Background

The risk-score of the HER2DX genomic test (HER2DX risk-score) was first validated in an independent combined dataset of 268 patients (pts) with early-stage HER2+ breast cancer treated with neoadjuvant and adjuvant anti-HER2-based treatment (EBioMedicine 2022). Here, we report an updated survival analysis with longer follow-up, with a special focus on the value of HER2DX risk-score beyond pathological complete response (pCR).

Methods

A dataset of 268 pts with early-stage HER2+ disease obtained from a combined cohort of 3 neoadjuvant studies was used for an independent validation of the standardized HER2DX risk-score. The dataset was composed of 147 pts from Hospital Clinic, 84 pts from PAMELA trial and 37 pts from the Padova University cohort. All pts received chemotherapy and 1-year of trastuzumab; 56% of pts received dual HER2 blockade; and 9% pts with residual disease received adjuvant T-DM1. The Kaplan-Meier method and stratified Cox models were used to estimate hazard ratios (HRs) to evaluate the association between HER2DX and disease-free survival (DFS).

Results

Median follow-up was 73.2 months (vs. 52.7 months in the prior report). HER2DX score as a continuous variable was significantly associated with DFS (HR=1.97, p=0.003) and overall survival (HR=2.23, p=0.009). According to the prespecified cut-off, the HER2DX low-risk group had longer DFS than high-risk (7-year 94.6% vs. 77.5%; HR=0.4, p=0.002). HER2DX risk-score was significantly (HR=1.90; p=0.003) associated with DFS independently of pCR status and hormone receptor status. Within pts with a pCR (n=118), the HER2DX low-risk group had longer DFS than the high-risk (5-year 96.5% vs. 88.4%; HR=0.33, p=0.178). Within pts without a pCR (n=148), the HER2DX low-risk group had longer DFS than the high-risk (5-year 95.6% vs. 85.3%; HR=0.20, p=0.004), and it was significantly (HR=2.03; p=0.006) associated with DFS independently of hormone receptor and adjuvant T-DM1.

Conclusions

The HER2DX risk-score determined in baseline pre-treatment core-biopsies provides prognostic information beyond pCR status in patients with early-stage HER2+ breast cancer treated with neoadjuvant and adjuvant anti-HER2 treatment.

Background

In the APT and ATEMPT trials, adjuvant paclitaxel and trastuzumab (TH) and T-DM1 were found associated with excellent long-term outcomes for patients (pts) with small, node-negative HER2-positive breast cancer (HER2+ BC), respectively. HER2DX risk score was found associated with outcomes in both trials, separately.

Methods

We conducted a retrospective analysis combining pts included in the APT and ATEMPT trials with available HER2DX data. Co-primary endpoints were associations of HER2DX with relapse-free interval (RFI) and invasive disease-free survival (iDFS). The HER2DX risk-score was evaluated i) as a continuous variable (0-100), ii) using the predefined cut-off (50), and iii) using an exploratory optimal cut-off (32). The Kaplan-Meier method and stratified Cox regression models were used to evaluate the association between HER2DX and outcomes.

Results

In total, 471 pts receiving TH (n=324) or T-DM1 (n=147) were included in the study, most having stage I (n=432, 92%) and hormone receptor positive disease (n=335, 75%). The median follow-up was 6.7 years (10.8 and 5.8 for APT and ATEMPT, respectively). The median HER2DX risk-score was 13.9 (IQR 4.7 - 27.0), with 5.5% and 18.3% of the pts having HER2DX high-risk disease according to the predefined and optimal cut-off, respectively. HER2DX risk score as a continuous variable was associated with RFI (HR per 10-units: 1.39, 95%CI: 1.09-1.78; p=0.009) but not with iDFS (HR per 10-units: 1.18, 0.98-1.42; p=0.09). Using the predefined cut-off (50), pts with HER2DX high-risk disease had higher RFI risk (HR: 7.33, 2.29-23.47, p<0.001), but the effect on iDFS was non-significant (HR: 2.78, 0.97-7.95, p=0.057). In multivariable analysis of RFI, HER2DX remained statistically significant after adjustment for hormone receptor status and tumor size (HR: 7.89, 2.06-30.22, p=0.003). The optimal cut-off (32) distinguished pts with low-risk (7-year RFI of 98.2%; 96.7%-99.6%) from high-risk disease (7-year RFI: 88.7%; 80.4%-97.8%) [delta of 9.5%], including in multivariable analysis for RFI (HR: 6.87, 2.22-21.27, p<0.001) and iDFS (HR: 2.81, 1.26-6.23, p=0.01).

