The role of serum thymidine kinase 1 activity in neoadjuvant-treated HER2-positive breast cancer: biomarker analysis from the Swedish phase II randomized PREDIX HER2 trial

BACKGROUND

Thymidine kinase 1 (TK1) plays a pivotal role in DNA synthesis and cellular proliferation. TK1 has been studied as a prognostic marker and as an early indicator of treatment response in human epidermal growth factor 2 (HER2)-negative early and metastatic breast cancer (BC). However, the prognostic and predictive value of serial TK1 activity in HER2-positive BC remains unknown.

METHODS

In the PREDIX HER2 trial, 197 HER2-positive BC patients were randomized to neoadjuvant trastuzumab, pertuzumab, and docetaxel (DPH) or trastuzumab emtansine (T-DM1), followed by surgery and adjuvant epirubicin and cyclophosphamide. Serum samples were prospectively collected from all participants at multiple timepoints: at baseline, after cycle 1, 2, 4, and 6, at end of adjuvant therapy, annually for a total period of 5 years and/or at the time of recurrence. The associations of sTK1 activity with baseline characteristics, pathologic complete response (pCR), event-free survival (EFS), and disease-free survival (DFS) were evaluated.

RESULTS

No association was detected between baseline sTK1 levels and all the baseline clinicopathologic characteristics. An increase of TK1 activity from baseline to cycle 2 was seen in all cases. sTK1 level at baseline, after 2 and 4 cycles was not associated with pCR status. After a median follow-up of 58 months, 23 patients had EFS events. There was no significant effect between baseline or cycle 2 sTK1 activity and time to event. A non-significant trend was noted among patents with residual disease (non-pCR) and high sTK1 activity at the end of treatment visit, indicating a potentially worse long-term prognosis.

CONCLUSION

sTK1 activity increased following neoadjuvant therapy for HER2-positive BC but was not associated with patient outcomes or treatment benefit. However, the post-surgery prognostic value in patients that have not attained pCR warrants further investigation.

TRIAL REGISTRATION

ClinicalTrials.gov, NCT02568839. Registered on 6 October 2015.

Combined assessment of metabolic response and tumor infiltrating lymphocytes as a predictor of outcomes following neoadjuvant therapy for HER2-positive breast cancer: Results from the randomized PREDIX HER2 trial

593

Background: Abundance of tumor infiltrating lymphocytes (TIL) is prognostic in early HER2-positive breast cancer (BC). Response to neoadjuvant therapy (NAT) according to positron emission tomography combined with computed tomography (PET-CT) has been shown to predict pathologic complete response (pCR). There is paucity of data regarding long-term prognostication using PET-CT and the potential value of the combined assessment of both these biomarkers. Methods: PREDIX HER2 (NCT02568839) is a prospective randomized phase 2 trial that compared standard NAT (docetaxel, trastuzumab, pertuzumab) with trastuzumab emtansine, in patients with HER2-positive BC. Overall, 202 patients were included (197 evaluable) and the primary efficacy analysis showed no difference in pCR or event-free survival (EFS) between the two groups (Hatschek, JAMA Oncology 2021). Assessment with fluorine 18–labeled fluorodeoxyglucose PET-CT was performed at baseline and after 2 and 6 treatment cycles, and SUVmax was evaluated as a continuous variable. TILs were assessed at baseline biopsies according to guidelines from the International TIL Working Group (J.H.). The aim of this secondary analysis was to investigate the combined assessment of TIL and PET-CT as an early predictor of response to NAT. Results: Overall, 112 patients underwent baseline PET-CT and 109 after C2, whereas 173 had baseline TIL. In multivariable analysis, baseline SUVmax did not predict pCR (ORadj= 1.04, 95% CI 0.97-1.12, p = 0.259) or EFS (HRadj= 1.07, 95% CI 0.98-1.17, p = 0.117). In contrast, higher SUVmax at C2 predicted lower pCR (ORadj= 0.65, 95% CI 0.48-0.87, p = 0.005) and worse EFS (HRadj= 1.18, 95% CI 1.04-1.34, p = 0.01). Baseline TIL > 10% (median cut-off) provided additional prognostic information to clinical parameters (stage and hormone receptor expression) and C2 SUVmax (LR-Δχ2 = 7.19, p = 0.007; ORadj= 3.52, 95% CI 1.37 – 9.06, p = 0.009). 75% of patients with high TIL and C2 SUVmax < 2.49 achieved pCR, compared with 13.8% of those with low TIL and high C2 SUVmax and 39.1%-41.3% for the intermediate groups (p = 0.001). Conclusions: SUVmax after two cycles of NAT for HER2-positive BC is an independent predictor of both short- and long-term outcomes. A combined assessment with TIL may facilitate early selection of good responders for de-escalation and poor responders for alternative treatment strategies. Clinical trial information: NCT02568839.

