Napovedovanje koncentracij takrolimusa v polni krvi bolnikov po presaditvi pljuč z metodami strojnega učenja

Marzidovšek, Špela (author), Grabnar, Iztok (mentor), Bürmen, Božena (co-mentor)

Takrolimus je eden izmed najpogosteje uporabljenih zaviralcev kalcineurina in se uporablja za preprečevanje zavrnitve organa po presaditvah. Vključen je v številne režime zdravljenja, vendar ima zelo ozko terapevtsko okno in zahteva terapevtsko spremljanje zdravila. Prenizke krvne koncentracije lahko vodijo do zavrnitve organa, medtem ko previsoke povečajo verjetnost pojava neželenih učinkov, npr. nefrotoksičnosti. V farmakokinetiki izkazuje zelo visoko intra- in interindividualno variabilnost, na katero vplivajo številni genetski, klinični in demografski dejavniki ter interakcije med sočasno uporabljenimi zdravili. V okviru magistrske naloge smo želeli preveriti možnost uporabe metod strojnega učenja za napovedovanje koncentracij takrolimusa v polni krvi pri pacientih po presaditvi pljuč, kar bi lahko olajšalo načrtovanje režima odmerjanja. V sodelovanju z Univerzitetnim kliničnim centrom Ljubljana smo izvedli retrospektivno raziskavo, v katero smo vključili 54 pacientov po presaditvi pljuč. V analizo smo vključili podatke o koncentracijah takrolimusa v polni krvi in pripadajočih odmerkih, demografske dejavnike, sočasno uporabljena zdravila, biokemične dejavnike in rezultate genotipizacije. Za namene napovedovanja smo uporabili javno dostopne zbirke programov strojnega učenja v odprtokodnem programskem okolju R, primerjali smo rezultate med umetno nevronsko mrežo in nizom odločitvenih dreves (XGBoost) na dveh različno pripravljenih skupinah podatkov. Najboljše rezultate smo dosegli z metodo XGBoost, in sicer R2 0,52 , RMSE 2,48, MAE 1,63 in MAPE 18,57 %, 67,03 % napovedi pa je bilo znotraj 20 % odstopanja od izmerjene koncentracije takrolimusa v polni krvi. Za optimizacijo analize in izboljšanje natančnosti napovedovanja bi bilo v prihodnosti smiselno razviti individualiziran računalniški program in izboljšati način zbiranja podatkov o pacientih. Oba razvita modela sta bila statična, medtem ko so naši podatki dinamični, saj smo paciente spremljali skozi daljše obdobje zdravljenja, zato bi bilo smiselno preveriti možnost uporabe metod, ki temeljijo na časovnih zaporedjih.

• Language: Slovenian
• Keywords: farmakokinetika, presaditev pljuč, strojno učenje, takrolimus
• Work type: Master's thesis/paper
• Organization: FFA - Faculty of Pharmacy
• Year: 2022
• PID: 20.500.12556/RUL-143532
• Publication date in RUL: 24.12.2022

Secondary language

• Language: English
• Title: Prediction of tacrolimus whole blood concentrations in lung transplant patients using machine learning methods

Abstract

Tacrolimus is one of the most widely used calcineurin inhibitors for prevention of organ rejection after transplantation. It is included in numerous therapeutic regimens; however, it's use is limited by its narrow therapeutic window, which calls for therapeutic drug monitoring. Insufficient plasma concentrations can lead to organ rejection, whereas excessive concentrations lead to an increase in adverse effects, such as nephrotoxicity. Tacrolimus has high intra- and interindividual pharmacokinetic variability, which is influenced by numerous genetic, clinical and demographic factors, along with drug-drug interactions. In the context of this master’s thesis, we aimed to evaluate the possibility of using machine-learning methods for predicting tacrolimus whole blood concentration in patients post lung transplantation, which could improve the dosing regimens design. In collaboration with University Medical Centre Ljubljana we performed a retrospective study, which included 54 patients post lung transplantation. Our analysis included tacrolimus concentrations in whole blood and corresponding doses, demographic factors, comedication, biochemical parameters and genotyping results. For tacrolimus concentration prediction, we used publicly accessible machine learning packages in open-source software environment R, and compared results between an artificial neural network model and gradient-boosted trees (XGBoost) on two differently prepared data sets. We achieved better results using XGBoost, which were R2 0,52 , RMSE 2,48, MAE 1,63 in MAPE 18,57 %, 67,03 % of predicted values were within 20 % deviation from measured whole blood tacrolimus concentrations. Going forward, an individualized software programme and optimised data collection would be needed to improve our data analysis and prediction accuracy. Both developed models were static, and because our data is dynamic as we followed the patients over a longer time period, machine learning methods appropriate for time series analysis might yeild better results.

• Keywords: lung transplantation, machine learning, pharmacokinetics, tacrolimus