OT? New AI predicts cancer survival from selfie

Investment Analysis Clubs Macro Economic Trends and Risks
1

I’m not sure this is OT because medical costs might be reduced if a free tool like a selfie can help a doctor predict survival. (Or ability to tolerate harsh treatment like radiation.)

New AI tool can predict cancer survival by a single patient selfie, study suggests

System can estimate a person’s biological age and predict chances of cancer survival from a single photograph
Andrew Gregory Health editor, The Guardian, Mon 19 Oct 2026

Doctors have developed an artificial intelligence tool that can predict a cancer patient’s chances of survival from a single photograph of their face.

The AI system was custom-built by experts at Mass General Brigham, a nonprofit, integrated academic healthcare system in Massachusetts, to estimate biological age and directly predict survival outcomes in patients diagnosed with cancer.

It is one of the most comprehensive demonstrations to date of how AI analysis of a selfie could help assess a patient’s health and potentially guide their medical care, and was trained on data from more than 40 million facial photographs across a range of ethnicities and ages…

Among cancer patients treated with radiotherapy in the US, the people who, according to FAHR-FaceAge, looked five or more years older than their actual age had a 21% higher risk of dying during the average four-year follow-up than the rest of the group. The patients who looked five or more years younger had a 17% lower risk…

Researchers found lifestyle factors were associated with facial ageing. For example, heavy alcohol use, drug use and smoking were linked to looking biologically older in AI analysis of facial photographs… [end quote]

Sometimes I see women who appear to be about my age and I say to myself, “What have you done to yourself?” because their skin and physique are so aged. Extreme stress is also aging. On the other hand, wealthier people can afford supplements (like collagen), cosmetic creams and vitamins which enhance skin condition.

Doctors are supposed to be able to evaluate patients from experience. But I guess that AI could be more quantitative.

Wendy

5 Likes
2

This app that calculates your age based on a selfie has been around for a while.

intercst

3 Likes
3

It says I look 60 so I think it’s great. :wink:

Wendy

4 Likes
4

Alcohol and cigarettes: AI can register it. The doctors register it as well. Anyone does.

Met a woman on Sunday who said she was going to be 70. She was a rare person who actually is older than me can doesn’t look it. Amazed. No cigarettes or alcohol.

The percentages in the study are major margins, but not that extreme.

Looking forward to getting on Mounjaro. Bring my odds for any of this stuff way down.

5

An interesting speculation…as all press releases are designed to be.

Yes, it is one…complete with the SBPR “Miranda” warning by one of the co authors that usually accompanies these advertisements for a publication that’s a preliminary study only…

“To the best of our knowledge this is the first demonstration that an AI model can predict cancer survival from a simplephotograph”.

Here’s the problem/potential danger with the flurry of interest generated by these uncritically enthusiastic pressers…they do tend to get picked up and determined to be a valid piece of evidence in advance of any peer review. My usual example is of Andrew Wakefield’s fraudulent anti-MMR vaccine paper (also published in The Lancet) but this publication is no different. Given that the opinion of Google AI overview is so valued, so often, by so many, I typed “Can a single photograph predict cancer survival?” The AI overviewer answered with a confident YES followed by a bit of blather. Stepping over this to wade through the sources revealed the various iterations of the article in The Grauniad etc…CNN, Facebook, WaPo, probably the MailOnline.

Edit: SBPR has been recognised by academic institutions for at least a couple of decades now. One of my fave medical writers, Ben Goldacre, highlighted the issue very frequently in his Bad Science column in The Grauniad back in the early aughts and his subsequent book. Heaps of articles and studies done on the topic…but it still persists.

Press Releases by Academic Medical Centers: Not So Academic? | Annals of Internal Medicine https://share.google/f24UzorGVM0HlNAEb

1 Like
6

Does this mean the Wakefield paper wasn’t peer reviewed prior to publication?

DB2

7

No cigarettes or alcohol.
A lifetime of aerobic and weight-bearing exercise, 4 times per week.
Plenty of sleep.
Low stress.
High-protein, low-carb diet with plenty of fresh fruits and vegetables. Home-cooked food, no junk.
1 - 2 tablespoons of collagen per day after age 50.

