By Rylan DeStefano
An artificial intelligence model called IRON (Integrated Radiogenomics for Ovarian Neoadjuvant therapy) shows an 80% accuracy rate in the prediction of therapy outcomes for patients with ovarian cancer. The artificial intelligence model focuses specifically on the volumetric reduction of tumor lesions within this patient population, according to a press release.
This tool surpassed the efficiency of present clinical methods, as IRON uses an approach that focuses on a patient’s liquid biopsy (cancer-specific information that can be observed via blood tests), overall health characteristics such as age, health status, tumor markers and disease images that were captured during CT scan images.
A recent research study, which was featured in Nature Communications, focused on 134 patients who had been diagnosed with high-grade ovarian cancer. Dr. Evis Sala, chair of Diagnostic Imaging and Radiotherapy at the Faculty of Medicine and Surgery of the Catholic University and Director of the Advanced Radiology Center at the Policlinico Universitario A. Gemelli IRCCS ran the study and the AI model was created by Professor Sala’s team at the University of Cambridge.
“We compiled two independent datasets with a total of 134 patients (92 cases in the first dataset, 42 in the second independent test set),” Sala and Dr. Mireia Crispin Ortuzar from Cambridge said in the press release.
When it comes to a precise accuracy rate in high-grade ovarian carcinoma, therapy response predictions result to about 50% accuracy. As there were a few significant biomarkers used for this type of cancer, the IRON model was able to predict chemotherapy responders more accurately.
Notably, this is not the first study showing that artificial intelligence has the potential to predict cancer outcomes. Earlier this year, research found that artificial intelligence may help determine which patients with prostate cancer were more likely to benefit from hormone therapy plus radiation. Additionally, another group of researchers also showed that artificially intelligence may play a role in diagnosing sarcopenia in patients with head and neck cancer.
Within the study, demographic information, treatment details, blood biomarkers (CA-125, which could indicate the growth of ovarian cancer) and circulating tumor DNA (ctDNA, which is bits of cancer DNA that can be observed on a blood test) were collected from patients. CT scans were also found and characteristics of the tumor were obtained throughout the CT scans.
Where disease spread in the initial stage was investigated within omental and pelvic/ovarian regions. Omental deposits showed more of a response compared to pelvic disease when it came to neoadjuvant (presurgical) therapy. Before and after therapy responses, tumor mutations and CA-125 had correlation to the overall disease burden, according to the press release.
Analyzation of CT scans showed that six patient subgroups, classified by biological and clinical characteristics became indicators of therapy responses. The efficiency of the model was shown throughout the independent patient sample shown within the study.
“From a clinical perspective, the proposed framework addresses the unmet need to early identify patients unlikely to respond to neoadjuvant therapy and may be directed to immediate surgical intervention,” Sala emphasized. “The tool could be applied to stratify the risk of each individual patient in future clinical research…”
This article was published by Cure.com