Dataset Information

Deep learning identifies morphological features in breast cancer predictive of cancer ERBB2 status and trastuzumab treatment efficacy.

ABSTRACT: The treatment of patients with ERBB2 (HER2)-positive breast cancer with anti-ERBB2 therapy is based on the detection of ERBB2 gene amplification or protein overexpression. Machine learning (ML) algorithms can predict the amplification of ERBB2 based on tumor morphological features, but it is not known whether ML-derived features can predict survival and efficacy of anti-ERBB2 treatment. In this study, we trained a deep learning model with digital images of hematoxylin-eosin (H&E)-stained formalin-fixed primary breast tumor tissue sections, weakly supervised by ERBB2 gene amplification status. The gene amplification was determined by chromogenic in situ hybridization (CISH). The training data comprised digitized tissue microarray (TMA) samples from 1,047 patients. The correlation between the deep learning-predicted ERBB2 status, which we call H&E-ERBB2 score, and distant disease-free survival (DDFS) was investigated on a fully independent test set, which included whole-slide tumor images from 712 patients with trastuzumab treatment status available. The area under the receiver operating characteristic curve (AUC) in predicting gene amplification in the test sets was 0.70 (95% CI, 0.63-0.77) on 354 TMA samples and 0.67 (95% CI, 0.62-0.71) on 712 whole-slide images. Among patients with ERBB2-positive cancer treated with trastuzumab, those with a higher than the median morphology-based H&E-ERBB2 score derived from machine learning had more favorable DDFS than those with a lower score (hazard ratio [HR] 0.37; 95% CI, 0.15-0.93; P?=?0.034). A high H&E-ERBB2 score was associated with unfavorable survival in patients with ERBB2-negative cancer as determined by CISH. ERBB2-associated morphology correlated with the efficacy of adjuvant anti-ERBB2 treatment and can contribute to treatment-predictive information in breast cancer.

SUBMITTER: Bychkov D

PROVIDER: S-EPMC7890057 | biostudies-literature | 2021 Feb

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Deep learning identifies morphological features in breast cancer predictive of cancer ERBB2 status and trastuzumab treatment efficacy.

Bychkov Dmitrii D Linder Nina N Tiulpin Aleksei A Kücükel Hakan H Lundin Mikael M Nordling Stig S Sihto Harri H Isola Jorma J Lehtimäki Tiina T Kellokumpu-Lehtinen Pirkko-Liisa PL von Smitten Karl K Joensuu Heikki H Lundin Johan J

Scientific reports 20210217 1

The treatment of patients with ERBB2 (HER2)-positive breast cancer with anti-ERBB2 therapy is based on the detection of ERBB2 gene amplification or protein overexpression. Machine learning (ML) algorithms can predict the amplification of ERBB2 based on tumor morphological features, but it is not known whether ML-derived features can predict survival and efficacy of anti-ERBB2 treatment. In this study, we trained a deep learning model with digital images of hematoxylin-eosin (H&E)-stained formali ...[more]