I am age 72. I’m also a cancer survivor who has had several major surgeries, including open-heart surgery 2 years ago. Healing and resilience is an important part of health. Even if the AI age evaluator is junk science, all the lifestyle factors I practice are evidence-based (not to mention common sense and personal experience).

Wendy

260411 Wendy Portrait
260411 Wendy Portrait1409×1776 474 KB

4 Likes
8

Well, yes.

I’ll assume you’re not JAQing off again and give a quick breakdown of how publication works the way it does and why the adage “Publication is not validation” holds.

Leaving out what we know now about Wakefield’s shenanigans, the route that a publication such as his (a case report not an actual study…lower bar for journal acceptance, granted, but still) The paper together with the compelling data is first submitted for editorial review and is read a checked over well by a team of reviewers to see if it seems kosher. Basically does ot go far enough to satisfy the basic tenets of the Scientific Method, I suppose…like, do the results and tentative conclusions match the original hypothesis sort of thing. This is not considered peer review.

That happens if/when the publication is accepted and the journal circulated among the scientific/biomedical community…that collective of folk with sufficient knowledge, skill and expertise in the field to actually be able to sit down, pore over the document and criticique any areas of weakness/smell fish if there’s a hint of any etc. This group does not include the copy writer s n the institution/journal’s advertising department (that’s what it is in reality)

The rest becomes history…either as part of the biomedical archive, The Journal of Irreproducible Results…or even as the evidence produced for the General Medical Council when considering the future of a medical license (in Wakefield’s case)

He legacy lingers on in the numbers of measles cases we have right now.

3 Likes
9

I’ll use my own experience to suggest we not lose sight of the notion that appearances can be deceptive. After a lifetime of rude good health and a consistent demonstration of good custodianship of my body and all the healthy lifestyle choices I’ve made I’ve been given a different perspective. Not only did that Righteous Path fail to protect me from significant ASCVD, that healthy BMI and good body composition that’s on display when walking into my primary care office for an annual physical served as a smokescreen and actually encouraged low level thinking and complacency when looking at my lipid panel. I’m convinced that 20 lbs or so of extra lard wrapped round my belly (with or without accompanying metabolic disruption) that so called “mildly elevated” LDL-C … and I think we’re all aware now that a value of 130 or so mg/dL is absolutely NOT mildly elevated!! … would not have been dismissed the way it was for so long. My intervention cardiologist and new electrophysiologist (seen on Monday) agree!

1 Like
10

BMI? Until fatties have a BMI over 35, their mortality risk profile does not rise much.

11

Riiight. Keep telling yourself that. Next you’ll be calling an A1C of 5.9 or so “only” prediabetes…and claiming it’s not a pathogic state (yep, folk do believe that, it seems)

12

As you know, I had a similar experience.

Perhaps an “older than” facial test is more useful than a “normal” or “younger than” facial test. The normal and young-looking people may or may not have an underlying problem. But the “older than” people definitely do have a higher risk of an underlying problem that should be investigated.

When I see a woman who seems to be around my age (this happened when I was in my mid-50s) with lots of wrinkles and sallow, poor-quality skin my first reaction is “Stop what you’re doing to harm yourself and go to the doctor.” So this may be a one-way test.

I have noticed that vegans over age 55 seem to be very wrinkly. I hypothesize that this is due to the lack of proline (from collagen) in their diets so their skin wrinkles. But they usually look healthy overall so the skin quality deterioration seen in smokers and drinkers is different. Just my observation.

Wendy

13

I tried it and played around with different lightings and expressions and it had a 11 year range… from 8 years younger to 3 years older.

2 Likes
14

VE,

I was hoping someone in charge with a high BMI would pop his clogs as my aunt says. No luck in the BMI front. It really does take an outrageously high BMI to get in serious trouble.

15

Well, I won’t waste my time arguing. I’m reasonably sure anyone remotely interested has done their own due diligence on this topic. Or, at least veered towards a Google AI overview with a prompt along the lines of "Is there significant pathology associated with overweight/excess bodyfat category but below obesity.

16

Thank you for not wasting our time! Much appreciated! :+1:

The Captain