{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"# **Histopathologic Cancer Detection** \n**Identification of Metastatic Tissue in Histopathologic Scans of Lymph Node Sections**","metadata":{}},{"cell_type":"markdown","source":"**Library used**: Keras, PyTorch\n\n**Language**: Python\n\nBy Jenna Ward, Gaurav Samudra, Lokanath Pedna, and Leopold Kamga Tchomgwi","metadata":{}},{"cell_type":"markdown","source":"","metadata":{}},{"cell_type":"markdown","source":"# **Project Description**","metadata":{}},{"cell_type":"markdown","source":"Metastasis is the principal cause of death for 66 to 90% of patients with cancer (Chaffer CL, Weinberg RA. A perspective on cancer cell metastasis. Science. 2011;331(6024):1,559–64.). Metastasis is a term used to describe the spread of cancer cells from the primary tumor, into other organs or tissues of the body. The conventional diagnostic procedure, which relies on a manual detection is very challenging and time consuming for pathologists as they must review multiple scans of lymph node sections before any conclusion. This technique also involves human error.\nWith the emergence of digital pathology, it is possible today to utilize convolutional neural networks (CNN) to detect metastatic cancer and classify histopathology images with a high accuracy. In this project, we created an algorithm to identify metastatic cancer in small image patches taken from larger digital pathology scans. Our data set contains 220,025 training images and 57,456 test images. The images are in the tif format and are 96x96 pixels, color. This is a binary classification problem. The images are labeled as 0 or 1, where 0=No Cancer Tissue and 1 = Has Cancer Tissue(s). To evaluate our classification model's performance, the Area Under the Curve (AUC) was used.","metadata":{}},{"cell_type":"markdown","source":"","metadata":{}},{"cell_type":"markdown","source":"# **Import Packages**","metadata":{}},{"cell_type":"code","source":"import numpy as np\nimport pandas as pd\nimport matplotlib.pyplot as plt\nimport matplotlib.image as mpimg\nimport os","metadata":{"execution":{"iopub.status.busy":"2021-12-12T22:20:44.543888Z","iopub.execute_input":"2021-12-12T22:20:44.544663Z","iopub.status.idle":"2021-12-12T22:20:44.571512Z","shell.execute_reply.started":"2021-12-12T22:20:44.544544Z","shell.execute_reply":"2021-12-12T22:20:44.570866Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# **Exploratory Data Analysis**\nIn this section, we start looking at some basic statistics like the amount of images, the label distribution, and the image size. ","metadata":{}},{"cell_type":"markdown","source":"## **Number of images in the training and test set**","metadata":{}},{"cell_type":"markdown","source":"","metadata":{}},{"cell_type":"code","source":"# Load the training data into a DataFrame named 'train'. \n# Print the shape of the resulting DataFrame. \n\ntrain = pd.read_csv(f'../input/histopathologic-cancer-detection/train_labels.csv', dtype=str)\ntest = pd.read_csv(f'../input/histopathologic-cancer-detection/sample_submission.csv', dtype=str)\n\nprint('Training Set Size:', train.shape)\nprint('Testing Set Size:', test.shape)","metadata":{"execution":{"iopub.status.busy":"2021-12-12T22:20:44.573278Z","iopub.execute_input":"2021-12-12T22:20:44.573547Z","iopub.status.idle":"2021-12-12T22:20:45.381079Z","shell.execute_reply.started":"2021-12-12T22:20:44.573507Z","shell.execute_reply":"2021-12-12T22:20:45.378380Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"**There are 220,025 training images and 57,458 test images in the dataset.