Comparison of mutation profile between primary phyllodes tumors of the breast and their paired local recurrences

Introduction

Phyllodes tumor of the breast (PTB) is a rare neoplasm (it constitutes from 0.2 to 2.0% of all breast neoplasms in women), composed of elements of glandular epithelium, myoepithelium and stroma, which determines the biology of this cancer [1]. The World Health Organization classification distinguishes three PTB subtypes (benign, borderline and malignant) [2]. These subtypes present a morphological continuum from benign to malignant [1]. The classification is based on a combination of several histological features including: the degree of stromal atypia, stromal overgrowth, mitotic activity and type of tumor border. For this reason, histopathological diagnosis of PTB is difficult and based only on histological features, it is not always possible to assess the course of the disease and the risk of cancer progression [3]. Currently, none of the biomarkers tested and described in the literature have been used in everyday clinical practice. Pathogenesis and molecular biological characteristics of PTB are unknown and no PTB prognostic factors have been identified. Generally, accepted treatment approach for patients with all of PTB types is breast conserving surgery with a margin >1 mm, which is sufficient to prevent local recurrence of PTB [4, 5]. There are no indications for resection of axillary lymph nodes or sentinel lymph node biopsy, because metastases to the lymph nodes occur very rarely [4, 6]. There are also no indications for hormonal therapy in patients with PTB. Adjuvant radiotherapy is recommended in borderline and malignant PTB, if the resection margin is less than 1 cm [4, 5, 7].
Most patients with diagnosed PTB have a good prognosis, with local recurrence risk from 17% in the benign type, up to 27% in the malignant type [8]. Distant metastases occur even in 22% of malignant PTB cases [8, 9]. The histological subtype of PTB is the only prognostic factor in these patients [6]. However, in some individuals, histological features do not allow for adequate assessment of the course of the disease (prediction of local recurrence, distant metastases or survival). This research has been undertaken to search for gene mutations related to the process of local recurrence in breast phyllodes tumor. If the onset of recurrence is caused by specific molecular changes, these changes should be present in most PTB recurrences and possibly in their respective primary tumors.

Material and methods

Patients

Between 1990 and 2013, 159 patients diagnosed with PTB were treated at the Oncology Center in Krakow. In this group, 16 patients developed local recurrence. In 4 patients, after re-evaluation of pathology slides recurrent tumors were classified as sarcomas. Two patients were operated outside the Oncology Center and obtaining tissue material from the primary tumor was impossible. This pilot study was carried out on a group of 10 selected patients. There were 4 malignant primary PTBs (in one case there were two clearly separated components: benign and malignant), 2 borderline PTBs, 3 mixed types (benign/borderline) and 1 benign PTBs (Table I). In 4 cases progression was observed (Table I).
The study has been reported to the Ethical Committee at the Regional Medical Chamber in Krakow and received positive decisions (number: L.DZ.OIL/KBL/1/220/71220). This is a retrospective study using archived preserved tissues. There were no direct contact with patients, no modification of diagnostic or treatment procedures. None of personal patients’ data were revealed and no specific patient consent was needed.

Next-generation sequencing

DNA was extracted from formalin-fixed paraffin-embedded tissues. Before extraction the tissue slides were evaluated by a pathologist to confirm the diagnosis and to select the paraffin blocks most suitable for molecular analysis. In the case of primary tumor with two clearly separated components (benign and malignant), for each component DNA was isolated separately. DNA was obtained from tissue sections using a semi-automatic method with Maxwell® RSC Instrument (Promega). The Maxwell® RSC DNA FFPE Kit (Promega) was used to isolate genomic DNA using paramagnetic particles. The amount of DNA was assessed fluorimetrically using the Qubit™ dsDNA HS Assay Kit and Qubit 3.0 device (ThermoFisher Scientific). Next-generation sequencing (NGS) was performed on IonTorrent platform. Two hundred and seven regions in 50 genes that are most often mutated in solid tumors were sequenced with Cancer Hotspot Panel v2. Libraries for sequencing were prepared automatically on Chef instrument using Ion Ampliseq™ Cancer Hotspot Panel v2 Chef-Ready Kit. Sequencing was performed using S5 instrument, Ion 510™ & Ion 520™ & Ion 530™ Kit and Ion 530™ Chip Kit. S5 Torrent Server Torrent Suite 5.10.1 software was applied for coverage analysis and to identify genes variants. Results (vcf files) were annotated using wANNOVAR (http://wannovar.wglab.org). Variants for which satisfactory sequencing conditions were obtained and which are estimated to occur in the population with a frequency below 1% were qualified for further analysis.
Variants occurring in the population with a frequency above 1% are classified as polymorphisms. With the help of the Varsome (https://varsome.com) and ClinVar (https://www.ncbi.nlm.nih.gov/clinvar) databases variants were classified into 5 classes: pathogenic, probably pathogenic, of unknown significance, probably benign and benign. The workflow of the experiment is presented in Figure 1.

