Dose Titration Algorithm Tuning (DTAT) — MTDi vs 'the' MTD

[dcnorris]

• General Topic: DTAT, as per below
• Presenter: David C. Norris, MD (Precision Methodologies, LLC)
• Location: YouTube Live Webinar https://www.youtube.com/watch?v=ANUeRqYO-o4
• Time: Friday April 7, 1pm-2pm EDT (10am-11am PDT)
• Background Resources for the Audience: Please read the newly-published version 2 of my paper.

Despite advances in Bayesian adaptive designs and model-based dose-finding, oncology dose-finding studies remain conceptually in the thrall of ‘the’ maximum tolerated dose (‘the’ MTD). This concept stands opposed to the long-recognized heterogeneity of cancer patients’ pharmacokinetics and pharmacodynamics (PK/PD), and to the diversity of their individual values and goals of care. Under this conceptual yoke, these dose-finding studies constitute a significant choke-point in drug development, where a severe discount may be applied to the potential value in new therapeutics through the hobbling of subsequent ‘efficacy’ trials by inadequate individual-level dosing.

Strangely, Bayesian innovation in dose-finding studies has proceeded apace without issuing a meaningful challenge to the inherently frequentist conception of an MTD as determined by whole-cohort frequencies of dose-limiting toxicities (DLTs). Thus, even as Bayesianism has made progress toward the ethical imperative of efficient use of data in such studies, it has neglected to confront the distinct ethical dimension of individualism. This seems a great irony, as the dynamic learning model of Bayesianism is equally suited, and indeed equally essential, to solving the latter problem.

Dose Titration Algorithm Tuning (DTAT) aims to break this impasse by introducing a new abstraction—the dose titration algorithm (DTA) together with its tuning parameters—to replace the fallacious abstraction of ‘the’ MTD. Crucially, this new abstraction has the capacity to embody objectively the knowledge we wish to acquire in dose-finding studies, a capacity which ‘the’ MTD most emphatically lacks. Indeed, the DTAT principle reveals ‘dose-finding study’ as a misnomer, the proper concept being rather that of dose titration algorithm-finding.

In a recently published F1000Research paper, I illustrate the DTAT principle through a simulation study in which a cytotoxic chemotherapy drug (modeled notionally on docetaxel, using published model parameters for PK and myelosuppression models) is titrated to a target neutrophil nadir. Connections are drawn with recursive filtering and optimal control, which have served important heuristic functions in the conception and development of DTAT. (The Tuning in DTAT in fact originates with the use of this term in connection with the practice of ‘tuning’ a Kalman filter.)

The webinar will span a range of issues, including:

• Evidence underscoring my vendetta against ‘the’ MTD
• Review of the DTAT principle as implemented in the paper
• Applicability of DTAT in immuno-oncology and molecularly targeted agents (MTAs)
• Philosophical and epistemological connections
• Open problems of DTAT, necessitatingr cross-disciplinary collaboration

I plan to leave a full 15 minutes for answering audience questions during and after the presentation. An engaged audience will contribute as much to this presentation as I will. I would be most interested to hear, for example, whether audience members believe an implementation in Stan or other more standard tools of pharmacometrics would be feasible.

[vjd]

@dcnorris thank you. Please update the issue with a date and time once you know you are ready so that I can update the calendar and send out an announcement!

[dcnorris]

@vjd Does one solicit interest and adjust webinar time for audience availability, or should I just set a time and roll with it? (This latter practice looks rather too much like the way we set doses! 8^)

[vjd]

:) just roll with it. We suggest lunch-time seminars on Friday so that we can have a bigger audience. We market the webinar so that many people can attend. Importantly, since the platform is youtube based, all webinars will be recorded and viewed multiple times based on user convenience.

[dcnorris]

Fantastic. I'll make it a 1-2pm late lunch Fri 3/31, then, on U.S. East Coast. That's a 10am-11am granola/yogurt break here on West Coast.

[vjd]

perfect 👍

[dcnorris]

I've just updated the proposed date to Fri, April 7 1-2pm EDT. I expect v2 will publish by this coming Friday, and that will give attendees a week to read it. (Will post link when v2 publishes.)

[dcnorris]

The proposal is now complete, with YouTube Live URL and a link to v2 of the DTAT paper.