Method and system for measuring a maturation stage using MRI

A computer-implemented method for providing a trained function for estimating the maturation stage of a biological organ, the method comprising:

• receiving an input dataset through a first interface, the input dataset including several quantitative MR (Magnetic Resonance) maps of different types for the organ in a representative population, each quantitative MR map including voxels having a quantitative value representing a quantitative measurement of a physical or physiological property, being a physical parameter, of a tissue of the biological organ for a subject of the representative population;
• receiving through a second interface an output dataset, the output dataset including, for each quantitative MR map, the age of the subject for which the quantitative MR map has been obtained;
• automatically creating an input training dataset, wherein a segmentation process is applied to each quantitative MR map, the segmentation process being configured to classify the different tissues included within the organ in order to segment the organ into different regions, each region characterized by a single organ tissue type, at least one characteristic quantitative value being determined for each region of each quantitative MR map from the voxels of the quantitative MR map, and the input training dataset being populated with the characteristic quantitative values obtained for each region of each quantitative MR map;
• training a function based on the input training dataset and the output dataset with a training computation unit, using the several quantitative MR maps as input to the trained function for calculation of the age, the function being trained to estimate the organ maturation stage from the characteristic quantitative value(s) determined for each region, or a set of preferred regions, and from the age of the subject for which the characteristic quantitative value has been obtained;
• the segmentation process being applied before applying the trained function;
• and
• providing the trained function with a third interface.

A training system for training a function for estimating the maturation stage of an organ, the training system comprising:

• a first interface configured for receiving an input dataset, the input dataset including several quantitative MR (Magnetic Resonance) maps of different types for the organ in a representative population, each quantitative MR map including voxels having a quantitative value representing a quantitative measurement of a physical or physiological property, being a physical parameter, of a tissue of the biological organ for a subject of the representative population;
• a second interface configured for receiving an output dataset, the output dataset including, for each quantitative MR map, an age of the subject for which the quantitative MR map has been obtained;
• a training computation unit configured for creating an input training dataset from the input dataset, wherein a segmentation process is applied to each quantitative MR map, the segmentation process being configured to classify the different tissues included within the organ in order to segment the organ into different regions, each region characterized by a single organ tissue type;
• said training computation unit additionally configured to train a function based on the input training dataset and the output dataset, the training using a machine learning algorithm configured for estimating the organ maturation stage, the several quantitative MR maps being applied as input to the trained function for calculation of the age, the function being configured to calculate a characteristic quantitative value for each region of the organ within each quantitative MR map, and the function being trained to estimate the organ maturation stage from the characteristic quantitative value determined for each region or a set of preferred regions;
• said training computation unit configured to apply the segmentation process before applying the trained function;
• and
• a third interface for providing the trained function.

A system for automatically measuring a maturation stage of a biological organ of a subject, the system comprising:

• a connection to an MRI (Magnetic Resonance Imaging) apparatus configured for acquiring several quantitative MR (Magnetic Resonance) maps of different types for the organ;
• a memory for storing each of the acquired quantitative MR maps;
• a control unit including a processor, said control unit being configured for automatically estimating the maturation stage of the organ of the subject;
• said control unit being configured for applying a function trained by a machine learning algorithm to the acquired quantitative MR maps in order to output an estimation of the maturation stage of the organ, the several quantitative MR maps being applied as input to the trained function for calculation of the age;
• said control unit being configured to apply to each of the acquired quantitative MR maps a segmentation process configured for segmenting said biological organ into different regions each characterized by a single organ tissue type;
• said trained function being configured to calculate a characteristic quantitative value for each region of the organ within each of said quantitative MR maps, said trained function being trained to estimate the organ maturation stage from the characteristic quantitative value determined for each region or a set of preferred regions;
• said segmentation process being applied before applying the trained function;
• and
• a display for displaying the estimated maturation stage obtained for the organ.