Deep learning trained on hematoxylin and eosin tumor region of Interest predicts HER2 status and trastuzumab treatment response in HER2+ breast cancer

2021 ◽  
Author(s):  
Saman Farahmand ◽  
Aileen I. Fernandez ◽  
Fahad Shabbir Ahmed ◽  
David L. Rimm ◽  
Jeffrey H. Chuang ◽  
...  
Keyword(s):  
Breast Cancer ◽  
Deep Learning ◽  
Treatment Response ◽  
Region Of Interest ◽  
Her2 Status ◽  
Trastuzumab Treatment ◽  
Her2 Breast Cancer ◽  
Tumor Region ◽  
Hematoxylin And Eosin

scholarly journals Deep learning trained on H&E tumor ROIs predicts HER2 status and Trastuzumab treatment response in HER2+ breast cancer

2021 ◽  
Author(s):  
saman farahmand ◽  
Aileen Fernandez ◽  
Fahad Shabbir Ahmed ◽  
David Rimm ◽  
Jeffrey H Chuang ◽  
...  
Keyword(s):  
Breast Cancer ◽  
Treatment Response ◽  
Cross Validation ◽  
Area Under The Curve ◽  
Standard Of Care ◽  
Her2 Status ◽  
Current Standard ◽  
Her2 Positive ◽  
Trastuzumab Therapy ◽  
Her2 Breast Cancer

The current standard of care for many patients with HER2-positive breast cancer is neoadjuvant chemotherapy in combination with anti-HER2 agents, based on HER2 amplification as detected by in situ hybridization (ISH) or protein immunohistochemistry (IHC). However, hematoxylin & eosin (H&E) tumor stains are more commonly available, and accurate prediction of HER2 status and anti-HER2 treatment response from H&E would reduce costs and increase the speed of treatment selection. Computational algorithms for H&E have been effective in predicting a variety of cancer features and clinical outcomes, including moderate success in predicting HER2 status. In this work, we present a novel convolutional neural network (CNN) approach able to predict HER2 status with increased accuracy over prior methods. We trained a CNN classifier on 188 H&E whole slide images (WSIs) manually annotated for tumor regions of interest (ROIs) by our pathology team. Our classifier achieved an area under the curve (AUC) of 0.90 in cross-validation of slide-level HER2 status and 0.81 on an independent TCGA test set. Within slides, we observed strong agreement between pathologist annotated ROIs and blinded computational predictions of tumor regions / HER2 status. Moreover, we trained our classifier on pre-treatment samples from 187 HER2+ patients that subsequently received trastuzumab therapy. Our classifier achieved an AUC of 0.80 in a five-fold cross validation. Our work provides an H&E-based algorithm that can predict HER2 status and trastuzumab response in breast cancer at an accuracy that is better than IHC and may benefit clinical evaluations.

scholarly journals Towards integration of 64Cu-DOTA-trastuzumab PET-CT and MRI with mathematical modeling to predict response to neoadjuvant therapy in HER2 + breast cancer

2020 ◽  
Vol 10 (1) ◽  
Author(s):  
Angela M. Jarrett ◽  
David A. Hormuth ◽  
Vikram Adhikarla ◽  
Prativa Sahoo ◽  
Daniel Abler ◽  
...  
Keyword(s):  
Breast Cancer ◽  
Targeted Therapy ◽  
Treatment Response ◽  
Growth Factor Receptor ◽  
Patient Specific ◽  
Mathematical Framework ◽  
Breast Cancer Patients ◽  
Data Types ◽  
Her2 Breast Cancer ◽  
Mri Scans

