ARTIFICIAL INTELLIGENCE IS CAUSING SEISMIC SHIFTS in cancer research and management. Now, this revolutionary A.I. tool promises instant national database cancer reporting detection. Today, we focus on this new AI tool and cancer incidence reporting.
I often hear about the perils of artificial intelligence.
“AI is a mirror, reflecting not only our intellect, but also our values and fears.” – Ravi Narayanan, VP of Insights and Analytics, Nisum.
In this brief essay, I will focus on how A.I. is helping us gather cancer diagnosis information years faster than we could historically.
Today’s Approach to Tracking Cancer in the U.S.
The National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) program stands as a cornerstone in our understanding of cancer trends and survival rates across the United States.
Yet, despite its authority, SEER’s data collection process has long been hampered by a significant bottleneck: manual updates.
The Problem
My cancer program has an extraordinarily capable crew of cancer registrars.
Yet, reporting is always delayed.
The reliance on human intervention results in a frustrating two-year lag between cancer incidence and its official reporting.
Consequently, we have delayed crucial insights into emerging trends or spikes in cancer rates, leaving researchers grappling with outdated information.
This delay not only hampers our ability to address areas of concern promptly but also undermines efforts to adapt strategies for improved cancer prevention and treatment swiftly.
A.I. Delivers a Win for Cancer Reporting
In collaboration with the National Cancer Institute (NCI), scientists from the Department of Energy’s Oak Ridge National Laboratory and Louisiana State University have unveiled a cutting-edge A.I. transformer.
This transformative technology can analyze vast volumes of pathology reports, offering researchers unparalleled insights into cancer diagnoses and management.
How A.I. Transforms Cancer Reporting
This A.I. system significantly enhances the accuracy and efficiency of cancer reporting by harnessing the power of long-sequencing capabilities.
Mayanka Chandra Shekar, a research scientist in the Computational Sciences and Engineering Division at Oak Ridge National Laboratory (ORNL; USA), offers her take:
“[AI can potentially] automate the process of extraction of specific cancer site information from these pathology reports and make it into structured data for nation level cancer incidence reporting.”
The approach provides me and other experts with invaluable data to advance their research and improve patient outcomes.
Details
The research team used the highly secure CITADEL framework on the Oak Ridge Leadership Computing Summit supercomputer.
Scientists at ORNL leveraged the specialized transformer model to analyze a staggering 2.7 million cancer pathology reports.
Dubbed Path-BigBird, this groundbreaking model extracts data from six Surveillance, Epidemiology, and End Results (SEER) cancer registries, marking a monumental leap forward in cancer research and data analysis.
Revolutionizing Cancer Reporting
With the current manual updating process for cancer registries, a frustrating two-year delay exists between cancer incidence and its official reporting.
This significant gap means that in the event of a national increase in cancer rates, researchers are left waiting for two years before identifying and addressing areas of concern.
This delay hampers timely responses and undermines efforts to adapt strategies for improved cancer prevention and treatment swiftly.
Into the Future – AI Tool and Cancer Reporting
Researchers have been testing this fancy new Path-BigBird model to dig out important information.
They believe this model can help us determine where cancer is happening more and even answer questions about it.
It’s a big deal because it could mean we spot cancer problems faster and help the people who need it most.
Now, the researchers are busy giving the model even more tasks, like finding special signs in the body, seeing if cancer returns, and ensuring we get all the right information about cancer cases.
In recent years, I have watched as our pathologists have been increasingly asked to standardize their cancer case (pathology) reporting.
Caution – AI Tool and Cancer Reporting
“By far, the greatest danger of AI is that people conclude too early that they understand it.” – Eliezer Yudkowsky.
I have some questions and concerns:
- Data Encryption: How do researchers encrypt the data to prevent unauthorized access? We need to ensure that even if someone intercepts the data, they won’t be able to interpret it without the proper decryption keys.
- Anonymization: Researchers should remove personal identifying information (PII) such as names, addresses, and social security numbers or anonymize the data before using it for training or analysis. This approach helps protect patients’ privacy while still allowing for meaningful analysis.
- Secure Computing Environments: Scientists can train and run models in secure computing environments that adhere to strict privacy and security protocols. This approach can include secure cloud platforms or on-premises servers with restricted access.
- Access Controls: Only authorized personnel should access the data and the trained models. Researchers should create access controls to ensure that only those with the proper credentials can view or manipulate the data.
- Compliance with Regulations: A.I. models used for medical data analysis must comply with relevant regulations such as HIPAA (in the United States) or GDPR (in the European Union). Researchers can help ensure patient privacy by complying with these regulations.
Still, I am cautiously optimistic that the approach will dramatically improve cancer reporting. What say you?
The researchers published their work in Clinical Cancer Informatics.
Artificial Intelligence May Render Some Medical Specialties Obsolete
ARTIFICIAL INTELLIGENCE (AI) IS ALREADY ENCROACHING in many medical specialties and may render some obsolete.medium.com
Thank you for reading “Revolutionary AI Tool Promises Instant Cancer Detection Reporting.” Please sign up to follow me.
