The AI model uses histopathological images for accurate diagnosis (Representational Image)
New Delhi:
Breast Cancer accounts for 13.6 percent of all cancer cases (male and female) in India, according to the 2022 World Cancer Report published by IARC (International Agency for Research on Cancer). Amongst women, it goes up to 26 percent of all cancer cases. In the United States, breast cancer accounts for about 30% of all new cancer cases amongst women.
New Research shows Artificial Intelligence (AI) can help fight this menacing disease. Early and accurate diagnosis can be pivotal for treatment amongst patients, and a newly developed AI system promises to do so with near perfect diagnosis.
A research paper titled ” Ensemble Deep Learning-Based Image Classification for Breast Cancer Subtype and Invasiveness Diagnosis from Whole Slide Image Histopathology” published in the Cancers Journal last month, details out an AI Model that classifies and identifies different types of breast cancer present in a patient, in addition to ruling out malignancy (cancer) in the first place by identifying benign tumors.
The study – done by researchers of Northeastern University, Boston along with Maine Health Institute for Research – has developed an AI model that analyses high resolution histopathological (tissue-level microscopic) whole slide images of breast tumor tissue.
The AI system, which outperforms earlier machine learning (ML) models in the domain by combining the predictions of other ML models, is capable of identifying and classifying the tumor into malignant (cancerous) or benign (non-cancerous) using historical data fed to the model during training.
It was trained on publicly available datasets called BreakHis (Breast Cancer Histopathological Database) and BACH (Breast Cancer Histopathology images). For BACH, microscopic breast tissue images were meticulously labelled by medical experts, categorising the images into four classes – Normal, Benign, In Situ Carcinoma and Invasive Carcinoma.
Exemplary microscopy images demonstrating the four classes in the BACH dataset (Image Source: Cancers 2024, 16(12), 2222)
And for BreakHis, which consists of 9,109 microscopic images of breast tumor tissue, it was used to categorise benign and malignant tumors further into 4 subclasses each- malignant tumors into Ductal carcinoma, Lobular carcinoma, Mucinous carcinoma and Papillary carcinoma, and benign tumors into Adenosi, Fibroadenoma, Phyllodes tumor, and Tubular adenoma.
Representative microscopy images of malignant and benign breast tissues from the BreakHis dataset (Image Source: Cancers 2024, 16(12), 2222)
Put together, the ensemble ML model has an accuracy of 99.84 percent. Such a performance metric during the research and development stage shows optimistic promise for the real-world application of the technology.
“The AI can’t miss a tumor in the biopsy and won’t be exhausted after diagnosing 10 or 20 people,” said Saeed Amal to Northeastern Global News. Amal is a professor of bioengineering at Northeastern university and is leading the ensemble model project.
Apart from diagnosis, AI systems have also made progress in prognosis and predictions related to breast cancers. For example, AI can now predict the neoadjuvant chemotherapy (NAC) response of breast cancer using Hematoxylin and eosin (common stains in tissue imaging) images of pre-chemotherapy needle biopsies. The AI systems responsible for the same have an accuracy of 95.15 percent and have been detailed out in a paper titled “Development of multiple AI pipelines that predict neoadjuvant chemotherapy response of breast cancer using H&E-stained tissues,” published in May 2023 in the Journal of Pathology.
Apart from this, AI has also made significant progress in identifying lymph node metastasis (spreading of cancer cells through lymphatic nodes) and evaluation of hormonal status which is important for breast cancer treatment. These and many more advances made by AI interventions over the years in the fight against breast cancer have been stated in a review paper published in Diagnostic Pathology in February.
