Machine Learning Modeling of Wet Granulation Scale-up Using Particle Size Distribution Characterization Parameters
- 24 Downloads
Optimal particle size distribution (PSD) is an important factor in wet granulation in order to achieve appropriate powder flow, compactibility, and content uniformity. Parameters like D50 and surface area (SA) are used to define PSD but both are only able to compare separate fractions of a granulate. In this work, we made an attempt to characterize PSD of a final dry granulate blend and suggest novel parameters (determination coefficient R2 and trend line slope of a PSD model) to quantitatively describe PSD.
The significance of these parameters was tested using machine learning. Laboratory-scale samples were used for training and commercial-scale samples for testing a model. Several machine learning techniques were used to further examine the importance of these input variables using a large data set from wet granulation scale-up study.
The Gradient Boosted Regression Trees (GBRT) algorithm had the lowest root mean square error (RMSE) values for the several responses studied (tablet tensile strength, tablet diameter and thickness, compaction work, decompaction work, and net work). The GBRT model for tablet tensile strength had an R2 model value of 0.87 and was not overfitted. The importance of input variables R2 and a was proven by the stepwise regression model’s p value (0.0003) and GBRT importance score (0.37 and 0.44, respectively). The GBRT model was the most successful in predicting decompaction work (R2 model = 0.97) with the least regularization effect.
The proposed parameters can be used in PSD characterization and applied in critical quality attributes (CQA) prediction and wet granulation scale-up.
KeywordsParticle size distribution Machine learning modeling Wet granulation Scale-up
We thank Wesley Stringer, Con Psalios, Lalit Khera, Snehal Malani, Danijel Canji, and Emanuel Millen for their support. We thank operators from Probiotec Ltd., Ursula Basilio, and Charlie Cutajar, for their assistance in commercial scale runs.
Aleksandar Kovačević was partially supported by the Grant No. III-47003 provided by the Ministry of Education, Science and Technological Development of the Republic of Serbia.
Compliance with Ethical Standards
Conflict of Interest
The authors declare that they have no conflict of interest.
- 2.Leuenberger H, Betz G. Scale-up in the field of granulation and drying. In: Levin M, editor. Pharmaceutical process scale-up. 2nd ed. New York: Informa Healthcare; 2005. p. 199–236.Google Scholar
- 3.Heng JY, Thielmann F, Williams DR. The effects of milling on the surface properties of form I paracetamol crystals. Pharm Res. 2006;23(8):1918–27.Google Scholar
- 4.Mullarney MP, Leyva N. Modeling pharmaceutical powder-flow performance using particle-size distribution data. Pharm Technol. 2009;33(3):126–34.Google Scholar
- 5.Virtanen S, Antikainen O, Räikkönen H, Yliruusi J. Granule size distribution of tablets. J Pharm Sci. 2010;99(4):2061–9.Google Scholar
- 6.Sun CC, Himmelspach MW. Reduced tabletability of roller compacted granules as a result of granule size enlargement. J Pharm Sci. 2006;95(1):200–6.Google Scholar
- 7.Sandler N, Wilson D. Prediction of granule packing and flow behavior based on particle size and shape analysis. J Pharm Sci. 2010;99(2):958–68.Google Scholar
- 8.Anand C. Optimization of granule size and disintegrants on formulation of rapid dispersible tablets of tolfenamic acid. J Drug Deliv Ther. 2013;3(2):31–4.Google Scholar
- 9.Prescott JK, Hoosfeld R. Maintaining product uniformity and uninterrupted flow to direct-compression tableting presses. Pharm Technol. 1994;18(6):98.Google Scholar
- 10.Shipar MAH, Wadhwa A, Varughese C, Kaur N, Niranjani T. Effect of particle size on tablet weight and hardness variation. Toronto: Toronto Institute of Phamaceutical Technology; 2014.Google Scholar
- 11.Rohrs BR, Amidon GE, Meury RH, Secreast PJ, King HM, Skoug CJ. Particle size limits to meet USP content uniformity criteria for tablets and capsules. J Pharm Sci. 2006;95(5):1049–59.Google Scholar
- 12.Otsuka M, Yamane I. Prediction of tablet properties based on near infrared spectra of raw mixed powders by chemometrics: scale-up factor of blending and tableting processes. J Pharm Sci. 2009;98(11):4296–305.Google Scholar
- 15.Sun C, Grant DJ. Effects of initial particle size on the tableting properties of L-lysine monohydrochloride dihydrate powder. Int J Pharm. 2001;215(1):221–8.Google Scholar
