Differential Second Primary Malignancy Occurrence After Breast Cancer According to HER2 Status: a SEER Population-Based Analysis

Background: The survival improvement in breast cancer (BC) renders the long-termsurvivorsan increasedprobability of second primary malignancy (SPM), and thus excess mortality. Although previous evidence has indicated various predictorsofSPM, little is known whether SPM incidencevaries by HER2 status of rst BC. Methods: Based on BC patients registered between 2010-2018 in the NCI SEER database, we utilized standardized incidence ratio (SIR) and Poisson regression to quantify SPM occurrence compared with the general population. Then, adjusted for competing death risk, cumulative incidence function and Gray’s test were adopted to estimate the probability of SPM. Subsequent proportional subdistribution hazards regression was executedto identify the HER2 status impact on SPM risk. Finally, survival analysiswas performed. Results: A total of 409,796 rst BC patients were includedand 18,283 were identied with at least one SPM. The SIR of SPM after HER2+ BC was signicantly lower than HER2- BC (1.03 vs 1.13; RR, 0.92; 95% CI, 0.88-0.96; p<0.001). But the predominantly declining SPM risk was only observed for second BC (RR, 0.89; 95% CI, 0.82-0.96; p=0.003) and lung and bronchus cancer (RR, 0.84; 95% CI, 0.74-0.95; p=0.007). Further competing risk analysis veried the protective effect of HER2 positivity status on SPM occurrence.The 5-year cumulative incidence of SPM following HER2+ and HER2- BC were 5.16% and 4.09%, respectively (p<0.001). In addition, among patients suffering from SPM,HER2 positivity status contributed tobetter overall survival. Conclusion: After considering intrinsic incremental risk with age and adjusting for competing risk of death, our study demonstrated that HER2+ BC patients had lower SPM occurrence, remarkable for second BC and lung and bronchus cancer. The disparity implies the relation between SPM occurrence and therapeutic along with genetic factors underlying BC HER2 status.


Background
In GLOBOCAN 2020, breast cancer (BC) has become the most frequently occurring cancer and the leading cause of cancer death in women [1]. During the past decades, dramatic survival bene t has been achieved by profound understanding of BC biology, widely applied early screening, and rapid development in the systemic therapies [2][3][4]. One milestone of these achievements was the recognition of overexpressed human epidermal growth factor receptor 2 (HER2) as a major driver in 18-20% of BC [5,6]. Fortunately, anti-HER2 therapies represented by trastuzumab has become the standard treatment for HER2-positive (HER2+) BC patients and substantially improved their prognosis [7][8][9]. However, the favorable survival outcome renders BC patient a higher probability of second primary malignancy (SPM) occurrence [10]. A recent case-control study also reported excess risk for SPM among BC survivors [11], with an adjusted standardized incidence ratio of 12.94.
Since SPM can lead to excess mortality in BC patients [12,13], SPM occurrence and its related risk factors raised the awareness of clinicians. For example, SPM risk decreased with age [14], with an exception of lung cancer that is aging-related [15]. Negative hormone receptor was correlated with increased risk of SPM [16][17][18], which probably shared the same etiologic factors with rst BC. It is also reported that postoperative treatment could in uence the SPM occurrence. The effect of chemotherapy remains controversial [18][19][20], although several literatures reached the agreement that utilization of DNAdamaging chemicals led to increased incidence of SPM [21]. Radiotherapy may subject patients to a greater risk for SPM, especially lung cancer, esophageal cancer and contralateral BC [22,23]. Conversely, hormonal therapy exhibited a protective effect from SPM occurrence [24], though patients treated with tamoxifen had an elevated risk of corpus uteri cancer [25]. In addition, genetic alteration, particularly BRCA1/2 mutation were proven to be risk factors for SPM occurrence [24,26]. Despite massive efforts in exploring predictors of increased SPM risk, few studies have considered the effect of HER2 status [24]. It is noteworthy that no literatures reported the association between anti-HER2 therapies and SPM incidence.
This study aimed to investigate the differential SPM occurrence after rst primary BC according to HER2 status, using surveillance, epidemiology, and end results (SEER) database. We used standardized incidence ratio (SIR) to pro le and compare the SPM incidence between BC patients with different HER2 status. To adjust for competing death risk, subsequent competing risk analysis was conducted to explore the role of HER2 in SPM risk. We further assessed the survival outcomes according to HER2 status and whether SPM exists or not.

