Oral etoposide and trastuzumab in HER2-positive metastatic breast cancer: a retrospective study at Institut Curie Hospitals.

Background: The TOP2A and ERBB2 genes are co-amplied in about 40% of HER2 positive (HER2+) breast cancers. Oral etoposide (VP16), an inhibitor of topoisomerase-II (encoded by TOP2A), has demonstrated clinical activity in metastatic breast cancer (MBC). However, the clinical benet of oral VP16 combined with trastuzumab (VP16-T) in HER2+ MBC has not been evaluated. Methods: Patients treated at Institut Curie Hospitals with VP16-T for HER2+ MBC were retrieved by an in-silico search. Trained medical oncologists retrospectively assessed progression-free survival (PFS), overall survival (OS), response rate, prolonged PFS (dened as a duration of at least 6 months), 6 months clinical benet rate and toxicity. Co-amplication of ERBB2 and TOP2A was assessed by shallow whole genome sequencing on tumor tissue whenever available. Results: Forty-three patients received VP16-T after a median number of six prior treatment lines for HER2+ MBC. Median PFS and OS were 2.9 months (95% CI [2.4-4.7]) and 11.3 months (95% CI [8.3-25.0]), respectively. Three patients had a complete response while 12/40 (30%) had a clinical benet. Only 3 patients stopped treatment for toxicity. Median PFS in the population with and without TOP2A/ERBB2 co-amplication was respectively 4.7 months (95% CI [2.3-NA]) and 2.9 months (95% CI [1.2-NA]; p=0.36). Conclusion: Our analysis suggests a favorable ecacy and toxicity prole for VP16-T in patients with heavily pretreated HER2+ MBC.

Oral etoposide (VP16) is an inhibitor of topoisomerase II. Oral VP16 has demonstrated clinical activity in heavily pre-treated patients with HER2-negative MBC compared to other active chemotherapies such as capecitabine, paclitaxel, eribulin or anthracyclines (17). Although not currently recommended in MBC guidelines, the use of oral VP16 could be relevant in heavily pre-treated MBC, with the advantage of an oral administration, a low cost and manageable toxicity. Moreover, while the cardiac toxicity of anthracyclins is overlapping that of anti-HER2 targeted agents (12), oral VP16 has no reported cardiac toxicity, thus allowing combination therapy.
To the best of our knowledge, there are no studies evaluating the e cacy of oral VP16 in combination with trastuzumab, a combination used in our institute as a palliative, late line therapy for HER2 + MBC patients. Here, we report a retrospective evaluation of the e cacy and safety of oral VP16 combined with trastuzumab in HER2 + MBC and the predictive value of TOP2A/ERBB2 co-ampli cation.

