An analysis of HER2 amplification in cervical adenocarcinoma: correlation with clinical outcomes and the International Endocervical Adenocarcinoma Criteria and Classification

Abstract Few studies have explored HER2 status in cervical adenocarcinoma, particularly in the gastric‐type adenocarcinoma (GAC), a nonhuman‐papillomavirus‐related subtype with poor clinical outcomes. In this study, we investigated HER2 expression and amplification by immunohistochemistry (IHC) and fluorescence in situ hybridization (FISH) in 209 well annotated cervical adenocarcinomas diagnosed using the International Endocervical Adenocarcinoma Criteria and Classification. IHC identified HER2 protein expression in 57.4% (123/209) of adenocarcinomas, of which 62 were IHC 1+ (negative), 38 2+ (equivocal) and 23 3+ (positive). HER2 amplification was found in 13 cases (6.2%) including 10 with IHC 3+ and 3 with IHC 2+. Among all the major histotypes of cervical adenocarcinoma, HER2 amplification was most common in GAC cases with a frequency of 14.7% (5/34). Moreover, HER2 amplification was more frequently associated with 2018 International Federation of Gynecology & Obstetrics (FIGO) stage III/IV, perineural involvement and ovarian spread (p < 0.05) while IHC 3+ was more common in patients with lymphovascular invasion and ovarian involvement (p < 0.05). Survival analysis indicated that FIGO stage III/IV, GAC, and p53 overexpression were associated with poor disease‐specific survival and tumor recurrence (p < 0.05). In conclusion, HER2 amplification was present in a subset of adenocarcinomas, and more common in GAC, pointing to a potential benefit from trastuzumab treatment. HER2 overexpression does not identify gene amplification status in cervical adenocarcinoma; therefore, FISH is suggested for both IHC positive and equivocal cases. Further investigation on more cases with longer follow‐up times is required to consolidate these findings.


Introduction
Cervical carcinoma is the fourth most frequent cancer worldwide in women, with an estimated 570 000 new cases in 2018 [1]. Adenocarcinoma is the second most common histotype, accounting for 10-25% of all cervical carcinomas [2], and both the relative incidence and absolute number of endocervical adenocarcinomas have been increasing over the last two or three decades. Compared with squamous cell carcinoma, cervical adenocarcinoma has been found to have a poorer prognosis although there are no differences in the treatment strategy between them [3,4]. Moreover, cervical adenocarcinomas are a heterogeneous group of tumors that may harbor prognostic differences. An international group of gynecological pathologists has recently proposed a new pathogenetic scheme, the International Endocervical Adenocarcinoma Criteria and Classification (IECC), which is based on the presence or absence of high-risk (HR) human papillomavirus (HPV) infection-related morphological features [5].
The IECC approach appears to be valuable for prognostic assessment [6,7]. In particular, gastric-type adenocarcinoma (GAC), the predominant type of non-HPV-related adenocarcinoma, has demonstrated poorer clinical outcomes compared with other histotypes of cervical adenocarcinoma [8][9][10][11][12]. It is a pressing requirement to identify biomarkers that have potential therapeutic significance to improve the prognosis in cervical adenocarcinomas.
Our knowledge of human epidermal growth factor receptor 2 (HER2) amplification has expanded from breast carcinoma to other human carcinomas including gastroesophageal cancer and uterine serous carcinoma because HER2-positive cancers may benefit from anti-HER2 monoclonal antibody therapy (trastuzumab) [13][14][15]. Several studies have investigated HER2 expression by immunohistochemistry (IHC) in cervical carcinomas including a small number of adenocarcinomas [16][17][18]. However, some important relevant aspects remain unresolved in cervical adenocarcinomas, such as the clinical significance, the precise prevalence and the histotype distribution of HER2 overexpression, and the concordance between HER2 immunohistochemical results and gene amplification. To address these issues, we performed a systematic analysis of HER2 expression and gene amplification in a large cohort of cervical adenocarcinomas.

Patients and pathological review
The hospital's Institutional Review Board approved this study (IRB: 20170139; October 2017). In total, 209 adenocarcinomas were retrieved from the archives of the Department of Surgical Pathology, Women's Hospital, School of Medicine, Zhejiang University, PR China, between January 2014 and June 2019. The clinical and demographic features of the cases are given in supplementary material, Table S1. Rare subtypes including the serous type and the endometrioid type were excluded from the present study if they had concurrent carcinomas in the uterus, the Fallopian tube and/or the ovary. The clinical data were extracted from the electronic medical records before de-identification. Tumor stage was re-assessed according to the 2018 International Federation of Gynecology & Obstetrics (FIGO) cervical cancer staging system [19]. Two authors (LB and SH) independently reviewed the archival H&E slides for detailed morphological analysis. All cases were reclassified according to the IECC [5]. The pattern of invasion in endocervical adenocarcinomas, usual-type, was determined according to a three-tier pattern-based histopathologic system [20]. The patients were followed up by chart review and telephone communication until June 2020.

