Ovarian cancer, described by Audra Moran, head of the Ovarian Cancer Research Alliance (OCRA), as “rare, underfunded, and deadly,” remains a significant health challenge.
Early detection is critical, as most ovarian cancers originate in the fallopian tubes and often spread before symptoms manifest or the disease is diagnosed.
Now, advances in artificial intelligence (AI) are showing promise in transforming the early detection of ovarian cancer and other life-threatening conditions.
Dr. Daniel Heller, a biomedical engineer at Memorial Sloan Kettering Cancer Center, is at the forefront of developing innovative AI-driven blood tests for ovarian cancer.
His team utilizes carbon nanotubes—microscopic tubes 50,000 times smaller than a human hair—that emit fluorescent light when exposed to blood samples.
These nanotubes are engineered to respond to specific molecules in the blood, creating unique light patterns. Interpreting these complex patterns, however, is beyond human capability.
Machine-learning algorithms bridge this gap. By training AI systems with blood samples from patients with ovarian cancer, as well as those with other gynecological diseases, researchers can teach the algorithms to identify subtle molecular differences.
Despite limited data—owing to the rarity of ovarian cancer—the AI has achieved higher accuracy than existing cancer biomarkers in early tests.
Dr. Heller’s vision is to develop a diagnostic tool that can quickly help doctors determine whether a gynecological disease is cancerous and, if so, identify its type.
He estimates that such tools could be available within three to five years, contingent on larger datasets and further validation.
AI’s potential in medical diagnostics extends beyond ovarian cancer. For instance, pneumonia, a dangerous condition for cancer patients, is caused by over 600 pathogens, requiring numerous tests to identify the specific organism.
California-based company Karius employs AI to streamline this process, using microbial DNA to rapidly pinpoint the causative pathogen.
The AI compares patient samples against a vast database of microbial DNA containing tens of billions of data points, producing results within 24 hours. This approach not only speeds up diagnosis but also reduces testing costs and ensures timely, targeted treatments.
The strength of AI lies in its ability to detect complex patterns in data that might involve multiple biomarkers rather than a single indicator.
Dr. Slavé Petrovski, a researcher at AstraZeneca, developed an AI platform called Milton that identifies over 120 diseases with a success rate exceeding 90%.
By analyzing biomarkers from extensive datasets, Milton exemplifies AI’s capacity to decode intricate biological information. Similarly, Dr. Heller’s ovarian cancer research relies on AI to interpret nanotube-based sensor data, even though researchers may not fully understand the underlying connections between biomarkers and cancer.
However, challenges remain. Data scarcity and fragmentation hinder progress, particularly in ovarian cancer research. Patient data is often siloed within hospitals, limiting its availability for algorithm training.
Dr. Heller likens current AI research in this field to a “Hail Mary pass,” as algorithms are trained on small datasets from only a few hundred patients.
Recognizing this limitation, OCRA is funding efforts to build large-scale patient registries with electronic medical records, enabling researchers to train AI models on more comprehensive data.
Despite these hurdles, the potential of AI in healthcare is immense. By leveraging vast datasets, AI can accelerate diagnostics, improve accuracy, and enhance patient outcomes.
For ovarian cancer, earlier detection could dramatically improve survival rates, offering hope in a field long overshadowed by challenges.
As Moran aptly observes, “We’re still in the wild west of AI now.” Yet, with continued innovation and collaboration, AI is poised to revolutionize healthcare diagnostics, from cancer detection to infectious disease identification.
For patients and physicians alike, this technology represents a beacon of hope, offering the possibility of earlier, more precise interventions and better long-term outcomes.
