open access resource
for quantitative prediction
biomedical properties of
magnetic nanoparticles for
MRI and hyperthermia treatment
for quantitative prediction
biomedical properties of
magnetic nanoparticles for
MRI and hyperthermia treatment
Medical treatment using magnetic nanoparticles have already shown perspective results in medical practice.
As an example, Feridex® and Resovist® have been approved as MRI contrast agents for
liver imagining; number of other magnetic nanoparticles for MRI have undergone various stages of clinical trials.
Moreover, magnetic nanoparticles have recently been used in medical trials for hyperthermia treatment glioblastoma,
prostate and pancreatic cancer. This was made possible thanks to the fact that performance of magnetic nanoparticles in
biomedical applications can be controlled adjusting size, shape, surface, composition of a material and parameters of
applied electromagnetic field.
.png)
The developed DiMag recourse gives access to Machine Learning models for prediction biomedical performance of magnetic nanoparticles on different levels of user requests: LGBM Regressor and ExtraTrees Regressor models are used for predicting Specific Absorption Rate (SAR) value (base, progressive and advanced) and r1/r2 relaxivities (base, progressive and advanced) respectively with sufficient accuracy: R2 = 0.86 for prediction SAR value, 0.72 for r1 relaxivity and 0.71 for r2 relaxivity. Moreover, the DiMag recourse contents a built-in expandable database of magnetic nanoparticles for MRI contrast enhancement (r1/r2 relaxivities) as well as hyperthermia treatment with links to sources of information and interactive visualization tool.\
Also, we consider it necessary to point out some clarifications regarding the correct working of the DiMag:
· Created algorithms works correctly only with obviously magnetic systems. Misusing ML models will lead to the incorrect predictions;
· If you screen several nanoparticles in a row, is better to reload the web page using "Ctrl + F5" combination;
· If you have "Error loading layout", please, wait a bit time and reload the web page;
· Error 500 means that at least one of the elements in your chemical formula is not included in the database which was used for training models. Please, offer you sample for extension abilities of ML algorithms.
The developed DiMag recourse gives access to Machine Learning models for prediction biomedical performance of magnetic nanoparticles on different levels of user requests: LGBM Regressor and ExtraTrees Regressor models are used for predicting Specific Absorption Rate (SAR) value (base, progressive and advanced) and r1/r2 relaxivities (base, progressive and advanced) respectively with sufficient accuracy: R2 = 0.86 for prediction SAR value, 0.72 for r1 relaxivity and 0.71 for r2 relaxivity. Moreover, the DiMag recourse contents a built-in expandable database of magnetic nanoparticles for MRI contrast enhancement (r1/r2 relaxivities) as well as hyperthermia treatment with links to sources of information and interactive visualization tool.\
Also, we consider it necessary to point out some clarifications regarding the correct working of the DiMag:
· Created algorithms works correctly only with obviously magnetic systems. Misusing ML models will lead to the incorrect predictions;
· If you screen several nanoparticles in a row, is better to reload the web page using "Ctrl + F5" combination;
· If you have "Error loading layout", please, wait a bit time and reload the web page;
· Error 500 means that at least one of the elements in your chemical formula is not included in the database which was used for training models. Please, offer you sample for extension abilities of ML algorithms.