HER2 amplification and HER2 low expression in endometrial carcinoma: prevalence across molecular, histological and clinicopathological risk groups

With increased understanding of the molecular landscape of endometrial carcinoma, tailored biomarker-driven cancer treatment is replacing “one size fits all” solutions that were employed in previous decades. The mainstay of adjuvant treatment of uterine cancer is chemotherapy or radiotherapy or their combination. Recently, the response of endometrial carcinoma to immunotherapy combined with chemotherapy has been shown to be strongly dependent on molecular subgroup [1, 2]. In the recurrent platinum-pretreated setting, immunotherapy can be potentiated with a multi-tyrosine kinase inhibitor lenvatinib [3]. In phase II clinical trials human epidermal growth factor 2 (HER2)-targeted therapies have shown promising therapeutic activity in endometrial carcinoma [4,5,6,7]

The transmembrane glycoprotein HER2/ErbB2 is a member of the ErbB family of protein kinases coded by the HER2/ERbB2 gene. HER2 is the only member of the ErbB family that has no known ligands. HER2 signaling is activated via heterodimerization with other ErbB family receptors or via homodimerization following HER2 overexpression [8]. Overexpression is mainly due to gene amplification or mutation [9]. Classicly, HER2-positivity is defined by the presence of HER2 amplification leading to protein overexpression. Less frequently, HER2 signaling can be overactivated by point mutations in the absence of overexpression on immunohistochemistry (IHC). Signaling through the HER2 receptor promotes cell proliferation and opposes apoptosis [10]. HER2 overexpression and its association with more aggressive tumor phenotype is recognized in several solid tumors, most notably in breast cancer [11].

The HER2 signaling pathway can be blocked by an anti-HER2 antibody. A randomized phase II study showed that HER2-targeting antibody trastuzumab combined with carboplatin and paclitaxel improved survival in patients with advanced and recurrent HER2-amplified uterine serous carcinoma compared with carboplatin and paclitaxel alone [4, 5]. Antibodies can also be linked with chemotherapy drugs to deliver cytotoxic molecules selectively to cancer cells expressing specific surface antigens, such as HER2. Trastuzumab-deruxtecan, an antibody drug conjugate (ADC), has recently been approved in the United States for use in patients with unresectable or metastatic HER2-positive (IHC 3 + ) solid tumors who have received prior systemic treatment and have no satisfactory alternative treatment options [12]. The approval is based on a phase II trial where positive overall response rates were demonstrated also in 40 cases of endometrial carcinoma. Response rates correlated with intensity of HER2 expression [6]. Definite guidelines for HER2 testing and interpretation in endometrial carcinoma are waiting to be established [13]. In the trastuzumab trial, HER2 positivity was defined as immunohistochemical 3+ score or HER2 amplification confirmed by fluorescence in situ hybridization (FISH) according to the ASCO/CAP 2007 scoring guidelines for breast cancer [5] (Supplementary Table 1). DESTINY-PanTumor-02 trial investigating trastuzumab-deruxtecan applied ASCO/CAP 2016 scoring guideline for gastric cancer with 2+ and 3 + IHC as inclusion criteria [6, 7] (Supplementary Table 1).

While therapy based solely on anti-HER2 antibodies is considered effective only on tumors with HER2 gene amplification, antibody–drug conjugates may also target cells expressing low levels of HER2 in the absence of gene amplification. In the uterus, HER2 amplification has mainly been reported in serous carcinoma and carcinosarcoma [14,15,16]. HER2 low expression appears to be more widespread with regard to histotypes and other clinicopathologic subgroups but substantial differences exist in the reported frequencies [17, 18]. Definite guidelines for HER2 testing and interpretation in endometrial carcinoma are waiting to be established [13]. The aim of this study was to investigate the prevalence of HER2 amplification and HER2 low expression in a large and well-characterized cohort of endometrial carcinoma.

