#AGBTPH – Kenna Mills Shaw, Precision oncology decision support: Building a tool to deliver the right drugs(s) at the right time(s) to the right patient(s).

[I have to catch a flight to the airport, so can’t stay for the whole talk…. d’oh]

@kennamshaw

Very narrow definition of precision medicine:  Use NGS to find patients who may respond better to one drug or another, or be resistant to a class of drugs: just matching patients to drugs.

Precision medicine is completely aspirational for patients.  We still do a bad job of figuring out how to match patients with drugs.  Right now, we don’t do it well – or at all.

We’re all bad at it, actually.

  • which patients should get tested?
  • use data to impact care
  • demonstrating data changes oucome
  • deciding how much of genome to sequence
  • how do we pay for it?

Why was MD anderson bad at it?

Patients concerned about, are those who have exhausted standard therapies, for instance.

Drop in cost leads to increases in data generation.  We all suck at using this data to impact outcome for patient.   MD Anderson was only able to impact 11% of patients with potentially actionable information.

Whole exome at other institutes were getting 5% (Beltran et al)

There are only 125 “actionable” genes.

NGS is not sufficient or necessary to drive personalized medicine.

Why?

  • solid tumours, behind liquid tumours because it’s hard to get the DNA.
  • Accessibility  – timing of data
  • Attitudes of doctors as well.

Leukaemia docs also use the molecular signature as well as other data to clarify.  Solid tumour docs do not.

Ignoring copy number, only 40% of patients have actionable variants.  (goes way up with copy number.)

Clinical trials categorized by type of match – even broadly, that’s 11% of patients.  Lack of enrolment not due to lack of available matched trials.

[Ok… time to go… alas, can’t stay to see the end of this talk.]

Leave a Reply

Your email address will not be published. Required fields are marked *