Conclusions

The HER2DX risk score is associated with the risk of recurrence among pts with small, node negative HER2+ breast tumors treated with adjuvant TH or T-DM1.

Background

HER2DX ERBB2 mRNA score captures the dynamic range of ERBB2 expression in both HER2-negative and HER2+ breast cancer (BC). In a previous study (Prat et al. EBiomedicine 2022), ERBB2 mRNA predicted clinical HER2+ status with high performance in the validation dataset (n=353; AUC=0.96). Here, we evaluated the ability of HER2DX ERBB2 mRNA to predict HER2+, HER2-low, and HER2-0 status.

Methods

Standardized HER2DX was evaluated centrally on formalin-fixed paraffin-embedded tumors from 1,149 samples with HER2+ (n=1,029) vs. HER2-negative status (n=120). We trained a new ERBB2 cutoff (by means of decision trees) to predict HER2-low (n=73) versus HER2-0 (n=47) status, and tested in an independent validation dataset of 110 HER2-negative tumors with known HER2-low (n=60) and HER2-0 (n=50) status. To quantify the diagnostic performance: area under the ROC curve (ROC AUC), global accuracy, positive (PPV) and negative predictive value (NPV) were calculated.

Results

In the HER2+ and HER2-negative dataset (n=1,149), the ROC AUC of ERBB2 mRNA expression to predict HER2+ status was 0.98. The mean ERBB2 expression (in log base 2) in HER2+ and HER2-negative disease was 1.41 and -2.41, respectively (14.1-fold difference). Of note, 0.8% of HER2-negative cases were identified as ERBB2-positive and 9.9% of HER2+ cases were identified as ERBB2-negative. In the HER2-negative training dataset (n=120), the ROC AUC of ERBB2 mRNA expression to predict HER2-low vs. HER2-0 status was 0.80, and a new cutoff was identified with 77.5% accuracy (68.5% PPV and 84.9% NPV). In the independent validation HER2-negative dataset (n=110), the ROC AUC of ERBB2 mRNA expression was 0.77. With the new ERBB2 cutoff, the overall accuracy to predict HER2-low and HER2-0 cases was 65.5% (66.7% PPV and 64.9% NPV). Finally, the mean ERBB2 mRNA expression (in log base 2) in HER2-low and HER2-0 disease was -1.86 and -2.72, respectively (1.8-fold difference).

Conclusions

The standardized HER2DX ERBB2 mRNA score predicts the clinical status of HER2 (positive, low and zero) in BC.

Background

HER2DX is a genomic test based on the expression of 27 genes tracking 4 signatures (luminal, proliferative, immune and HER2 amplicon) that provides prognostic (HER2DX risk score) and predictive information (HER2DX pathologic complete response [pCR] score) in HER2+ early breast cancer (BC). Here, we report the initial results of the first ongoing decision impact study of HER2DX at Hospital Clinic of Barcelona.

Methods

We conducted an observational, prospective, pilot, unicentric study, since Nov/21 (ongoing), to analyze the impact of HER2DX in clinical practice in early-stage HER2+ BC. Any medical oncologist of the Breast Unit could order the test. A survey was completed by the treating physician before and after receiving the result of HER2DX. The main objective was to assess the % of change in the therapeutic plan after obtaining the HER2DX report. Secondary objectives included 1) assess changes in the physician’s confidence before and after the test and 2) analyze the association of the HER2DX pCR score with the pathological response after neoadjuvant therapy (NAT). Descriptive statistics were used.