Study Variables
Patients were classi ed into 4 groups based on the diagnosis age: <45, 45-59, 60-74, ≥ 75 years. Fourgrade system was utilized to acquire BC grades according to microscopic examination of tumor tissue. A combined stage group was derived based on clinical and pathologic information composed of primary tumor(T), regional lymph nodes(N) and distant metastasis(M). The TNM classi cation followed the TNM manuals published by the American Joint Committee on Cancer (AJCC) and the International Union Against Cancer (UICC), the most up-to-date edition at that time. Hormone receptor (HR) status was obtained via the immunohistochemistry test of estrogen or progesterone receptor (ER/PR) [27]. HER2 status was de ned via the algorithm for deriving HER2 summary variable, according to immunohistochemistry and in situ hybridization tests [28]. The timing of the SPM and death occurrence and cause of death were also provided.
Furthermore, SEER*Stat computed the SIR to quantify the SPM incidence in rst BC patients. SIR is the observed incidence of SPM among BC survivors to the expected incidence [29,30], based on speci c cancer incidence rates for the general population in SEER 18 Registries (excluding AK). The 95% CI for SIR was estimated assuming a Poisson distribution for the observed SPM numbers. SIRs of SPM following HER2-positive (HER2+) and HER2-negative (HER2-) BC patients were acquired overall and strati ed by HR status, diagnosis age and malignancy site.

Statistical analysis
According to HER2 status of rst BC, patient characteristics and most importantly, whether suffering from SPMs were analyzed. Pearson's chi-square test was employed to compare the variable distribution.
Due to the natural different morbidity in different cohorts, SIRs reliably manifested SPM incidence in BC survivors. With the expected numbers as an offset, Poisson regression estimated the relative risk (RR) of HER2 status, HR status and age on SPM. Since latency after rst BC diagnosis varied and a large proportion of patients died before SPM occurrence, competing risk analysis was adopted to verify the risk of SPM after HER2-and HER2 + BC [31,32]. We utilized Cumulative incidence function (CIF) to describe the occurrence, and Gray's test to assess the statistical difference of occurrence probability.
Subsequently, the hazard ratios (HRs) and 95% CIs of developing SPM after HER2 + vs. HER2-BC were calculated using proportional subdistribution hazards regression, which adjusts for competing risk of death and other signi cant characteristics. Regarding survival, overall survival (OS) and breast cancerspeci c survival (BCSS) were calculated from the date of diagnosis of BC to the date of death, due to any causes and BC speci cally. Kaplan-Meier survival analyses were utilized to estimate the OS and BCSS of BC patients based on whether SPMs exist and their intrinsic HER2 status. Log-rank tests were adopted to determine the difference signi cance.
All statistical analyses were performed using R version 4.0.5 software. Statistical signi cance was set at two-sided, with p value < 0.05 de ned as statistically signi cant.

Results
Patient characteristics 409,796 BC patients diagnosed between 2010-2018 were selected in this study, with a median follow-up period for 44 months. HER2 + BC (N = 64,740) accounted for 15.80% of all patients, 3.54% of which (N = 2293) were reported to develop SPMs, while HER2-BC (N = 345,056) had a higher SPM incidence of 4.63% (N = 15990). Table 1 outlines the clinicopathologic characteristics and occurrence of SPM, grouped by HER2 status at rst primary BC diagnosis. All factors were statistically signi cantly associated with HER2 status (p < 0.001) based on Pearson's chi-square test. For HER2 + BC patients, 11592 (17.91%) women were diagnosed aged < 45 and 26282 (40.60%) aged between 45-59, accounting for signi cantly higher proportions than those with HER2-BC. Abbreviations: HR hormone receptor, HER2 human epidermal growth factor receptor 2 Besides, HER2 + BC were more likely to have advanced grade and stage, and negative HR status. However, SPMs were remarkably less common in HER2 + BC patients, despite their clinicopathologic characteristics re ect more aggressive clinical courses.
Pro le of SPM incidence 18,283 patients were diagnosed with at least one SPM after 2 months of rst BC diagnosis, with 19,297 SPMs overall, which was signi cantly more than the 17,346.64 expected cases based on the rates in the general population (SIR = 1.11; 95% CI 1.10-1.13). As listed in Table 2, compared with the general population, SPM incidence was

