[AI x Medicine] A story about how Open Evidence became so smart it got a little scary
Recently, I saw a patient with a rare cancer that had recurred for the second time.
It is something that happens from time to time in the world of otolaryngology:
a type of cancer for which there is no clear standard treatment.
Because it is a rare disease, it is often the case that:
the number of cases is small
the evidence is weak
there are no clear guidelines
and so on.
The initial treatment was quite aggressive.
In the initial treatment,
the tumor was resected quite extensively under nasal endoscopy
dissection was performed from the posterior nasal septum to near the skull base
the medial orbital wall was also treated
It was a very aggressive surgery. I remember it being difficult.
After that,
chemoradiotherapy
immunotherapy
Considering the stage,
honestly, I think it was a very tough battle.
- There was also a neck recurrence, so a neck dissection was performed.
Even so, the patient did their best and
lived for a long time.
Before I knew it, five years had passed.
Before I knew it,
five years had passed since the final treatment.
Even at the outpatient clinic,
we were talking about how "we've made it this far."
It was that kind of conversation.
There was a sense that
it was almost time to graduate from follow-ups.
But,
it's always at times like that when you find it.
A second recurrence.
Maybe immunotherapy next.
I think about it while looking at the images.
The next treatment will probably
be an immune checkpoint inhibitor.
However, since it's a rare disease,
I'm a bit hesitant.
Treatment strategies for recurrent cases
Evidence for re-administration
The role of genetic testing
I have fragmented knowledge in my head, but
it's not neatly organized.
So, I asked AI,
Open Evidence
about it.
Regarding treatment for a second recurrence of a rare cancer.
Then
current treatment strategies
the role of immunotherapy
reference papers
were
all organized and presented quite neatly.
In the past,
this would have required searching for 10 to 20 papers comprehensively on PubMed,
a task that would have taken a considerable amount of time.
Consulting the university just in case
Just in case,
I also consulted with a professor at the university.
The answer that came back was
almost the exact same content.
That was when I thought,
It really is amazing.
However, there is also a sense of fear
The more excellent AI becomes,
the scarier it can also be.
One thing is
that people will stop memorizing things.
Another is
that people will stop reading papers.
I still feel it is quite dangerous
to feel like you understand something just from an AI summary.
Doctors in the AI era
Doctors from now on will be
doctors who can use AI.
Doctors who don't use AI
might find themselves
at a significant disadvantage.
Medicine is
from a world of "how much you can remember"
to
a world of "how well you can ask questions"
I feel it is gradually changing.
I feel it is gradually changing.