Women with endometrial carcinoma treated with surgery at Helsinki University Hospital between January 4th, 2007, and February 2nd, 2023, with adequate tumor samples for a tissue microarray (TMA) were included in the study. A retrospective cohort of all 824 women surgically treated in our institution between 2007 and 2012 and a prospectively recruited cohort of 413 women starting on September 23rd, 2014, were included. This study was approved by the Helsinki University Hospital institutional review board and the National Supervisory Authority for Welfare and Health.

Clinicopathological data were collected from the institution’s electronic medical records and stage was determined according to the 2009 International Federation of Gynecology and Obstetrics (FIGO) [19] Risk groups were defined as in the 2020 joint guidelines of the European Society of Gynecological Oncology (ESGO), European SocieTy for Radiotherapy and Oncology (ESTRO), and European Society of Pathology (ESP) [20]. Molecular subgroups were identified using the Proactive Molecular Risk Classifier for Endometrial Cancer (ProMisE) [21]. Molecular classification by ProMisE was accessible for 1009 cases. Lymphovascular space invasion (LVSI) was graded as: none (no LVSI), focal (a single focus of LVSI around the tumor), and substantial (multifocal or diffuse LVSI or the presence of tumor cells in 5 or more lymphovascular spaces) [20]. Data on LVSI were unavailable for 15 patients, and the depth of myometrial invasion (MI) for 1 patient. Causes of death were mainly abstracted from medical records and missing data complemented from death certificates when necessary.

Standard surgical procedures included total hysterectomy and bilateral salpingo-oophorectomy. Routine regional lymphadenectomy was performed until December 2011 and was then omitted in low-grade (grades 1 and 2) endometrioid carcinomas with superficial (<50%) MI (selective lymphadenectomy). Adjuvant therapy decisions were guided by multidisciplinary tumor board conclusions, supported by international guidelines in effect at the time of diagnosis.

Chromogenic HER2 IHC was performed on tissue microarray (TMA) slides using Ventana (Tucson, Arizona, USA) 4B5 monoclonal antibody according to the recommendations of the manufacturer. HER2 gene status was assessed by dual color chromogenic in situ hybridization (CISH) performed according to the manufacturer’s instructions using Roche Ventana (Tucson, Arizona, USA) HER2 Dual ISH DNA Probe Cocktail Assay and Ventana BenchMark Ultra Plus staining instrument. Silver detection kit was used to detect HER2 probes and Red detection kit for Chromosome 17 probes.

Immunohistochemical stainings were scored according to a 4-tiered scoring system recommended by the American Society of Clinical Oncology (ASCO)/College of American Pathologist (CAP) guidelines for gastric cancer resection specimens. The same scoring system is used in the ongoing DESTINY-PanTumor trial (NCT04482309) which includes also endometrial carcinomas. Accordingly, 3+ was defined as strong, complete, basolateral or lateral staining in ≥10% of the carcinoma cells. 2+ score was assigned when weak to moderate staining was seen in ≥10% of tumor cells and 1+ when the intensity was faint/bairly perceptible. Negative staining was defined as no staining or membranous staining in <10% of the cells. For quality control, external control samples expressing various levels of staining intensity are routinely used in our laboratory.

HER2 CISH was evaluated for all the specimens with 2 + /3+ staining. Slides were interpreted using a bright field microscope with 40× objective. The HER2/CEN-17 ratio was calculated and cases presenting HER2/CEN-17 ≥ 2.0 (or more than 6 HER2 signals/cell) were assigned into HER2 amplified category (referred to as “ISH positive” in the text). For subclonal alterations, a cut-off of >10% of the tumor cells presenting HER2 amplification defined CISH positivity. Non-tumoral cells exhibiting one-to-one ratio of HER2 and CEN-17 signals served as an internal control for staining success. Finally, all the cases were classified according to HER2 IHC and ISH status as HER2 amplified (ISH positive), HER2 low 2+ (IHC 2 + ISH negative), HER2 low 1+ (IHC 1 + ) and HER2 negative (IHC 0) (Fig. 1).

HER2 immunohistochemistry (a–c) and dual color chromogenic in situ hybridization (d–f): 1+ non amplified (a, d), 2+ amplified (b, e), 3+ amplified (c, f); red=chromosome 17, silver=HER2.

