Ispitivanja uticaja dizajna modela i parametara 3D štampe na brzinu rastvaranja lekovite supstance iz tableta dobijenih tehnikom deponovanja istopljenog filamenta
Investigation of the influence of model design and 3D printing parameters on model drug dissolution from tablets obtained by fused deposition modelling
Докторанд
Obeid, SamihaМентор
Ibrić, SvetlanaЧланови комисије
Parojčić, JelenaMedarević, Đorđe
Kovačević, Jovana
Метаподаци
Приказ свих података о дисертацијиСажетак
3D štampanje lekova predstavlja napredan pristup za obezbeđenje personalizovane terapije u skladu sa
potrebama individualnih pacijenata. Mogućnost primene različitih tehnika 3D štampe, izbor pogodnih
materijala i prevazilaženje postojećih izazova predmet su intenzivnih naučnih istraživanja. Cilj ovog
naučnog istraživanja je ispitivanje mogućnosti primene tehnologije 3D štampe u proizvodnji čvrstih
farmaceutskih oblika dobijenih tehnikom deponovanja istopljenog filamenta (engl. Fused Deposition
Modelling, FDM). Posebna pažnja posvećena je ispitivanje mogućnosti pripreme filamenata sa
diazepamom i amplodipinom kao model lekovitim supstancama tehnikom ekstruzije topljenjem (engl.
Hot Melt Extrusion, HME) s ciljem njihove primene kao materijala za punjenje u 3D FDM štampi.
Uticaj dizajna modela i parametara 3D štampe na brzinu rastvaranja lekovite supstance iz odštampanih
farmaceutskih oblika analiziran je primenom naprednih alata za mašinsko učenje.
Metodom ekstruzije topljenjem (HME), ...uz primenu polivinil alkohola (PVA) kao osnovnog polimera,
bez, kao i uz dodatak natrijum-skrobglikolata i/ili hipromeloze bilo je moguće izraditi filamente
ujednačenog prečnika, glatke površine i odgovarajućih mehaničkih karakteristika pogodnih za 3D
štampu.
Ispitivanje uticaja dizajna modela i parametara 3D štampe na brzinu rastvaranja lekovite supstance iz
štampanih tableta pokazalo je da se optimizacijom odnosa površine i zapremine (SA/V) štampanih
objekata, gustine punjenja i obrasca štampe može postići ciljani profil oslobađanja lekovite supstance.
Promenom gustine punjenja postiže se promena mase tablete, a time i doza aktivne supstance po
tableti, bez promene dimenzija tablete. Ovo je od izuzetnog značaja prilikom prilagođavanja doze
lekovite supstance potrebama individualnog pacijenta, jer omogućava da se isključivo podešavanjima
softvera i parametara štampe podesi doza leka u tabletama istog oblika i veličine, izrađenim od istog
filamenta. Najbrže oslobađanje leka je postignuto korišćenjem cik-cak obrasca štampe, smanjenjem
debljine zida tablete i uz dodatak najtrijum-skrobglikolata.
Primenom samoorganizovane mape (SOM) i neuronske mreže tipa višeslojnog perceptrona (MLP) kao
naprednih alata za duboko učenje procenjen je uticaj SA/V odnosa i parametara štampanja (gustina
punjenja i obrasca štampe) na oslobađanje diazepama iz štampanih tableta. MLP je obučen korišćenjem
back propagation algoritma i imao je tri sloja (sa strukturom mreže 2-3-5). Dobijeni rezultati su
pokazali da veći SA/V odnos, manja gustina punjenja (manje od 50%) i cik-cak obrazac štampe dovode
do bržeg oslobađanja lekovite supstance. Poređenje predviđenih i eksperimentalno dobijenih profila
rastvaranja diazepama iz ispitivanih formulacija pokazalo je da razvijeni veštačke neuronske mreže
(engl. Artificial neural networks, ANN) model može da uspešno predvidi profil oslobađanja leka.
