Breakthrough AI Model for Breast Cancer Diagnosis

Breast cancer is a menacing disease that accounts for a significant percentage of all cancer cases. In India, it comprises 26 percent of all cancer cases among women, while in the United States, it accounts for about 30 percent of all new cancer cases in women. However, a recent breakthrough in artificial intelligence (AI) has given hope in the fight against breast cancer.

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 unveils an AI model that can accurately classify and identify different types of breast cancer. This model also has the capability to rule out malignancy by identifying benign tumors.

The study was conducted by a team of researchers from Northeastern University, Boston, along with the Maine Health Institute for Research. They developed an AI model that analyzes high-resolution histopathological whole slide images of breast tumor tissue. What sets this AI system apart from earlier machine learning models is its ability to combine the predictions of other models, resulting in a more accurate diagnosis.

To train the AI model, the researchers used publicly available datasets called BreakHis (Breast Cancer Histopathological Database) and BACH (Breast Cancer Histopathology images). The BACH dataset consists of meticulously labeled microscopic breast tissue images categorized into four classes: Normal, Benign, In Situ Carcinoma, and Invasive Carcinoma. The BreakHis dataset, on the other hand, contains 9,109 microscopic images of breast tumor tissue, further categorized into four subclasses each for benign and malignant tumors.

The ensemble machine learning model achieved an impressive accuracy of 99.84 percent during the research and development stage. This high level of accuracy offers promising potential for real-world applications. Saeed Amal, a professor of bioengineering at Northeastern University and leader of the ensemble model project, highlights the advantages of this AI system. “The AI can’t miss a tumor in the biopsy and won’t be exhausted after diagnosing 10 or 20 people,” he said.

While accurate diagnosis is crucial for effective treatment, AI systems have also made progress in prognosis and predictions related to breast cancer. For example, AI can now predict the response to neoadjuvant chemotherapy (NAC) using images of pre-chemotherapy needle biopsies stained with Hematoxylin and eosin. The AI systems responsible for this prediction have an accuracy of 95.15 percent, as detailed 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.

Additionally, AI has shown significant advancements in identifying lymph node metastasis, which is the spreading of cancer cells through lymphatic nodes, and evaluating hormonal status, which is crucial for breast cancer treatment. These remarkable achievements, along with many others made by AI interventions over the years, have been documented in a review paper published in Diagnostic Pathology in February.

The breakthrough in AI’s role in breast cancer diagnosis and treatment holds immense promise for improving patient outcomes. With its ability to accurately classify tumors, predict treatment response, and evaluate disease progression, AI is poised to revolutionize the field of breast cancer research and clinical care. As technology continues to advance, we can expect even greater strides in the fight against this devastating disease.