Full size image

Chi-squared test was used for comparison of categorical variables, and analysis of variance and Kruskal-Wallis test for comparison of continuous variables after testing for normality by Shapiro-Wilk test. Multivariable Cox regression analysis was used to assess the association of HER2 status with disease-specific survival (DSS). DSS times were calculated as the time from surgery to death from endometrial carcinoma using the Kaplan-Meier method. Differences between groups were compared using the log rank test. Statistical significance was set at p < 0.05. Data were analysed using the Statistical Package for the Social Sciences version 29 software (IBM Corp., Armonk, NY, USA).

Clinicopathological data on the included 1239 patients with a median follow-up time of 73 months (range 0–136) are presented in Table 1. HER2 IHC and CISH provided results for 1239 endometrial cancer specimens. In the IHC stainings, 1.0% of the cases were assigned score 3+ (n = 13), 4.4% score 2+ (n = 55) and 14.7% score 1+ (n = 182). The remaining 987 cases were scored negative. As expected, all the 3+ cases presented HER2 amplification by CISH. Of the IHC 2+ cases, 20% (n = 11/55) were ISH positive and 80% (n = 44/55) were ISH negative (HER2 low 2 +). Two amplified cases presented weak (equivocal) immunohistochemical staining possibly due to technical reasons. Of clinicopathological factors, MI and LVSI were not associated with HER2 expression status. In contrast, increasing patient age (p < 0.001), body mass index (BMI) (p = 0.003), and stage III-IV (p = 0.013) were associated with HER2 expression status (Table 2).

Relative frequencies of HER2 categories varied significantly between histological types (p < 0.001; Table 2). HER2 was amplified in 13.1% (n = 8/61) of serous carcinomas, but also in 11.5% (n = 3/26) of carcinosarcomas, 11.1% (n = 5/45) of clear cell carcinomas, but only in 0.9% (n = 10/1078) of endometrioid carcinomas, eight of which were high-grade endometrioid (4.5%; Table 2). HER2 low 2+ expression was also common in nonendometrioid histologies, 10.5% (n = 17/161). HER2 negativity was most common in low-grade endometrioid carcinoma (83.4%, n = 752/901). In the subgroup analyses of nonendometrioid carcinomas DSS was affected by HER2 expression status (p = 0.048), with fewest events occuring in the HER2 low 2+ cases (Supplementary Fig. 1B).

Significant differences were also noted across molecular subgroups (Fig. 2a). Approximately 80% of cases in the NSMP, MMRd and POLEmut subgroups were HER2 negative while 68.3% were HER2 negative in the p53abn subgroup. HER2 amplification was mainly observed in the p53abn subgroup (16.9%, n = 16/89). There were also two HER2 amplified NSMP cases of which both had clear cell histology and one MMRd case of high-grade endometrioid histology. The MMRd case also presented cytoplasmic p53 expression indicating concomitant TP53 mutation. Both NSMP cases presented 1–2 + IHC and low copy number amplification of HER2. HER2 low 2+ expression was observed between 1.7 and 5.3% of NSMP, MMRd, and POLEmut in ascending order while in p53abn it was found in 6.7% (Fig. 1a). HER2 low 1+ was more frequent being observed in 14.2% of NSMP, 14.7% of MMRd, and 22.5% in p53abn. In the subgroup analysis, HER2 expression status did not affect DSS within the p53abn group, p = 0.929 (Supplementary Fig. 1C).

Distribution of HER2 expression in molecular subgroups (a) and ESGO-ESTRO-ESP 2020 risk groups (b). NSMP no specific molecular profile, MMRd mismatch repair deficient, POLEmut POLE ultramutated, p53abn p53 abnormal, Low ESGO-ESTRO-ESP low risk category, IM/hIM ESGO-ESTRO-ESP intermediate and high-intermediate risk category, High ESGO-ESTRO-ESP high risk category, Adv-Met ESGO-ESTRO-ESP advanced-metastatic risk category.