PMID: 33597560

Similar Datasets

Project description:Targeted therapy using monoclonal antibodies conjugated to toxins is gaining space in the treatment of cancer. Here, we report the anti-tumor effect of a new antibody drug conjugate (ADC) delivering a HDAC inhibitor to ErbB2+ solid tumors. Trastuzumab was partially reduced with tris [2-carboxyethyl] phosphine (TCEP) and conjugated to ST7464AA1, the active form of the prodrug HDAC inhibitor ST7612AA1, through a maleimide-thiol linker to obtain the Antibody Drug Conjugate (ADC) ST8176AA1. The average drug/antibody ratio (DAR) was 4.5 as measured by hydrophobic interaction chromatography (HIC). Binding of ST8176AA1 to ErbB2 receptor and internalization in tumor cells were investigated by enzyme-linked immunosorbent assay (ELISA), surface plasmon resonance (SPR), cytofluorimetry, and High Content Screening (HCS) Imaging. The biological activity of the ADC was evaluated in vitro and in vivo by measuring cell proliferation/cell cycle, apoptosis/DNA damage, tubulin, and histone acetylation and modulation of Epithelial/Mesenchymal Transition (EMT) markers. Receptor binding and internalization of ST8176AA1 were confirmed to be similar to trastuzumab. Higher anti-tumor activity of ST8176AA1 compared to trastuzumab was observed in vitro in tumor cell lines. Such higher activity correlated with increased acetylation of histones and alfa-tubulin as a consequence of HDAC inhibitor-mediated epigenetic modulation that also induced increased expression of ErbB2 and estrogen receptor in triple negative breast cancer cells. Consistently with in vitro data, ST8176AA1 exhibited higher tumor growth inhibition than trastuzumab in xenograft models of ovary and colon carcinoma and in two patient-derived xenograft (PDX) models of pancreatic carcinoma. Immunohistochemistry analysis of tumor masses showed lower expression of the proliferation marker Ki67 and higher expression of cleaved caspase-3 in mice treated with the ADC compared to those treated with trastuzumab and results correlated with increased acetylation of both histones and tubulin. Collectively, present data indicate that ADC ST8176AA1 can target epigenetic modulation to ErbB2+ tumors. Interestingly, the amount of HDACi estimated to be delivered at the ST8176AA1 effective dose would correspond to ~1/1,000 of ST7612AA1 effective dose. Therefore, ST8176AA1 is an attractive new therapeutic candidate because it exhibits increased anti-tumor potency compared to trastuzumab by exerting epigenetic modulation at a much safer dose compared to standard HDACi-based therapeutic protocols.

Project description:ImportanceERRB2 (formerly HER2)-positive advanced breast cancer (ABC) remains typically incurable with optimal treatment undefined in later lines of therapy. The chimeric antibody margetuximab shares ERBB2 specificity with trastuzumab but incorporates an engineered Fc region to increase immune activation.ObjectiveTo compare the clinical efficacy of margetuximab vs trastuzumab, each with chemotherapy, in patients with pretreated ERBB2-positive ABC.Design, setting, and participantsThe SOPHIA phase 3 randomized open-label trial of margetuximab plus chemotherapy vs trastuzumab plus chemotherapy enrolled 536 patients from August 26, 2015, to October 10, 2018, at 166 sites in 17 countries. Eligible patients had disease progression on 2 or more prior anti-ERBB2 therapies and 1 to 3 lines of therapy for metastatic disease. Data were analyzed from February 2019 to October 2019.InterventionsInvestigators selected chemotherapy before 1:1 randomization to margetuximab, 15 mg/kg, or trastuzumab, 6 mg/kg (loading dose, 8 mg/kg), each in 3-week cycles. Stratification factors were metastatic sites (≤2, >2), lines of therapy (≤2, >2), and chemotherapy choice.Main outcomes and measuresSequential primary end points were progression-free survival (PFS) by central blinded analysis and overall survival (OS). All α was allocated to PFS, followed by OS. Secondary end points were investigator-assessed PFS and objective response rate by central blinded analysis.ResultsA total of 536 patients were randomized to receive margetuximab (n = 266) or trastuzumab (n = 270). The median age was 56 (27-86) years; 266 (100%) women were in the margetuximab group, while 267 (98.9%) women were in the trastuzumab group. Groups were balanced. All but 1 patient had received prior pertuzumab, and 489 (91.2%) had received prior ado-trastuzumab emtansine. Margetuximab improved primary PFS over trastuzumab with 24% relative risk reduction (hazard ratio [HR], 0.76; 95% CI, 0.59-0.98; P = .03; median, 5.8 [95% CI, 5.5-7.0] months vs 4.9 [95% CI, 4.2-5.6] months; October 10, 2018). After the second planned interim analysis of 270 deaths, median OS was 21.6 months with margetuximab vs 19.8 months with trastuzumab (HR, 0.89; 95% CI, 0.69-1.13; P = .33; September 10, 2019), and investigator-assessed PFS showed 29% relative risk reduction favoring margetuximab (HR, 0.71; 95% CI, 0.58-0.86; P < .001; median, 5.7 vs 4.4 months; September 10, 2019). Margetuximab improved objective response rate over trastuzumab: 22% vs 16% (P = .06; October 10, 2018), and 25% vs 14% (P < .001; September 10, 2019). Incidence of infusion-related reactions, mostly in cycle 1, was higher with margetuximab (35 [13.3%] vs 9 [3.4%]); otherwise, safety was comparable.Conclusions and relevanceIn this phase 3 randomized clinical trial, margetuximab plus chemotherapy had acceptable safety and a statistically significant improvement in PFS compared with trastuzumab plus chemotherapy in ERBB2-positive ABC after progression on 2 or more prior anti-ERBB2 therapies. Final OS analysis is expected in 2021.Trial registrationClinicalTrials.gov Identifier: NCT02492711.