**","metadata":{}},{"cell_type":"markdown","source":"Lets update the dataset to include filename extensions.","metadata":{}},{"cell_type":"code","source":"train['id'] = train['id'].apply(lambda x: f'{x}.tif')","metadata":{"execution":{"iopub.status.busy":"2021-12-12T22:20:45.382403Z","iopub.execute_input":"2021-12-12T22:20:45.382710Z","iopub.status.idle":"2021-12-12T22:20:45.487675Z","shell.execute_reply.started":"2021-12-12T22:20:45.382666Z","shell.execute_reply":"2021-12-12T22:20:45.486738Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"test['id'] = test['id'].apply(lambda x: f'{x}.tif')","metadata":{"execution":{"iopub.status.busy":"2021-12-12T22:20:45.489110Z","iopub.execute_input":"2021-12-12T22:20:45.489335Z","iopub.status.idle":"2021-12-12T22:20:45.519384Z","shell.execute_reply.started":"2021-12-12T22:20:45.489308Z","shell.execute_reply":"2021-12-12T22:20:45.518509Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## **Label Distribution**","metadata":{}},{"cell_type":"code","source":"(train.label.value_counts() / len(train)).to_frame().sort_index().T","metadata":{"execution":{"iopub.status.busy":"2021-12-12T22:20:45.521210Z","iopub.execute_input":"2021-12-12T22:20:45.521423Z","iopub.status.idle":"2021-12-12T22:20:45.573173Z","shell.execute_reply.started":"2021-12-12T22:20:45.521397Z","shell.execute_reply":"2021-12-12T22:20:45.572175Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"labels_count = train.label.value_counts()\n\nplt.pie(labels_count, labels=['No Cancer', 'Cancer'], startangle=180, \n        autopct='%1.1f', colors=['#00ff99','#FF96A7'], shadow=True)\nplt.figure(figsize=(16,16))\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2021-12-12T22:20:45.574538Z","iopub.execute_input":"2021-12-12T22:20:45.574868Z","iopub.status.idle":"2021-12-12T22:20:45.742359Z","shell.execute_reply.started":"2021-12-12T22:20:45.574825Z","shell.execute_reply":"2021-12-12T22:20:45.741690Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"#Histogram label distribution\nimport seaborn as sns\nsns.countplot(train.label, edgecolor = 'black',\n              palette = sns.color_palette())\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2021-12-12T22:20:45.743423Z","iopub.execute_input":"2021-12-12T22:20:45.743982Z","iopub.status.idle":"2021-12-12T22:20:47.148543Z","shell.execute_reply.started":"2021-12-12T22:20:45.743946Z","shell.execute_reply":"2021-12-12T22:20:47.147011Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"The first diagram shows that 40.5% of images in the dataset belong to cancerous class and 59.5% of images are Non-cancerous.\nAnd the second diagram displays the number of observations belonging to class 0 (Non_cancerous) and to class 1 (cancerous).","metadata":{}},{"cell_type":"markdown","source":"## **Images Size**","metadata":{}},{"cell_type":"code","source":"print('Training Images:', len(os.listdir('../input/histopathologic-cancer-detection/train/')))\n\nfor i in range(10):\n  img = plt.imread('../input/histopathologic-cancer-detection/train/' + train.id[i])\n  print('Images shape', img.shape)","metadata":{"execution":{"iopub.status.busy":"2021-12-12T22:20:47.150020Z","iopub.execute_input":"2021-12-12T22:20:47.150372Z","iopub.status.idle":"2021-12-12T22:20:51.059917Z","shell.execute_reply.started":"2021-12-12T22:20:47.150327Z","shell.execute_reply":"2021-12-12T22:20:51.059017Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print('Test Images:', len(os.listdir('../input/histopathologic-cancer-detection/test/')))\n\nfor i in range(10):\n  img = plt.imread('../input/histopathologic-cancer-detection/test/' + test.id[i])\n  print('Images shape', img.shape)","metadata":{"execution":{"iopub.status.busy":"2021-12-12T22:20:51.060964Z","iopub.execute_input":"2021-12-12T22:20:51.061176Z","iopub.status.idle":"2021-12-12T22:20:52.202064Z","shell.execute_reply.started":"2021-12-12T22:20:51.061149Z","shell.execute_reply":"2021-12-12T22:20:52.201285Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"As we can see, the images are 96x96 pixels and are full color. Each pixel is made up of three channels (R,G,B), with each channel representing a color.","metadata":{}},{"cell_type":"markdown","source":"## **Visualization**","metadata":{}},{"cell_type":"markdown","source":"**Here, we display the following items in this visualization notebook: [CANCER_DETECTION_VISUALIZATION_NOTEBOOK](https://www.kaggle.com/jennaward6/jw-cancerdetection-v01-with-visualizations).