Patients and clinical data
The research project was submitted and approved by the Internal Research Committee of the Institut Curie (No. DATA200187). A waiver of informed consent was obtained because of the retrospective nature of the study.
Patients treated with oral VP16 and trastuzumab were retrieved by an in-silico search in the database of Institut Curie Hospitals (Paris and Saint Cloud, France). Computerized medical les were then manually inspected by experienced medical oncologists. Inclusion criteria were: HER2 + MBC female patients treated with oral VP16 in combination with trastuzumab, regardless of the treatment line. HER2 + tumors were de ned according to the 2018 American Society of Clinical Oncology/College of American Pathologists guidelines (3). Trastuzumab could have been received prior to oral VP16 and continued after VP16-T treatment. All oral VP16 administration regimens were included in the study. The dose of 50mg or 75mg per day for 10-14 days out of 21 was de ned as the standard oral VP16 regimen (28).
The primary objective was to evaluate the progression-free survival (PFS) in HER2 + MBC patients treated by VP16-T. PFS was de ned as the period from initiation of combination therapy to disease progression or death for any cause, whichever came rst.
Secondary objectives were to evaluate overall survival (OS), progression-free survival (PFS) under the prior treatment line, response rate, clinical bene t, toxicity and predictive value of TOP2/ERBB2 coampli cation. OS was calculated from the start of treatment until death from any cause or until the last date the patient was known to be alive. The response rate was measured as the ratio between patients experiencing a partial or complete response as best response, using RECIST 1.1 criteria, and patients who had a measurable disease at treatment start (29). Clinical bene t at 24 weeks was de ned as a PFS > 24 weeks and/or objective tumor response. Toxicities were retrospectively classi ed according to the National Cancer Institute's Common Criteria for Toxicity (version 5.0).
TOP2A/ERBB2 co-ampli cation TOP2A/ERBB2 co-ampli cation was analyzed by shallow Whole Genome Sequencing (sWGS) using Formalin-Fixed Para n-Embedded (FFPE) tumor tissue (30-32) from an available tumor tissue (from metastasis or primary tumor). All slides have been reviewed by a pathologist, to ensure a minimum tumor cellularity of 30%. Between 5 and 50 ng (when available) of tumor DNA were processed with the precapture kit XT-HS2 (Agilent) according to the manufacturing protocol. First, DNAs were fragmented with the ME220 sonicator, reparated, adenylated and ligated with the duplex molecular barcode and the Illumina paired-end sequencing elements during 1h. Then, unique dual sample indexes were added by 14 cycles of PCR ampli cation. The libraries were quali ed and quanti ed by the HS Qubit kit and TapeStation 4200 (Agilent) with the D1000 DNA ScreenTape analysis kit prior to pooling in one single tube. The nal pool was nally quanti ed by qPCR on the 7500 Real-Time PCR System (Thermo Fisher Scienti c). 100 pb paired-end shallow sequencing was performed at Institut Curie core sequencing facility, using an Illumina Novaseq6000.
Sequencing les were pre-processed as indicated in Eeckhoutte et al, 2020 (33). Details are available upon request. Pre-processed alignment les were analyzed by counting and normalizing the number of aligned reads in xed window of 50kb with QDNAseq (34).QDNAseq associates contiguous windows considered to be in the same copy number level in genomic segments. The middle of TOP2A and ERBB2 loci were used to extract from QDNAseq their respective xed window and genomic segment values. QDNAseq outputs were then processed with shallowHRD (33), which extracts a minimal copy number alteration (CNA) cut-off.
The TOP2A/ERBB2 co-ampli cation status was de ned when associated xed window and segment values of both genes were over 4-fold of the CNA cut-off. The absence of TOP2A/ERBB2 co-ampli cation status was de ned when the xed window and segment values of ERBB2 were over 4-fold of the CNA cutoff and those of TOP2A less than 4-fold of the CNA cut-off. Samples were classi ed as "not interpretable" in case of discrepancy between window and segment values for one gene or if no ampli cation of ERBB2 was retrieved by sWGS.

Statistics
Quantitative variables are presented with their median, minimum and maximum. Qualitative variables are presented with the number and percentage. Missing data (not available = NA) are excluded from the denominator for the calculation of percentages. Median follow-up was determined by the inverted Kaplan-Meier method (35). Median values for PFS and OS (with their 95% con dence intervals [CI]) were estimated using the Kaplan-Meier method. All statistical analyses were performed using R 3.6 (36).

Results
Patients and treatment 2,003 patients treated for HER2+ MBC were retrieved by in silico screening of the Institut Curie electronic medical les. Among those patients, 43 met inclusion criteria and were analyzed as part of this retrospective study: their characteristics are shown in Table 1. The median age at diagnosis of primary breast cancer was 47 years (22-80 years). The median age at diagnosis of MBC was 51 years (22-83 years). Synchronous BC metastases were diagnosed in 14 (33%) patients (de novo stage IV). Patients had received a median number of six prior treatment lines (range 0-12) at the time of receiving VP16-T regimen. Thirty-ve patients (81%) had visceral metastases. Oral VP16 regimen was administered at the above-de ned standard doses to 31 patients (72%). Median duration of VP16-T treatment was 2.9 months (0.2 -14.6 months). VP16-T was stopped for disease-progression (n=35 patients, 81%), toxicity (n=3 patients, 7%), therapeutic break (n=3 patients, 7%) or unknown cause (n=2 patients, 5%). Table 1 Patients' characteristics  Figure 1B). Forty patients were eligible for response rate assessment using RECIST 1.1 (Suppl le 1). Four patients (10%) had a partial or complete response to VP16-T. A complete response was observed in 3 patients who received VP16-T given as rst, second and thirteenth line of treatment respectively. One patient had a partial response. Overall, 12 out of 40 evaluable patients (30%) had a clinical bene t at 24 weeks (24 weeks clinical bene t rate: 30%; Figure 2).
Clinical bene t is de ned by either an objective tumor response (N=4 patients) and/or a PFS under VP16-T longer than 6 months (N=8 patients).
The different systemic treatments administered immediately prior to VP16-T are detailed in Suppl le 2 (one patient received VP16-T as rst line treatment Of note, the median number of prior treatment lines in these 6 patients was 5 (range 0-12), as for the overall study population. All patients had previously received taxanes and 63% had previously received anthracyclines. No signi cant difference in response rates, nor in PFS, was found between patients who had previously received or not anthracyclines (data not shown).
Brain metastases were observed in 22 of 40 evaluable patients and in 6 of 12 patients with prolonged PFS. Among these 6 patients with brain metastases and prolonged PFS, only one experienced a disease progression of her brain metastases while receiving VP16-T.

