Félicien Le Louedec

Pharmacometrics.

[Félicien Le Louedec @FelicienLL]

New release: mapbayr 0.10.0 is on CRAN! https://t.co/wrSTxRdDKt
Define your model in mrgsolve and make Maximum A Posteriori Bayesian estimation of individual parameters. This new version comes with useful features to help you in your Model-Informed Precision Dosing projects.

[InsightRX @InsightRX]

Researchers from @MSKCancerCenter & InsightRX collaborated on a recent study that proposed a novel hybrid #pharmacokinetic / #pharmacodynamic /machine learning approach, applied to predict absolute neutrophil counts during chemotherapy: https://t.co/yKMA9D81ej | @ASCPT_ClinPharm

[Félicien Le Louedec @FelicienLL]

The package mapbayr is developed on github where suggestions, issues and any input from the Pharmacometrics community are welcome. Feel free to ask any question about your project! https://t.co/lN4AhhN1jl

[Félicien Le Louedec @FelicienLL]

New release: mapbayr 0.10.0 is on CRAN! https://t.co/wrSTxRdDKt
Define your model in mrgsolve and make Maximum A Posteriori Bayesian estimation of individual parameters. This new version comes with useful features to help you in your Model-Informed Precision Dosing projects.

[Félicien Le Louedec @FelicienLL]

2) Prediction-corrected Visual Predictive Checks. Overlap the observations from your dataset with stochastic simulations from your model and identify if the latter is coded correctly and/or suitable for MIPD. Just call `mapbayr_vpc()`. More options in the documentation.

[Félicien Le Louedec @FelicienLL]

https://t.co/sCSFfOckMu

[Félicien Le Louedec @FelicienLL]

Corruption attempt on my person with Hex Stickers from @mattfidler at the #PAGEmeeting… Sorry but I think I will stick with my NONMEM / mrgsolve / mapbayr framework 🤓. Still, kudos to the @nlmixr team for their amazing job 👍

[Félicien Le Louedec @FelicienLL]

Let’s Go ! The 31st PAGE Meeting us just about to start!

[Félicien Le Louedec @FelicienLL]

@idg_pharma @lemondefr Ben… si?

[Ron Keizer @ronpirana]

It was an honor to talk about #AI in #PrecisionDosing at the 3rd Annual World Precision Dosing Seminar, organized by @CRE_RESPOND @jasonroberts_pk @browsingcotta. The talk is now online: https://t.co/4GCQWdL0IC

[Mike K Smith @MikeKSmith]

Really happy to have my presentation accepted for #positconf2023. I'll be talking about open-source in pharma and how things in clinical trial reporting have changed over the last 30 years...

[Félicien Le Louedec @FelicienLL]

Code a model and simulate in mrgsolve --> MAP-estimate with mapbayr (https://t.co/lN4AhhN1jl) --> correct the bias with ML infrastructure (i.e. xgboost/tidymodels).

[Félicien Le Louedec @FelicienLL]

Our colleagues from the @Woillard4 team (@CHULimoges) just published a nice paper in Pharmaceutical Research about combining Machine Learning and classical MAP Bayesian analyses in the context of dose individualization of isavuconazole.

https://t.co/GXytf0cIPg

[Félicien Le Louedec @FelicienLL]

Of note, they used mapbayr to obtain MAP estimates of PK parameters, from simulated data in mrgsolve. It is a nice example of streamlined analyses fully performed in R, and should be inspiring for future researches on improving model-informed precision dosing practices!

[Félicien Le Louedec @FelicienLL]

This is an encouraging study for the generalization of dose-banding to other cytotoxics. Read more at https://t.co/tMAeABrnoC

[Félicien Le Louedec @FelicienLL]

New publication in @BritJClinPharm! In this study @florent_puisset reports the impact of dose-banding of paclitaxel on clinical outcomes.
Cytotoxic drugs are often dosed in mg/m², but the relationship between body surface area and clearance is not so close.

[Félicien Le Louedec @FelicienLL]

The benefits are numerous: easer, cheaper and safer to prepare etc... And in our study, we show that there is not more toxicity in patients for whom the dose was rounded up, and same efficacy for for whom the dose was rounded down.👍

[Félicien Le Louedec @FelicienLL]

6) Visit us on https://t.co/lN4AhhNz8T to report bugs, suggest features, and ask for any question related to your project!

[Félicien Le Louedec @FelicienLL]

mapbayr 0.9.0 is now on CRAN! https://t.co/r2332MMl6A
Define your model in @mrgsolve and perform MAP-Bayesian estimations. Many nice new features to help you in your model-informed precision dosing projects! #rstats

[Félicien Le Louedec @FelicienLL]

5) Select the ETAs to estimate thanks to `mapbayest(select_eta = ...)`. Run times increase exponentially with the number of ETAs to estimate, so it is sometimes useful to ignore non-identifiable ETA or if your model have a lot of OMEGA due to inter-occasion.