Full size image

Most HER2 amplified cases were of the high risk or advanced metastastic ESGO-ESTRO-ESP disease category (n = 22/23, 95.7%) (Fig. 2b). However, also of the HER2 low 2+ group 64.3% of cases (n = 22/35) were in the higher risk categories.

In all cases in the cohort, HER2 expression status was associated with DSS, p = 0.010 (Supplementary Fig 1A). However, in multivariable endometrial cancer-specific survival analysis, HER2 expression was not associated with DSS (p = 0.619) while ESGO-ESTRO-ESP risk group was (p < 0.001) (Table 3).

In this study, we investigated the expression of HER2 in a large molecularly and clinicopathologically characterized cohort of endometrial carcinoma cases. As recent clinical trials have presented new treatment options for patients with tumors expressing HER2 in the absence of gene amplification, also cases with low levels of expression (IHC 1 + , 2 +) with negative ISH were investigated.

The prevalence of HER2 amplification and low levels of HER2 expression in endometrial carcinoma is not clear as notable differences exist in the reported frequencies [17, 18]. In our study 21.3% of the cases presented HER2 expression of some level (1 + , 2 + , 3 +) and HER2 amplification was observed in 2% of the cases. Our study corroborates the findings of Van Dijk et al. who reported similar albeit slightly higher frequencies of HER2 expression in a more selected cohort of high-risk endometrial carcinoma [17]. They used the same HER2 antibody (4B5) for IHC stainings and similar scoring methods. Krakstad et al. reported considerably higher (87%) frequencies of HER2 expression in their endometrial carcinoma series using different staining and scoring methods (HercepTest[Agilent]). The result is surprising, since HercepTest scoring algorithm which only counts complete membrane staining could be expected to lead to a lower frequency of positive cases compared to ASCO/CAP gastric scoring method which allows for basolateral incomplete staining pattern with the same 10% cut-off for positivity. Given the heterogeneous nature of HER2 expression in endometrial carcinoma, the number of TMA cores (3 cores in Krakstad’s study vs 4 cores in our study) may influence the representativity of the TMA but is not likely to explain the unexpected difference in the results in whole. The overall interantibody agreement rate between HercepTest (Agilent Dako) and PATHWAY anti-HER2 (4B5) (Ventana) antibodies is relatively low (57.8%), but staining with 4B5 appears to yield more positive scorings [22]. Importantly, interobserver variability in HER2 scoring is considerable [22]. Notably, in the DESTINY-PanTumor02 trial, central review confirmed only 64% of the local 3 + IHC scores [6]. These findings undermine the importance of regular training of pathologists. Whether computer-assisted image analysis will offer reliable tools for HER2 scoring remains to be seen.

Official scoring criteria for endometrial carcinoma remain to be established. It is known that endometrial carcinoma often presents incomplete (basolateral) HER2 staining pattern akin to gastric carcinomas [23]. Therefore, scoring algorithms for breast carcinoma based on complete membranous pattern may not have optimal concordance with HER2 amplification status or clinically meaningful HER2 expression in endometrial cancer. In endometrial cancer, IHC positivity and HER2 amplification are also often more focal than in breast cancer. A 30% cut-off for 3+ positivity appears to have stronger correlation between IHC and FISH amplification data compared with a 10% cut-off [23] (Supplementary Table 1).

