Study: AI can predict pancreatic cancer three years ahead of human doctors AI algorithms can screen for pancreatic cancer and predict whether patients will develop the disease up to three years before a human doctor can make the same diagnosis, according to research published in Nature on Monday. Pancreatic cancer is deadly; the five-year survival rate averages 12 percent. Academics working in …
A similar announcement was made a bit less than 2 years ago, IIRC. The topic is memorable because my father was taken by pancreatic cancer in under 6 months after diagnosis, and any help in the battle against this lethal disease is welcome.
But this is the second announcement that they have ideas about computer-aided diagnosis...and, still, nothing actually available to the public.
Stop making promises you never seen to keep. Roll the damn programs out, even at an experimental level, because the odds are so great that, even if this isn't completely effective, a few lives saved is better than the current status quo.
Someone I know has been on the opposite end of this - they presented with symptoms of gallstones, were told they likely had pancreatic cancer on the back of a blood test but they finally found, after several weeks of serious anxiety at a potentially fatal diagnosis, they simply had had gallstones all along.
Whereas late diagnosis is the principal issue with this disease, false positives can have a devastating effect too. If you identify 1000 people at high risk, but 680 don't go on to develop the disease, that's a lot of people who have been needlessly worried. This could be a bit of a moral maze.
Not to mention the often devastating side effects of the treatments.
Surgery is damaging by definition - very likely to cause diabetes, for a start. Radiotherapy and chemotherapy are not much fun either.
Quite a few people choose not to be treated because for them, the sideeffects outweigh the likely benefit :(
If it turns out that the diagnosis was wrong and the treatment was unnecessary, a lot of physical harm has happened on top of the mental harm.
Great care and a lot of good science is needed before any diagnosis system gets rolled out.
Agreed, an excessive number of false positives and causing unnecessary alarm, investigation or surgery is bad. But if the system was to be enhanced with further potential indicators, however inconclusive on their own, false positives could be reduced to a more acceptable level. Maybe some genetic predispositions or correlations.
My wife spent 6 months undergoing tests to diagnose the cause of what appeared to be indigestion and back pain, nothing found and no treatment was helping. At that point the doctors said it can only be psychosomatic so stop taking the pile of ineffective medications. That was when things deteriorated massively so another gastroenterologist had a look and gamma camera imaging showed the pancreatic cancer - far too late for there to be any possibility of a positive outcome, died 3 months later at 49.
After I had prostate cancer some years later, which had spread to the lungs, my DNA was sequenced for inclusion in the UK genomic research database. Hopefully that kind of information will contribute to earlier diagnoses or even preventative medications and lifestyle changes. The problem then is that we'll all live into our 90s - with all the concomitant infirmities, incontinence, Parkinson's, osteoporosis, arthritis, failing eyesight and hearing, dementia...
As someone who is currently undergoing treatment for inoperable pancreatic cancer I don't know why it took so long for the diagnosis. I found out the same day I visited a gastro-intestinal specialist after a high contrast CT scan. The CT scan also showed that it had spread to two lymph nodes making it inoperable with a 2% chance of survival after 5 years. A biopsy later confirmed the original diagnosis.
Richard: As to treatment i know that it affects people differently but my original chemo treatment using two different aggressive drugs every other week was not devastating. It was debilitating, mainly serious fatigue and immune system suppression but that was just about it. No nausea or other serious side effects. It included a couple of side effect suppression drugs administered before the chemo which were effective. I also had a couple of anti-nausea/vomiting drugs prescribed for home use that I never needed. The treatment itself was effective and brought my cancer under control. I am now on a different treatment with a study drug that uses a gene mutation I have (PALB2) to attack the cancer. It is currently used for other types of cancer and has the advantage of being an oral medication that I can take at home instead of having to go to the hospital twice a month for blood tests and intravenous chemo. I do still have to go for blood tests once a month with a CT scan once every two months for the study. So far the new treatment has also proved to be effective. This should add some (many, I hope) years to my life.
What I am trying to say is that people should at least try the currently available treatments. Over the years the treatments and treating the side effects has gotten much better. Unfortunately it will not always be effective or cause severe side effects but not trying them first leads to the ultimate devastating.
Problem with trying to “save a few lives” is that most deadly diseases is like that. If we throw more resources (money and man power) at the issue, things can be better but we can’t do this for everything. And if we do it for just one, what other illnesses are going to get less resources?
What we should celebrate is that these “new” ideas seems to get more traction. AI is yet another tool that could be used for many other diseases and with more breakthroughs, the systems will be able to help a broad range of people in the future. But for now, we must just hope for good results and that it can be implemented in the real world in not too far into the future.
I’m sorry for you loss. Losing love ones to unfair things like this sucks.
Once a week, we read reports that scientists present, promise or propose a breakthrough to save our lives from fiendish diseases. At the same time, you don't get a doctor's appointment for problems that have been understood and curable for a century, for lack of doctors and money.
As somebody who is possibly at a high risk, this sounds great . But it also reminds me of claims over the last 15 years promising dementia analysis and cures.
Even the data seems suspect. The pancreatic cancer results per are sixfold higher in Denmark than the US?
And the linked abstract notes that cross application of the results to US date was problematic.
This has all the hallmarks of a well-intentioned study deriving more attention due to the magic mention of AI