**\n- Sample images of original dataset (96x96)\n- Sample images for cropped dataset (32x32)\n- Example of images for Class = 1   \n- Example of images for Class = 0\n- Look at some augmented images\n- Visualize Filters\n- Visualize Feature Maps\n- Cancerous Feature Map\n- Non Cancerous Feature Map\n- Class Activation Map\n- Heatmap Function\n- Distribution of Pixel Channels","metadata":{}},{"cell_type":"markdown","source":"","metadata":{}},{"cell_type":"markdown","source":"# **Technical Challenges**","metadata":{}},{"cell_type":"markdown","source":"For most of us, it was the first time working on the kaggle platform. So, it took us some time to familiarize ourselves and learn how to navigate the platform.\nThe second challenge was that, it was not possible to work or edit the same notebook on kaggle at the same time with our collaborator(s), otherwise you get an error when you try to open the notebook. \nOne of the biggest challenges we had was related to the management of the GPU hours. Kaggle provides a limited amount of hours for each user every week. As a result, some of us ran out of GPU hours since most of our models could take up to 6 hours to train depending on the number of epochs selected. ","metadata":{}},{"cell_type":"markdown","source":"# **Model Architecture Summary**","metadata":{}},{"cell_type":"markdown","source":"","metadata":{}},{"cell_type":"markdown","source":"\nIt was challenging to find the best model for this problem. We started our journey by creating very basic models. The idea was to find a model that will be used as a baseline for more complex models. At first, we explored the 96x96 size images from Kaggle and the pre-cropped 32x32 images dataset provided by the instructor. It turned out that, the model created with the pre-cropped 32x32 images yielded a lower accuracy, while the model created with the full size images led to an acceptable result (~75-80% accuracy). Later on, by simply twisting the CNN architecture of the model that uses the original 96x96 size images, we were able to significantly increase the AUC from 0.8 to 0.9885. In the subsequent model, we used the original 96x96 images and cropped them down to the middle 32x32 using Cropping2D Layer. This technique yielded an AUC value of 0.9457. It was noticed that the pre-cropped dataset was consuming noticably more time than the original dataset. In the next step of our journey, we explored more advanced techniques such as image augmentation, transfer learning, and ensemble learning. Image augmentation is a technique that consists of artificially creating new training data by applying image transformation, such as scaling, rotations, and translations to the original training images. We applied this technique on the original 96x96 images and this gave us an AUC score of 0.9589 on the validation dataset. In the subsequent model, we applied transfer learning on the same original 96x96 size images. In transfer learning, a pre-trained model developed for a task is re-used as the starting block for a model on a second task. In our case, we used the pre-trained VGG16 model and the ResNet50 model. We achieved an AUC of 0.9815 and 0.9956 for VGG16 and ResNet50 respectively. Ensemble learning is a technique, where multiple models or CNN are combined to generate one optimal