Competing risk analysis
When estimating the SPM incidence, occurrence of death can compete and should be taken into account by competing risk analysis. Fig. 3 depicted the cumulative incidence of SPMs and deaths via CIF. The 5- year cumulative incidence of SPMs following rst HER2-and HER2+ BC were 5.16% and 4.09%, respectively. HER2-BC patients had a markedly higher cumulative SPM incidence than HER2+ BC patients (p<0.001), and on the contrary, lower signi cantly death incidence (p<0.001). To understand whether the difference of SPM cumulative incidence is solely due to the HER2 status, subdistribution hazard function was performed, as shown in Table 3. Univariate regression analysis indicated that not only HER2 status, but also age, grade, stage and HR status were correlated with SPM risk. Multivariate regression analysis further screenedHER2 status and diagnosis age as independent risk factor for SPM. HER2 positive status signi cantly reduced the total SPM risk (sdHR, 0.86; 95% CI, 0.82-0.90; p<0.001), in accordance with aforementioned results. However, the inconsistence was that SPM risk increased as patients diagnosed at older age (p<0.001), and HR status did not differ statistically in SPM risk (p=0.28).
In addition, multivariate proportional subdistribution hazards regression was

Kaplan-Meier survival analysis
Finally, the impact of SPM and HER2 status on OS and BCSS was analyzed (Fig. 4)