- 16.Eichie F, Kudehinbu A. Effect of particle size of granules on some mechanical properties of paracetamol tablets. Afr J Biotechnol. 2009;8(21):5913–6.Google Scholar
- 17.Zuurman K, Riepma K, Bolhuis G, Vromans H, Lerk C. The relationship between bulk density and compactibility of lactose granulations. Int J Pharm. 1994;102(1–3):1–9.Google Scholar
- 18.Riepma K, Vromans H, Zuurman K, Lerk C. The effect of dry granulation on the consolidation and compaction of crystalline lactose. Int J Pharm. 1993;97(1–3):29–38.Google Scholar
- 19.Sweat ML, Parker AS, Beaudoin SP. Compressive behavior of high viscosity granular systems: effect of particle size distribution. Powder Technol. 2017;311:506–13.Google Scholar
- 20.Wikberg M, Alderborn G. Compression characteristics of granulated materials II. Evaluation of granule fragmentation during compression by tablet permeability and porosity measurements. Int J Pharm. 1990;62(2–3):229–41.Google Scholar
- 21.Johansson B, Nicklasson F, Alderborn G. Effect of pellet size on degree of deformation and densification during compression and on compactability of microcrystalline cellulose pellets. Int J Pharm. 1998;163(1):35–48.Google Scholar
- 22.Alderborn G, Nyström C. Studies on direct compression of tablets. IV. The effect of particle size on the mechanical strength of tablets. Acta Pharm Suec. 1982;19(5):381–90.Google Scholar
- 23.Alderborn G, Borjesson E, Glazer M, Nystrom C. Studies on direct compression of tablets. XIX: the effect of particle size and shape on the mechanical strength of sodium bicarbonate tablets. Acta Pharm Suec. 1988;25(1):31–40.Google Scholar
- 24.Vromans H, De Boer A, Bolhuis G, Lerk C, Kussendrager K. Studies on tabletting properties of lactose, part I. the effect of initial particle size on binding properties and dehydration characteristics of lactose. Acta Pharm Suec. 1984;22(3):163–72.Google Scholar
- 25.Šantl M, Ilić I, Vrečer F, Baumgartner S. A compressibility and compactibility study of real tableting mixtures: the effect of granule particle size. Acta Pharma. 2012;62(3):325–40.Google Scholar
- 26.Johansson B, Alderborn G. The effect of shape and porosity on the compression behaviour and tablet forming ability of granular materials formed from microcrystalline cellulose. Eur J Pharm Biopharm. 2001;52(3):347–57.Google Scholar
- 27.Gao JZ, Jain A, Motheram R, Gray D, Hussain M. Fluid bed granulation of a poorly water soluble, low density, micronized drug: comparison with high shear granulation. Int J Pharm. 2002;237(1):1–14.Google Scholar
- 28.Herting MG, Kleinebudde P. Roll compaction/dry granulation: effect of raw material particle size on granule and tablet properties. Int J Pharm. 2007;338(1):110–8.Google Scholar
- 29.Badawy SIF, Hussain MA. Effect of starting material particle size on its agglomeration behavior in high shear wet granulation. AAPS PharmSciTech. 2004;5(3):16–22.Google Scholar
- 30.Badawy SIF, Lee TJ, Menning MM. Effect of drug substance particle size on the characteristics of granulation manufactured in a high-shear mixer. AAPS PharmSciTech. 2000;1(4):55–61.Google Scholar
- 31.Fichtner F, Rasmuson Å, Alderborn G. Particle size distribution and evolution in tablet structure during and after compaction. Int J Pharm. 2005;292(1):211–25.Google Scholar
- 32.Hilden J, Schrad M, Kuehne-Willmore J, Sloan J. A first-principles model for prediction of product dose uniformity based on drug substance particle size distribution. J Pharm Sci. 2012;101(7):2364–71.Google Scholar
- 33.Aksu B, Paradkar A, de Matas M, Özer Ö, Güneri T, York P. A quality by design approach using artificial intelligence techniques to control the critical quality attributes of ramipril tablets manufactured by wet granulation. Pharm Dev Technol. 2013;18(1):236–45.Google Scholar
- 34.Hand DJ. Principles of data mining. Drug Saf. 2007;30(7):621–2.Google Scholar
- 35.Petrović J, Chansanroj K, Meier B, Ibrić S, Betz G. Analysis of fluidized bed granulation process using conventional and novel modeling techniques. Eur J Pharm Sci. 2011;44(3):227–34.Google Scholar
- 36.Ibrić S, Djuriš J, Parojčić J, Djurić Z. Artificial neural networks in evaluation and optimization of modified release solid dosage forms. Pharmaceutics. 2012;4(4):531–50.Google Scholar
- 37.Huang C-Y. Ku MS. prediction of drug particle size and content uniformity in low-dose solid dosage forms. Int J Pharm. 2010;383(1):70–80.Google Scholar
- 38.MathWorks. Linear Regression. 1994-2019 The MathWorks, Inc., www.mathworks.com; 2019. https://au.mathworks.com/help/matlab/data_analysis/linear-regression.html#bswinlz. Accessed 04 Feb 2019.
