Louisiana doctor uses AI software to detect pancreas cancers | Louisiana Health
The American Cancer Society predicts that 66,440 people in 2024 will be diagnosed with pancreatic cancer, and 51,750 people will die of pancreatic cancer by the end of the year.
In Louisiana, cancer mortality was highest for lung, colorectal, pancreatic and breast cancers between 2016 and 2020, according to the Louisiana Tumor Registry at LSU New Orleans School of Public Health. Although pancreatic cancers accounted for 7.3% of cancer-related deaths in Louisiana, there still isn’t a general population screening program for the cancer.
EON technology is artificial intelligence software that Our Lady of the Lake Regional Medical Center in Baton Rouge is looking to bridge the gap. The software works by analyzing CT scans (a noninvasive imaging X-ray) across the hospital system, whether patients were admitted for pancreatic concerns or not. For example, if a patient is in a car wreck and gets an abdominal CT scan, the EON software will analyze the scan for indicators of a pancreatic cyst and flag it for the oncology team for a follow-up.
Dr. Mo Al-Efishat, a surgical oncologist at the Lake in Baton Rouge who specializes in advanced pancreatic tumors or cysts, answered questions on the new technology and its effectiveness in preventing the development of pancreatic cancers.
What is EON technology used for at Our Lady of the Lake, and how has it been implemented?
The technology behind it is called computational linguistics. The technology is up to a 98% accuracy by combining artificial intelligence and computational linguistics — a more sophisticated analysis compared to traditional methods that merely read the text and identify cysts.
In the body of the report of the CAT scan, there will be a mention of a cyst that oftentimes goes either unnoticed or does not get the appropriate follow-up. We know that 20% of pancreatic cancers originate from some of these cysts — what we call mucinous cysts. Once we identify a cyst, we schedule an MRI or CT scan every six to 12 months. At this point, the program has done its job for the patient, and we take over.
The program is always running, and it runs across our entire health care system. Any CAT scan or MRI that occurs across the system in the different hospitals will be brought to our queues. Then, we can filter and decide which cysts need to be flagged for a follow-up with our pancreatic oncology team.
The key is early detection. That’s one of the best ways of preventing pancreatic cancers — through detecting these cysts and making sure that we intervene before they get a chance to become cancerous.
How important is it to get this early screening for pancreatic cancers?
This is critical for pancreas cancer, even for the earliest stage of pancreas cancer. When discovered, the cure rates are in the range of maybe 30% and 40%. We say “cure” to mean that the patient lives beyond five years.
I had the results of a surgical biopsy after I did a robotic procedure for a patient. The pathology result came back as high grade dysplasia — the step right before cancer.
This is when we high-five and say we did the right operation, at the right time, for the right reason.
We don’t want to intervene when the cyst is still in its infancy and not showing any signs of transformation or change. We also don’t want to intervene when the patient already has pancreatic cancer. Once diagnosed with pancreatic cancer, the survival rate goes from 100% down to 30%.
Overall, in pancreatic cancer the survival is less than 10%. The majority of patients with pancreatic cancer in the community present at later stages where the tumor is either unresectable (unremovable by surgery) or has already spread (meaning it is stage four and is incurable).
There’s no mammogram or colonoscopy like we have for breast cancer and colon cancer. There is a screening program for individuals who have family history of pancreatic cancer who meet certain criteria, but that’s just a minority of the population who are getting screened.
So, we have a very aggressive cancer with no screening program. That’s why it is critical to have some form of program to follow the pancreas cyst. The AI technology is one way to prevent pancreatic cancer.
Are there visible symptoms that would signal to someone to get their pancreas checked or have a pancreatic cancer screening?
The common symptom of turning yellow, or yellow eyes, is actually dependent on where the cyst is in the pancreas. If the cyst is in the body or tail of the pancreas, even if it grows and becomes cancerous, the patient will not turn yellow.
The body and the tail of the pancreas are notorious for having cysts or tumors that can grow to a big size without causing symptoms. We know that if you have pancreas cancer in the body and the tail, your prognosis (or how long you’re going to live) is worse than if you have it in the head of the pancreas.
Sometimes there are very subtle symptoms. If a patient who is 50 years old has been healthy with a normal weight and nothing changes in their diet, all of a sudden becomes diabetic, that is a worrisome sign that there might be something going on in the pancreas.
Whenever I meet with patients, I check their hemoglobin to make sure that, even if they do have diabetes, it’s not getting worse. I also ask patients whether they have abdominal pain that is severe or constant — that can be a sign of what we call pancreatitis, which is the inflammation of the pancreas.
Another uncommon sign is that sometimes, if the tumor or the mass is big enough and blocking the pancreas, the patient may have some form of diarrhea. That is a sign that the pancreas is not secreting enough juice.
We always ask about family history. If you have a family history of pancreatic cancer, that puts you at higher risk. We will follow you or watch you closer than a patient who did not have a family history.
What new technologies or studies around pancreatic cancer treatment are you excited to see develop?
When I was doing my basic science research, we asked: “How do these cysts become cancerous? What drives a cyst to go from being normal to low grade dysplasia, high grade dysplasia and then cancer?”
We have shown over and over that pancreatic cancer is driven by inflammation. Now, multiple institutions are enrolling in a trial where patients with pancreas cysts are given an anti-inflammatory agent to see if we can slow down, or even prevent, progression to cancer. That will be a game changer.
link