model. In our case, we constructed two models using this technique. First, we combined three models (simple model, VGG16, ResNet50) with logistic regression. In the second model, we combined the same three previously mentioned models, but this time with weighted average. According to our results, the second model with weighted average produced a better score (public score of 0.9537 and a private score of 0.9298) than the model with logistic regression. For our last model in this competition, we implemented PyTorch and the result was surprisingly better than all the models, where Keras was used. We got a public score of 0.9642 and a private score of 0.9263. To conclude, it is possible to create a good model that can detect metastatic cancer and classify histopathologic images without using advanced deep learning or machine learning techniques. However, using techniques like transfer learning (with VGG16) and ensemble learning can significantly improve the model's accuracy and performance.\n\nYou will find detailed information (such as the code, traning and validation learning curves, etc.) about our work/journey in the following notebook: **[TRAINING NOTEBOOK](https://www.kaggle.com/gauravsamudra/team-4-training-overview-notebook).**\n\nSome of our best models are summarized in the table below.\n","metadata":{}},{"cell_type":"markdown","source":"| Model Architecture  \t| Image Size  | Epochs  \t| Private Score  \t| Public Score   | Notebook  | \n|---\t|---\t|---\t|---\t|---\t|---\t|\n| Simple Model  | 96x96  \t| 30  \t| [0.8288](https://www.kaggle.com/leopoldtchomgwi/lt-cancer-detection-v01-submission-revised)  \t| 0.8669  \t|[[LT] Cancer Detection Simple Model](https://www.kaggle.com/leopoldtchomgwi/lt-cancerdetection-v01-with-balanced-target-dist)   \t|\n| VGG16  \t| 96x96  | 60  \t| [0.9376](https://www.kaggle.com/lokanathpatro/lp-cancer-detection-models-submission?scriptVersionId=81777457)  \t| 0.9449  \t| [[LP] Cancer Detection VGG16 Model](https://www.kaggle.com/lokanathpatro/lp-cancer-detection-vgg16-model)  |\n| ResNet50  \t| 96x96  | 40  \t| [0.9029](https://www.kaggle.com/lokanathpatro/lp-cancer-detection-models-submission?scriptVersionId=81778571)  \t| 0.8798  \t| [[LP] Cancer Detection ResNet50 Model](https://www.kaggle.com/lokanathpatro/lp-cancer-detection-resnet50-model)  |\n| Ensemble VGG16, ResNet50, Simple Model <br/> with Logistic Regression | 96x96  | 40   | [0.8370](https://www.kaggle.com/lokanathpatro/lp-cancer-detection-ensemble-model-with-lr?scriptVersionId=81787013) | 0.8569  \t|  [[LP] Cancer Detection Ensemble Model with LR](https://www.kaggle.com/lokanathpatro/lp-cancer-detection-ensemble-model-with-lr)\t|\n| Ensemble VGG16, ResNet50, Simple Model <br/> with Weighted Average | 96x96  | 40  |  [0.9298](https://www.kaggle.com/lokanathpatro/lp-cancer-detection-models-submission?scriptVersionId=81797079) \t|  0.9537 \t|  [[LP] Cancer Detection Models Submission](https://www.kaggle.com/lokanathpatro/lp-cancer-detection-models-submission?scriptVersionId=81797079)  \t|\n| PyTorch  \t| 96x96  | 40  |  [0.9263](https://www.kaggle.com/lokanathpatro/lp-cancer-detection-cnn-model-with-pytorch?scriptVersionId=81689603) \t| 0.9642  \t| [[LP] Cancer Detection CNN Model with PyTorch]( https://www.kaggle.com/lokanathpatro/lp-cancer-detection-cnn-model-with-pytorch)  |\n| Cropping2D Layer Model  \t| 96x96  | 25  |  [0.8368](https://www.kaggle.com/gauravsamudra/gs-cancersubmission-v1?scriptVersionId=79504395) \t| 0.8958  \t| [[GS] Cancer Detection with Cropping2D Layer](https://www.kaggle.com/gauravsamudra/gs-cancerdetection-v01?scriptVersionId=79421784)  |\n| Model With Pre-Cropped Images  \t| 32x32  | 25  |  [0.7224](https://www.kaggle.com/leopoldtchomgwi/lt-cancer-detection-cropped-images-submission?scriptVersionId=79658223) \t| 