Artificial intelligence model predicts treatment response in ovarian cancer
Click Here to Manage Email Alerts
A trained artificial intelligence model predicted treatment outcomes before surgery of women with high-grade serous ovarian cancer, according to study results.
Researchers presented the findings of the pilot study in a plenary session during the Society of Gynecologic Oncology 2022 Annual Meeting on Women’s Cancer.
Methods
The pilot study included 435 still-frame images of the diaphragm, omentum, peritoneum and pelvis from pretreatment laparoscopic surgical videos of 113 women with high-grade serous ovarian cancer. Researchers trained an artificial intelligence (AI) model to use the images to detect morphologic disease patterns and predict outcomes in two predefined populations of women, those with an excellent response (PFS of 12 months or longer) or poor response (PFS of 6 months or less) to standard treatment. They grouped images into three categories: training (70%), validation (10%) and testing (20%).
Key findings
Results showed the model identified the 53% of women who experienced an excellent response to treatment, resulting in an accuracy of 93% and a sensitivity of 100%. However, the model misclassified some patients with a poor response to treatment, resulting in a specificity of 63%.
“The observed lack of specificity may be related to the lower number of images corresponding to patients with partial response to therapy (n = 167) compared with patients with excellent response (n = 268), as this impacts the development of neural networks utilized for machine learning in AI,” Deanna Glassman, MD, researcher at The University of Texas MD Anderson Cancer Center, and colleagues wrote.
Implications
“This pilot study is an exciting frontier in surgical innovation that shows how we can use machine learning to enhance our clinical approach to treating patients with gynecologic cancers,” Glassman said in a press release.
“A major implication of our study is that the AI model could identify patients who are likely to have a poor response to traditional therapies, enabling clinicians to alter surgical plans and goals, and providing opportunities for tailoring therapeutic strategies in those patients,” Glassman added. “The concept of using an AI model trained on laparoscopic images requires additional validation studies, but in the future it could be extended to other gynecologic cancers to identify patterns of disease, predict treatment outcomes, and distinguish between viable and necrosed malignant tissue at the time of interval debulking surgery.”
Collapse
Glassman D, et al. Abstract 9. Presented at: Society of Gynecologic Oncology 2022 Annual Meeting on Women’s Cancer (hybrid meeting); March 18-21, 2022; Phoenix.
Click Here to Manage Email Alerts