Project description:Trastuzumab resistance is a severe problem in the treatment of ErbB2-amplified cancer. Although trastuzumab plus pertuzumab is able to partly overcome trastuzumab resistance in ErbB2-overexpressing cancer, its antitumor efficacy remains limited. The present study investigated the antitumor activity of the combination of trastuzumab with H2-18, which is an antibody targetinsg ErbB2 domain I. Cell proliferation and inhibition experiments indicated that H2-18 and trastuzumab synergistically inhibited the proliferation of both the trastuzumab-sensitive gastric cancer cell line, NCI-N87 and the trastuzumab-resistant gastric cancer cell line, NCI-N87-TraRT. Furthermore, H2-18 plus trastuzumab inhibited the growth of NCI-N87-TraRT cells more effectively than trastuzumab plus pertuzumab, both in vitro and in vivo. Compared with trastuzumab plus pertuzumab, H2-18 plus trastuzumab had a potent ability to inhibit NCI-N87-TraRT cells to form colonies. Notably, H2-18 plus trastuzumab was more effective in inducing cell death than trastuzumab plus pertuzumab. The in vivo studies demonstrated that H2-18 plus trastuzumab effectively inhibited the growth of both NCI-N87 and NCI-N87-TraRT xenograft tumors. Further experiments revealed that in NCI-N87-TraRT cells, H2-18 plus trastuzumab was comparable to trastuzumab plus pertuzumab in the inhibition of phosphorylated (p-)HER3, p-AKT and p-ERK. However, compared with trastuzumab plus pertuzumab, H2-18 plus trastuzumab effectively activated ROS production and the phosphorylation of JNK and c-jun in NCI-N87-TraRT cells. Therefore, the superior antitumor efficacy of H2-18 plus trastuzumab over trastuzumab plus pertuzumab may be mainly attributable to the potent cell death-inducing activity. In addition, the in vitro and in vivo antitumor effect of the combination of H2-18, trastuzumab and pertuzumab were further investigated. The results revealed that H2-18 plus trastuzumab plus pertuzumab exhibited a maximal antitumor effect among all the anti-ErbB2 monoclonal antibody combinations tested. In summary, H2-18 plus trastuzumab may have potential as an effective strategy to overcome the resistance to trastuzumab in ErbB2-amplified gastric cancer cell lines.

