The age-specific differences in histopathological tumor characteristics and TNM classification of breast carcinomas in Quality assured mamma diagnostic (QuaMaDi) program in the state of Schleswig–Holstein in Germany

Our project was granted ethical clearance by the local ethics committee of the University Hospital in Kiel, Germany, in agreement with the Helsinki Declaration (D470/17).

From the electronic database of the Department of Pathology, University Hospital Schleswig–Holstein, Campus Kiel, we retrieved all patients, who were part of QuaMaDi between 01.01.2005 and 31.12.2016. Only women were included if (1) the referral indicated “QuaMaDi” and (2) a biopsy specimen was submitted to the Dept. of Pathology. Patients were excluded if they were (1) of male gender or (2) if only cytological specimens were submitted. The following patient characteristics were retrieved from the electronic database: age, gender, date of biopsy, biopsy site, number of biopsy samples, histological diagnoses, and the B-classification (since 2007). The following additional data were retrieved for non-invasive and invasive cancers: histological type of the neoplastic lesion, anatomical site, immunoreactivity score (IRS) for the estrogen (ER) and progesterone receptor (PR) (in-situ lesions and invasive cancer), and additionally Elston and Ellis grading (since 2009), HER2 status and Ki67-index (since 2012) for invasive cancers. If patients had undergone surgery at the Dept. of Gynecology and Obstetrics, University Hospital Kiel, the following data were retrieved: date of surgery, age of the patient at the time of surgery, adjuvant or neoadjuvant therapy regimen, histological tumor type, pTNM-classification, grading, and tumor regression (according to Sinn regression score, Sinn et al. 1994). All patient-related data were pseudonymized after inclusion into the study.

Tissue specimens were fixed in formalin and embedded in paraffin. Deparaffinized sections were stained with hematoxylin and eosin. The grading system of Elston and Ellis was applied since 2009 to invasive cancers (Elston and Ellis 1991). The pTNM-stage of all study patients was determined according to sixth and seventh editions of the UICC guidelines. All biopsy and resection specimens had been examined by trained and board-certified surgical pathologists.

According to national guidelines for the diagnosis and treatment of breast cancer, the histopathological lesions in core (NCB) and vacuum-assisted biopsies (VAB) were categorized according to the B-classification (Ellis et al. 2004). In brief: category B1 denotes normal breast tissue or insufficient sampling, B2 benign lesions, B3 benign lesions of unknown biological potential, B4 lesions suspicious for malignancy, B5a in situ carcinomas, B5b invasive breast cancer, B5c indiscernible whether non-invasive or invasive, and B5d malignant tumor of divergent histology (e.g., malignant phylloides tumor) (S1).

Immunohistochemical staining was done using a Bondmax automated slide staining system (Leica Biosystems, Wetzlar, Germany), the Polymer Refine Detection Kit (Leica Biosystems) and antibodies directed against ER (dilution 1:150), PR (both Leica Biosystems Newcastle, Newcastle, United Kingdom; dilution 1:100), Her2/neu (1:100) and anti-Ki-67 antibody (both Thermo Fisher Scientific, Fermont, USA; dilution 1:300). For chromogenic in situ hybridization the ZytoDot^® 2C SPEC ERB2/CEN 17 Probe Kit (ZytoVision GmbH, Bremerhaven, Germany) was used.

The ER and PR status were assessed according to Remmele and Stegner (1987). In brief: Category A documented the intensity of nuclear staining as absent (0), weak (1), moderate (2) or strong (3). Category B documented the percentage of stained tumor cells as absent (0), < 10% (1), 10–50% (2), 51–80% (2), and > 80% positive nuclei (4). The immunoreactivity score (IRS) was calculated according to the formula: Category A (immunostaining intensity) × Category B (proportion of positive cells of the tumor). An IRS of 0 and 1 was interpreted as receptor negative. IRS ≥ 2 were considered receptor positive. The range of 2–3 represented a weak, 4–8 a moderately strong, and 9–12 a strong expression of the hormone receptors. Over time the IRS was abandoned, because scores between 1 and 3 could indicate both receptor negative and positive status. Hence, the percentage of positive tumor cells in the specimen was crucial for the assessment. Since 2010, each tumor was assessed according to ASCO/CAP guidelines for ER and PR testing (Hammond et al. 2010). Hormone receptor status for ER and PR was considered positive if ≥ 1% of tumor cells were immunoreactive. Breast carcinoma was hormone receptor negative if < 1% of tumor cells stained (with positive internal control). The IRS was still optionally reported in the pathological findings. Finally, the hormone receptor status was determined by using the documented IRS in the pathological report. Since an IRS of 1, 2 and 3 could indicate both a negative and positive hormone receptor status, assignment to the categories hormone receptor positive and negative was based on the pathologist’s evaluation as documented in the pathology report. Ambiguous reports were re-examined during this study by a board-certified surgical pathologist. All IRS values ≥ 4 were assigned to positive hormone receptor status.