Results

We report the results after the inclusion of the first 89 pts (as of 2nd of Feb/23). Median age was 53 years (range 30-79) and 52% of pts were postmenopausal. Most pts had T1-2 tumors (87%), negative nodes (64%), grade 2 (56%) or 3 (41%), ductal histology (87%), hormone receptor positive (65%), median Ki67 of 35% (range 4-90%) and median tumor-infiltrating lymphocytes of 10% (range 0-90%). 78% of pts received NAT and 22% upfront surgery. A change in the treatment plan before and after the HER2DX result was observed in 49 of 87 (56%) cases. De-escalation of therapy was observed in 59% of pts (less intense chemotherapy [ChT] in 57% of them) and escalation in 41% of pts (more intense ChT in 65% of them). The confidence in the decision improved in 67% of cases. Among 56 evaluable pts treated with NAT, HER2DX pCR score was significantly associated with pCR (81% in pCR-medium/high and 32% in pCR-low; odds ratio=9.3, p=0.001), independently of the rest of variables.

Conclusions

In this first pilot and prospective study, HER2DX impacted clinical care in early-stage HER2+ BC.

Background

The prognostic value of neutrophil/lymphocyte ratio (NLR) for HER2-positive metastatic breast cancer (MBC) is not well studied. This study aims to evaluate the prognostic role of baseline NLR in HER2-positive MBC patients treated with trastuzumab/pertuzumab.

Methods

The clinical data of 780 patients from CLEOPATRA were applied from VIVLI platform, and 248 HER2-postive MBC were collected from six local hospitals. Propensity score matching (PSM) and inverse probability of treatment weighting (IPTW) analyses were used to control bias. The associations between clinicopathological factors, NLR and progression-free survival (PFS) and overall survival (OS) were analyzed by univariate and multivariate analyses.

Results

After PSM or IPTW adjustment, the subgroups were similar. Low baseline NLR was prognostic with better PFS and OS in the TH group in raw, PSM and IPTW models. Upon IPTW, low NLR, versus high NLR, was associated with improved PFS (HR 1.35, 95% CI 1.07-1.70, P = 0.012) and OS (HR 1.47, 95% CI 1.12-1.94, P = 0.006) in the TH group. In the THP group, low baseline NLR was also correlated with better PFS, but not for OS in the three models. After IPTW, patients with low NLR were associated with better PFS (HR 1.52, 95% CI 1.20-1.93, P = 0.001) comparing that with high NLR. Multivariate analyses showed that low baseline NLR was a predictor for PFS and OS in TH group, and PFS in the THP group in the three models. In the real-world study, low baseline NLR was a predictor of better PFS among patients with trastuzumab plus docetaxel or trastuzumab plus pertuzumab plus docetaxel therapy (P = 0.025 and 0.009 respectively).

Conclusions

Low baseline NLR is associated with better survival outcome among HER2-positive MBC receiving docetaxel plus trastuzumab or docetaxel plus trastuzumab plus pertuzumab as first-line therapy. Re-analyses of prospective randomized studies are needed to verify the role of baseline NLR in HER2-positive MBC treated with trastuzumab/pertuzumab.

Clinical trial identification

Many studies have shown that higher neutrophil/lymphocyte ratio (NLR) was correlated with worse survival in triple-negative breast cancer (BC), but there were some controversies in HER2-positive metastatic BC (MBC). Our previous study with a total of 843 early HER2-positive BC proved that low baseline NLR was associated with better survival outcome in patients receiving trastuzumab therapy. This exploratory analysis of 780 HER2-positive MBC from CLEOPATRA trial shows that low baseline NLR is significantly associated with better progression-free survival (PFS) and overall survival (OS) among patients receiving docetaxel plus trastuzumab, and with PFS among those receiving docetaxel plus trastuzumab plus pertuzumab. Another cohort analysis of 248 HER2-positive MBC from six local hospitals also shows the similar results. These findings suggest that HER2-positive MBC with low high NLR might benefit less from treatment of docetaxel plus trastuzuma/pertuzumab, and need intensive treatment for those metastatic disease.