Project description:ImportanceTrastuzumab has been the standard of care for the treatment of patients with ERBB2-positive breast cancer; however, cardiac events have been reported. This long-term follow-up study provides clinical evidence supporting the similarity of a trastuzumab biosimilar (SB3) to reference trastuzumab (TRZ).ObjectiveTo compare cardiac safety and efficacy between SB3 and TRZ for patients with ERBB2-positive early or locally advanced breast cancer after up to 6 years of follow-up.Design, setting, and participantsThis prespecified secondary analysis of a randomized clinical trial, conducted from April 2016 to January 2021, included patients with ERBB2-positive early or locally advanced breast cancer from a multicenter double-blind, parallel-group, equivalence phase 3 randomized clinical trial of SB3 vs TRZ with concomitant neoadjuvant chemotherapy who completed neoadjuvant and adjuvant treatment.InterventionsIn the original trial, patients were randomized to either SB3 or TRZ with concomitant neoadjuvant chemotherapy for 8 cycles (4 cycles of docetaxel followed by 4 cycles of fluorouracil, epirubicin, and cyclophosphamide). After surgery, patients continued SB3 or TRZ monotherapy for 10 cycles of adjuvant treatment per previous treatment allocation. Following neoadjuvant and adjuvant treatment, patients were monitored for up to 5 years.Main outcomes and measuresThe primary outcomes were the incidence of symptomatic congestive heart failure and asymptomatic, significant decrease in left ventricular ejection fraction (LVEF). The secondary outcomes were event-free survival (EFS) and overall survival (OS).ResultsA total of 538 female patients were included (median age, 51 years [range, 22-65 years]). Baseline characteristics were comparable between the SB3 and TRZ groups. Cardiac safety was monitored for 367 patients (SB3, n = 186; TRZ, n = 181). Median follow-up was 68 months (range, 8.5-78.1 months). Asymptomatic, clinically significant LVEF decreases were rarely reported (SB3, 1 patient [0.4%]; TRZ, 2 [0.7%]). No patient experienced symptomatic cardiac failure or death due to a cardiovascular event. Survival was evaluated for the 367 patients in the cardiac safety cohort and an additional 171 patients enrolled after a protocol amendment (538 patients [SB3, n = 267; TRZ, n = 271]). No difference was observed in EFS or OS between treatment groups (EFS: hazard ratio [HR], 0.84; 95% CI, 0.58-1.20; P = .34; OS: HR, 0.61; 95% CI, 0.36-1.05; P = .07). Five-year EFS rates were 79.8% (95% CI, 74.8%-84.9%) in the SB3 group and 75.0% (95% CI, 69.7%-80.3%) in the TRZ group, and OS rates were 92.5% (95% CI, 89.2%-95.7%) in the SB3 group and 85.4% (95% CI, 81.0%-89.7%) in the TRZ group.Conclusions and relevanceIn this secondary analysis of a randomized clinical trial, SB3 demonstrated cardiac safety and survival comparable to those of TRZ after up to 6 years of follow-up in patients with ERBB2-positive early or locally advanced breast cancer.Trial registrationClinicalTrials.gov Identifier: NCT02771795.

Project description:ERBB2 is overexpressed in up to 20-30% of human breast cancers (BCs), and it is associated with aggressive disease. Trastuzumab (Tz), a humanized monoclonal antibody, improves the prognosis associated with ERBB2-amplified BCs. However, the development of resistance remains a significant challenge. Carnosic acid (CA) is a diterpene found in rosemary and sage, endowed with anticancer properties. In this in vitro study, we have investigated whether Tz and CA have cooperative effects on cell survival of ERBB2 overexpressing (ERBB2+) cells and whether CA might restore Tz sensitivity in Tz-resistant cells.We have studied BC cell migration and survival upon CA and Tz treatment. In particular, migration ability was assessed by transwell assay while cell survival was assessed by MTT assay. In addition, we have performed cell cycle and apoptosis analysis by high-resolution DNA flow cytometry and annexin-V, resazurin and sytox blue staining by flow cytometry, respectively. The expression of proteins involved in cell cycle progression, ERBB2 signaling pathway, and autophagy was evaluated by immunoblot and immunofluorescence analysis. Cellular structures relevant to the endosome/lysosome and autophagy pathways have been studied by immunofluorescence and transmission electron microscopy.We report that, in ERBB2+ BC cells, CA reversibly enhances Tz inhibition of cell survival, cooperatively inhibits cell migration and induces cell cycle arrest in G0/G1. These events are accompanied by ERBB2 down-regulation, deregulation of the PI3K/AKT/mTOR signaling pathway and up-regulation of both CDKN1A/p21WAF1 and CDKN1B/p27KIP1. Furthermore, we have demonstrated that CA impairs late autophagy and causes derangement of the lysosomal compartment as shown by up-regulation of SQSTM1/p62 and ultrastructural analysis. Accordingly, we have found that CA restores, at least in part, sensitivity to Tz in SKBR-3 Tz-resistant cell line.Our data demonstrate the cooperation between CA and Tz in inhibiting cell migration and survival of ERBB2+ BC cells that warrant further studies to establish if CA or CA derivatives may be useful in vivo in the treatment of ERBB2+ cancers.