Long-term outcomes of adjuvant denosumab in breast cancer: Fracture reduction and survival results from 3,425 patients in the randomised, double-blind, placebo-controlled ABCSG-18 trial

507

Background: State-of-the-art adjuvant endocrine therapy with aromatase inhibitors (AI) compromises bone health in postmenopausal patients with hormone receptor-positive (HR+) breast cancer, increasing fracture incidence. Adjuvant treatment with the anti-RANK ligand denosumab (Dmab) counteracts these side effects and may improve outcomes. We here report the final long-term outcomes of the ABCSG-18 trial (ClinicalTrials.gov NCT00556374). Methods: In this prospective, double-blind, placebo-controlled, phase 3 trial, 3,425 postmenopausal patients with early HR+ breast cancer on AI therapy were randomised in 58 trial centers between 2006 and 2013 to receive either Dmab 60 mg or placebo s.c. every 6 months (q6m). The primary endpoint was time to first clinical fracture, secondary disease outcome-related endpoints were disease-free survival (DFS), bone-metastasis free survival (BMFS), and overall survival (OS). In addition to the main endpoint analyses reported previously, exploratory long-term follow-up was conducted in ABCSG-18. Main time-to-event analyses were based on stratified Cox models. Sensitivity analyses accounting for treatment cross-over associated with a late Dmab open-label phase, as well as for receiving any anti-resorptive agents were performed. Results: For this final protocol-defined analysis, median follow-up is 8 years (Q1,3: 6, 9.6), and all patients had ended their randomly assigned double-blind treatment (Dmab 60mg s.c. q6m n = 1711; placebo s.c. q6m n = 1709) for a median of 5 years. DFS was improved in the Dmab group versus the placebo group (309 versus 368 DFS events, hazard ratio (HR) 0.83, 95% CI 0.71-0.97, p = 0.016), resulting in an absolute 9-year DFS difference of 3.5% (79.4% vs 75.9%, respectively). When censoring for late cross-over and use of anti-resorptive agents, the DFS difference was confirmed (HR 0.82, p = 0.010). BMFS was improved by 19 per cent (HR 0.81, 95% CI 0.65-1.00, p = 0.047) in the Dmab group, and OS was improved by 20 per cent in the uncensored analysis (127 versus 158 OS events, HR 0.80, 95% CI 0.64-1.01, p = 0.065), and 26 per cent after censoring (HR 0.74, 95% CI 0.58-0.94, p = 0.013). The previously reported marked reduction in clinical fractures persisted even long-term, with 201 fractures in the Dmab and 255 fractures in the placebo group (HR 0.76, 95% CI 0.63-0.92, p = 0.004). No new toxicities for this (low) bone-protective dose of adjuvant Dmab were reported, particularly no ONJ occured.. Conclusions: Adjuvant Dmab 60mg every 6 months during AI therapy is safe, and markedly reduces treatment-induced clinical fractures even in the long-term. DFS, BMFS, and OS are improved in this descriptive final long-term analysis of ABCSG-18. Adjuvant denosumab should be considered for routine clinical use in postmenopausal patients with HR+ breast cancer. Clinical trial information: NCT00556374.

Event-free survival by residual cancer burden after neoadjuvant pembrolizumab + chemotherapy versus placebo + chemotherapy for early TNBC: Exploratory analysis from KEYNOTE-522

503

Background: KEYNOTE-522 (NCT03036488) tested the benefit from adding pembrolizumab (pembro) to chemotherapy (chemo) in patients (pts) with early TNBC. The primary results showed statistically significant and clinically meaningful improvements in pCR and EFS with pembro.Prior studies have shown the prognostic value of the residual cancer burden (RCB) method to quantify the extent of residual disease after neoadjuvant chemo. In this exploratory analysis, we assessed EFS by RCB in KEYNOTE-522. Methods: 1174 pts with previously untreated, nonmetastatic, stage T1c/N1-2 or T2-4/N0-2 TNBC were randomized 2:1 to pembro 200 mg Q3W or placebo (pbo) given with 4 cycles of paclitaxel + carboplatin, then 4 cycles of doxorubicin or epirubicin + cyclophosphamide. After definitive surgery, pts received pembro or pbo for 9 cycles or until recurrence or unacceptable toxicity. Dual primary endpoints are pCR and EFS. RCB was assessed by the local pathologist at the time of surgery. The association between RCB categories (RCB-0, -1, -2, -3, corresponding to increasingly larger residual cancer) and EFS was assessed based on a Cox regression model with treatment as a covariate. Results: Median follow-up was 39.1 months at data cutoff (23 MAR 2021). Pembro shifted RCB to lower categories across the entire spectrum (Table). The HRs (95% CI) for EFS were 0.70 (0.38 - 1.31) for RCB-0 (equivalent to pCR), 0.92 (0.39 - 2.20) for RCB-1, 0.52 (0.32 - 0.82) for RCB-2, and 1.24 (0.69 - 2.23) for RCB-3. The most common EFS event in both arms was distant recurrence, which occurred in fewer pts in the pembro arm in all RCB categories. Conclusions: Increased RCB score was associated with worse EFS. Pts with residual disease had lower RCB values in the pembro arm, including fewer pts with RCB-3. Pembro + chemo prolonged EFS vs chemo alone in the RCB-0, -1, and -2 categories; the small sample size limits interpretation in the RCB-3 category. The small subset of pts with extensive residual disease (RCB-3) in both arms, 5.1% and 6.7%, respectively, had a poor prognosis. These results highlight the importance of neoadjuvant treatment with pembro for improving survival in pts with early TNBC, and identified a subset of pts for whom additional therapies will be needed. Clinical trial information: NCT03036488. [Table: see text]

Tumor-infiltrating lymphocytes (TILs) dynamics in breast cancer patients receiving neoadjuvant therapy: A systematic review and meta-analysis

e12620

Background: Increased baseline tumor-infiltrating lymphocytes (TILs) are associated with improved pathological complete response rates and better prognosis in HER2+ and triple negative breast cancer (TNBC) patients receiving neoadjuvant therapy (NAT). However, the role of TILs dynamics/change (ΔTILs) at the neoadjuvant setting remains unclear, thus a meta-analysis of the published studies was carried out. Methods: Medline, Embase, Cochrane Library and Web of Science Core Collection were searched for studies reporting on TILs expression in paired invasive breast cancer patient tissue samples before and after NAT. Data were extracted by two investigators (Y.Z., E.T.) and discordances were resolved by a third (I.Z.). Outcomes included pooled TILs rates pre- & post-treatment (also per subtype), pooled rates of ΔTILs and direction of change after NAT as well as correlation of ΔTILs with survival outcomes. Heterogeneity was assessed using the I2 statistic. Results: Of 1569 identified entries, 22 studies fulfilled the criteria and provided adequate data for the outcomes of interest. Overall, a significantly decreased level of TILs was observed after NAT in paired samples (pooled OR = 1.60, 95% CI: 1.12-2.30, p = 0.01; TILs as categorical variable). Regarding pooled rates of ΔTILs, a change was observed after NAT, irrespective of BC subtype. Among the different subtypes, the effect of NAT on TILs was most prominent in HER2+ disease with a direction towards decreased TILs to be more common (pooled ΔTILs rates: 14.4% increased vs 46.2%, decreased). In TNBC, bi-directional TIL kinetics were noted (pooled ΔTILs rates: 41.6% increased vs 37.1% decreased). An increase in ΔTILs in TNBC was associated with better disease-free/relapse-free survival in univariate analysis (HR = 0.59, 95% CI: 0.37–0.95, p = 0.03). Substantial between-study heterogeneity was observed in most analyses. Conclusions: The first to our knowledge meta-analysis on TILs dynamics during NAT in BC informs about differences in matched pre- and post-treatment patient samples and the prognostic implications of ΔTILs in TNBC. The potential clinical utility of the longitudinal assessment of immune response during neoadjuvant therapy warrants further investigation in prospective trials.

Serum thymidine kinase 1 and its kinetics in HER2-positive breast cancer: Results from the Swedish phase II PREDIX HER2 trial

e12598

Background: Thymidine kinase 1 (TK1) plays a pivotal role in DNA synthesis and cellular proliferation. TK1 has been studied as a prognostic marker and as an early indicator of treatment response in HER2-negative early and metastatic breast cancer (BC). Here we present the first report of TK1 in HER2-postitive BC. Methods: In the PREDIX HER2 trial, 202 patients with HER2-positive BC were randomized to 6 cycles of neoadjuvant trastuzumab, pertuzumab and docetaxel or trastuzumab emtansine every three weeks followed by surgery and adjuvant epirubicin and cyclophosphamide. Serum was prospectively collected from all participants at multiple timepoints: At baseline, after cycles 2, 4 and 6, at end of adjuvant therapy and then annually for 5 years and in case of recurrence. TK1 activity was measured by DiviTum assay (Biovica, Sweden), blinded to treatment allocation, patient characteristics and outcome. TK1 activity was correlated with baseline characteristics, pathologic complete response (pCR) and event-free survival (EFS). Results: Baseline TK1 activity as categorical variable was not associated with tumor size or hormone receptor expression. An increase of TK1 activity was seen in the majority of cases after treatment (mean TK1 at baseline = 111.7, after cycle 2 = 1257.0, after cycle 4 = 1157.0, after cycle 6 = 1178.0). Neither baseline TK1, on-treatment TK1 nor its change from baseline to cycle 2 were significantly associated with pCR in multivariable logistic regression analysis (Table). Baseline TK1 activity was not predictive for differential benefit to the study treatments ( pinteraction = 0.19). After a median follow-up of 52.5 months, 21 patients had EFS events. There was no significant correlation between baseline TK1 activity and EFS in multivariable Cox regression analysis. Conclusions: Serum TK1 activity in HER2-positive BC increased following treatment with neoadjuvant therapy but was not correlated to pCR rates or EFS. [Table: see text]

Adherence to adjuvant endocrine therapy assessed by data from prescription renewals and medical records from tamoxifen in Swedish patients with cytochrome P450 2D6 (CYP2D6) genotyped early breast cancer

542

Background: Suboptimal adherence to adjuvant endocrine treatment (AET) is an important clinical concern. A correlation between CYP2D6 activity and tamoxifen discontinuation has been suggested. The main aim of this study was to investigate the consistency between prescription refill data and reports from medical records on adherence to AET in early breast cancer. We also studied if there was an association between menopausal status, CYP2D6 activity, estimated risk for recurrence and adherence to AET. Methods: 1235 pre-and postmenopausal Swedish breast cancer patients operated 2006 – 2014, genotyped for CYP2D6, who initiated adjuvant tamoxifen treatment, were included in the study. Information on AET was retrospectively collected from both medical records and the Swedish Prescribed Drug Registry. Consistency was defined as dispensed doses of AET divided by AET intake documented in medical records. Adherence was calculated for patients with at least 4.5 years of follow up and was defined as Medical Possession Rate (MPR) ≥ 80 %. Subgroup analyses were performed based on menopausal status, recurrence-risk and CYP2D6 activity. Results: In 84% of the patients the consistency of AET between the sources of information was within 80-125%. Consistency < 80% was most frequent in premenopausal/high risk patients and CYP2D6 Poor Metabolizers (PM). Among 899 patients with at least 4.5 years follow up, adherence to tamoxifen was 72% according to prescription refill data, compared to 77% as reported by medical records. When including aromatase inhibitors adherence increased to 82% and 88%. Adherence did not differ by menopausal status or risk for recurrence. CYP2D6 PM had poorer adherence (54%) to tamoxifen compared to patients with the highest CYP2D6 activity (83%). Conclusions: Consistency between medical records and dispensing data on adherence was better than anticipated. Adherence to AET was good, 82% when including switch to aromatase inhibitors. Surprisingly, CYP2D6 PMs had low adherence to tamoxifen, despite a reduced risk of side effects according to previous studies. Further work is needed to clarify the impact of CYP2D6 activity on adherence to tamoxifen.

PD-1 protein and gene expression in early breast cancer: Prognostic implications

545

Background: We have previously shown the prognostic value of PD-L1 protein and gene expression in early breast cancer (BC), however, the prognostic role of PD-1 expression remains unclear. Methods: The prognostic value of PD-1 in early BC was investigated using three different approaches: i) evaluation of PD-1 at the protein (IHC, immunohistochemistry in tissue microarrays) and mRNA levels in a retrospective patient cohort of 586 patients treated for early BC in Stockholm, Sweden between 1997-2005, ii) systematic review and trial-level meta-analysis of studies published in Medline, Embase, Cochrane Library and Web of Science Core Collection libraries on the prognostic value of PD-1 IHC expression, and iii) pooled analysis of transcriptomic data from 39 publicly available datasets for the prognostic capacity of PD-1 gene expression. Univariate and multivariable Cox regression models were used. Results: In the retrospective study cohort, PD-1 protein was significantly associated with biologically high-risk characteristics. PD-1 protein, but not gene expression, was correlated with improved overall survival (OS) (adjusted HR = 0.73, 95% CI 0.55 – 0.96, p = 0.023 and adjusted HR = 0.88, 95% CI 0.68 – 1.13, p = 0.307, respectively). In the trial-level meta-analysis, 4736 entries were initially identified and 15 studies, including our original cohort, fulfilled the predefined eligibility criteria. PD-1 IHC expression was not prognostic in unselected patients. However, a significant correlation to improved disease-free survival was seen within the triple-negative subtype (pooled multivariate HR = 0.57, 95% CI 0.29 – 0.90, p = 0.02). In the pooled gene expression analysis, PD-1 gene expression was associated with improved OS in the entire population (adjusted HR = 0.89, 95% CI 0.80 – 0.99, p = 0.025) and in basal-like (adjusted HR = 0.77, 95% CI 0.63 – 0.95, p = 0.014) tumors. Conclusions: PD-1 expression at the RNA and protein levels represent promising prognostic factors, especially in the triple-negative and basal-like subtypes. Standardization and further validation are needed prior to clinical implementation.

PREDIX II HER2: Improving pre-operative systemic therapy for human epidermal growth factor receptor 2 (HER2) amplified breast cancer (BC)

TPS605

Background: Neo-adjuvant systemic therapy (NAT) is the standard of care for most patients with early HER2-amplified and triple negative breast cancer (BC). Increasing the rate of pathological complete response (pCR) is highly meaningful for those patients, as pCR is strongly predictive for improved long-term disease-related outcomes. Clinical and preclinical evidence support the hypothesis that pCR-rates may be augmented by the addition of checkpoint inhibitors, such as monoclonal antibodies targeting the Programmed Death Ligand receptor 1 (PD-L1), to standard systemic NAT. Studies in different BC patient cohorts (e.g., IMPassion130, PANACEA, KATE2) have indicated that PD-L1 protein expression on tumor-infiltrating lymphocytes (TIL’s) is a predictive marker for checkpoint inhibitor efficacy. Methods: We have initiated a phase II open-label, 2:1 randomized clinical trial where women with early HER2-amplified, PD-L1+ BC (cT2-3 and/or cN+) are treated with standard NAT (composed of anti-HER2 antibodies with a chemotherapy backbone of sequentially taxanes + carboplatin and epirubicin + cyclophosphamide [EC]) +/- atezolizumab during EC. N = 190 patients will be accrued in nine centers in Sweden to be able to demonstrate a 20% increase in pCR-rate, with a power of 80% and a two-sided alpha of 10%. Firstly, a prescreening is performed to select patients with a PD-L1 expression of > 1% on TIL’s. Important exclusion criteria are significant organ dysfunction and (with some exceptions) active auto-immune diseases. Extensive translational side-studies are performed to explore predictive markers for treatment efficacy, including clinicopathologic studies, molecular imaging and microbiome analyses, as well as monitoring of acute and chronic treatment-related toxicity, objective cognitive function and quality of life. As of February 11th, 4 patients have been prescreened and 1 enrolled in the trial. The clinical trial registry number is NCT03894007. Clinical trial information: NCT03894007 .

One-year follow-up of health-related quality of life in the Swedish PREDIX HER 2 trial, evaluating docetaxel, trastuzumab sc, pertuzumab versus trastuzumab emtansine as neoadjuvant treatment of HER2 positive breast cancer

590

Background: Neoadjuvant therapy combining docetaxel, trastuzumab and pertuzumab (DTP) was compared to trastuzumab emtansine (T-DM1) in the randomised phase II PREDIX HER2 trial. Patients, ≥18 years with HER2 positive breast cancer, ≥20mm or with verified lymph node metastases, were randomised to six courses of DTP (Standard group) or T-DM1 (Experimental group) before surgery. After operation patients in the Standard arm received two, and those in the Experimental arm four courses of EC. Since there were no differences in proportions of complete response at surgery and in the event-free survival between the groups, health-related quality of life (HRQoL) is of special interest. Methods: HRQoL was measured, using European Organization for Research and Treatment of Cancer (EORTC) QLQ-C30 + EORTC QLQ-BR23, before randomisation, after six courses before surgery, at three months and one year after surgery. Results: Response rate for the questionnaires varied between 87% and 99% for the 198 included patients. There were no between-group differences at baseline. Results after six courses revealed statistically significant differences (p≤0.01), favouring the Experimental group on 12 out of 21 of the EORTC QLQ-C30 and BR23 variables (Physical-, Role-, and Social- functioning, Global quality of Life, Fatigue, Dyspnea, and Diarrhea, Body image, Sexual functioning, Sexual enjoyment, Systemic therapy side effects and Upset by hair loss). Three months after surgery, however, statistically significant differences in favour of the Standard group were found for six variables (Physical functioning, Nausea/vomiting, Dyspnea, Constipation, Systemic therapy side effects, Upset by hair loss). No other between group differences were found with one exception: lower levels of Breast symptoms in the Experimental group. One possible explanation is that patients in the Experimental group were still on chemotherapy at that assessment point, whereas the majority in the Standard group had terminated their treatment. No differences were found between the groups at the one-year after surgery follow-up, where the levels on most variables had returned to baseline values. Conclusions: HRQoL was better in the Experimental group during neoadjuvant treatment. Three months after surgery, however, HRQoL was in favour or the Standard arm. HRQoL returned to baseline levels for most variables at the one-year follow-up and no between-group differences were found. Clinical trial information: NCT02568839 .

Docetaxel, trastuzumab, pertuzumab versus trastuzumab emtansine as neoadjuvant treatment of HER2-positive breast cancer: Results from the Swedish PREDIX HER2 trial identifying a new potential de-escalation standard?