Background

HER2 loss on residual disease (RD) is frequent after NAT. Subtype switch has been also reported, with HER2-Enriched (HER2E) tumors converting frequently to non-HER2E. However, the association between HER2 immunohistochemistry (IHC) status and intrinsic subtype (IS) has never been described.

Methods

Between Feb/08 and Mar/22, 82 patients (pts) with HER2+ BC who underwent NAT at Hospital Clinic of Barcelona and had RD with matched IHC data were included. HER2 loss was defined as HER2 IHC 0/1+ or 2+/ISH not amplified on RD. Research-based PAM50 subtyping was performed with the nCounter platform. Associations between HER2 loss, IS dynamics, clinicopathological characteristics and event-free survival (EFS) were assessed.

Results

At baseline, 61% (n=50) of tumors were HER2 3+ and 83% (n=68) were hormone receptor (HR) positive. All pts received NAT with trastuzumab, 98% with chemotherapy, 52% with pertuzumab; and 24% received adjuvant T-DM1. HER2 loss was identified in 46% of BC (24 IHC 0, 12 IHC 1+, 2 IHC 2+/ISH-).

IS was assessed on 73 baseline and 72 RD samples (67 paired). Distribution of IS at baseline was: HER2E 42%, LumA 31%, LumB 14%, normal-like 9%, basal-like 4%; on RD was: HER2E 17%, LumA 36%, LumB 8%, normal-like 33%, basal-like 5%. ERBB2 mRNA levels significantly decreased after NAT (p=0.001). An IS switch was observed in 40% (n=27) of samples and was not associated with HER2 loss (p=0.455). However, HER2 loss was numerically more frequent among BC that switched from HER2E to non-HER2E (58%) than in BC that remained HER2E (23%) (p=0.082).

In a multivariate regression analysis including baseline IHC, ERBB2 mRNA, IS, HR status, time from NAT to surgery and administration of dual HER2 blockade, only ERBB2 mRNA was significantly associated with HER2 loss (p=0.003). At a median follow up of 61.0 months, 12 EFS events were recorded. After adjusting for T-DM1 use, none of the variables assessed, including HER2 loss, was associated with EFS.

Conclusions

HER2 loss was associated with decrease in ERBB2 mRNA levels, was more frequent in tumors switching from HER2E to non-HER2E subtype, and was not associated with EFS. Further validation on large cohorts is warranted.

Background

HER2DX is a 27-gene prognostic (risk-score) and predictive (pathological complete response [pCR]-score) assay in early-stage HER2+BC based on clinical data and the expression of 4 gene signatures (immune, proliferation, luminal differentiation and HER2 amplicon). Here we aim to further validate the ability of HER2DX to predict pCR.

Methods

Standardized HER2DX was evaluated centrally on FFPE tumor biopsies from the BionHER study, in which patients(pts) with stage I-III HER2+BC were treated with neoadjuvant THPx16weeks; tumor biopsies were obtained pre-treatment(D1) and 8 days later(D8), following the loading-dose of HP, prior to adding paclitaxel. Primary aim was to test the ability of HER2DX pCR-score to predict pCR (ypT0/isN0). Secondary objective were to test the ability of HER2DX pCR-score to predict pCR independently of hormone receptor (HR) status. HER2DX was also evaluated at D8. Logistic regression and receiver-operator curve (ROC) analysis were assessed.