AbstractWhile targeted therapies exist for human epidermal growth factor receptor 2 positive (HER2 +) breast cancer, HER2 + patients do not always respond to therapy. We present the results of utilizing a biophysical mathematical model to predict tumor response for two HER2 + breast cancer patients treated with the same therapeutic regimen but who achieved different treatment outcomes. Quantitative data from magnetic resonance imaging (MRI) and 64Cu-DOTA-trastuzumab positron emission tomography (PET) are used to estimate tumor density, perfusion, and distribution of HER2-targeted antibodies for each individual patient. MRI and PET data are collected prior to therapy, and follow-up MRI scans are acquired at a midpoint in therapy. Given these data types, we align the data sets to a common image space to enable model calibration. Once the model is parameterized with these data, we forecast treatment response with and without HER2-targeted therapy. By incorporating targeted therapy into the model, the resulting predictions are able to distinguish between the two different patient responses, increasing the difference in tumor volume change between the two patients by > 40%. This work provides a proof-of-concept strategy for processing and integrating PET and MRI modalities into a predictive, clinical-mathematical framework to provide patient-specific predictions of HER2 + treatment response.

Survival by HER2 receptor status in stage IV breast cancer: SEER 2010-2012.

2017 ◽  
Vol 35 (15_suppl) ◽  
pp. 1032-1032 ◽  
Author(s):  
Alexandra Thomas ◽  
Seema Ahsan Khan ◽  
Charles Lynch ◽  
Mary Chen Schroeder
Keyword(s):  
Breast Cancer ◽  
De Novo ◽  
Reference Group ◽  
Cox Model ◽  
Stage Iv ◽  
The United States ◽  
Her2 Status ◽  
Her2 Receptor ◽  
Her2 Breast Cancer ◽  
Stage Iv Breast Cancer

1032 Background: Therapeutic advances have altered the course of once highly lethal HER2+ breast cancer (BC). We report survival in a recent population-based cohort by HER2 status, overall, and within hormone receptor(HR)+ BC. Methods: Surveillance, Epidemiology, and End Results Program data were queried to identify women diagnosed 2010-2012 with Stage IV BC as first cancer. Patients were grouped by HER2 and HR status. Kaplan Meier estimates of 3-yr observed survival (OS) were compared with log-rank tests. A multivariate cox model was fitted for the HER2+ cohort. Results: 3-yr OS for HER2+(any HR), HR+/HER- and triple-negative (TN) BC was 52.3%, 48.4% and 16.0% respectively (p<0.01 HER2+(any HR) vs TNBC; p=0.20 HER2+(any HR) vs HR+/HER2-). Across registries, OS for HER2+(any HR) BC ranged from 29.2% to 61.7% (p=0.05). On Cox model, survival in HER2+(any HR) BC was associated with age 50+ (Hazard ratio (HR) 1.84, 95% CI 1.45-2.34), HR+ status (HR 0.70, 0.58-0.84), high histologic grade (HR 1.30, 0.58-0.84), surgery (HR 0.40, 0.33-0.49), separated marital status (HR 1.72, 1.4-2.13), year 2012 (HR 0.81, 0.64-1.04), and registry (varies by reference group). For HR+ BC, OS also differed by HER2 status: 55.3% for HR+/HER+ and 48.4% for HR+/HER2- (p<0.01). 3-yr OS by HER2 status for women presenting with HR+ BC is shown (Table). Conclusions: Survival in de novo Stage IV HER2+ BC in the United States exceeds that in HER2- BC, with median survival >3 yrs. Survival was significantly better for HR+/HER+ BC than HR+/HER- BC. Disparate OS in HER2+ BC suggest opportunities may remain to fully realize advances in HER2-directed therapy. Given recent therapeutic advances, the trend of HER2+ survival gains from 2010 to 2012 will likely continue. [Table: see text]

scholarly journals Hormone receptor status does not alter the effect of trastuzumab in breast cancer

2016 ◽  
Vol 23 (5) ◽  
pp. 349-355 ◽  
Author(s):  
Tanja Ignatov ◽  
Holm Eggemann ◽  
Elke Burger ◽  
Serban Dan Costa ◽  
Atanas Ignatov
Keyword(s):  
Breast Cancer ◽  
Hormone Receptor ◽  
Performance Status ◽  
Her2 Status ◽  
Study Cohort ◽  
Her2 Overexpression ◽  
Adjuvant Trastuzumab ◽  
Treatment Benefit ◽  
Cancer Data ◽  
Trastuzumab Treatment