501

Background: Neoadjuvant therapy produces high rates of pathological complete response (pCR) and is the standard of care in HER2 positive breast cancer; however, the optimal treatment regimen remains to be established. Methods: In this randomized phase II study patients ≥18 years with HER2 positive breast cancer > 20mm or verified lymph node metastases were randomized to 6 courses of docetaxel, trastuzumab and pertuzumab (DTP, group A) or trastuzumab emtansine (T-DM1, group B), q 21 days. The protocol allowed switch to the competing treatment upon lack of response or drug-related severe toxicity. Patients received postoperative epirubicin+cyclophosphamide, trastuzumab for a total of one year and endocrine therapy. Accrual was completed in October 2018 after randomization of 202 patients, data on pCR were available for 190 at the time for this abstract submission. Median age, 52 years (26-74), menopausal status, histological type and grade were well balanced between the treatment groups. 62.6% of the tumors were hormone receptor (HR) positive. Results: Primary endpoint was pathological objective response. 190 patients completed the protocol-specified preoperative treatment. pCR was achieved in 45.3% of patients, 46.4% in patients treated with DTP and 44.1% with T-DM1 (chi-sq., p = 0.75). In HR-positive tumors, pCR was obtained in 35.3% of patients, 35.9% in group A vs. 34.6% in group B (p = 0.87); in HR-negative tumors, the overall pCR rate was 62.0%, 66.7% in group A vs. 57.9% in group B (p = 0.45). Severe (grade 3/4) toxicity was reported at 68 occasions related to DTP, compared with 16 related to T-DM1, 26 vs. 3 caused by febrile neutropenia. Significantly better quality of life was reported by patients treated with T-DM1. Conclusions: Our data on TDM-1 demonstrates similar efficacy and less toxicity, in particular for patients with HER2 and HR positive cancers, being a potential new standard for neoadjuvant therapy. Clinical trial information: NCT02568839.

Prognostic implications of PD-L1 expression in breast cancer at the protein and mRNA levels

e14284

Background: Conflicting data have been reported on the prognostic value of PD-L1 expression per immunohistochemistry (IHC) in breast cancer (BC). There is a paucity of data on the role of PD-L1 gene expression (GE). Methods: Medline, Embase, Cochrane Library and Web of Science Core Collection were searched and data were extracted independently by two researchers. Outcomes included pooled PD-L1 positivity in tumor cells, immune cells or both, per subtype and per antibody used; and the prognostic value of PD-L1 positivity for DFS and OS. Heterogeneity was assessed using the Q and I2 statistics. A pooled GE analysis of 39 publicly available transcriptomic datasets was also performed. Results: Of the initial 4184 entries, 38 retrospective studies fulfilled the inclusion criteria. The overall pooled PD-L1 positivity rate in tumor cells was 24%, 33% in immune cells and 25% in both; highest positivity was reported with Dako 28-8 clone. PD-L1 IHC expression in tumor cells was prognostic for shorter DFS (HR = 1.36, 95% CI 1.02 – 1.83, p < 0.04) and OS (HR = 1.66; 95% CI 1.09 – 2.50, p = 0.02); there was significant heterogeneity. PD-L1 IHC expression in immune cells was associated with better DFS (HR = 0.61; 95% CI 0.51 – 0.73, p < 0.001) and OS (HR = 0.53, 95% CI 0.39 – 0.73, p < 0.001) in TNBC. In addition, higher PD-L1 GE predicted better survival in multivariate analysis in the entire population (HR = 0.70, 95% CI 0.60 – 0.82, p < 0.001 for DFS and HR = 0.84, 95% CI 0.75 – 0.93, p = 0.001 for OS) and in basal-like tumors (HR = 0.55, 95% CI 0.38 – 0.80, p = 0.001 for DFS and HR = 0.63, 95% CI 0.50 – 0.79, p < 0.001 for OS), pinteraction 0.124 for DFS and 0.005 for OS. Conclusions: The largest to our knowledge meta-analysis on IHC PD-L1 expression in BC informs on PD-L1 positivity rates and its prognostic value. Standardization is needed prior to routine implementation. PD-L1 GE is a promising prognostic factor, especially in basal-like BC.

