StackBC: Deep learning and transfer learning techniques based stacking approach for accurate Invasive Ductal Carcinoma classification using histology images

2022 ◽  
pp. 1-12
Author(s):  
Amin Ul Haq ◽  
Jian Ping Li ◽  
Samad Wali ◽  
Sultan Ahmad ◽  
Zafar Ali ◽  
...  

Artificial intelligence (AI) based computer-aided diagnostic (CAD) systems can effectively diagnose critical disease. AI-based detection of breast cancer (BC) through images data is more efficient and accurate than professional radiologists. However, the existing AI-based BC diagnosis methods have complexity in low prediction accuracy and high computation time. Due to these reasons, medical professionals are not employing the current proposed techniques in E-Healthcare to effectively diagnose the BC. To diagnose the breast cancer effectively need to incorporate advanced AI techniques based methods in diagnosis process. In this work, we proposed a deep learning based diagnosis method (StackBC) to detect breast cancer in the early stage for effective treatment and recovery. In particular, we have incorporated deep learning models including Convolutional neural network (CNN), Long short term memory (LSTM), and Gated recurrent unit (GRU) for the classification of Invasive Ductal Carcinoma (IDC). Additionally, data augmentation and transfer learning techniques have been incorporated for data set balancing and for effective training the model. To further improve the predictive performance of model we used stacking technique. Among the three base classifiers (CNN, LSTM, GRU) the predictive performance of GRU are better as compared to individual model. The GRU is selected as a meta classifier to distinguish between Non-IDC and IDC breast images. The method Hold-Out has been incorporated and the data set is split into 90% and 10% for training and testing of the model, respectively. Model evaluation metrics have been computed for model performance evaluation. To analyze the efficacy of the model, we have used breast histology images data set. Our experimental results demonstrated that the proposed StackBC method achieved improved performance by gaining 99.02% accuracy and 100% area under the receiver operating characteristics curve (AUC-ROC) compared to state-of-the-art methods. Due to the high performance of the proposed method, we recommend it for early recognition of breast cancer in E-Healthcare.

2020 ◽  
pp. 1-16
Author(s):  
Deepika Kumar ◽  
Usha Batra

Breast cancer positions as the most well-known threat and the main source of malignant growth-related morbidity and mortality throughout the world. It is apical of all new cancer incidences analyzed among females. However, Machine learning algorithms have given rise to progress across different domains. There are various diagnostic methods available for cancer detection. However, cancer detection through histopathological images is considered to be more accurate. In this research, we have proposed the Stacked Generalized Ensemble (SGE) approach for breast cancer classification into Invasive Ductal Carcinoma+ and Invasive Ductal Carcinoma-. SGE is inspired by the stacking model which utilizes output predictions. Here, SGE uses six deep learning models as level-0 learner models or sub-models and Logistic regression is used as Level – 1 learner or meta – learner model. Invasive Ductal Carcinoma dataset for histopathology images is used for experimentation. The results of the proposed methodology have been compared and analyzed with existing machine learning and deep learning methods. The results demonstrate that the proposed methodology performed exponentially good in image classification in terms of accuracy, precision, recall, and F1 measure.


2021 ◽  
Author(s):  
Deepa B G ◽  
S. Senthil

Abstract Breast Cancer (BC) is the common type of cancer found in women which is caused due to the abnormal growth of cells in the breast. An early BC detection helps to increase the survival rate of the patient and 80% BC type was Invasive Ductal Carcinoma (IDC) .In this work, a deep learning-based IDC prediction model is proposed with multiple classifiers and CNN (Convolutional Neural Network). The developed deep learning method used a sequential Keras model like conv2D, Maxpooling2D, Dropout, Flatten and Dense. The multiple classifiers are LR (Logistic Regression), RF (Random Forest), K-NN (K-Nearest Neighbors), SVM (Support Vector Machine), Linear SVC, GNB (Gaussian NB) and DT (Decision Tree). The CNN model generated by using SkLearn, Keras and Tensor flow libraries, and results are organized by MatPlot libraries. At the classification stage, a helper function was defined, and Google Colab online browser platform used for developing the proposed model. The performance is analysed in terms of Accuracy, Precision, Recall, F1-score and Support.


2020 ◽  
Vol 38 (15_suppl) ◽  
pp. e12565-e12565
Author(s):  
Lauren Eisenbud ◽  
Tsering G. Lama Tamang ◽  
Caleb Cheng ◽  
Ibe Ifegwu ◽  
Tianyi Tang ◽  
...  

e12565 Background: DCIS is usually treated with resection followed by 5 years of adjuvant endocrine therapy for hormone receptor (HR) + DCIS. Endocrine therapy is not used in HR- DCIS. Although DCIS is considered a precursor lesion to invasive breast cancer, the different molecular subtypes confer variable clinical outcomes. The host immune response plays a key role in breast cancer progression and response to therapy. However, relative to invasive breast cancer, the immune milieu of DCIS is less understood. This retrospective study compares the clinical outcomes and tumor microenvironment of HR+ and HR- DCIS in order to identify clinical and immunological features in HR- DCIS that may predict an increased risk of recurrence or progression to invasive breast cancer. Methods: A single institution retrospective chart review was performed to identify patients diagnosed with DCIS between 2012 and 2017. A clinico-pathologic data set, as well as the PD-L1 expression of the DCIS and TILs were collected and correlated with various outcomes. Results: Our cohort consisted of 20 cases of HR- DCIS and 50 cases of HR+ DCIS. Overall, 56% were Caucasian, 20% Asian, 18% Hispanic, and 6% African American. Of the HR- patients, 70% were Caucasian, 15% Hispanic, and 15% Asian. Of the 17 HR- patients with available HER2 data, 76% had HER2+, and 24% triple negative (TN) DCIS. 18% of the HR+ patients and 38% of the HR- patients were PD-L1+. 25% of the HR-/HER2+ patients, and 75% of the TN patients were PD-L1+. 6% of the HR+ patients developed recurrent disease, 2 with DCIS and 1 with invasive ductal carcinoma. 20% of the HR- patients had recurrent disease, all of whom were HER2+. Of the HR- patients that recurred, 2 recurred with metastatic disease, 1 with ipsilateral invasive ductal carcinoma, and 1 with DCIS. All 7 patients that recurred had original DCIS pathology showing a high nuclear grade. Our future results at the time of the meeting will expand on this cohort. Conclusions: This retrospective analysis showed that HR- DCIS conferred higher rates of local and distant recurrence. Therefore, there is a need for treatments to reduce the recurrence rates of HR- DCIS. There are ongoing clinical trials for the high risk, HR-/HER2+ DCIS subtype. TN DCIS is also an aggressive phenotype. Given the high rate of PD-L1 positivity we detected in TN DCIS, immune-based therapy may be useful in the adjuvant setting to reduce the risk of recurrence in this cohort of patients.