Since 2012, the Ki-67 index was assessed according to recommendations of the International Ki-67 in Breast Cancer Working Group (Dowsett et al. 2011). After prior overview of the entire tissue specimen, the evaluation of the Ki-67 proliferation index was carried out using at least 3 high-power (× 40 objective) fields. The fields were selected to reflect the staining of the whole specimen. The invasion front of the tumor was evaluated. Only nuclear staining was assessed. The Ki-67 proliferation index is defined as the percentage of positively stained cells among the total number of malignant cells scored. A minimum of 100 cells was analyzed and determined by single cell counting using a counting device. If distinct clusters of positively stained cells were seen in the preparation, they were included in the overall evaluation. Finally, the Ki-67 index was categorized into the following groups: ≤ 10%, 10–20%, 20–25%, 25–50%, > 50% and unclassifiable. Ranges and no exact value of Ki-67 were assigned as unclassifiable.

Scoring of each breast carcinoma was assessed on core biopsy specimens according to the ASCO/CAP recommendations for HER2 testing (Wolff et al. 2007, 2013). The immunostaining intensity and extent as well as percentage of positive tumor cells was evaluated. On slide positive controls were used in each case. During study period, the recommendations of ASCO/CAP were constantly updated.

If the HER2 Score was 2 + a reflex test on the same biopsy specimen was performed using the chromogenic in situ hybridization according to the ASCO/CAP recommendations for HER2 testing (Wolff et al. 2007, 2013). The amplification was determined by examining average HER2 copy number and since 2007 the HER2/CEP 17 ratio. During study period, the recommendations of ASCO/CAP were constantly updated. The in situ hybridization status used in this study are listed.

The classification of the molecular subtypes was based on the St. Gallen recommendations (Goldhirsch et al. 2011) with IHC analysis of ER, PR, HER2, Ki-67 and CISH analysis of HER2. The molecular subtypes were defined as follows: luminal A (ER positive, HER2 negative, Ki-67 < 20%), luminal B (either ER positive, HER2 positive and Ki-67 ≥ 20% or ER positive, HER2 negative and Ki-67 ≥ 20%), HER2 positive (ER negative, HER2 positive, any Ki-67) and triple-negative (ER negative, HER2 negative, any Ki-67).

During the entire study period, the Dept. of Pathology regularly participated successfully in external quality assurance schemes. These comprised the annual participation in the Quality Assurance Initiative for Pathology (QuIP) for ER, PR, Ki-67, HER2, and HER2 CISH. The proficiency test results are published and available on the homepage of the Dept. of Pathology (https://​www.​patho.​uni-kiel.​de/​krankenversorgun​g/​qualitätssicheru​ng). In addition, the Dept. of Pathology voluntarily participated in the HER2-Monitor of the Hannover Medical School. There, the internal positivity rate of HER2 is constantly compared with the nationwide average. Furthermore, the histological examinations were carried out by trained and board-certified surgical pathologists and any ambiguous finding was re-assessed by at least one additional board-certified surgical pathologist (four-eye principle).

SPSS version 25.0.0.2 (IBM Corp., Armonk, NY, USA) was used for statistical analyses. To test for correlation between non-ordinal variables, we used Fisher´s exact test. When testing for correlation between variables of nominal scale, we used Pearson´s chi square test. We assumed a significance level of 0.05. The p-values were calculated by two-sided T-tests and a p-value ≤ 0.01 was considered highly significant.

In total, 7.111 biopsy specimens and 2.887 corresponding resection specimens obtained from 6.845 women were assigned to the QuaMaDi program. The median age of the patients was 55 years (range from 15 to 94 years for biopsy specimens; 26 to 94 years for resection specimens). 6.586 women participated in the QuaMaDi program once, 252 twice and 7 three times. In 782 (11.4%) women, more than one biopsy (different localization within the breast or different anatomical sites) was obtained per referral. We categorized our patients into three age groups: pre- (< 50 years), peri- (50–69 years) and post-screening group (≥ 70 years): 2.685 (37.7%) of 7.111 histological referrals were obtained from women < 50 years (median age 44 years; age range 15–49), 2.935 (41.3%) from women between 50 and 69 years (median age 60; age range 50–69), and 1.491 (21.0%) from women ≥ 70 years (median age 75; age range 70–94).

First, we correlated the three different age groups with various pathological patient characteristics, i.e., anatomical localization (C-code), B-category, tumor type (M-category), the presence of precursor lesions, tumor grade, receptor status and Ki-67 index and molecular subtypes in patients with a single biopsy. As summarized in Table 1 significant differences between the three age groups were found for all parameters except the Ki-67 index.

The percentage of invasive tumors (B5b) increased from 30.8% in the pre- to 73.2% in the post-screening group, while the percentage of benign lesions (B2) decreased from 50.5 to 11.7%. The proportion of benign lesions with uncertain biological potential (B3) halved from 8.3 to 3.7% (Table 1). While the risk of detecting cancer increased continuously with age, a substantial number of women aged < 50 years, i.e., 35% had an in-situ carcinoma (B5a) or an invasive carcinoma (B5b).

Invasive carcinoma NST was found most commonly in the pre-screening group (81.3%), while invasive lobular (17.5%) and mucinous (4.2%) carcinomas were found most commonly in the post-screening group (Table 1). Conclusively, although invasive carcinoma NST was the most common tumor type in each age group, better differentiated carcinomas, such as mucinous carcinomas, were more likely to occur with increasing age.

Likewise, the detection rate of precursor lesions increased significantly with patient age (Table 1). 29% of the high-risk lesions and in situ carcinomas occurred in women < 50 years. The proportion of well-differentiated carcinomas (G1) also increased with age, whereas the proportion of poorly differentiated carcinomas (G3) decreased significantly from 18.9% to 11.8% (Table 1). Regardless of age, moderately differentiated carcinoma (G2) was diagnosed in more than half of the cases. The results show that younger women (< 50 years) were more likely to have more aggressive, poorly differentiated carcinomas.

There were highly significant differences for the ER- and HER2 status: The proportion of ER-positive carcinomas increased from 81.7% in the pre- to 88.1% in the post-screening group. The percentage of HER2-positive carcinomas halved from 16.9 to 8.4%. No significant differences between the three age groups were demonstrated for the PR status and Ki-67 index. Overall, the results indicate that elderly women (≥ 70 years) show a higher frequency of ER positive and HER2-negative carcinomas, whereas those under 50 years tended to have ER negative and HER2-positive carcinomas, underscoring the higher frequency of more aggressive breast carcinomas in the younger age group.

The significant differences in the hormone receptor status mirrored in the four molecular subtypes, i.e., luminal A, luminal B, HER2 positive and triple-negative. The proportion of triple-negative breast carcinomas decreased significantly from 28.2% in the pre- to 19.1% in the post-screening group (Table 1). In contrast, the proportion of Luminal A carcinomas doubled from 17.1% in the pre- to 31.2% in post-screening group. Luminal B occurred most frequently in all age groups, accounting for more than 40%. Overall, women < 50 years exhibited 30.5% triple-negative and HER2-positive carcinomas.

A total of 7.111 clarification diagnoses consisting of both single and multiple biopsies were categorized into the three age groups. The comparison of the three age groups showed a highly significant difference. The proportion of invasive carcinomas increased from 32.6% in the pre-screening to 74.6% in the post-screening group (S2). In particular, a linear increase of malignant breast lesions of 2.4% per year of age was observed in the age range between 43 and 77 years (Fig. 1). In addition, the ratio of benign to malignant lesions increased from an initial 1:0.64 for women < 50 years to 1:1.99 for women between 50 and 69 years. A markedly higher ratio of benign to malignant lesions, i.e., 1:4.87, occurred for women ≥ 70 years. Overall, 54.5% of breast carcinomas appeared beyond the screening age. In particular, 23.9% of the diagnosed carcinomas were detected in women < 50 years.

We identified 2.887 resection specimens of QuaMaDi, of which 2.454 cases had been assigned to primary surgical therapy (non-neo group) and 433 had received neoadjuvant therapy (neo group) (Table 2). Comparing the three age groups, we found significant differences regarding tumor type, local tumor growth (pT category), nodal spread (pN category), lymph vessel invasion (pL category), tumor grade and resection status (R category) in patients who were primarily resected and had not received neoadjuvant treatment (Table 2). While in women of the neo group, significant differences were only found for the pT category. In all three age groups, the tumor stage pT1c was found most commonly in the resection specimens of the non-neo group. Overall, 70.3% of the untreated carcinomas in the pre-, 75.2% in the peri- and 63.3% in the post-screening group had a tumor size < 2 cm (pT1). In the neoadjuvant treatment regimen, a pathologic complete remission (defined as the absence of tumor residuals in the breast and axilla) was achieved in one-fifth of the patients in both the pre- and peri-screening groups (ypT0). These results show that the majority of the carcinomas were detected at an early, more prognostically favorable stage of the disease in all three age groups. However, higher and prognostically less favorable tumor stages were found with a higher probability in women < 50 and ≥ 70 years.

All women who had participated in QuaMaDi presented most commonly without lymph node metastases (54.1–74.5%; Table 2). Regarding the lymph node status, women in the pre-screening group were more likely to present with lymph node metastases in the non-neo group (Table 2) further substantiating that these women suffer from more aggressive tumors.

We examined the concordances of morphology in 2.561 and grading in 983 eligible cases with corresponding biopsy and resection specimens. Regarding histological diagnoses, 2.152 (84.0%) diagnoses obtained in biopsy specimens were concordant with the diagnoses in the resection specimens. In 115 cases (4.5%), no (invasive) carcinoma was identified in the biopsy, but invasive carcinoma was found in the resection specimen (S3). Most discordant morphologies were apparent between invasive carcinoma NST and lobular carcinoma.

Likewise, an identical tumor grade was given in 760 of 983 (77.3%) corresponding biopsy and resection specimens (S4). A higher tumor grade had been documented in 40 (4.0%) biopsy specimens and a lower tumor grade in 183 (18.6%) compared with the corresponding resection specimens. Among these, two carcinomas were graded as well-differentiated carcinoma in biopsy but presented as poorly differentiated carcinoma in resection specimens. The results show that there was a tendency for undergrading in biopsies. However, overall high agreement was achieved. Tumor grading in biopsy specimens was representative in more than three quarter of the cases.