Results

HER2DX was evaluated in 49 pts of 52 (94%). cT1-2 disease represented 85% of cases (mean tumor size was 29mm), cN0 59%, and 67% were HR+. Among them, 46 of 49 (94%) pts had undergone surgery to date. The overall pCR rate was 45.6% and the rate of both pCR and ypT1miN0 was 55%. The % of HER2DX low-, medium- and high-pCR groups was 30.6%, 40.8% and 28.6%, respectively. HER2DX pCR-score (as a continuous variable [CV]) was significantly associated with pCR (odds ratio [OR]=4.25, p=0.001). The pCR rates in HER2DX pCR-high, pCR-medium and pCR-low groups were 75.6%, 40% and 13.3% (-high vs -low OR=18.33, p=0.004), respectively. The AUC ROC of HER2DX pCR score (as a CV) and pCR status was 0.813. HR status was significantly associated with pCR score (OR=0.43, p=0.008). HER2DX pCR score was significantly associated with pCR independently of HR status (OR=3.89, p=0.010), which lost its statistically significance in the presence of HER2DX pCR-score (OR=0.87), p=0.756).

Conclusions

The 27-gene HER2DX genomic test predicts pCR following neoadjuvant THP in HER2+ BC. Updated data, including D8 HER2DX results, will be presented at the conference.

Background

HER2DX is a 27-gene assay for early-stage HER2+ breast cancer based on clinical data and the expression of 4 signatures (immune, proliferation, luminal, and HER2). Here we evaluated the utility of HER2DX in HER2+ IBC for the first time.

Methods

Standardized HER2DX was evaluated centrally on baseline pre-treatment tumors from a prospective phase II clinical trial (NCT01796197), in which patients with HER2+ IBC were treated with neoadjuvant paclitaxel, trastuzumab and pertuzumab (THP) x 16 weeks. The primary aim of this correlative analysis was to evaluate the pCR rates (ypT0/isN0) according to HER2DX pCR-score pre-defined cutoffs (i.e., low, medium, and high). Secondary objectives were to compare the HER2DX scores and signatures in IBC vs. stage II-III HER2+ non-IBC treated with neoadjuvant THP x 12 weeks on trial (NCT03716180). Descriptive statistics were used. Means between the two groups were compared using a Student´s t-test.

Results

HER2DX was evaluated in 23 patients with HER2+ IBC (Table 1). In IBC, the pCR rates in HER2DX pCR-high, pCR-med and pCR-low groups were 75.0% (3/4; 95% CI 19.4-99.4), 45.5% (5/11; 16.7-76.7) and 25.0% (2/8; 3.2-65.1), respectively. HER2DX high-risk in IBC represented 95.7% of the patients. Compared to non-IBC, IBC had a lower % of HER2DX pCR-high disease, higher % of HER2DX pCR-medium, and similar % of HER2DX pCR-low disease. Among the 4 HER2DX signatures, the HER2 amplicon was expressed at lower values in IBC vs. non-IBC (p=0.004), despite a similar % of HER2 3+ disease.

Conclusions

HER2DX pCR-high designation may be less frequent in HER2+ IBC compared to HER2+ non-IBC. Lower mRNA expression of HER2 amplicon genes, and higher tumor burden at diagnosis, might explain the differences observed in the distribution of HER2DX scores and signatures between HER2+ IBC and non-IBC.

Clinical trial identification

NCT01796197

Background

HER2DX is a genomic test for early HER2+ breast cancer clinically available since January 2022 as a Laboratory Developed Test. Based on the expression of 27 genes (nCounter) plus tumor and nodal staging, it predicts the risk of relapse (risk-score), the probability of achieving a pathological complete response (pCR-score), and the individual levels of ERBB2 expression (ERBB2-score). Here, we report the analytical validation of HER2DX performed in a central laboratory (CLab) having as a reference the development lab (DLab).

