New artificial intelligence technology predicts effectiveness of breast cancer chemotherapy

University of Waterloo engineers have developed artificial intelligence (AI) technology to predict whether women have breast cancer. (CREDIT: Creative Commons)

Breast cancer is the second most common type of cancer in women in Canada and the United States, accounting for more than 25% of all new cancers in women. Thus, tremendous research has been done and progress has been made in improving screening methods and processes to proactively detect the presence of breast cancer in individuals at risk. However, more than 48,000 Canadian and American women are projected to die from breast cancer in 2022.

Although the exact path of treatment for patients with breast cancer largely depends on the stage of the breast cancer, surgery is usually performed to prevent further development of breast cancer and to remove cancerous tissue. However, some non-metastatic breast cancers are inoperable.

Recently, a type of treatment called neoadjuvant chemotherapy is being used more and more because it can shrink a large tumor before surgery (so that the tumor becomes resectable). However, neoadjuvant chemotherapy is expensive, time consuming, and can expose patients to radiation as well as other significant side effects such as reduced fertility.

The current process for recommending neoadjuvant chemotherapy is based on the expert but human judgment of the medical oncologist and/or radiation oncologist as to whether the patient will live longer and benefit from the treatment. With potential biases and high uncertainty in clinical judgment, there is a possibility that some erroneous recommendations will result in some patients later developing a preventable harmful advanced cancer or being exposed to unnecessary radiation.

Similar stories

In response, engineers at the University of Waterloo have developed artificial intelligence (AI) technology to predict whether women with breast cancer will benefit from chemotherapy before surgery.

A new AI algorithm, part of the open-source Cancer-Net initiative led by Dr. Alexander Wong, could help ineligible candidates avoid serious chemotherapy side effects and pave the way for better surgical outcomes for those who qualify.

“Determining the right treatment for a particular breast cancer patient is now very difficult, and it is very important to avoid unnecessary side effects from using treatments that are unlikely to bring real benefit to that patient,” said Wong, professor of systems engineering. .

Radiation therapy for breast cancer. (CREDIT: Creative Commons)

“An AI system that can help predict whether a patient will respond well to a given treatment gives clinicians the tool they need to prescribe the best personalized treatment for a patient to improve recovery and survival.”

Led by Amy Tai, a PhD student at the Vision and Image Processing (VIP) Lab, the AI ​​software was trained on breast cancer images taken with a new magnetic resonance imaging method invented by Wong and his team called synthetic. correlated diffusion imaging (CDI).

Sample slice illustrating visual differences between ADC, CDI, DWI, and T2w prior to neoadjuvant chemotherapy for a patient undergoing pCR. In this case, pCR prediction was correct for CDI and DWI (b = 800), but not for other modalities. (CREDIT: Cancer-Net BCa)

With knowledge gained from CDI images of old breast cancer cases and information about their outcomes, AI can predict whether preoperative chemotherapy will benefit new patients based on their CDI images.

Preoperative treatment, known as neoadjuvant chemotherapy, can reduce the size of a tumor to make surgery possible or easier, as well as reduce the need for major surgeries such as mastectomy.

Clinical support workflow for pCR prediction using Synthetic Correlated Diffusion Imaging (CDI) Volumetric Depth Radiomology Characteristics. (CREDIT: Cancer-Net BCa)

“I’m pretty optimistic about this technology because deep learning AI can see and discover patterns that are related to whether a patient will benefit from a given treatment,” said Wong, director of the VIP Lab and Canadian Research Center. in artificial intelligence and medical imaging.

Recently at Med-NeurIPS at NeurIPS 2022, a major international conference on artificial intelligence, a paper was presented on the Cancer-Net BCa project: Complete Pathological Response Prediction in Breast Cancer Using Synthetic Correlated Diffusion Imaging Volumetric Deep Radiomological Characteristics.

A new AI algorithm and a complete set of breast cancer CDI imaging data have been made publicly available as part of the Cancer-Net initiative so that other researchers can help advance the field.

For more science news, visit our New Innovations section at The bright side of the news.

Note: Materials provided above by the University of Waterloo. Content can be edited for style and length.

Do you like such pleasant stories? Get Brighter Side of News Newsletter.