back to article Risk-based algorithm could improve cancer screenings

An algorithm capable of estimating the risk that a particular patient will develop prostate cancer over the next five years should be used in a national screening program in the UK, one of the software's creators has said. One in eight men in the United Kingdom will be diagnosed with prostate cancer in their lifetime, …

  1. GlenP Silver badge

    Interesting*, and well done to the team for making it clear it's an Algorithm not jumping on the so-called AI bandwagon.

    *As a 50-something male I'm in the danger zone, although the last PSA test was, thankfully, negative.

  2. Screwed

    Swings and Roundabouts

    In another area of medicine, it is becoming recognised that some people require a different medicine regime. The science has indicated that this can be due to a specific genetic factor (a particular SNP on a gene).

    The positive is that people with this SNP are seeing better chances of getting the different medicine. Some consultants accept that it is justified despite its higher cost.

    Most of the patients have had a private gene test - it has not advanced to the point of being widely available on the NHS.

    The negative is that those who do not have that SNP are less likely to get it. Despite genetic research, being a fairly young field, is still discovering new SNPs which have the same indication for the other treatment.

    The statistics are clear across the population. But population statistics do not apply to the individual. That patient either does, or does not, need the other treatment. The SNP should be seen more as an automatic route, and an explanation. But others should not be excluded from being assessed on a clinical basis, maybe given a trial of the other medicine to see if it helps.

    1. Fruit and Nutcase Silver badge

      Re: Swings and Roundabouts

      This may be of interest - some background on genetic sequencing work...

      BBC R4 - The Life Scientific

      Shankar Balasubramanian on decoding DNA

  3. Mike 137 Silver badge


    I haven't been able to check the paper, but my crude calculations suggest that, given the sample size, there would be 105 false positives for 0.5% FP rate. That's against 571 actual cases. I'd also be interested in the false negative rate.

  4. cantankerous swineherd

    want a positive psa test? go for a 2 hour bike ride.

  5. Chris the bean counter

    Not a solution

    Usually fortunately the cancer is benign, most people diagnosed will die of something else, so diagnosis can cause unnecessary worry and unnecessary interventions can have devastating side effects.

    The need is for a test to identify whether the cancer is benign or malignant. Preferably without an autopsy which can cause damage.

    Some tests are being developed. A small test reported last month where people who had been diagnosed with the cancer were injected with a sugary solution that contained a marker that can be detected by an MRI scan (cancer loves sugar). This was effective at distinguishing malignant cancers that need intervention and benign cancers that simply need monitoring.

    1. Jimmy2Cows Silver badge

      Re: Preferably without an autopsy which can cause damage

      Typo (i.e. you meant "biopsy"), or subtle understatement...?

  6. Jimmy2Cows Silver badge

    Smart poop

    It's a pity this stuff (and other excretion diagnostics) can't be built into a toilet to give real-time alerts to, ahem, users.

    1. HildyJ Silver badge

      Re: Smart poop

      Panasonic has a urine analysis toilet in production and use in China.

      Stanford and MIT, among others, are working on more advanced concepts that could also analyze poop.

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like

Biting the hand that feeds IT © 1998–2022