Methods

Intra- and inter- labs HER2DX scores were compared by testing a set of FFPE samples previously analyzed at DLab. All scores ranged from 0 to 100 and pre-defined cut-offs were used to get HER2DX-groups. Repeatability and reproducibility were evaluated from different tissue sections, RNA, or FFPE blocks. Simulations (n=1·10⁶) were used to calculate diagnostic values (sensitivity, specificity, positive and negative predictive values, and accuracy). Robustness was measured by evaluating the interference of non-tumor tissue and by using different RNA quantities. Differences due to the nCounter instrument, Tagset lot, and Tagset defrost cycles were also evaluated. HER2DX performance using a RNAseq platform (Illumina Exome Panel) was compared to nCounter.

Results

Repeatability analysis within CLab in 10 samples showed a maximal standard error among the 3 scores of 0.94 (scale 0-100). Reproducibility starting from RNA was analyzed in 20 samples, and the probabilities of +/- 5 units difference in the risk-score, pCR-score, and ERBB2-scores were 0.5%, 5.2%, and 0.2%, respectively. Starting from FFPE blocks, correlation of scores between labs were >0.97. In the simulation, all diagnostic values were >90% for the three scores, showing a high association in HER2DX-group classification between labs. Robustness at low tumor content samples (10%) or low RNA quantity (100 ng) was acceptable for the assay. No significant differences were observed across different nCounter instruments, Tagset lots and defrost cycles at CLab. Data from RNAseq and nCounter platforms were highly correlated in 30 analyzed RNA samples.

Conclusions

Analytical validation of HER2DX has proven it to be suitable for its intended purpose.

Background

Recently, a HER2-targeting antibody-drug conjugates (HER2-ADC) was approved to treat HER2-low breast cancer (BC), currently defined by the immunohistochemical (IHC) HER2-expression ASCO/CAP scores of 1+ or 2+ without HER2/ERBB2 amplification. While the HER2 IHC assay was optimized to detect protein overexpression, concerns exist regarding the use of this assay to reliably detect HER2-low BC. Here, we aimed at investigating the correlation between the IHC classification and the tumor cell expression levels of ERBB2 mRNA at the single cell level.

Methods

We retrospectively analyzed 22 untreated BC samples with single cell RNA-sequencing and centralized HER2 IHC data. IHC staining for HER2 (Agilent, GA0485, RTU) was performed and scored according to ASCO/CAP 2018 guidelines. Single cell data were retrieved from the original publication (PMID: 33958794) and only the cancer cells (n= 31016) were retained. ERBB2 expression was considered among cells where ERBB2 could be detected (non-zero normalized expression).

Results

We identified 1 HER2-zero (with no staining on tumor cells), 16 HER2-low (4 HER2-1+, 12 HER2-2+ FISH-negative) and 5 HER2-positive samples. The median percentage of cells expressing ERBB2 per patient was 38% (8%, 34%, 46% in HER2-zero, low, and positive respectively). Median ERBB2 expression was 0.94 (range: [0.13-5.27]) overall (0.48, 0.90, 2.8 in HER2-zero, low, and positive, respectively). Intra-patient heterogeneity of ERBB2 expression was observed with a median IQR of 0.63 overall (0.49, 0.63, 0.94 in HER2-zero, low, and positive, respectively). 36% (27/74) of the ERBB2 expressing tumor cells in the HER2-zero tumor had an ERBB2 expression value equal or greater than the 25^th percentile of the ERBB2 expression observed in HER2-low cells. When considering the ER IHC status, the percentage of expressing cells was higher in ER-positive samples but the level of ERBB2 expression was similar.

Conclusions

Single cell data might provide more granularity into the tumor-specific expression levels of HER2. Future research is needed to investigate whether single-cell ERBB2 expression could serve as a predictive biomarker for HER2-ADC.

Background

HER2-positive breast cancer is a heterogeneous disease, presenting tumor and microenvironment features which can impact prognosis and treatment response. Here, we aimed at better understanding the heterogeneity of this disease by performing spatial transcriptomics (ST) on HER2-positive breast cancer samples.