Fluorescence in situ hybridization
A dual-probe fluorescence in situ hybridization (FISH) assay for HER2 and chromosome enumeration probe 17 (CEP17) was carried out using a PathVysion HER2 DNA probe kit (Abbott Molecular Inc., Des Plaines, IL, USA) according to the manufacturer's protocol. Tumor nuclei were counterstained with 4 0 ,6-87 HER2 amplification in cervical adenocarcinoma diamidino-2-phenylindole (DAPI) (Abbott Molecular Inc., Des Plaines, IL, USA) after hybridization. The HER2 and CEP17 signals per nucleus were recorded. The results from 20 interphase nuclei from invasive carcinomatous cells were reported as the ratio of total HER2 signals to total CEP17 signals. HER2 testing was eventually interpreted as amplification or no-amplification according to the updated American Society of Clinical Oncology/College of American Pathologists (ASCO/CAP) guideline for breast carcinoma (2018) [23].

Statistical methods
The SPSS 16.0 software package (SPSS Inc., Chicago, IL, USA) was used for statistical analysis. Nonparametric tests (Wilcoxon test or Mann-Whitney U test) were used to determine the clinicopathological significance of HER2 expression/amplification. The Kaplan-Meier method was performed to compare mortality and recurrence between groups, and the log-rank test was used to analyze the statistical significance of the differences between groups. Multivariate survival analysis was carried out by the Cox's proportional regression hazard (HR) method (conditional forward). We adjusted multiple parameters that were found to be significant in the univariate model. The statistical threshold was set at 0.05 (two-sided).

90
H Shi et al overexpression were associated with poor diseasespecific survival in cervical adenocarcinomas (p < 0.05) ( Figure 3). Moreover, these factors except p53 overexpression were also correlated with tumor recurrence (p < 0.05). HER2 amplification or IHC3+ was not an indicator for adverse clinical outcomes in patients with endocervical adenocarcinomas (p > 0.05). However, the Kaplan-Meier plot showed a marginal trend towards a worse overall survival in patients with HER2 amplification ( Figure 3). Multivariate survival analysis including these characteristics indicated that GAC, FIGO stage and p53 overexpression were strong, independent markers for disease-specific survival and tumor recurrence (Table 3).

Discussion
Numerous targeted therapies against aberrant molecules have been developed in the past three decades. Among these, the most successful might be the anti-HER2 therapy, trastuzumab, which has been approved by FDA for the treatment of breast, gastric and gastroesophageal carcinomas with HER2 gene amplification or protein overexpression. HER2 gene amplification and/or protein overexpression have been discovered in some cancers in the female genital tract, such as uterine serous carcinoma and ovarian mucinous carcinoma [24,25]. These findings suggest the potential value of trastuzumab therapy in these cancers. HER2 protein overexpression has been reported in cervical carcinomas, mainly of squamous cell type, with a frequency varying from 1 to 21% using different methods and criteria [26]. However, few studies have so far explored HER2 expression in cervical adenocarcinomas. In an early study, c-erb-B2 (HER2) expression was detected in 34 of 44 (77%) cervical adenocarcinomas, with strong membrane staining in 11 of 44 (25%), which was associated with advanced tumor stage and poor prognosis [17]. More recently, Halle et al [18] reported HER2 overexpression in 22 of 70 (30%) cervical adenocarcinomas; here, there was no correlation with poor clinical outcome. Our study may represent the largest clinicopathological investigation on HER2 protein expression in cervical adenocarcinomas to date. We noted the frequent presence of incomplete membrane (basolateral or 'U'-shaped) staining and intra-tumoral heterogeneity in cervical adenocarcinoma, leading to the adoption of the amended scoring system for gastric adenocarcinoma [21]. However, even with the modified criteria, the incidence of HER2 overexpression (IHC 3+) in our series was at the low end Invasive adenocarcinomas, NOS were recategorized into non-HPV (n = 4) and HPV-related (n = 9) adenocarcinomas based on p16 staining and/or HPV RNA ISH detection. ‡ The pattern of invasion is only assessed in endocervical adenocarcinoma, usual type.

HER2 amplification in cervical adenocarcinoma
of what has been reported previously [17,18]. The reasons for these conflicting results may be largely associated with certain technical issues, such as tissue fixation (duration of, and time between surgical removal and, fixation), tissue processing protocol, different antibodies to HER2, IHC staining protocols, and stringent criteria for the assessment of HER2 staining; and to a lesser extent, with the genetic background of cervical adenocarcinomas in Chinese populations.
To the best of our knowledge, HER2 gene amplification has not been extensively analyzed in a large series of cervical adenocarcinoma with detailed histopathological assessment and clinical follow-up. Our study showed that the frequency of HER2 amplification by FISH was 6.2% in cervical adenocarcinomas. A chromogenic in situ hybridization analysis has shown that true HER2 amplification was rare in invasive cervical squamous cell carcinomas (5/136 cases, 3.7%) using tissue microarrays [16]. Recent integrated