2009 ◽  
Vol 29 (4) ◽  
pp. 400-403
Author(s):  
Shu-rong SHEN ◽  
Jun-yi SHI ◽  
Xian SHEN ◽  
Guan-li HUANG ◽  
Xiang-yang XUE

2013 ◽  
Vol 99 (1) ◽  
pp. 39-44
Author(s):  
Claudia Maria Regina Bareggi ◽  
Dario Consonni ◽  
Barbara Galassi ◽  
Donatella Gambini ◽  
Elisa Locatelli ◽  
...  

Aims and background Often neglected by large clinical trials, patients with uncommon breast malignancies have been rarely analyzed in large series. Patients and methods Of 2,052 patients diagnosed with breast cancer and followed in our Institution from January 1985 to December 2009, we retrospectively collected data on those with uncommon histotypes, with the aim of investigating their presentation characteristics and treatment outcome. Results Rare histotypes were identified in 146 patients (7.1% of our total breast cancer population), being classified as follows: tubular carcinoma in 75 (51.4%), mucinous carcinoma in 36 (24.7%), medullary carcinoma in 25 (17.1%) and papillary carcinoma in 10 patients (6.8%). Whereas age at diagnosis was not significantly different among the diverse diagnostic groups, patients with medullary and papillary subtypes had a higher rate of lymph node involvement, similar to that of invasive ductal carcinoma. Early stage diagnosis was frequent, except for medullary carcinoma. Overall, in comparison with our invasive ductal carcinoma patients, those with rare histotypes showed a significantly lower risk of recurrence, with a hazard ratio of 0.28 (95% CI, 0.12–0.62; P = 0.002). Conclusions According to our analysis, patients with uncommon breast malignancies are often diagnosed at an early stage, resulting in a good prognosis with standard treatment.


2021 ◽  
Author(s):  
Süleyman UZUN ◽  
Sezgin KAÇAR ◽  
Burak ARICIOĞLU

Abstract In this study, for the first time in the literature, identification of different chaotic systems by classifying graphic images of their time series with deep learning methods is aimed. For this purpose, a data set is generated that consists of the graphic images of time series of the most known three chaotic systems: Lorenz, Chen, and Rossler systems. The time series are obtained for different parameter values, initial conditions, step size and time lengths. After generating the data set, a high-accuracy classification is performed by using transfer learning method. In the study, the most accepted deep learning models of the transfer learning methods are employed. These models are SqueezeNet, VGG-19, AlexNet, ResNet50, ResNet101, DenseNet201, ShuffleNet and GoogLeNet. As a result of the study, classification accuracy is found between 96% and 97% depending on the problem. Thus, this study makes association of real time random signals with a mathematical system possible.


2009 ◽  
Vol 27 (30) ◽  
pp. 4939-4947 ◽  
Author(s):  
Heather A. Jones ◽  
Ninja Antonini ◽  
Augustinus A.M. Hart ◽  
Johannes L. Peterse ◽  
Jean-Claude Horiot ◽  
...  

Purpose To investigate the long-term impact of pathologic characteristics and an extra boost dose of 16 Gy on local relapse, for stage I and II invasive breast cancer patients treated with breast conserving therapy (BCT). Patients and Methods In the European Organisation for Research and Treatment of Cancer boost versus no boost trial, after whole breast irradiation, patients with microscopically complete excision of invasive tumor, were randomly assigned to receive or not an extra boost dose of 16 Gy. For a subset of 1,616 patients central pathology review was performed. Results The 10-year cumulative risk of local breast cancer relapse as a first event was not significantly influenced if the margin was scored negative, close or positive for invasive tumor or ductal carcinoma in situ according to central pathology review (log-rank P = .45 and P = .57, respectively). In multivariate analysis, high-grade invasive ductal carcinoma was associated with an increased risk of local relapse (P = .026; hazard ratio [HR], 1.67), as was age younger than 50 years (P < .0001; HR, 2.38). The boost dose of 16 Gy significantly reduced the local relapse rate (P = .0006; HR, 0.47). For patients younger than 50 years old and in patients with high grade invasive ductal carcinoma, the boost dose reduced the local relapse from 19.4% to 11.4% (P = .0046; HR, 0.51) and from 18.9% to 8.6% (P = .01; HR, 0.42), respectively. Conclusion Young age and high-grade invasive ductal cancer were the most important risk factors for local relapse, while margin status had no significant influence. A boost dose of 16 Gy significantly reduced the negative effects of both young age and high-grade invasive cancer.


Sign in / Sign up

Export Citation Format

Share Document