NIH Researchers Unveil AI Tool to Predict Oncology Patients’ Response to Immunotherapy
Scientists at the National Institutes of Health have designed an artificial intelligence (AI) tool that can predict if a patient’s cancer will respond to immune checkpoint inhibitors using clinical data. This artificial intelligence model may assist physicians in determining whether immunotherapy medications are effective in the treatment of different cancers.
To determine the effectiveness of their machine-learning model, the researchers carried out a proof-of-concept study led by scientists from Memorial Sloan Kettering Cancer Center and the National Cancer Institute’s Center for Cancer Research. Their model was built and assessed using data from independent data sets, including more than 2,800 patients with 18 different types of solid tumors who were treated using immune checkpoint inhibitors.
Currently, the FDA has approved a pair of predictive biomarkers to help identify patients who could be candidates for immune checkpoint inhibitor treatment. This is a type of immunotherapy medication that assists immune cells in eliminating cancer cells.
The predictive biomarkers include PD-L1, a tumor cell protein that limits immune response and tumor mutational burden, which is the number of mutations in cancer cell DNA. However, these biomarkers, despite being revolutionary, don’t always give accurate predictions of responses to immune checkpoint inhibitors.
This new tool was designed to give forecasts based on clinical features collected from patients, including systemic therapy history, type of cancer, levels of blood albumin, patient age and blood neutrophil-to-lymphocyte ratio, which is an inflammation marker. In addition, the tool considers tumor mutational burden, which is evaluated using sequencing panels.
The scientists determined that the model could accurately forecast the likelihood that a patient would respond to an immune checkpoint inhibitor as well as how long they would live before the illness came back as well as overall.
In their report, the scientists stated that the tool could also identify patients with low tumor mutational burden who could gain benefit from treatment with immunotherapy. They then noted that bigger prospective studies were required to further assess the machine-learning model in clinical settings.
The study was spearheaded by Yingying Cao and Tian-Gen Chang of Dr. Ruppin’s group. It was coled by Dr. Luc G. T. Morris of Memorial Sloan Kettering Cancer Center and Dr. Eytan Ruppin of the Center for Cancer Research. Other researchers include Saugato Rahman Dhruba, Cristina Valero, Hannah J. Sfreddo, Diego Chowell, Seong-Keun Yoo, and Se-Hoon Lee. The study’s findings were reported in “Nature Cancer.”
The AI model, called Logistic Regression-Based Immunotherapy-Response Score (LORIS), is available for use at loris.ccr.cancer.gov.
With such screening platforms becoming available, it is bound to become easier to determine which patients could benefit from the immune therapies commercialized by various companies such as Scinai Immuntherapeutics Ltd. (NADAQ: SCNI), thereby saving vital time and money.
NOTE TO INVESTORS: The latest news and updates relating to Scinai Immunotherapeutics Ltd. (NASDAQ: SCNI) are available in the company’s newsroom at https://ibn.fm/SCNI
Please see full terms of use and disclaimers on the BioMedWire website applicable to all content provided by BMW, wherever published or re-published: http://BMW.fm/Disclaimer
Scientists at the National Institutes of Health have designed an artificial intelligence (AI) tool that can predict if a patient’s cancer will respond to immune checkpoint inhibitors using clinical data. This artificial intelligence model may assist physicians in determining whether immunotherapy medications are effective in the treatment of different cancers.
To determine the effectiveness of their machine-learning model, the researchers carried out a proof-of-concept study led by scientists from Memorial Sloan Kettering Cancer Center and the National Cancer Institute’s Center for Cancer Research. Their model was built and assessed using data from independent data sets, including more than 2,800 patients with 18 different types of solid tumors who were treated using immune checkpoint inhibitors.
Currently, the FDA has approved a pair of predictive biomarkers to help identify patients who could be candidates for immune checkpoint inhibitor treatment. This is a type of immunotherapy medication that assists immune cells in eliminating cancer cells.
The predictive biomarkers include PD-L1, a tumor cell protein that limits immune response and tumor mutational burden, which is the number of mutations in cancer cell DNA. However, these biomarkers, despite being revolutionary, don’t always give accurate predictions of responses to immune checkpoint inhibitors.
This new tool was designed to give forecasts based on clinical features collected from patients, including systemic therapy history, type of cancer, levels of blood albumin, patient age and blood neutrophil-to-lymphocyte ratio, which is an inflammation marker. In addition, the tool considers tumor mutational burden, which is evaluated using sequencing panels.
The scientists determined that the model could accurately forecast the likelihood that a patient would respond to an immune checkpoint inhibitor as well as how long they would live before the illness came back as well as overall.
In their report, the scientists stated that the tool could also identify patients with low tumor mutational burden who could gain benefit from treatment with immunotherapy. They then noted that bigger prospective studies were required to further assess the machine-learning model in clinical settings.
The study was spearheaded by Yingying Cao and Tian-Gen Chang of Dr. Ruppin’s group. It was coled by Dr. Luc G. T. Morris of Memorial Sloan Kettering Cancer Center and Dr. Eytan Ruppin of the Center for Cancer Research. Other researchers include Saugato Rahman Dhruba, Cristina Valero, Hannah J. Sfreddo, Diego Chowell, Seong-Keun Yoo, and Se-Hoon Lee. The study’s findings were reported in “Nature Cancer.”
The AI model, called Logistic Regression-Based Immunotherapy-Response Score (LORIS), is available for use at loris.ccr.cancer.gov.
With such screening platforms becoming available, it is bound to become easier to determine which patients could benefit from the immune therapies commercialized by various companies such as Scinai Immuntherapeutics Ltd. (NADAQ: SCNI), thereby saving vital time and money.
NOTE TO INVESTORS: The latest news and updates relating to Scinai Immunotherapeutics Ltd. (NASDAQ: SCNI) are available in the company’s newsroom at https://ibn.fm/SCNI
Please see full terms of use and disclaimers on the BioMedWire website applicable to all content provided by BMW, wherever published or re-published: http://BMW.fm/Disclaimer