Prognostic value of PD-L1 gene expression with Recurrence Score and 70-gene signature in patients with ER+/HER2- early breast cancer

550

Background: We have previously demonstrated that PD-L1 mRNA expression can serve as prognostic biomarker in breast cancer (BC). In ER+/HER2- BC, RS and 70-gene signature are used to predict the risk of recurrence and benefit from chemotherapy. Methods: Discovery cohort (cohort 1) included 302 patients diagnosed with primary ER+/HER2- BC (1997-2005) in Stockholm health care region. Gene expression profiling has been performed using DNA microarrays (GSE48091) while information regarding tumor characteristics, treatment and follow-up have been obtained. TCGA’s dataset including 590 ER+/HER2- patients, was used as validation cohort (cohort 2). Kaplan–Meier estimates and Cox regression univariate/multivariable analyses were performed using breast cancer-specific survival(BCSS) and progression-free interval (PFI) as endpoints in cohorts 1 and 2, respectively. Gene signature scores were calculated using the R genefu package. Likelihood ratio (LR) tests and concordance indices (c-indices) were used to assess each score’s added prognostic value. Results: PD-L1 mRNA expression (treated as a continuous variable) was independently associated with better BCSS in cohort 1 (HR = 0.72; 95% CI = 0.58-0.90;p = 0.003) and with better PFI in cohort 2 (HR = 0.67; 95% CI = 0.50-0.90; p = 0.008) in the multivariable analysis. PD-L1 provided significant additional prognostic information beyond that of both RS alone (LR-Δχ2= 9.6; p = 0.002 and LR-Δχ2= 9.7; p = 0.002, in cohorts 1 and 2, respectively), and 70-gene signature score alone (LR-Δχ2= 10.4; p = 0.001 and LR-Δχ2= 9.2; p = 0.002 in cohort 1 and 2, respectively). C-indices for PD-L1 + RS vs RS were 0.65 vs 0.60 (cohort 1) and 0.66 vs 0.60 (cohort 2), and for PD-L1 + 70-gene vs 70-gene were 0.65 vs 0.59 (cohort 1) and 0.64 vs 0.54 (cohort 2), respectively. Conclusions: PD-L1 gene expression was correlated with better outcomes and can provide added prognostic value to RS and 70-gene signature scores in ER+/HER2- BC.

Health-related quality of life in the Swedish PREDIX HER2 trial, evaluating docetaxel, trastuzumab, pertuzumab versus trastuzumab emtansine as neoadjuvant treatment of HER2-positive breast cancer

583

Background: Neoadjuvant therapy combining docetaxel, trastuzumab and pertuzumab (DTP) was compared to trastuzumab emtansine (T-DM1) in the randomized phase 2 PREDIX HER2 trial. Patients, ≥18 years with HER2 positive breast cancer, ≥20mm or with verified lymph node metastases, were randomized to six courses of DTP (Standard arm) or T-DM1 (Experimental arm). Primary endpoint was pathological objective response to primary medical therapy at post-treatment surgery. Health related quality of life (HRQoL) was a secondary outcome, and is of specific interest as there was no difference between the randomization groups regarding the main endpoint (results presented in a separate abstract sent to ASCO 2019, Bergh et al.). Methods: Of 202 randomized patients, 190 are available for evaluation at this point. HRQoL was measured, using EORTC QLQ-C30 + EORTC QLQ-BR23, at baseline before randomization and after six courses. Results: No differences between the randomization arms were found at baseline. Results after six courses, based on 163 patients (86%) and adjusted to baseline values, revealed statistical significant differences (p≤0.01), favoring the experimental T-DM1 arm on 7 out of 15 of the EORTC QLQ-C30 variables (Physical functioning, Role functioning, Social functioning, Global quality of Life, Fatigue, Dyspnea, and Diarrhea). For the breast cancer specific questionnaire (EORTC-BR23), the experimental arm scored statistically significantly better on 5 out of 7 subscales (Body image, Sexual functioning, Sexual enjoyment, Systemic therapy side effects and Upset by hair loss). All of the statistical significant differences were of moderate or large clinical significance (≥10 scale scores). No differences between the randomization arms were found for the remaining HRQoL variables. Conclusions: The experimental arm reported better HRQoL than the control arm after six courses. Trastuzumab emtansine may be a useful treatment alternative due to better HRQoL and lower toxicity. Clinical trial information: NCT02568839.

Tailored dose-dense chemotherapy in combination with trastuzumab as adjuvant therapy for HER2-positive breast cancer: A secondary analysis of the phase III PANTHER trial

