Project Title: Histopathological image analysis for Breast Cancer prediction
Project Description:
Breast cancer is the second leading cause of cancer-related deaths among women in Nepal. Every year, more than 2000 new cases of breast cancer are diagnosed in the country (GLOBOCAN 2018 estimates). Majority of these women are under 50 years of age which severely impacts the economically active population as well as places a substantial burden on the Nepalese healthcare system.
Early diagnosis of breast cancer can increase the chance of successful treatment and survival. The diagnostic procedure often involves manual visual inspection of microscopic histopathology image section of suspected cancerous tissue biopsy obtained from the patient. This is a very tedious and time-consuming step, the accuracy of which largely depends on the expertise and experience of highly qualified pathologist. However, the diagnostic interpretation often varies between one pathologist to the other, especially in borderline or complicated cases. This in turn may have a great effect in the patient’s treatment decision. Therefore, an automated histopathology image-based breast cancer diagnostic system is required to aid the pathologist speed up the diagnosis and further improve the quality of diagnosis.
To address the challenges above, we propose (1) to develop a machine learning based histopathology image classification model to efficiently and accurately predict breast cancer. For this, we will train the machine learning models with Hematoxylin-eosin (H&E) histopathology images from biopsy confirmed breast cancer cases. (2) Furthermore, we will develop histopathology image classification model to classify different subtypes of breast cancer. For this, we will utilize Immunohistochemical (IHC) stained images generated using different diagnostic protein markers commonly used in breast cancer diagnosis.
We believe, with the use of artificial intelligence, the pathologist will be able to make cancer diagnosis more accurate, efficient, and consistent.
Background on Breast Cancer Screening and Diagnostics
Breast cancer is one of the most prevalent cancers among women as well as one of the leading cause of cancer-related deaths in women world wide including in Nepal. Breast cancer may occur in both women and men, though prevalence in men is rare.
Breast Cancer Screening
When a breast cancer is suspected, the patient is referred for a breast cancer screening tests. There are many different screening test available but the most popular techniques are
- physical examination of breast by a medical doctor or nurse to check for presence of lumps in breast,
- Mammogram (i.e. x-ray of breast) can help identify tumors that are small,
- Breast ultrasound is used to test if a new breast lump is a solid mass or a fluid-filled cyst, and
- MRI of breast may be used as a screening test for women who have a high risk of breast cancer. This is especially performed when any of the above screening test are inconclusive.
Recently, many genetic tests as also available to assess the risk of getting breast cancer in the patient. The most popular among them are tests to check for the presence of mutations in BRCA1 or BRCA2 gene. If any woman (or man) has a mutation in BRCA1 or BRCA2 gene inherited genetically from the parents, she (or he) is highly likely to develop breast cancer (or many other cancers as current research suggests) in her (or his) life time. If the patient's family members (blood related) has had a history of breast cancer in the past then this further elevates the risk of developing breast cancer in the patient. The very famous case of a Hollywood actress, Angelina Jolie, illustrates this case scenario.
Breast Cancer Diagnosis
Breast Tissue Biopsy
If any of the above screening test (Mammogram, Ultrasound, or MRI) suspects positive for breast cancer then the patient is further referred for a biopsy test. Here, the a very small section of the tissue sample (or sometimes fluid) is obtained from the cancer suspected area of the breast using different techniques. The removed tissue is examined under a microscope and further tested to check for the presence of breast cancer. A biopsy is the only definitive way to make a diagnosis of breast cancer.
There are three different ways to obtain a tissue biopsy sample which may directly effect the quality and quantity of the biopsy sample obtained for downstream pathological or computational analysis. In any computational analysis or machine learning studies, the knowledge of how the biopsy are obtained is essential for correct interpretation of the results. These three types of biopsy techniques are as follows:
- Fine Needle Aspiration (FNA and FNABx): In fine needle aspiration, a thin needle is inserted into an area of abnormal-appearing tissue or body fluid. This is often guided by ultrasound to pinpoint the suspected cancerous area of the breast. This technique results in very minute amount of tissue sample as compared to the other two biopsy techniques below.
- Core Needle Biopsy: Core needle biopsy is the procedure to remove a small amount of suspicious tissue from the breast with a larger “core” (meaning “hollow”) needle. This is often guided by ultrasound to pinpoint the suspected cancerous area of the breast. While performing this technique, the patient is typically under local anesthesia.
- Surgical Biopsy: This is also known as “wide local excision,” “wide local surgical biopsy,” “open biopsy,” or “lumpectomy”. Surgical biopsy is performed to remove the cancerous tissues lumps from the breast (i.e. only a portion of the breast is removed). While performing this technique, the patient is typically under local anesthesia. This is also a first treatment option for some women with early-stage breast cancer.
The biopsy tissue samples obtained by the above procedures are sent to the pathology team for further assessment of the tissue to confirm if the patient is breast cancer positive or negative. If confirmed, further treatment decision and treatment plan/route is determined on the basis of the results of this histopathological tests. Most of the histopathological images obtained for machine learning studies are either core needle biopsy or surgical biopsy.
Histopathology Tests
The biopsy tissue is quickly frozen to preserve and harden it. The frozen biopsy tissue is then fixed into a wax-like blocks known as Formalin-Fixed Paraffin-Embedded (FFPE) tissues. These FFPE tissue blocks are then cut into very small slices (few micro meter thick) and fixed into glass slides. Most cells are colorless and transparent, and therefore histological sections have to be stained (using colored dyes) in some way to make the cells visible. Different staining technique and protein markers are used as needed to be further examined under the microscope. At this stage, the stained tissue slides are also digitized using a specialized scanning machine (or using a specialized camera fitted on top of the microscope ).