Toxicity
Toxicity has been retrospectively assessed for 42 patients (Table 2). Oral VP16 was discontinued due to toxicity in 3 patients: two for grade 3 nausea/vomiting, one for febrile neutropenia. Nauseas (grade 2 and 3) were observed in 14% of cases. Grade 1 alopecia was recorded in only 1 patient. No diarrhea, mucositis or allergies were observed.

Discussion
To our knowledge, no studies have evaluated the e cacy of oral VP16 and trastuzumab combination in HER2 + MBC. We show that this combination achieves clinically meaningful PFS with about one third of prolonged PFS (de ned as PFS greater than or equal to 6 months), a clinical bene t in a third of patients, and 3 complete responses. PFS and OS were 2.9 months and 11.3 months, respectively. These results were obtained in a heavily pre-treated population with a median number of six prior treatment lines for MBC. Moreover, most of our patient population displayed unfavorable clinical features, such as visceral metastases. Limitations of our study are related to its limited size and retrospective nature. However, this study is the rst speci cally analyzing outcome and toxicity of oral VP16 associated with trastuzumab for HER2 + MBC.
Oral VP16 is a metronomic chemotherapy, de ned as the regular administration of a minimally toxic dose of treatment over an extended period of time. In advanced breast cancer, metronomic chemotherapy has been shown to provide disease control with a lower incidence of adverse events compared to conventional chemotherapy at the maximum tolerated dose (37,38). In the years 1994 to 2000, oral VP16 showed interesting clinical activity in patients with MBC after multiple lines of treatment (39)(40)(41)(42)(43). More recently, a study by Cabel et al. (17) showed survival rates with oral VP16 comparable to other treatment lines including capecitabine, paclitaxel, eribulin or anthracycline (median PFS of 3. Interestingly, the median PFS in the control arm of TH3RESA is similar to that observed with VP16 and trastuzumab in our report. Presence of a co-ampli cation of TOP2A and ERBB2 on chromosome 17 suggests a biological interest to combine oral VP16 and trastuzumab in HER2 + MBC. In keeping with prior reports, our sWGS analysis retrieved a TOP2A/ERBB2 co-ampli cation in 35% of cases. TOP2A/ERBB2 co-ampli cation was numerically, but not statistically, more frequent in patients bene ting from VP16-T. The limited number of patients analyzed prevents any de nitive conclusion about the predictive value of the co-ampli cation. Of note, other non-genetic mechanisms may also modulate the response to topoisomerase 2 inhibitors (20), such as epigenetic mechanisms modulating DNA accessibility (58).

Conclusions
Finally, our retrospective study suggests oral VP16 and trastuzumab may be considered as a treatment option in heavily pre-treated HER2 + MBC patients. This combination yields to prolonged responses in some patients and has the advantage of an oral administration, limited cost and acceptable toxicity. -Consent for publication: Not applicable.
-Availability of supporting data: The datasets used and analyzed during this study are available upon request from the corresponding author.
-Competing interests: The author declared no con ict of interest. -Funding: This research received no speci c grant from any funding agency in the public, commercial, or not-forpro t sectors.
-Authors' contributions: CC, MC, FCB and FL made substantial contributions to conception and design, and revising the manuscript, and gave nal approval for publication. CC, MC, AE, SB, MHS, FCB and FL contributed to acquisition of data, analysis and interpretation of data. CC, MC, AE, FCB and FL drafted the manuscript. CC, MC, and AE performed data analysis for the study. CC, MC, FCB and FL participated in manuscript preparation and revision. All other authors made substantial contributions to the acquisition of data, revising the manuscript, and nal approval. -Acknowledgements: Not applicable.