Overexpression of human epidermal growth factor receptor 2 (HER2) predicts response to anti-HER2 therapy in breast cancer. We investigated whether hormone receptor (HR) status influences the treatment benefit of trastuzumab in patients with breast cancer. Data from 8338 patients with primary nonmetastatic breast cancer from the cancer registry of Saxony-Anhalt Germany were analyzed. A total of 5554 patients were eligible for analysis. The median follow-up of the study was 6 years. Of the 5554 patients investigated, 1128 (20.3%) showed HER2 overexpression and 656 (58.2%) of them received adjuvant trastuzumab. The 10-year overall survival (OS) in the study cohort according to HR, HER2 status, and trastuzumab treatment was as follows: 78.4% for HR−/HER2−, 85.0% for HR+/HER2−, 70.4% HR–/HER2+/TRA−, 71.4% for HR+/HER2+/TRA−, 80.9% for HR−/HER2+/TRA+, and 89.2% for HR+/HER2+/TRA+. Trastuzumab treatment improved OS in the HR− patients only in the first 3 years, whereas in the HR+ group the effect of trastuzumab was still apparent 5 years after diagnosis. Notably, the relative improvement in a patient outcome was higher for HR+ patients. Nevertheless, matching for age, histological type, tumor stage, tumor grade, and performance status between patients with HR− and HR+ tumors demonstrated that the survival effect of trastuzumab was not affected by HR status; P=0.890. Trastuzumab treatment improves patients’ survival regardless of HR status and should be offered to all HER2+ patients.

scholarly journals Breast Tumor Cellularity Assessment using Deep Neural Networks

2019 ◽  
Author(s):  
Alexander Rakhlin ◽  
Aleksei Tiulpin ◽  
Alexey A. Shvets ◽  
Alexandr A. Kalinin ◽  
Vladimir I. Iglovikov ◽  
...  
Keyword(s):  
Breast Cancer ◽  
Deep Learning ◽  
Automatic Assessment ◽  
Appropriate Treatment ◽  
Image Patches ◽  
Expert Pathologist ◽  
Hematoxylin And Eosin ◽  
Treated Breast ◽  
Intra Class Correlation Coefficient ◽  
Whole Slide Images

AbstractBreast cancer is one of the main causes of death world-wide. Histopathological cellularity assessment of residual tumors in post-surgical tissues is used to analyze a tumor’s response to a therapy. Correct cellularity assessment increases the chances of getting an appropriate treatment and facilitates the patient’s survival. In current clinical practice, tumor cellularity is manually estimated by pathologists; this process is tedious and prone to errors or low agreement rates between assessors. In this work, we evaluated three strong novel Deep Learning-based approaches for automatic assessment of tumor cellularity from post-treated breast surgical specimens stained with hematoxylin and eosin. We validated the proposed methods on the BreastPathQ SPIE challenge dataset that consisted of 2395 image patches selected from whole slide images acquired from 64 patients. Compared to expert pathologist scoring, our best performing method yielded the Cohen’s kappa coefficient of 0.69 (vs. 0.42 previously known in literature) and the intra-class correlation coefficient of 0.89 (vs. 0.83). Our results suggest that Deep Learning-based methods have a significant potential to alleviate the burden on pathologists, enhance the diagnostic workflow, and, thereby, facilitate better clinical outcomes in breast cancer treatment.

scholarly journals Deep Learning-Based Breast Cancer Diagnosis at Ultrasound: Initial Application of Weakly-Supervised Algorithm Without Image Annotation Original Research

2021 ◽  
Author(s):  
Jaeil Kim ◽  
Hye Jung Kim ◽  
Chanho Kim ◽  
Jin Hwa Lee ◽  
Keum Won Kim ◽  
...  
Keyword(s):  
Breast Cancer ◽  
Deep Learning ◽  
Image Annotation ◽  
Characteristic Curve ◽  
External Validation ◽  
Region Of Interest ◽  
Breast Cancer Diagnosis ◽  
Original Research ◽  
Internal Validation ◽  
Weakly Supervised

