An analysis of invasive micropapillary breast carcinoma based on hormone receptor and epidermal growth receptor 2 status

This research aimed to explore the role of combining hormone receptor (HR) with epidermal growth factor receptor 2 (Her-2) status of invasive micropapillary breast carcinoma (IMPC). 1291 IMPC patients were retrieved from the Surveillance, Epidemiology, and End Results (SEER) database of the US National Cancer Institute. The X 2 test and Fisher’s exact probability test were used to evaluate differences between qualitative variables. Univariate and multivariate Cox proportional hazards regressions were performed to assess the independent prognostic factors on cancer-specic survival (CSS) and overall survival (OS). The nomogram models were made to predict 3-year survival rate of IMPC using R software.


Background
IMPC was a rare special type of breast cancer, 1 composed of small, hollow or morula-like clusters of cancer cells, surrounded by clear stromal spaces. 2 Several studies have illuminated that IMPC had higher rates of positive axillary lymph nodes, lager tumor size, more advanced histological grade than IDC. 3 4 5 6 Despite its highly aggressive clinical performance, a study found that IMPC had a better long-term survival outcome comparing with IDC. 5 A meta-analyses found that IMPC patients were more likely to have local regional recurrence(LRR) than IDC, but the OS and CSS were not signi cantly different between IMPC and IDC. 7 A population based study showed that hormone positive IMPC had a better survival. 8 And prior studies have compared prognosis based on HR or Her-2 status between IMPC and IDC. 9 However, they did not combine these two factors together to analyze its impact on survival of IMPC. Here, in this study, we divided IMPC into four molecular subtypes according to hormone receptor (including estrogen receptor (ER) and progesterone receptor (PR)) and epidermal growth factor receptor 2(Her-2) status to analyze the features of IMPC in a more reasonable way.

Data collection
The SEER*Stat (version 8.3.5) Case Listing session was used to get 1291 cases of IMPC (8507/3: Ductal carcinoma, micropapillary). In consideration of the information on breast cancer molecular subtype was not available until 2010, we selected IMPC patients from 2010 to 2015 from the Surveillance, Epidemiology, and End Results(SEER), version 8.3.5. The deadline for follow-up time was November 2016. The morphological classi cation of breast cancer was Ductal carcinoma, micropapillary according to Oncology, version 3(ICD-O-3). All patients included in this study were female, and patients with unknown subtype, ethnicity, T stage, N stage, M stage and grade were excluded from this study.

Clinicopathological factors
In this study, we used the 7th AJCC version to evaluate TNM (T: tumor size; N: axillary lymph nodes metastasis; M: distant metastasis) and stage. IMPC was classi ed into four molecular types according to the expression of hormone receptor(HR) and epidermal growth factor receptor 2 (HER2): HR+/Her-2-, HR-/Her-2+, HR-/HER2-, HR+/HER2+. According the criterion of SEER database, the de nition of hormone receptor positive was as follows: either ER positive or PR positive, and if 1% or greater cells stained positive, the test results of ER or PR was considered positive. HER2 expression was examined by any of following ways in laboratories: uorescence in situ hybridization (FISH), immunohistochemistry (IHC) or chromogenic in situ hybridization (CISH). Her-2 positive meant it was ampli ed or overexpression in any of these three ways. Surgery type was sorted into three types: mastectomy, breast-conserving surgery, and no surgery or not recommended. And we got the permission to obtain chemotherapy and radiotherapy data of patients from the SEER.
Statistical analysis SPSS Statistics software (version 23.0) was used to analyze the data. X 2 test was used to analyze the characteristics of IMPC. The Kaplan-Meier method was used to construct the OS and CSS of IMPC. The Cox proportional hazards regression model was used to determine the independent prognostic factors for OS and CSS. Hazard ratios (HRs) with corresponding 95% CI were used to show the effect of factors on OS and CSS. A p value of < 0.05 was considered statistically signi cant. To present the prognostic factors of IMPC more clearly, we use nomogram models to show factors that have impacts on OS and CSS of IMPC.