Results

Assessment of NGS data

Next-generation sequencing (NGS) was performed for 10 pairs consisting of primary and recurrent tumor. The most important parameters of the run are presented in Table I. The age of the blocks ranged from 5.6 to 24.7 years. For all samples there was enough of DNA to perform sequencing. However, in the case of two pairs with the oldest blocks (aged 23-24 years) the data obtained were of low quality and mutation analysis for these samples was not possible (Table I). DNA from these samples was too fragmented, especially from primary tumor. These data were characterized by too low number of reads, low uniformity (65-88%), percent reads on target (33-86%) and low target base coverage at 100× (i.e. the percentage of bases for which 100 readings were obtained, the recommended minimum required for analysis is 98%) (Table I). For all other cases, coverage at 100× was 100%, so even for 17-year-old blocks it was possible to obtain material that allowed analysis. However, it should be noted that for blocks aged 13-17 years a large number of observed changes should be considered as sequencing errors (low sequencing parameters) and changes typical for the formalin-fixed material (change G>A and T>C).

Comparison of primary and recurrent tumor

Eight primary tumor – recurrence pairs (16 samples) were qualified for the analysis of mutation status. In 4 out of 8 pairs, probably pathogenic/pathogenic variants, according to Varsome and/or ClinVar databases, were detected. In three of them mutations were found in CDKN2A gene (Table II). In two pairs, the same variant creating the STOP codon and causing synthesis of the truncated protein was detected. The other variant in CDKN2A gene was a non-synonymous change, resulting in insertion of another amino acid. In the case of the PTEN gene, there was a change of the reading frame, resulting in truncated protein, while in the TP53 gene non-synonymous change was found (Table II). Moreover, in one patient, two clearly separated components: benign and malignant were identified in the primary tumor. Both components were isolated and tested separately. Pathogenic/probably pathogenic variants in the CDKN2A and TP53 genes were detected only in the malignant component. The same mutations were also detected in the recurrent tumor. Only in one pair, a mutation that was not present in the primary tumor (involving the EGFR gene) was detected in the recurrence.

Discussion

Breast phyllodes tumor is a rare breast neoplasm and most patients with diagnosed PTB have a good prognosis. Metastatic disease is observed less frequently than local recurrences, however, it is associated with worse prognosis [10]. Recurrences of malignant PTB are found in 23-30% of patients, while in benign and borderline PTBs recurrences appear in 10-20% and 14-25% of cases, respectively [1]. The frequency of local recurrence is higher in patients with malignant and borderline PTB than in benign cases [9]. Breast conserving therapy is proposed in all cases of local PTB recurrences, if a margin of healthy tissue can be achieved with good cosmetic result [11]. Benign PTB is successfully treated with surgery alone. In borderline and malignant type, with a tumor-free margin < 1 cm (0.3-0.8 cm), adjuvant radiotherapy is recommended [4]. Only in the cases of extensive recurrences, simple mastectomy may be necessary to obtain a negative margin.
In our study, we discovered mutations in primary tumour and also in matched local recurrence in CDKN2A (in 3 pairs), PTEN (1 pair) and TP53 (1 pair) genes. Additionally, in one pair only in the recurrent lesion mutation in the EGFR gene was observed. According to COSMIC database (https://cancer.sanger.ac.uk/cosmic) a frequently mutated gene in phyllodes tumours is MED12 (55%). Unfortunately, Cancer Hotspot Panel v2 does not asses mutation in MED12 gene. However, mutations of MED12 gene were less frequently observed in malignant phyllodes tumors than in fibroadenomas, benign and borderline phyllodes tumors [1]. Moreover, the presence of MED12 mutation was associated with longer disease-free survival, whereas its absence was related to a higher likelihood of recurrence [1].
In our study, we observed pathogenic variants in PTEN, TP53, EGFR and CDKN2A genes. The first three of them are reported by COSMIC in “Top 20 genes” for phyllodes tumours: TP53 (mutation in 12%), EGFR (6%), PTEN (4%). In our study, most frequently mutated gene was CDKN2A, in which mutation were observed both in primary tumors and local recurrences. Mutations or homozygous deletions in CDKN2A gene in phyllodes tumour are also reported by other authors [12, 13, 14]. Tan et al. studied a group of 20 cases of phyllodes tumour, including 11 cases with no recurrence and 9 with local recurrence/metastases. In this group, homozygous deletion at 9p21 involving CDKN2A gene was reported in two cases with progression, while in two other cases with progression single copy loss at 9p21 was noted. None of these aberrations was observed in cases that did not develop recurrences [13]. In our study, we detected pathogenic variants in the CDKN2A gene in as many as three out of eight pairs of primary and recurrent tumors. Tan et al. and our findings may suggest that CDKN2A gene is involved in the development of primary PTB and local recurrences/metastases.
It is worth to mention that all mutations detected in the paired samples affected the suppressor genes:
• CDKN2A – encodes a protein that is a CDK4-dependent cyclin kinase inhibitor and a P53 stabilizing protein,
• PTEN – encodes a protein that negatively regulates the PI3K/AKT signaling pathway,
• TP53 – encodes a protein involved in, among others, a mechanism of controlling cell arrest in the cell cycle, apoptosis and DNA repair.
The allelic frequency of variants detected in the sequenced material from the primary tumor and recurrence (40-60%) suggests that they might be hereditary changes. In order to confirm this hypothesis in the future research, an attempt will be made to assess the presence of the detected variants in archival material derived from normal tissue (from the paired samples) by Sanger sequencing.
Only in one pair, in the recurrence but not in the primary tumor, pathogenic variant in the EGFR gene was found. This might be a possible molecular hallmark of progression. In order to determine if this variant is associated with progression, the frequency of this mutation should be examined in a larger group of malignant and borderline tumors.

Conclusions

The CDKN2A gene might be involved in the development of phyllodes tumor of the breast and its local recurrences/metastases.
Mutations in the EGFR gene might be a possible molecular hallmark of progression.
Funding: Polish National Science Centre Grant “Miniatura 2”, number: 2018/02/X/NZ5/02868.
The authors declare no conflict of interest.

References

1. Tan BY, Acs G, Apple SK, et al. Phyllodes tumours of the breast: a consensus review. Histopathology 2016; 68: 5-21.
2. Hoon Tan P, Ellis I, Allison K, et al. The 2019 WHO classification of tumours of the breast. Histopathology 2020.
3. Tan PH, Thike AA, Tan WJ, et al. Predicting clinical behaviour of breast phyllodes tumours: a nomogram based on histological criteria and surgical margins. J Clin Pathol 2012; 65: 69-76.
4. Mitus JW, Blecharz P, Jakubowicz J, et al. Phyllodes tumors of the breast. The treatment results for 340 patients from a single cancer centre. Breast 2019; 43: 85-90.
5. Mitus JW, Blecharz P, Wysocki W. The appropriate width of the tumour-free margin in surgery of phyllodes tumour of the breast. Oncology and Radiotherapy 2019; 1: 013-014.
6. Mitus JW, Blecharz P, Reinfuss M, et al. Changes in the clinical characteristics, treatment options, and therapy outcomes in patients with phyllodes tumor of the breast during 55 years of experience. Med Sci Monit 2013; 19: 1183-1187.
7. Tremblay-LeMay R, Hogue JC, Provencher L, et al. How Wide Should Margins Be for Phyllodes Tumors of the Breast? Breast J 2017; 23: 315-322.
8. Yom CK, Han W, Kim SW, et al. Reappraisal of conventional risk stratification for local recurrence based on clinical outcomes in 285 resected phyllodes tumors of the breast. Ann Surg Oncol 2015; 22: 2912-2918.
9. Mitus JW, Blecharz P, Walasek T, et al. Treatment of Patients with Distant Metastases from Phyllodes Tumor of the Breast. World J Surg 2016; 40: 323-328.
10. Chaney AW, Pollack A, McNeese MD, et al. Primary treatment of cystosarcoma phyllodes of the breast. Cancer 2000; 89: 1502-1511.
11. Mitus JW, Jakubowicz J, Walasek T, et al. Are patients with phyllodes tumor of the breast good candidates for breast conserving therapy? A single center experience. Eur J Gynaecol Oncol 2018; 39: 641-644.
12. Nozad S, Sheehan CE, Gay LM, et al. Comprehensive genomic profiling of malignant phyllodes tumors of the breast. Breast Cancer Res Treat 2017; 162: 597-602.
13. Tan WJ, Lai JC, Thike AA, et al. Novel genetic aberrations in breast phyllodes tumours: comparison between prognostically distinct groups. Breast Cancer Res Treat 2014; 145: 635-645.
14. Liu SY, Joseph NM, Ravindranathan A, et al. Genomic profiling of malignant phyllodes tumors reveals aberrations in FGFR1 and PI-3 kinase/RAS signaling pathways and provides insights into intratumoral heterogeneity. Mod Pathol 2016; 29: 1012-1027.