Abstract Conventional deep learning (DL) algorithm requires full supervision of annotating the region of interest (ROI) that is laborious and often biased. We aimed to develop a weakly-supervised DL algorithm that diagnosis breast cancer at ultrasound without image annotation. Weakly-supervised DL algorithms were implemented with three networks (VGG16, ResNet34, and GoogLeNet) and trained using 1000 unannotated US images (500 benign and 500 malignant masses). Two sets of 200 images (100 benign and 100 malignant masses) were used for internal and external validation sets. For comparison with fully-supervised algorithms, ROI annotation was performed manually and automatically. Diagnostic performances were calculated as the area under the receiver operating characteristic curve (AUC). Using the class activation map, we determined how accurately the weakly-supervised DL algorithms localized the breast masses. For internal validation sets, the weakly-supervised DL algorithms achieved excellent diagnostic performances, with AUC values of 0.92–0.96, which were not statistically different (all Ps > 0.05) from those of fully-supervised DL algorithms with either manual or automated ROI annotation (AUC, 0.92–0.96). For external validation sets, the weakly-supervised DL algorithms achieved AUC values of 0.86–0.90, which were not statistically different (Ps > 0.05) or higher (P = 0.04, VGG16 with automated ROI annotation) from those of fully-supervised DL algorithms (AUC, 0.84–0.92). In internal and external validation sets, weakly-supervised algorithms could localize 100% of malignant masses, except for ResNet34 (98%). The weakly-supervised DL algorithms developed in the present study were feasible for US diagnosis of breast cancer with well-performing localization and differential diagnosis.

scholarly journals Molecular Subtypes of Breast Cancer: A Review for Breast Radiologists

2020 ◽  
Author(s):  
Karen S Johnson ◽  
Emily F Conant ◽  
Mary Scott Soo
Keyword(s):  
Breast Cancer ◽  
Molecular Subtypes ◽  
Disease Free Survival ◽  
Response To Treatment ◽  
Her2 Status ◽  
Imaging Features ◽  
Luminal A ◽  
Luminal B ◽  
Her2 Breast Cancer ◽  
Er Expression

Abstract Gene expression profiling has reshaped our understanding of breast cancer by identifying four molecular subtypes: (1) luminal A, (2) luminal B, (3) human epidermal growth factor receptor 2 (HER2)-enriched, and (4) basal-like, which have critical differences in incidence, response to treatment, disease progression, survival, and imaging features. Luminal tumors are most common (60%–70%), characterized by estrogen receptor (ER) expression. Luminal A tumors have the best prognosis of all subtypes, whereas patients with luminal B tumors have significantly shorter overall and disease-free survival. Distinguishing between these tumors is important because luminal B tumors require more aggressive treatment. Both commonly present as irregular masses without associated calcifications at mammography; however, luminal B tumors more commonly demonstrate axillary involvement at diagnosis. HER2-enriched tumors are characterized by overexpression of the HER2 oncogene and low-to-absent ER expression. HER2+ disease carries a poor prognosis, but the development of anti-HER2 therapies has greatly improved outcomes for women with HER2+ breast cancer. HER2+ tumors most commonly present as spiculated masses with pleomorphic calcifications or as calcifications alone. Basal-like cancers (15% of all invasive breast cancers) predominate among “triple negative” cancers, which lack ER, progesterone receptor (PR), and HER2 expression. Basal-like cancers are frequently high-grade, large at diagnosis, with high rates of recurrence. Although imaging commonly reveals irregular masses with ill-defined or spiculated margins, some circumscribed basal-like tumors can be mistaken for benign lesions. Incorporating biomarker data (histologic grade, ER/PR/HER2 status, and multigene assays) into classic anatomic TNM staging can better inform clinical management of this heterogeneous disease.

Sign in / Sign up

Export Citation Format

Share Document