- 39.Chen S, Ai X, Dong T, Li B, Luo R, Ai Y, et al. The physico-chemical properties and structural characteristics of artificial soil for cut slope restoration in southwestern China. Sci Rep. 2016;6:20565.Google Scholar
- 40.MathWorks. Stepwise Regression. 1994–2019 The MathWorks, Inc., www.mathworks.com. 2019. https://au.mathworks.com/help/stats/stepwise-regression.html. Accessed 21 Mar 2019.
- 41.Hoerl AE, Kennard RW. Ridge regression: biased estimation for nonorthogonal problems. Technometrics. 1970;12(1):55–67.Google Scholar
- 42.Ogutu JO, Schulz-Streeck T, Piepho H-P. Genomic selection using regularized linear regression models: ridge regression, lasso, elastic net and their extensions. BMC Proc. 2012: BioMed Central;6.Google Scholar
- 43.James G, Witten D, Hastie T, Tibshirani R. An introduction to statistical learning. New York: Springer; 2013.Google Scholar
- 44.Vestal B. Re-analysis of mortality rates using linear regression models; 2012.Google Scholar
- 45.Tibshirani R. Regression shrinkage and selection via the lasso. J R Stat Soc Ser B Methodol. 1996:267–88.Google Scholar
- 46.Friedman J, Hastie T, Tibshirani R. Regularization paths for generalized linear models via coordinate descent. J Stat Softw. 2010;33:1–22.Google Scholar
- 47.Zou H. The adaptive lasso and its oracle properties. J Am Stat Assoc. 2006;101(476):1418–29.Google Scholar
- 48.Zou H, Hastie T. Regularization and variable selection via the elastic net. J R Stat Soc Ser B Stat Methodol. 2005;67(2):301–20.Google Scholar
- 49.Trevor H, Robert T, JH F. The elements of statistical learning: data mining, inference, and prediction. New York: Springer; 2009.Google Scholar
- 50.Quinlan JR. Simplifying decision trees. Int J Man Mach Stud. 1987;27(3):221–34.Google Scholar
- 51.Friedman J. Greedy Function Approximation: A Gradient Boosting Machine http://www.salford-systems.com/doc/GreedyFuncApproxSS.pdf; 1999. Accessed 25 Mar 2019.
- 52.Carty DM. An analysis of boosted regression trees to predict the strength properties of wood composites; 2011.Google Scholar
- 53.Fernández-Delgado M, Cernadas E, Barro S, Amorim D. Do we need hundreds of classifiers to solve real world classification problems? J Mach Learn Res. 2014;15:3133–81.Google Scholar
- 54.Jacobs MC. Regression trees versus stepwise regression. UNF Graduate Theses and Dissertations. 145.: University of North Florida. 1992.Google Scholar
- 55.MathWorks. PredictorImportance. MathWorks; 2018. https://au.mathworks.com/help/stats/compactregressionensemble.predictorimportance.html. Accessed 25 Mar 2019.
- 56.Caruana R, Niculescu-Mizil A. An empirical comparison of supervised learning algorithms. In: Proceedings of the 23rd international conference on machine learning. New York City: ACM; 2006.Google Scholar
- 57.Bardin M, Knight P, Seville J. On control of particle size distribution in granulation using high-shear mixers. Powder Technol. 2004;140(3):169–75.Google Scholar
- 58.Breneman B. Effect of size reduction parameters in pharmaceutical manufacturing process. San Luis, Obispo: California Polytechnic State University; 2011.Google Scholar
- 59.Schenck LR, Plank RV. Impact milling of pharmaceutical agglomerates in the wet and dry states. Int J Pharm. 2008;348(1):18–26.Google Scholar
- 60.Rojas JJ, Aristizabal J, Henao M. Screening of several excipients for direct compression of tablets: a new perspective based on functional properties. Rev Ciênc Farm Básica Apl 2013;34(1):17–23. https://portal.issn.org/resource/issn/2179-443X
- 61.Garr J, Rubinstein M. The effect of rate of force application on the properties of microcrystalline cellulose and dibasic calcium phosphate mixtures. Int J Pharm. 1991;73(1):75–80.Google Scholar
- 62.Zhang Y, Wrzesinski A, Moses M, Bertrand H. Comparison of Superdisintegrants in orally disintegrating tablets. Pharm Technol. 2010;34(7):54–65.Google Scholar
- 63.Thompson B. Stepwise regression and stepwise discriminant analysis need not apply here: a guidelines editorial. Thousand Oaks: Sage Publications; 1995.Google Scholar
- 64.Trevor Hastie JQ. Glmnet Vignette. 2014. https://web.stanford.edu/~hastie/glmnet/glmnet_alpha.html. Accessed 21 Jan 2019.