H Shi et al
genomic studies have indicated a greater tendency for the occurrence of HER2 amplification and recurrent mutations in cervical adenocarcinoma than in squamous cell carcinoma [27,28]. Moreover, Halle et al [18] found that HER2 genomic alterations rather than protein overexpression were associated with poor disease-specific survival in patients with cervical carcinoma despite the presence of a link between gene copy number and protein status. In the current study, exclusively on cervical adenocarcinomas, patients with HER2 amplification showed a marginal trend towards a worse overall survival. Furthermore, HER2 amplification was significantly associated with several potential indicators of poor prognosis, such as high FIGO stage, perineural invasion and ovarian involvement. Therefore, the patients with HER2 amplification can be expected to benefit from the current trastuzumab treatment.
We observed discrepancies of potential clinical relevance between HER2 gene amplification and protein expression, although there was a high concordance between the two tests. IHC 3+ neither detected all amplified cases nor invariably correlated with gene amplification, as reported previously in some cancer types including cervical cancer (mostly squamous cell carcinoma) and gastric cancer [18,29]. According to the ASCO/CAP guideline [23], IHC 3+ is determined to be HER2 (amplification) positive in breast cancer. However, our results clearly demonstrated that IHC 3+ alone did not confirm, and IHC 2+ did not exclude, the presence of HER2 amplification in cervical adenocarcinoma. Therefore, we feel that, unlike breast cancer, tests for both IHC and FISH (for cases with IHC 2+ and 3+) are required for cervical adenocarcinoma patients in clinical trials of trastuzumab treatment, which is primarily aimed towards patients with HER2 amplification. Likewise, such methodology has been recommended in clinical trials for trastuzumab treatment of gastric and gastroesophageal junction adenocarcinomas [14,29]. The discrepancies between HER2 amplification by FISH analysis and IHC results may suggest the presence of other factors activating HER2 in cervical adenocarcinomas, such as recurrent somatic mutations, HPV integration, transcriptional upregulation of HER2 without gene amplification, or polysomy leading to false positive IHC [24,27,28]. These tumor biology-related molecular abnormalities indicate a potential for anti-HER2 therapy other than trastuzumab, such as pertuzumab and trastuzumab emtansine (T-DM1), and pan-HER inhibitors afatinib and neratinib [30,31].
We systematically analyzed the distribution of HER2 amplification and clinicopathological features across various subtypes of cervical adenocarcinomas using IECC, a new pathogenetic classification for invasive endocervical adenocarcinoma [5]. We found that non-HPV-related adenocarcinomas and GAC had a poor overall survival and shorter relapse time. These results add more evidence to support the prognostic value of IECC for endocervical adenocarcinoma and the aggressive nature of GAC [6][7][8][9]. The distribution of HER2 amplification among histotypes of cervical adenocarcinomas has not yet been fully investigated, although IHC studies demonstrated that endocervical adenocarcinomas of usual type had a HER2 positive frequency of 1.8% [5], which was much lower than in GAC [32]. It was estimated that 5-15% of cervical GAC may harbor HER2 amplification [32]. In this study, we observed that the frequency of HER2 amplification was 14.7% (5/34) in GAC, substantially higher than in other histotypes. Nevertheless, the precise level remains inconclusive because HER2 amplification has been investigated in very few cases owing to its low incidence. Nakamura et al [32] detected only one GAC with HER2 amplification from four cases with equivocal protein expression that were successfully analyzed by dual color in situ hybridization. Recently, Garg et al [33] identified HER2 amplification by FISH in 2 of 12 cervical GACs. Taken together, the relatively high HER2 amplification rate in cervical GAC, which was comparable to that reported in gastric cancers (12.2-20%) [14,29,34], may suggest the particular treatment implications for these patients on account of their poor clinical outcomes. However, we have to acknowledge that the prognostic and therapeutic significance of GAC should be further validated in clinical studies on more cases with long-term follow-up.
In this large case series, we demonstrated that p53 overexpression was an independent prognostic biomarker for cervical adenocarcinoma. Our findings are consistent with results from previous studies, which

HER2 amplification in cervical adenocarcinoma
showed that p53 overexpression exerted an adverse impact on overall survival in cervical carcinomas including adenocarcinoma [18,35]. In cervical carcinoma, it is well-known that loss of p53 is caused by proteasomal degradation induced by HPV E6 protein, while overexpression is related to gene mutations [28,36]. Severe dysregulation of cellular growth caused by TP53 mutation may be the underlying mechanism for poorer survival in cervical cancer patients with p53 overexpression [18].
The limitations of our study include a small number of HER2 amplified cases and GACs, the relatively short follow-up time, and lack of well-established evaluation criteria for HER2 expression in cervical adenocarcinoma. Further investigation of more HER2-amplified cases and GAC with longer follow-up time is required to develop more precise evaluation criteria for HER2 detection and to clarify the clinical significance of HER2 amplification in cervical adenocarcinomas.
In conclusion, our large case series identified FIGO stage, GAC histotype and p53 overexpression as adverse prognostic features in cervical adenocarcinomas. HER2 amplification was present in a subset of adenocarcinomas, and more commonly in GAC. Trastuzumab treatment is suggested as being beneficial for these patients although the prognostic value of gene amplification needs further validation. HER2 overexpression does not identify gene amplification status; therefore, our recommendation is to use FISH in IHC-positive and equivocal cases.