Results
In this study, we made comparisons in four groups using X 2 test: HR+/Her-2-vs. HR-/Her-2-, HR+/Her2-vs. HR+/Her-2+, HR-/Her-2 + vs. HR-/Her-2-, HR-/Her-2 + vs. HR+/Her-2+. The characteristics of four subtypes of IMPC were shown in Table 1. There were 964 (74.7%) IMPC patients belonging to HR+/Her-2-subtype, which accounted for the largest number of population in this study. HR-/Her-2-subtype was the smallest part which accounted for only 4.0%. Most patients were over 50 years old when they were diagnosed with IMPC (80.0%). 614 of 1291 (47.6%) IMPC patients had positive axillary lymph nodes when rst diagnosis which suggested IMPC had a high ratio of lymph node invasion. A mastectomy was performed in 590 (45.7%) patients and 642 (49.7%) patients received conserving surgery.  In addition, we made further analysis of the two surgery types (data showed in Table 2), and we found that comparing with HR+/Her-2-subtype, HR+/Her-2 + and HR-/Her-2-IMPC patients were more likely to receive mastectomy than conserving surgery.  Table 3. We also conducted two nomogram models to predict prognostic factors of invasive micropapillary breast carcinoma's 3-year survival rate (OS, Fig. 1, supplementary Fig. 1; CSS: Fig. 2, supplementary Fig. 2). The nomogram plot showed the calibration was ideal. to its high proportions of axillary lymph node metastases, nodal stage was not a prognostic factor on OS and CSS in this study. Multivariate analysis data was shown in Table 4. Among the four subtypes, HR-/Her-2-subtype IMPC had the worst OS (p < 0.001) and CSS (p < 0.001) (Fig. 3, 4).

Discussion
Breast cancer could be divided into four subtypes according to hormone status, epidermal growth factor 2 status and Ki-67 expression level. In recent years, it has been widely accepted that treatments for breast cancer should be based on different molecular subtypes. Most studies' explorations were based on nonspeci c breast cancer. But IMPC was a special and rare type of breast cancer, which behaviored more aggressively than invasive ductal breast carcinoma.
Previous clinical studies proved that breast cancer with ER and PR positive status was associated with better OS and CSS. 5 10 11 Several studies suggested that molecular subtype of breast cancer provided additional prognostic information regarding patient outcomes. 12 11 13 A population based study showed that hormone positive IMPC had a better survival. 8 Therefore, in this study we used breast cancer molecular type to analyze the impact of combination HR with Her-2 status on IMPC. Our results indicated that HR-/Her-2-subtype had the most unfavorable prognosis.
As was mentioned above, hormone positive status was associated with better OS and CSS, many studies separately analyzed the effects of hormone receptor and HER2 expression status on IMPC. 9 8 In fact, many previous research has proven that molecular subtype can affect prognosis on breast carcinoma. 12 14 15 Here in this study, multivariate analysis indicated that only HR-/Her-2-subtype had a worse prognosis on OS and CSS comparing to HR+/HER2-subtype, suggesting that Her-2 expression status might play an important role, especially when both hormone receptor and Her-2 were negative.
Comparing with other three subtypes, HR-/Her-2-IMPC had more aggressive behaviors, for it had larger tumor size and more advanced histological grade and stage. Tumors which performed more aggressively were more possible to receive surgery, chemotherapy and radiotherapy. And further analysis showed that HR-/Her-2-and HR+/Her-2 + subtype had more proportions to receive mastectomy than HR+/Her-2-subtype, this was might because that HR+/Her-2-status IMPC had better biological behaviors.
We conducted a nomogram models to predict 3 However, this study did not obtain data on speci c chemotherapy regimens which might affect prognosis. And we could not obtain data about whether Her-2 positive patients received targeted therapy (Herceptin) or not. Because the patients with Her-2 expression information only started from 2010, we selected breast cancer patients from 2010 to 2015, which might lead to insu cient follow-up time. It is necessary to follow up longer time to verify our ndings. When typing breast cancer, the factor of Ki-67 is lacking, a factor that is practical for most breast cancers. And it would be better to acquire data about local recurrence rate (LRR), recurrence-free interinterval (RFI) and distant relapse-free survival (DRFS) data to make a more complete analysis of the four subtypes of IMPC.

Conclusion
This population-based study provided evidence that HR-/Her-2-IMPC accounted the smallest part of four subtypes, but were more likely to have larger tumor size and more advanced histological grade. Furthermore, HR+/Her-2-IMPC had most unfavorable prognosis among four subtypes. The only author Yuting Sun agree to submit this manuscript.