0.7813  \t| [[LT] Cancer Detection with Pre-Cropped Images](https://www.kaggle.com/leopoldtchomgwi/lt-cancer-detection-cropped-im)  |\n| Model With Augmented Images  \t| 96x96  | 25  |  [0.8585](https://www.kaggle.com/jennaward6/team-4-cancer-detection-submit-jw-edit?scriptVersionId=80319158) \t| 0.9124  \t| [[JW] Cancer Detection with Augmented Images](https://www.kaggle.com/jennaward6/jw-cancerdetection-v01-with-visualizations?scriptVersionId=80269876)  |\n| Simple Model2  \t| 96x96  | 20  |  [0.9231](https://www.kaggle.com/leopoldtchomgwi/cancerdetection-submission-final2?scriptVersionId=81039203) \t| 0.9397  \t| [[LT] Cancer Detection_Simple Model](https://colab.research.google.com/drive/1_PvSVhBlYdwXEAqbb7a8RABTXgcU14ME?usp=sharing#scrollTo=yYkbG_C6KJ1Y)  |\n\n","metadata":{}},{"cell_type":"markdown","source":"# **Final Model Evaluation**","metadata":{}},{"cell_type":"markdown","source":"**[Final_Model_Evaluation_Notebook](https://www.kaggle.com/lokanathpatro/team-4-final-model-evaluation-notebook)**.","metadata":{}},{"cell_type":"markdown","source":"Transfer learning with VGG16 was selected as our final model. With this model, we achived an accuracy of 94.25% and got an AUC score of 0.9815 on the validation set. The difference between the public score (0.9449) and the private score (0.9376) was only 0.0073. Furthermore, when you see the first Confusion Matrix (see diagram below), Accuracy is about 94% but the False benign rate is around 3.45% which is two times higher than False Malignant. When a trusted model predicts Benign the case doesn't go for further review and hence False Benign rate is a very important parameter of trust after Accuracy. And our final model is NOT upto the mark.\n\nFor our submissions, anything more than 50% probability is predicted a Malignant and less than 50% is predicted and Benign.\n\nThe sceond confusion matrix shows that when we add a 10 basis point positive bias to our probabilities and then derive the predictions The accuracy stays same but the False Benign percentage drops significantly. If we stretch further and add 20 basis points Accuracy goes down along with False benign percentage.\n\n\n![image.png](attachment:9c6fc180-1dd6-495c-b82c-1cc405d73110.png)\n\n**Classification Report**\n\nThe report also indicates our model is correct 95% of the time, when it predicts a tumor is benign and 94% of the time, when it predicts a tumor is malignant.\n\nOur model also correctly identifies 96% of benign tissues and 92% of malignant tissues.\n\n![image.png](attachment:fd8dca3d-51ac-4202-b5b2-bb247b523c2d.png)","metadata":{},"attachments":{"fd8dca3d-51ac-4202-b5b2-bb247b523c2d.png":{"image/png":"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"},"9c6fc180-1dd6-495c-b82c-1cc405d73110.png":{"image/png":"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"}}},{"cell_type":"markdown","source":"The entire code and results of our final model can be found here: **[Final_Model_Evaluation_Notebook](https://www.kaggle.com/lokanathpatro/team-4-final-model-evaluation-notebook)**","metadata":{}},{"cell_type":"markdown","source":"","metadata":{}},{"cell_type":"markdown","source":"# **Conclusion**","metadata":{}},{"cell_type":"markdown","source":"At the end of our work, we were able to create several convolutional neural network models or algorithms that can be used to detect metastatic cancer in histopathological image scans. We used the pre-trained VGG16 model to build our best model. We achieved an accuracy of 94% and an AUC score of 0.98 with this model.  \nOn this Kaggle competition, our model achieved a private score of 0.9316 and a public score of 0.9449 on test data. We may obtain improved results by altering the network design/architecture and parameters.","metadata":{}}]}