553

Background: Dose-dense (DD) adjuvant chemotherapy improves outcomes in early breast cancer (BC). However, there are no data to inform on the combination of DD chemotherapy with trastuzumab for patients with HER2-positive disease. Methods: This is a protocol predefined secondary analysis of the randomized phase 3 PANTER trial. Women 65 years old or younger with node-positive or high-risk node negative BC were randomized 1:1 to either tailored according to hematologic nadirs and DD epirubicin/cyclophosphamide (4 cycles) followed by 4 cycles of docetaxel (tDD EC/D) or standard 3-weekly 5-fluorouracil/E/C (3 cycles) and D (3 cycles); Patients with HER2-positive disease received 1 year of adjuvant trastuzumab. In addition, HER2-positive and an equal number of matched for age, treatment group and institution, HER2-negative patients that were enrolled in Swedish sites were enrolled in a predefined study of cardiac safety and underwent echocardiography or MUGA and electrocardiography at baseline and at four and six years of follow-up. The primary endpoint was BC relapse-free survival (BCRFS). Results: There were 342 HER2-positive patients; 335 received at least one dose of trastuzumab, while 29 patients discontinued trastuzumab prematurely. BCRFS was not statistically significantly in favor of tDD EC/D (HR = 0.68, 95% CI 0.37 – 1.27, P= 0.231). Cardiac outcomes after four and six years of follow-up did not differ significantly between HER2-positive and HER2-negative patients, nor between tDD and standard treatment. Conclusions: To our knowledge, these are the only data on combining DD adjuvant chemotherapy and trastuzumab in BC. The combination decreased the risk for BC relapse by 32% compared to standard treatment, a statistically non-significant difference. Its efficacy and safety merit further evaluation as part of both escalation and de-escalation strategies. Clinical trial information: NCT00798070.

Long-term safety and survival outcomes from the Scandinavian Breast Group 2004–1 (SBG 2004-1) randomized trial of tailored dose adjuvant chemotherapy for early breast cancer

e12036

Background: Although adjuvant polychemotherapy (ACT) improves outcomes for early breast cancer (BC), the significant interpatient variability in pharmacokinetics result in differences in efficacy and both short and long term toxicities. Dose tailored ACT has been shown to improve outcomes, while tailored and dose dense therapy improved 5-year event free survival in the phase 3 PANTHER trial. Methods: The SBG 2004-1 trial was a randomized feasibility/phase II study which assessed dose tailored epirubicin and cyclophosphamide (EC) followed by docetaxel (T), compared to the same fixed dose schedule and to the TAC regimen (table 1). Women aged 18-65 years old, ECOG PS 0-1 with at least one positive lymph node were randomized 1:1:1. The primary endpoint of the study was the safety and feasibility of the administered treatment. The secondary endpoint of the study was to evaluate the dose-intensity of the treatment. Toxicity was graded according to CTC-AE version 3.0. The design and short-term safety results of the trial have been previously published. Here, we report safety and efficacy data after 10 years follow-up. Results: A total of 124 patients were included and >95% of patients without relapse were followed for 10 years.Five secondary malignancies but no myelodysplastic syndrome / acute myeloid leukemias were reported. Long term disease-free and overall survival results will be presented at the meeting. Conclusions: Long term safety results from this study in combination with efficacy results from the PANTHER trial imply the feasibility of a tailor-based approach for ACT in early BC. [Table: see text]

Regulation of PD-L1 in breast cancer

e23088

Background: Programmed death ligand 1 (PD-L1, CD274) is currently being investigated as a therapeutic target in breast cancer (BC), however its transcriptional regulation has not been elucidated. Methods: Three BC cells lines, one negative (MCF7) and two positive (MDA-MB-231, BT549) for PD-L1 were used. Protein expression of potential regulators of PD-L1, including Myc, STAT3, p-rpS6 (mTOR effector), Erk and p53 was assessed by western blotting and immunohistochemistry. Furthermore, transfections with small interfering RNAs (siRNAs) of selected proteins or plasmids and pharmacologic inhibition were performed. Fluorescent in situ hybridization (FISH) analysis was used for PD-L1 gene amplification studies. Publically available gene expression data derived from The Cancer Genome Atlas (TCGA) database (cBioportal) were analyzed for correlations between mRNA levels of PD-L1 and potential regulators. Results: PD-L1 gene amplification was not detected in BC cells expressing PD-L1, suggesting non-genetic regulation of PD-L1 expression. Additionally, the correlation between CD274 transcript levels and copy-number alterations (CNA) in primary BC was weak. Pharmacological treatment with either Myc (expressed in MDA-MB-231) or mTOR (p-rpS6 expressed in all three cell lines) inhibitors did not affect PD-L1 expression. STAT3 protein was expressed in all cell lines and STAT3 mRNA levels had a weak positive correlation with PD-L1 (Spearman’s rho = 0.25) in the TCGA dataset. Both gene silencing using STAT3 siRNA and pharmacologic inhibition of STAT3 activity resulted in decreased PD-L1 protein levels in BT549 and MDA-MB-231 cells respectively. Multiple DNA binding sites for STAT3 were identified in silico on the PD-L1 gene promoter suggesting that the observed regulatory effects are likely transcriptional. Conclusions: STAT3 can act as an inducer of PD-L1 expression, while Myc and mTOR seem to have no effect in PD-L1 regulation in BC.