October 24, 2024 |

Breast Cancer Survivor Pioneers AI Breakthroughs in Oncology

By Emily Blackburne

AI technology is transforming oncology and stands at the forefront of this revolution. Now an Assistant Professor at the CUNYGraduate Center and New York City College of Technology (City Tech), as well as the founder of the , her pioneering research in AI-driven oncology began in 2017 during a postdoctoral project at Weill Cornell Medical College. There, she developed an AI model to differentiate ���while simultaneously battling breast cancer herself.

Khosravi moved from Iran to New York City in 2017 to pursue her postdoctoral studies, but shortly after arriving, a routine health check revealed she had breast cancer. Despite undergoing two lumpectomies, the cancer recurred, leading to the decision for a double mastectomy. Over the next two years, she underwent extensive radiotherapy and surgeries to treat the disease.

Her personal battle was compounded by the challenges of maintaining her immigration status and work visa. Despite the intensity of her treatments, Khosravi remained dedicated to her research. ���At Weill Cornell, I���d receive treatment in one building and immediately continue my research in the next,��� she recalls. She vividly remembers walking long hospital corridors, attached to an IV drip, to reach her laptop. During this period, she authored two peer-reviewed publications on the application of AI in cancer diagnosis and treatment, using her research as a source of strength and perseverance.

Khosravi emphasizes the importance of AI in precision oncology, particularly for earlier and more accurate cancer detection. At Weill Cornell and later as an AI scientist at Memorial Sloan Kettering Cancer Center (MSKCC), she helped develop , an AI-driven model that integrates pathology and MRI data to improve prostate cancer diagnosis and treatment planning.

At CUNYCity Tech, Khosravi developed , a fully automated framework designed to be adaptable to various cancers, including breast cancer, where the combination of different data modalities is critical for accuracy. She also developed two deep learning models, , which offer improved computational efficiency and diagnostic accuracy over conventional medical imaging models. These models were tested on nine radiological datasets, including cases related to breast cancer and neoadjuvant therapy, and demonstrated robust performance in both accuracy and processing speed. These studies marked a significant improvement over existing methods, which are often computationally intensive and less accurate.

One of the key innovations Khosravi highlights is the use of AI for personalized treatment plans. ���AI models can handle multi-modality data integration, meaning they can combine medical imaging (such as MRI and CT scans), clinicopathologic data (such as tumor grade and hormone receptor status), and genomic information to provide personalized risk assessments and treatment predictions,��� Khosravi explains. For instance, AI can predict a patient���s response to chemotherapy, assess recurrence risk, or suggest optimal surgery options by analyzing this diverse set of data. This approach enables earlier detection, more accurate prognoses, and individualized preventative strategies���particularly for populations that may have been underrepresented in traditional clinical datasets.

Khosravi adds that AI can surpass human diagnostic capabilities in some areas. ���AI can detect patterns in data that are invisible to human eyes,��� she says. While treatment decisions ultimately rest with doctors and patients, AI assists in improving the precision of those decisions, minimizing unnecessary interventions, and reducing healthcare costs. This is particularly beneficial for underserved populations, such as immigrants and minority groups, who may lack access to comprehensive healthcare or clinical trials that cater to their specific demographics. By offering data-driven, individualized treatment plans, AI has the potential to bridge these gaps and address disparities in cancer care.

Khosravi���s dedication to cancer research is deeply personal. Her interest in the medical field began after her father passed away from cancer, and her own experience with breast cancer has only strengthened her resolve. ���I live for my family and the joy of life, and I���m driven by a deep desire to find solutions that not only advance science but also ease the pain and suffering of other patients, so they can live more fully, too,��� she shares. Supported by her husband during her battle, Khosravi is now using her personal and scientific insights to make lasting contributions to oncology.

While cancer continues to be a formidable challenge, Khosravi believes that AI offers a transformative path forward, one that could significantly alleviate patient suffering and improve long-term health outcomes for both current patients and those at risk.