Project description:ImportanceLittle is known regarding the outcomes associated with tucatinib combined with trastuzumab and capecitabine (TTC) after trastuzumab-deruxtecan exposure among patients with ERBB2 (previously HER2)-positive metastatic breast cancer (MBC).ObjectiveTo investigate outcomes following TTC treatment in patients with ERBB2-positive MBC who had previously received trastuzumab-deruxtecan.Design, setting, and participantsThis cohort study included all patients with MBC who were treated in 12 French comprehensive cancer centers between August 1, 2020, and December 31, 2022.ExposureTucatinib combined with trastuzumab and capecitabine administered at the recommended dose.Main outcomes and measuresClinical end points included progression-free survival (PFS), time to next treatment (TTNT), overall survival (OS), and overall response rate (ORR).ResultsA total of 101 patients with MBC were included (median age, 56 [range, 31-85] years). The median number of prior treatment lines for metastatic disease at TTC treatment initiation was 4 (range, 2-15), including 82 patients (81.2%) with previous trastuzumab and/or pertuzumab and 94 (93.1%) with previous ado-trastuzumab-emtansine) exposure. The median duration of trastuzumab-deruxtecan treatment was 8.9 (range, 1.4-25.8) months, and 82 patients (81.2%) had disease progression during trastuzumab-deruxtecan treatment, whereas 18 (17.8%) had stopped trastuzumab-deruxtecan for toxic effects and 1 (1.0%) for other reasons. Tucatinib combined with trastuzumab and capecitabine was provided as a third- or fourth-line treatment in 37 patients (36.6%) and was the immediate treatment after trastuzumab-deruxtecan in 86 (85.1%). With a median follow-up of 11.6 (95% CI, 10.5-13.4) months, 76 of 101 patients (75.2%) stopped TTC treatment due to disease progression. The median PFS was 4.7 (95% CI, 3.9-5.6) months; median TTNT, 5.2 (95% CI, 4.5-7.0) months; and median OS, 13.4 (95% CI, 11.1 to not reached [NR]) months. Patients who received TTC immediately after trastuzumab-deruxtecan had a median PFS of 5.0 (95% CI, 4.2-6.0) months; median TTNT of 5.5 (95% CI, 4.8-7.2) months, and median OS of 13.4 (95% CI, 11.9-NR) months. Those who received TTC due to trastuzumab-deruxtecan toxicity-related discontinuation had a median PFS of 7.3 (95% CI, 3.0-NR) months. Best ORR was 29 of 89 patients (32.6%). Sixteen patients with active brain metastasis had a median PFS of 4.7 (95% CI, 3.0-7.3) months, median TTNT of 5.6 (95% CI, 4.4 to NR), and median OS of 12.4 (95% CI, 8.3-NR) months.Conclusions and relevanceIn this study, TTC therapy was associated with clinically meaningful outcomes in patients with ERBB2-positive MBC after previous trastuzumab-deruxtecan treatment, including those with brain metastases. Prospective data on optimal drug sequencing in this rapidly changing therapeutic landscape are needed.

Dataset Information

Deep learning identifies morphological features in breast cancer predictive of cancer ERBB2 status and trastuzumab treatment efficacy.

Publications

Deep learning identifies morphological features in breast cancer predictive of cancer ERBB2 status and trastuzumab treatment efficacy.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets