predict endocrine tx (ET) and extended ET (EET) benefit. We examined the effect of BCI-H/I for EET benefit prediction in NSABP B-42, evaluating extended letrozole tx (ELT) in HR+ breast cancer patients (pts) who completed 5 yrs of ET.

方法：所有可獲取原發腫瘤組織的乳腺癌患者均可入組，主要研究終點為：無複發間隔（RFI）。次要終點是遠處複發（DR）、乳腺癌無瘤間隔（BCFI）和無病生存率（DFS）。采用 COX 比例風險模型分析，由於延長內分泌治療 ELT 對遠處複發的影響不成比例，進一步進行了時間相關的二次分析（≤4y，＞4y）。似然比檢驗評估 BCI-H/I 相互作用。

Methods: All pts with available primary tumor tissue were eligible. Primary endpoint was recurrence-free interval (RFI). Secondary endpoints were distant recurrence (DR), breast cancer-free interval (BCFI), and disease-free survival (DFS). Stratified Cox proportional hazards model was used. Due to a non-proportional effect of ELT on DR, time-dependent secondary analyses (≤4y, >4y) were performed. Likelihood ratio test evaluated treatment by BCI-H/I interaction.

結果：共分析 2179 例患者（60%患者 N0；62%患者僅使用過 AI；80%為 HER2-，45%為高 BCI-H/I，55%患者為低 BCI-H/I）。延長內分泌治療顯示）10 年 RFI 總體獲益為 1.6% （HR=0.77, 95% CI 0.57-1.05, p=0.10），其中（ （BCI-H/I-Low 組 1.1% [HR=0.69, 0.43-1.11, p=0.13），BCI-H/I-High組 2.4% [HR=0.83, 0.55-1.26, p=0.38]; interaction p=0.55）， 延長內分泌治療在兩組中 BCFI 無顯著差異（HR=0.53, 0.36-0.78, p=0.001; BCI-H/I-High： HR=0.85, 0.60-1.21, p=0.36，相互作用p=0.07）。DFS（BCI-H/I-Low：HR=0.75，0.58-0.95，p=0.017；BCI-H/I-高：HR=0.81，0.64-1.04，p=0.09；相互作用 p=0.62）。4 年以前，在高乳腺指數患者中延長內分泌治療 DR 獲益無統計學意義。4 年後，BCI-H/I-High 患者延長內分泌治療 DR 獲益 （HR： 0.29, 0.12-0.69, p=0.003）, 而 BCI-H/I-Low 患者中延長內分泌治療對 DR 不太可能獲益 （HR： 0.68, 0.33-1.39, p=0.28）（interaction p=0.14）。

Results: In 2,179 pts analyzed (60% N0; 62% AI only; 80% HER2-), 45% were BCI-H/I-High and 55% BCI-H/I-Low. ELT showed an absolute 10y benefit of 1.6% for RFI (HR=0.77, 95% CI 0.57-1.05, p=0.10) (BCI-H/I-Low: 1.1% [HR=0.69, 0.43-1.11, p=0.13]; BCI-H/I-High: 2.4% [HR=0.83, 0.55-1.26, p=0.38]; interaction p=0.55). There was no statistically significant ELT by BCI-H/I interaction for BCFI (BCI-H/I-Low: HR=0.53, 0.36-0.78, p=0.001; BCI-H/I-High: HR=0.85, 0.60-1.21, p=0.36; interaction p=0.07) or for DFS (BCI-H/I-Low: HR=0.75, 0.58-0.95, p=0.017; BCI-H/I-High: HR=0.81, 0.64-1.04, p=0.09; interaction p=0.62). Before 4y, there was no statistically significant ELT benefit on DR in either BCI-H/I group. After 4y, BCI-H/I-High pts had statistically significant ELT benefit on DR (HR: 0.29, 0.12-0.69, p=0.003), while BCI-H/I-Low pts were less likely to benefit (HR: 0.68, 0.33-1.39, p=0.28) (interaction p=0.14).

	BCI-H/I	≤ 4 y				> 4 y
		HR (95%CI)	4y abs. benefit (%)	P	HR (95%CI)	10y abs. benefit (%)	P
All	All	0.85 (0.51, 1.43)	0.4	0.55	0.47 (0.28, 0.81)	1.7	0.005
	L	0.47 (0.19, 1.17)	1.1	0.10	0.68 (0.33, 1.39)	0.4	0.28
	H	1.21 (0.63, 2.32)	-0.4	0.56	0.29 (0.12, 0.69)	3.6	0.003
N0	L	0.28 (0.03, 2.48)	0.7	0.22	1.05 (0.35, 3.14)	-1.0	0.93
	H	2.03 (0.62, 6.60)	-1.5	0.23	0.32 (0.06, 1.68)	1.8	0.16
N+	L	0.54 (0.19, 1.48)	2.4	0.22	0.49 (0.19, 1.28)	3.5	0.14
	H	0.94 (0.42, 2.09)	0.5	0.87	0.28 (0.10, 0.77)	5.6	0.009
AI Only	L	0.50 (0.15, 1.70)	0.8	0.26	0.70 (0.29, 1.68)	0.6	0.42
	H	1.06 (0.48, 2.37)	0.2	0.89	0.42 (0.15, 1.19)	3.2	0.09
Tam-AI	L	0.44 (0.11, 1.70)	1.7	0.22	0.63 (0.18, 2.24)	0.1	0.47
	H	1.56 (0.51, 4.76)	-1.5	0.43	0.16 (0.04, 0.76)	4.3	0.009

結論：BCI-H/I 作為延長內分泌治療對無複發間隔（RFI）獲益預測指標作用不確定，時間依賴性 DR 分析中，4 年後高 BCI-H/I-Highpts 對預測患者風險指導內分泌延長治療有統計學意義，而BCI-H/I-Lowpts 沒有。BCI-H/I-Low 患者中評估延長內分泌治療的 BCFI 獲益主要由第二原發性乳腺癌引起。在這項研究中，需要進一步的隨訪來進一步確定 BCI-H/I 的預測能力。臨床試驗信息：NCT00382070。

Conclusions: BCI-H/I prediction of ELT benefit on RFI was not confirmed. In time-dependent DR analyses, BCI-H/I-High pts had statistically significant benefit from ELT after 4y, while BCI-H/I-Low pts did not. Observed ELT benefit on BCFI in BCI-H/I-Low pts was primarily driven by second primary breast cancers. Additional follow-up is needed to further characterize BCI-H/I predictive ability in this study .