Obučena MLP mreža je omogućila uspostavljanje prostora za dizajn (engl. design space) formulisanih
3D štampanih tableta diazepama uz predviđanje kinetike oslobađanja leka u zavisnosti od gustine
punjenja i odnosa SA/V, što predstavlja značajan naučni doprinos ovog istraživanja. U slučaju tableta
sa amlodipinom, samoorganizovane mape (SOM) su korišćene da se opiše uticaj ekscipijenasa i
obrazaca štampe na oslobađanje amlodipina iz štampanih tableta. Samoorganizovane mape su pokazale
da je najbrže oslobađanje amlodipina postignuto kada su korišćeni cik-cak obrazac štampe, uz dodatak
natrijum-skrobglikolata, dok dodatak hipromeloze nije značajno uticao na brzinu rastvaranja
amlodipina.
3D printing of drugs represents an advanced approach to provide personalized therapy according to the
needs of individual patients. The possibility of applying different 3D printing techniques, the selection
of suitable materials and overcoming existing challenges are the subject of intensive scientific research.
The goal of this scientific research is to investigate the possibility of applying 3D printing technology
in the production of solid pharmaceutical forms obtained by Fused Deposition Modelling (FDM).
Special attention was to examine the possibility of preparing filaments with diazepam and amlodipine
as model drug substances by Hot-melt extrusion (HME) with the aim of using them as feeding material
in FDM 3D printing. The influence of model design and 3D printing parameters on the dissolution rate
of drug substance from printed pharmaceutical forms was analyzed using advanced machine learning
tools.
By hot-melt extrusion (HME), it was possible to produce filaments with a unifor...m diameter, smooth
surface and suitable mechanical characteristics appropriate for 3D printing, using polyvinyl alcohol
(PVA) as the base polymer, with and without the addition of sodium starch glycolate and/or
hypromellose.
Examining the effect of model design and 3D printing parameters on the rate of dissolution of drug
substance from printed tablets showed that by optimizing the surface-to-volume ratio (SA/V) of printed
objects, infill density and infill pattern, a targeted drug substance release profile can be achieved.
By changing the infill density, a change in the mass of the tablet is achieved, and thus the dose of the
active substance per tablet, without changing the dimensions of the tablet. This is of extreme
importance when adjusting the dose of the drug substance to the needs of the individual patient,
because it allows to adjust the dose of the drug in tablets of the same shape and size, made of the same
filament, only by software settings and printing parameters. The fastest release of the drug was
achieved by using zigzag infill pattern, reducing the thickness of the tablet wall and with the addition of
sodium starch glycolate.
Using self-organizing map (SOM) and multi-layer perceptron (MLP) neural network as advanced deep
learning tools, the influence of SA/V ratio and printing parameters (infill density and infill pattern) on
the release of diazepam from printed tablets was evaluated. The MLP was trained using the back
propagation algorithm and had three layers (with a 2-3-5 network structure). The obtained results
showed that a higher SA/V ratio, a lower infill density (less than 50%) and a zigzag infill pattern lead
to a faster release of the drug substance. A comparison of the predicted and experimentally obtained
diazepam dissolution profiles from the investigated formulations showed that the developed Artificial
neural networks (ANN) model can successfully predict the drug release profile. The trained MLP
network enabled the establishment of a design space for formulated 3D printed tablets of diazepam
with the prediction of drug release kinetics depending on the infill density and the SA/V ratio, which
represents a significant scientific contribution of this research. In the case of amlodipine tablets, self-
organizing maps (SOMs) were used to describe the influence of excipients and infill patterns on the
release of amlodipine from printed tablets. Self-organized maps showed that the fastest release of
amlodipine was achieved when the zigzag infill pattern was used, with the addition of sodium starch
glycolate, while the addition of hypromellose did not significantly affect the dissolution rate of
amlodipine.