Methods

Spatial transcriptomics (Visium) was performed on 33 frozen HER2-positive breast cancer surgical samples, including 6 residual disease samples. H&E images of the ST slides were annotated for morphological structures at the single-cell/structure level (QuPath software). Clusters identified on integrated data (harmony R package) were characterized calculating gene expression signatures, including HER2DX gene modules, at the spot level, and by morphological annotation. Gene signatures were computed on pseudo-bulk RNA data as well.

Results

A total of 25 integrated clusters were identified (range 15-21 in each sample). As each spot/cluster represents a mixture of different cell types, using gene expression and morphological data we defined a total of 9 tumor-enriched clusters (of which 5 sample-specific), as well as 12 clusters mainly enriched for stroma, 3 for adipose tissue, and 1 for tumor-infiltrating lymphocytes. All samples presented more than 1 tumor cluster, in various proportions. Interestingly, when comparing tumor clusters, levels of HER2DX signatures depicting HER2 amplicon, luminal phenotype, proliferation, B cell infiltration, as well as signatures related to stroma activation, signaling pathways and metabolism differed, demonstrating heterogeneity in tumor-enriched areas. Of note, within the same sample, tumor clusters with high/low levels of the HER2DX modules and other signatures could co-exist, and samples presenting signature scores above/below the cohort median at the pseudo-bulk level (also influenced by the stroma composition) showed the co-presence of tumor clusters with high/low signature levels.

Conclusions

Our findings highlight the heterogeneity of HER2-positive breast cancer. Spatial transcriptomics may help in refining gene expression signatures computed on bulk RNA, and these results open to further analyses aimed at better understanding the tumor microenvironment in this disease.

Background

Triple-negative breast cancer (TNBC), a biologically diverse subtype of breast cancer, is associated with poor prognosis. Patients (pts) with TNBC who achieve pathologic complete response (pCR) after neoadjuvant chemotherapy (NAC) have improved disease free and overall survival. We investigated whether early circulating tumor DNA (ctDNA) measurements could predict response to NAC.

Methods

ctDNA was detected in 94 serial plasma samples (n) from 37 pts (N) with TNBC (stage I=2, II=19, III=16) with a tumor-informed ctDNA assay (Signatera^TM, bespoke mPCR-NGS). Pts received standard NAC; plasma was collected pre-NAC (n=30), 12 weeks after NAC initiation (mid-NAC, n=34), after NAC prior to surgery (n=15), and after surgery (n=15). Associations between ctDNA and ultrasound (US) imaging and response to NAC were evaluated. Circulating tumor cells (CTCs) were also assessed using CellSearch. pCR was defined as the absence of invasive tumor in the breast and axillary lymph nodes in surgical specimens. P values were measured using Student t-test, and correlation between variables used Pearson analysis.

Results

At diagnosis, ctDNA was detected in 90% (27/30) of patients, of whom 76% (19/25) had early ctDNA clearance by mid-NAC. Imaging at mid-NAC showed that 95% (18/19) of cases with undetectable ctDNA had evidence of partial or complete response (PR/CR), while 1 case had progressive disease. After completion of NAC, all 19 cases had PR, CR, or stable disease. Importantly, 58% (11/19) of cases who were ctDNA-negative at their mid-NAC blood draw achieved pCR, while none of the pts with ctDNA detectable at mid-NAC (N=6) had pCR. ctDNA clearance at mid-NAC was significantly associated with pCR (P=0.0304). ctDNA dynamics early in the treatment course correlated with pCR rates (P=0.009) and tumor volume reduction based on US (P=0.008). CTCs were detected in 36% (11/30) of pts pre-NAC, 47% (16/34) mid-NAC, 40% (6/15) post-NAC, and 20% (3/15) after surgery and were not associated with pCR.

Conclusions

Early clearance of ctDNA at mid-treatment but not CTCs during NAC was associated with a higher rate of pCR in TNBC. Personalized ctDNA monitoring during NAC is feasible and may help predict response and guide treatment.