Traditionally, most of the clinical HER2 studies in endometrial carcinoma have been restricted to serous carcinomas where HER2 amplification is most often encountered. Accordingly, trastuzumab is in clinical use only in HER2 amplified advanced or recurrent endometrial carcinomas of serous histotype in combination with platinum-based chemotherapy [4, 5]. As expected, in our cohort, amplification of HER2 was mostly noted in serous carcinomas (13%), but also substantially in carcinosarcomas and clear cell carcinomas (approximately 11%) and less frequently in high-grade endometrioid carcinomas. Importantly, HER2 amplification co-occurred almost invariably with the p53abn phenotype. Previous studies reported similar results [17, 24] These findings raise the question of whether therapeutic eligibility criteria for anti-HER2 treatment should be evaluated based on molecular subgroup rather than histotype. The antibody-drug conjugate trastuzumab-deruxtecan already has also indications for the HER2 low (IHC 1 + /2 + ISH-) group in breast cancer [25]. High drug-to-antibody ratio may explain the efficacy of the conjugated treatment observed also in cases devoid of HER2 amplification. In our cohort, the frequency of HER2 low 2+ was comparable to HER2 amplification, whereas HER2 low 1+ was more common. HER2 low expression was more frequently observed in tumors with high-risk histotype, but the cases appeared not to be restricted to p53abn molecular subgroup. Krakstad et al. reported similar findings [18]. Importantly, low HER2 expression was seen relatively frequently in clinicopathological high-risk cases where additional treatment options are mostly needed. In the phase II DESTINY-PanTumor02 Trial, an overall objective response rate of 57.5% was reported for the 40 patients with endometrial carcinoma. However, the benefit measured by objective response rate, progression-free survival and overall survival was greater in the 13 cases with tumors presenting HER2 IHC 3+ as compared to IHC 2+ [10]. Only few patients with IHC 0-1+ tumors were recruited. Thus, the predictive value of IHC 1+ or ISH negative IHC 2+ as compared to IHC 0 remains unclear. Trastuzumab-deruxtecan has also been used in a small phase II study on HER2 positive (ASCO/CAP gastroesophageal scoring) recurrent uterine carcinosarcoma with positive overall response rates regardless of HER2 expression levels [26]. Scoring algorithms with clinically meaningful thresholds may need to be formulated separately for anti-HER2 and ADC based therapies.

The prognostic value of HER2 expression in endometrial carcinoma has remained mostly unknown as it has been examined in high risk histotypes (serous, carcinosarcoma), which are known to have inherently a poor prognosis. In our results HER2 amplification was associated with aggressive histotypes and the p53abn molecular subgroup. Based on the multivariable analysis HER2 was not associated with risk of cancer-related death which implies that other factors are more important regarding prognosis. Surprisingly in nonendometrioid carcinomas DSS reached nominal statistical significance (p = 0.048) in favor of the HER 2+ low group, which cannot currently be explained.

Strengths of this study include the large size of our study cohort carrying relatively low selection bias, as well as the availability of comprehensive clinicopathological and follow-up data, all obtained from a single tertiary center. HER2 gene status of all the IHC 2+ and 3+ cases was assessed by CISH that is routinely performed in our laboratory. Various studies have reported high concordance between HER2 FISH and HER2 CISH in breast cancer specimens [20]. As an advantage, CISH allows the simultaneous observation of gene copies and tissue morphology of both tumor cells and normal cells functioning as internal control. HER2 expression is known to be heterogeneous in breast and uterine carcinomas. However, high concordance between TMA and whole slide results have been reported for three-core (97.6%) and four-core (98.8%) TMAs as regards HER2 immunostatus in breast cancer specimens [27]. To ensure representability of our TMA we included four biopsy cores from each tumor. As a significant advantage, use of a TMA permitted us to study a very large number of tumor samples including rare histotypes. Occasionally HER2-driven oncogenesis depends on a mutation in the HER2 gene, which leads to increased activity without a change in copy number or protein amount. The frequency of activating mutations of HER2 in endometrial carcinoma is unknown. Detection of these tumors would require next-generation sequencing and they would have remained undetected in our analysis. Currently there are no specific treatments recommended for this group of tumors.

In conclusion, based on HER2 amplification data in endometrial carcinoma, molecular subgroup (p53abn), instead of histotype, should be used as the criteria for analysing HER2 amplification status for future clinical trials of anti-HER2 therapies. In our large study cohort, HER2 low cases presented an even wider range of histotypes and molecular subgroups including many patients with high-risk uterine cancer. These patients could be candidates for targeted treatment with HER2 antibody-drug conjugates. Future clinical trials are needed to establish the clinical relevance of HER2 low status, treatment indications and guidelines for HER2 testing in endometrial carcinoma.