Cover of RXAIPO Pharma, AI, Public Opportunities

RXAIPO Pharma, AI, Public Opportunities

RXAIPO Pharma, AI, Public Opportunities

By: Mihir Otia
Listen for free

Season's Savings | 0.99 €/month for 3 months

5.99 €/month thereafter - terms apply. Cancel monthly.
Summary
AI, healthcare, investing and the future of wealth and careers. Faceless daily breakdowns of the trends that matter.Mihir Otia Science
Episodes
  • AI Found Who a Lung Cancer Drug Works For: 7.2 vs 4.0 Months
    14 Nov 2026

    An AI just did something pathologists couldn't: it picked out who did better on a lung-cancer drug.


    The trick wasn't counting the target. It was measuring where the target sits inside each cancer cell.


    New in Nature Medicine (11 Nov, peer reviewed), on Datroway, the AstraZeneca/Daiichi Sankyo antibody-drug conjugate:

    ▪️ The target, TROP2, is on ~90% of lung tumours, so the usual pathologist score barely separated responders

    ▪️ The AI splits each tumour cell's TROP2 into surface vs inside, giving a "normalised membrane ratio" (NMR)

    ▪️ On Datroway, NMR-positive patients went 7.2 months before progression vs 4.0 for NMR-negative (221 patients, phase 3 TROPION-Lung01)

    ▪️ On chemo, the same marker made no difference: 4.1 vs 4.4 months. It predicts benefit from the drug, not just a better outlook

    ▪️ Overall survival on Datroway: 17.0 vs 10.8 months


    The signal: precision medicine may not need new targets. It may need old ones measured properly, and the algorithm is becoming part of the drug. Roche is co-developing an AI-powered companion test.

    The noise: this is a retrospective, AstraZeneca-funded analysis, and the cut-off was tuned on the same data. The real test is TROPION-Lung17, a ~400-patient trial enrolling only NMR-positive patients.


    Would you trust an algorithm to decide who gets a cancer drug? Tell me below 👇

    Comment SIGNAL and I'll send you the full breakdown (sources included).

    Save this for the next time someone says AI in pathology is "just diagnosis".


    Sources: Garassino et al., Nature Medicine (published online 11 Nov 2026); Daiichi Sankyo (TROPION-Lung17, 13 Jan 2026); ApexOnco (2 Nov 2026). Not medical or investment advice.


    #HealthAI #AIinHealthcare #MedicalAI #DigitalPathology #ComputationalPathology #LungCancer #Oncology #PrecisionMedicine #Biomarkers #ADC #Pharma #Biotech #Pharmacist #HealthcareInvesting #RXAIPO

    Show more Show less
    11 min
  • AI: ~60% of Kids in a Brain Cancer Trial May Not Have Needed Full-Dose Radiation
    4 Nov 2026

    A Dana-Farber-led team built MANTIS, an AI that reads DNA methylation from children's brain tumours (a test already used for diagnosis) and predicts each child's survival over 10 years. Built on data from 2,786 patients and tested on 898 children from three clinical trials it never saw, it significantly beat current clinical risk markers. Re-analysing a randomised trial of lower-dose radiation (170 children aged 3–7), it identified ~60% (104 children) for whom the lower dose showed no detectable survival disadvantage, inside a trial where lower dose did worse overall. Published 3 Nov 2026 in Nature Genetics. Fine print: retrospective, small groups, radiation dose only, prospective trials needed. Not a reason to change any child's treatment.


    Get the daily signal on AI, healthcare and investing: newsletter link in the pinned comment.


    Source: Stefan et al., Nature Genetics (2026), DOI 10.1038/s41588-026-02815-4. Not medical or investment advice.


    #Shorts #HealthAI #AI #PrecisionMedicine #ChildhoodCancer #Biotech #Investing

    Show more Show less
    11 min
  • An AI just out predicted hearts experts
    3 Nov 2026

    An AI read 27,382 heart MRI scans, and its ejection-fraction measurements predicted who would die slightly but significantly better than expert clinicians' measurements, in 2–3 seconds per patient vs around 13 minutes. Published 2 Nov 2026 in The Lancet Digital Health (City St George's, University of London + UCL). The twist: it graded AI and doctors against survival, not against each other. Fine print: retrospective, marginal edge, clinician oversight still essential.


    Get the daily signal on AI, healthcare and investing: newsletter link in the pinned comment.


    Source: Joy et al., The Lancet Digital Health (2026), DOI 10.1016/j.landig.2026.101042. Not medical or investment advice.


    #Shorts #HealthAI #AI #Cardiology #MedicalImaging #HealthTech #Investing

    Show more Show less
    10 min
adbl_web_anon_alc_button_suppression_t1
No reviews yet