Discussion
In our study, we mainly focused on the association between SPM occurrence and HER2 status of rst BC. Signi cantly increased incidence of SPM was demonstrated in HER2-BC patients. In 2015, Raffaella et al. reported that HER2 positivity was associated with increased risk of secondary digestive system and thyroid cancer [24]. Incorporating 305 cases and 1,525 control patients, this study was limited by small sample sizes. For the rst time, two different methods were performed to explore the risk factors for SPM incidence. On the one hand, older age is associated with intrinsic incremental cancer risk. And HER2 status varied with age according to our baseline characteristics analysis. Therefore, SPM risk and effect of HER2 status were quanti ed using standardized incidence ratios (SIRs) and Poisson regression, by introducing age-speci c rates in the general population as external reference. On the other hand, we further executed competing risk analysis to verify the impacts of risk factors on SPM occurrence, considering death as a competing risk. Eventually, two approaches yielded similar result that remarkable SPM risk difference existed between HER2+ and HER2-BC patients.
However, apparent discrepancy was presented for age and HR status. In the latter proportional subdistribution hazards regression, the SPM risk increased when BC patients were diagnosed at older age. But if we take the natural incremental cancer risk with age into account, totally contrary result was obtained that SPM risk declined with age. It seemed more reliable and may be attributable to the genetic predisposition of cancer in young BC patients and probable prolonged survival period. We also discovered an elevated risk of SPM among HR-BC patients as reported by previous studies [14,17], which may due to common genetic determinants and limited treatment. Nevertheless, after adjusting for the death risk, no statistically signi cant result was observed between HR+ and HR-BC patients. We suspect that the non-negligible occurrence of death precluded the occurrence of SPMs, probably contributing to an overestimation of SPM occurrence in HR-BC patients. Age could also be regarded as a confounder, since HR negative status was more common in younger BC patients [33].
Although HER2-BC was not so aggressive as HER2+ BC, the higher incidence of SPM makes it a new concern for BC survivors. As above mentioned, compared with HER2+ BC, HER2-BC patients had superior BCSS in all patients, but worse OS in patients followed by SPM. To improve the OS of HER2-BC patients, it is important to nd out the potential contributing factors that are relevant to HER2 status.
In recent decades, HER2 status has been an established target in BC therapy.Biologically, HER2 exists as a driving factor in various tumor types derived from epithelia, represented byBC, non-small-cell lung cancer [34,35], gastric and gastroesophageal junction cancers [36,37]. It was further con rmed by the fact that HER2 overexpression or gene ampli cation were detected in a variety of human malignancies including lung[38], gastric [39], ovarian [40], biliary tract [41] and colorectal cancers [42].In therapeutic application, anti-HER2 therapies have already shown e cacy inHER2-positive advanced gastric cancers[36, 43,44]. Therefore, we speculate that when HER2+ patients undertook anti-HER2 treatment for BC, these targeting therapies would probably simultaneously extinguish and eliminate latent HER2+ transformed cells in other system. As a result, the SPM incidence decreased after HER2+ BC due to the HER2-targeted treatment. It is known that more than two decades have passed since the approval of the rst anti-HER2 targeted therapy trastuzumab by FDA in 1998. And HER2-targeted treatment has become the standard care in managing HER2+ BC patients all over the world. Hence, recent calendar year and HER2positivity diagnosis have been used as a crude proxy to support this hypothesis. Besides anti-HER2 therapy, HER2+ BC patients were more likely to received adjuvant chemotherapy [45,46], which may also contribute to the declining SPM incidence.
As for the intrinsic risk factors of SPM, shared genetic predisposition associated with BC HER2 status is inevitable. It has been reported that pathogenic germline variants were exhibited in 8% of adult cancer cases [47] . Genomic alterations are also reported to be incriminated towards HER2+ BC [48,49]. Moreover, several genes such as FCRLM1, BLK and IHGD were indicative of the development of SPMs among BC patients [48,49]. It led to the assumption that some unique intrinsic genetic variants existed in BC patients may be a natural inclination for the SPMs. Indirect support for the hypothesis can be gleaned from the ndings that higher proportion of HER2positivestatus was observed in BCpatients carrying BRCA2 mutations, and meanwhile, BRCA2 carriers were less likely to be diagnosed with SPMs[50]. In summary, the SPM risk reduction after HER2+ BC may be attributable to the widespread anti-HER2 treatment combined with chemotherapy, as well as genetic susceptibility. More investigation about therapeutic and inherent factors should be conducted to reveal the disparity between different BC HER2 status.
The differential SPM occurrencefollowing HER2+ and HER2-BC was comprehensively pro ledin this context. It was based on a large population cohort from SEER program, thus minimizing the sampling error and ensuring the quality of the data. However, there were still some limitations. First, some metastasis and relapse may be mistaken as SPM. Although we excluded the cases within 2 months since BC diagnosis to avoid synchronous cancers. Second, because of data incompleteness, we could not analyze effects of adjuvant treatment, especially anti-HER2 therapies. Third, data on some potential risk factors was not available, such as smoking, alcohol use and body mass index. More researches with detailed information on treatment delivery, genetic variants and other confounders are needed to reveal the disparity.
In conclusion, SPM incidence signi cantly declined after HER2+ BC, demonstrated by different quantitative tools including SIR and competing risk model. The OS deterioration in HER2-BC patients accompanied by SPM indicated the requirement to explain this disparity, with the aim of reducing SPM risk. We suggested that the differential SPM occurrence could be attributable to HER2-targeted therapy combined with chemotherapy, and inherent genetic factors, both of which are closely related to HER2 status of rst BC. Further studies are required to prove this plausible hypothesis.  Site-speci c Standardized incidence ratios (SIRs) and relative risks of developing second primary malignancies (SPM) after HER2+/HER2-BCs. Cumulative incidence curves for second primary malignancies and deaths in HER2+ and HER2-breast cancer patients.