Chemotherapy for T1 pN0M0 Breast Cancer Patients: A Propensity Score Matching Study Based on SEER Database and External Data


 Background

There is no definitive, unified view on chemotherapy for T1 pN0M0 breast cancer. Our study explored the effects of chemotherapy on T1 pN0M0 breast cancer.
Methods

75,139 patients diagnosed with T1 pN0M0 breast cancer were selected from the Surveillance, Epidemiology, and End Results (SEER) database. Multivariate Cox analyses were performed to investigate the effects of chemotherapy on T1a, T1b, and T1c pN0M0 breast cancer, various tumor grades, and four molecular subtypes. Propensity score matching (PSM) was used to eliminate confounding factors and further verify the results between chemotherapy and no chemotherapy. Finally, 545 T1pN0M0 breast cancer patients treated at the Northern Jiangsu People’s Hospital were included for external validation. Univariate and multivariate Cox analyses were used to confirm the role of chemotherapy in T1a, T1b, and T1c pN0M0 breast cancer. Survival curves were plotted using the Kaplan–Meier method for tumor grades and molecular subtypes.
Results

Chemotherapy demonstrated a statistically significant improvement in T1b and T1c breast cancer, not in T1a breast cancer. With T1b breast cancer, chemotherapy had effects on grade III and molecular subtypes hormone receptor+ [HR+]/human epidermal growth factor receptor 2+ [HER2+], HR-/HER2+, and HR-/HER2-. Chemotherapy was beneficial to overall survival for grade II/III and T1c breast cancer. After PSM, identical results were obtained. We also obtained similar results with external validation, except that chemotherapy made a difference in grade II and T1b breast cancer of external validation.
Conclusion

Partial T1 pN0M0 breast cancer patients with tumor grade III T1b pN0M0 except HR+/HER2-, those with tumor grade II and III T1c pN0M0 can obtain overall survival benefits from chemotherapy.


Introduction
Breast cancer is the most common female malignancy and the leading cause of cancer-related death worldwide. 1 The incidence of breast cancer, especially in the early stage, has been increasingly observed in recent decades because of the advancements in and the widespread use of imaging and longer life expectancy. [2][3][4][5] Small tumors without lymph node in ltration have been given much attention by physicians. These cases are considered to have a good prognosis even though patients undergo surgery without adjuvant therapy. T1 pN0M0 breast cancer has been shown to have a relatively low risk of death and recurrence. 6,7 According to the eighth edition of the American Joint Committee on Cancer Staging manual, T1 pN0M0 breast cancer includes tumors smaller than 2 cm without node involvement. These tumors are subdivided into three groups: T1a (≤ 0.5 cm), T1b (> 0.5 cm but ≤ 1.0 cm), and T1c (> 1.0 cm but ≤ 2.0 cm). 8 Breast cancer is a heterogeneous disease with distinct therapeutic responses consisting of four various molecular subtypes. 9 Considering the results of international reviews, adjuvant therapy for breast cancer can reduce deaths by approximately 25% across all risk groups. 10 Adjuvant systematic therapy includes chemotherapy, radiotherapy, endocrine therapy, and targeted therapy. Although chemotherapy is an effective treatment, it is also toxic to humans and has shortterm and long-term side effects that can result in death. 11 Patients with early stage breast cancer (EBC) can achieve only a small absolute percentage of survival advantages from chemotherapy. 12 Therefore, it is crucial to identify which patients with EBC require adjuvant chemotherapy. When considering chemotherapy, the possible risks are compared to the possible bene t. Chemotherapy can not only reduce the risk of recurrence and death but also reduce the damage caused by toxic effects and unnecessary expenses. Nevertheless, without advanced evidence, the possible exclusions from chemotherapy are controversial. Therefore, our study aimed to evaluate the effects of adjuvant chemotherapy on the overall survival of T1 pN0M0 breast cancer patients.

