AI-based fragmentomic approach could turn the tide for ovarian cancer
Ovarian cancer is often detected at a late stage, providing poor prognoses, while benign ovarian growths can lead to unnecessary surgeries. A new combined fragmentomic–protein biomarkers test could change all that.
An international collaboration of researchers led by Victor Velculescu (Johns Hopkins Kimmel Cancer Center, MA, USA) has integrated AI into a novel diagnostic test, which combines protein and DNA fragment detection, for ovarian cancer. The test could prove vital in catching ovarian cancer at an earlier stage, while also sparing women from invasive surgeries to confirm an ovarian cancer diagnosis.
Overall, 5-year survival rates for ovarian cancer are roughly 50%, a confronting figure that is challenging to reduce as most diagnoses occur during later stages of the disease, when mortality rates are significantly higher. Commenting on the challenge of diagnosing ovarian cancer early, co-first author Jamie Madena (Johns Hopkins Kimmel Cancer Center) bemoaned the “lack of specific symptoms early in the course of the disease or effective biomarkers.”
What’s more, initial screens for ovarian cancer involve an ultrasound to detect abnormal growths. This approach is unable to accurately distinguish benign from cancerous growths and so is followed by exploratory surgery to characterize any growths identified, putting patients through an invasive surgical procedure that may not be necessary.
And yet, there are some clues that could be used to aid early detection. Two biomarkers for the disease have been established, cancer antigen 125 (CA-125) and human epididymis protein 4 (HE4); however, their diagnostic powers are weak. Another clue is presented by the processes following cell death. When a healthy cell disintegrates, its DNA fragments into regular, predictable segments. Cancer cells, on the other hand, leave behind a chaotic DNA fragmentation pattern. This fragmentation pattern or profile, also known as a fragmentome, can be analyzed to determine the likelihood that an individual has cancer.
Biomarkers for predicting multiple sclerosis progression identified
Researchers have identified biomarkers that can predict disease progression in multiple sclerosis, which could lead to the development of more personalized treatment strategies.
The team recently explored this ‘fragmentomic’ approach with an AI-powered method named DELFI (DNA Evaluation of Fragments for early Interception), which analyzes fragmentation profiles to determine an individual’s likelihood of having lung cancer. In this previous study, the approach was shown to effectively detect lung cancer.
Taking this research one step further, the team set out to combine DELFI with liquid-biopsy-based detection of HE4 and CA-125 to evaluate the performance of a combined diagnostic test for ovarian cancer. The team used this approach, which they labeled DELFI-Pro, to analyze blood samples from 94 ovarian cancer patients, 203 patients with benign ovarian growths and 182 women with no known ovarian growths, all from the Netherlands and Denmark.
From this cohort, the team were able to detect 72%, 69%, 87% and 100% of ovarian cancer cases at stages 1, 2, 3 and 4 respectively using DELFI-Pro. This compared favorably to testing for CA-125 alone, which detected 34%, 62%, 63% and 100% across the same stage profile.
These results were then validated in a cohort from the USA consisting of 40 ovarian cancer patients, 50 benign ovarian growth patients and 22 people in the control group with no known ovarian growths. Here, 73% of all cancers were detected and 81% of high-grade serous ovarian carcinomas were detected. There were vanishingly few false positives reported and the DELFI-Pro test was able to distinguish benign from cancerous growths with high accuracy.
The improved ability to accurately diagnose ovarian cancer, particularly at the early stages of the disease could prove crucial to reducing the mortality rates for the disease. Combining this with the ability to distinguish benign from cancerous growths, the DELFI-Pro test could save huge numbers of patients from undergoing unnecessary surgeries while accelerating the path to treatment for those who do have ovarian cancer.
Extolling the benefits of the DELFI-Pro, Velculescu commented that it, “has the potential to be an affordable, accessible method for widespread screening for ovarian cancer.” However, before the test can deliver on his vision, the team next need to validate it in larger sample sizes.