Data acquisition and patient selection
The Surveillance, Epidemiology, and End Results (SEER) database was established in 1973 by the National Cancer Institute and is one of the largest tumor registration databases. Data were selected from 18 registries of the SEER program, which identi ed T1 pN0M0 female breast cancer patients from 2010 to 2014. Then, exclusion was performed for the following patients: those who did not undergo surgery; those with ductal carcinoma in situ and lobular carcinoma in situ; those who participated in less than 3 months of follow-up; those with unknown data; and those with other primary malignant tumors.
We collected the data of 75,139 EBC patients from the SEER database. During external validation, the same inclusion and exclusion criteria were used for patients treated at Northern Jiangsu People's Hospital from 2010 to 2015. Finally, 545 female T1 pN0M0 breast cancer patients were included in the study.
Our study was approved by the ethics committee of Northern Jiangsu People's Hospital, Clinical Medical School, A liated Hospital of Yangzhou University.
The following data of each patient were gathered: patient number, age, death status (yes/no), follow-up time, tumor size, grade, estrogen receptor (ER) status, progesterone receptor (PR) status, human epidermal growth factor receptor-2 (HER-2), surgical method, chemotherapy history, and radiation history. Because chemotherapy-related toxicities can lead to death, we chose overall survival (OS) as the endpoint instead of breast cancer-speci c survival. 13 An age of 60 years was regarded as the threshold for distinguishing young and old patients. 14 2.2 SEER methods T1 pN0M0 breast cancer patients were divided into three groups: T1a (n = 10,073); T1b (n = 24,951); and T1c (n = 40,115).
We conducted a descriptive analysis of the baseline clinical features of eligible patients and used the chi-square test to compare the characteristics of patients among the three groups. The multivariable Cox regression analysis was performed to calculate the effects of chemotherapy on T1a, T1b, and T1c. We also performed a multivariable Cox regression analysis of diverse tumor grades and molecular subtypes. To further con rm the speci c role of chemotherapy for those in the three groups and four molecular subtype subgroups, a multivariable Cox regression analysis of tumor grades was performed. Furthermore, propensity score matching (PSM) was performed to balance disparities between chemotherapy and no chemotherapy. The propensity score for chemotherapy status was calculated for each patient using multivariate logistic regression and considered the following imbalanced factors: tumor grade, surgery type, radiation record, molecular subtype, and age. PSM was conducted using a propensity score constructed using 1:1 nearest-neighbor matching within calipers without replacement. 15 After PSM, to obtain more accurate results, a multivariable Cox regression analysis was repeated to assess the usefulness of chemotherapy for the tumor grades and molecular subtypes of the three groups.

External validation methods
EBC patients comprised three groups: T1a (n = 98), T1b (n = 130), and T1c (n = 317). Because of a lack of data, we used the univariable Cox regression analysis to identify the signi cant variables and remove some inconsequential parameters. These meaningful indicators were incorporated in the multivariable Cox regression analysis to further evaluate the three groups. In each of the three groups, Kaplan-Meier survival curves of tumor grades and molecular subtypes were plotted and estimated log-rank test results were used to compare the control group (no chemotherapy) and the experimental group (chemotherapy).

Statistical methods
Data analyses were performed using R software version 3.5.1 (R Foundation for Statistical Computing, Vienna, Austria) and SPSS version 19.0 (IBM Corporation, Armonk, NY, USA). Two-tailed P < 0.05 was considered statistically signi cant.

Discussion
In spite of the dramatic increase in the number of EBC patients, [3][4][5]16 chemotherapy for T1 pN0M0 breast cancer remains controversial. Therefore, it is imperative to establish a safe, speci c, and effective chemotherapy strategy to guide treatment and improve the prognosis of these patients.
In addition to creating the strategy, we utilized PSM and external validation to verify the association between chemotherapy and signi cant survival advantages obtained by breast cancer patients with T1 pN0M0. Chemotherapy is recognized as a primary systematic adjuvant modality; however, it negatively in uences survival and reduces the quality of life because of its short-term toxicities, including alopecia, nausea, vomiting, and fatigue, and potential long-term effects, including myelosuppression, cardiovascular toxicity, neurotoxicity, marrow neoplasm, and cessation of menses and fertility. [17][18][19][20] EBC patients are expected to survive their cancer diagnosis; however, it is not appropriate to select breast cancer-speci c mortality as an endpoint. Therefore, adjuvant therapy has been crucial for observing improvements in the OS of breast cancer patients. 21,22 Mortality should be considered because it can be caused by toxicities related to chemotherapy. Furthermore, our study suggests that chemotherapy possibly accelerates death for some HR+/HER2-T1ab pN0M0 patients.
Postmastectomy radiation therapy is widely considered to reduce the risk of local recurrence and breast cancer mortality, especially for patients with locally advanced breast cancer who are at high risk because of large tumors and axillary lymph node involvement. 23 However, the majority of T1 pN0M0 breast cancer patients prefer to undergo breastconserving surgery instead of mastectomy, and adjuvant radiotherapy is an indivisible locoregional treatment for breastconserving surgery that achieves local control bene ts and OS advantages. [24][25][26][27][28][29] These results are consistent with the results of our study.
There was an obvious contradiction in the results obtained using the SEER database and those of external validation.
The external validation results indicated that patients with grade II T1b pN0M0 can acquire survival bene t from chemotherapy, but the SEER database results did not. There are two interpretations of this phenomenon. First, the data used for external validation are relatively limited; therefore, inevitable deviations may have occurred during statistical analyses. Second, the two cohorts of data were derived from China and the United States. Admittedly, the factors that affect a patient's lifetime vary from country to country and include cultural barriers, ethnic differences, and genetics. 30 Consequently, the nal conclusions refer to the results obtained using the SEER database.
Our research results, the guidelines of the National Comprehensive Cancer Network (NCCN), and the guidelines of the St.
Gallen International Breast Cancer Conference (BCC) are nearly identical. 31,32 The NCCN suggests the following: 1. For node-negative HR+/HER2-breast cancer, if the tumor is 0.5 cm or smaller, chemotherapy is not recommended. If the tumor is larger than 0.5 cm, then performing a 21-gene reverse-transcription polymerase chain reaction assay (Oncotype DX) is strongly considered. [33][34][35] a. If the recurrence score is ≥ 31, the risk of recurrence is high and chemotherapy is recommended.
b. If the recurrence score is 26-30, the risk of recurrence is moderate and the decision to perform chemotherapy is based on other clinical factors.
c. If the recurrence score is < 26, the risk of recurrence is low and chemotherapy is not recommended.
2. For node-negative HR+/HER2 + breast cancer, if the tumor is 1.0 cm or smaller, it is uncertain whether chemotherapy is required. However, chemotherapy is recommended for T1a category 2B, which means that there is an NCCN consensus that intervention is appropriate based on lower-level evidence. If the tumor is larger than 1.0 cm, chemotherapy is recommended.
3. For node-negative HR-/HER2 + breast cancer, chemotherapy is recommended. However, chemotherapy is recommended for category 2B when the tumor is smaller than 0.5 cm.
4. For node-negative HR-/HER2-breast cancer, if the tumor is smaller than 0.5 cm, chemotherapy is not recommended. However, chemotherapy is necessary for any other case.  40,41 We incorporated tumor grade, which is an independent prognostic indicator, to assess the effects of chemotherapy. [42][43][44] Patients, including those with TNBC, can be exempt from adjuvant chemotherapy if they have grade I/II T1b pN0 and grade I T1c pN0 breast cancer.
There were several limitations to our study. First, the SEER database was devoid of variables for genes and therapies, such as the 21-gene assay and schemes and dosages of chemotherapy and endocrine therapies. Second, because the endpoint was OS, age was a signi cant factor that could not be included to evaluate the effects of chemotherapy. Third, because we used a retrospective cohort population, inevitable selection bias might have affected the conclusions. Further large-scale, prospective, randomized, controlled trials are warranted to accurately identify the outcomes.

Conclusion
Our study found that chemotherapy is not useful and may even be detrimental to T1a pN0 breast cancer patients. Moreover, chemotherapy is recommended for tumor grade III T1b pN0 and grade II/III T1c pN0 breast cancer, but not with HR+/HER2-. Regarding the molecular subtype HR+/HER2-, in the absence of genetic testing, adjuvant chemotherapy is suggested for tumor grade II and grade III T1c pN0 breast cancer.

Declarations
Ethics approval and consent to participate The SEER Program collects data from population-based cancer registries with anonymous information. The SEER is a publicly available database and data extracted from SEER was deemed "non-human study" by the North Shore LIJ IRB committee.
This article is a retrospective review of patient data, and the study design was approved by the Medical Ethics Committee of Northern Jiangsu People's Hospital. This article does not contain any studies with animals or human participants performed by any of the authors.
For this type of study formal consent is not required, and the need for informed consent was waived by the ethics committee.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.