Methods, systems, apparatuses, and devices for facilitating a diagnosis of pathologies using a machine learning model

1 . A method for facilitating a diagnosis of pathologies using at least one machine learning model, the method comprising:

receiving, using a communication connection, at least one medical data associated with at least one user from at least one device;

receiving, using the communication connection, at least one indication corresponding to at least one condition from the at least one device;

identifying, using a processing unit, a plurality of clinician-annotated echocardiography reports and a plurality of echocardiography data associated with the plurality of clinician-annotated echocardiography reports from at least one data source based on the at least one indication;

extracting, using the processing unit, at least one category from the plurality of clinician-annotated echocardiography reports using at least one natural language processing model;

training, using the processing unit, one or more machine learning models using the at least one category and the plurality of echocardiography data;

selecting, using the processing unit, the at least one machine learning model from the one or more machine learning models based on the at least one medical data;

analyzing, using the processing unit, the at least one medical data using the at least one machine learning model, wherein the analyzing of the at least one medical data using the at least one machine learning model is further based on the selecting, wherein the at least one machine learning model comprises at least one artificial neural network comprising a plurality of layers, wherein the plurality of layers comprises at least one input layer, a plurality of output layers, and at least one middle layer between the at least one input layer and the plurality of output layers, wherein the at least one input layer is configured for taking a plurality of medical data as a plurality of inputs to the at least one input layer, wherein the at least one middle layer is configured for outputting a lower dimensional abstract vector space representation for each of the plurality of inputs by encoding the plurality of medical data to a lower dimensional abstract vector space, wherein the plurality of output layers is configured for classifying the lower dimensional abstract vector space representation to a plurality of outputs corresponding to a plurality of assessment parameters considered in the diagnosis of the pathologies, wherein each of the plurality of output layers is configured for communicating each of the plurality of outputs with at least one of the plurality of output layers, wherein each of the plurality of output layers is configured for adjusting each of the plurality of outputs based on the at least one of the plurality of outputs of at least one of the plurality of output layers, wherein the analyzing of the at least one medical data comprises inputting the at least one medical data to the at least one machine learning model;

obtaining, using the processing unit, one or more outputs corresponding to one or more assessment parameters from the at least one machine learning model for the diagnosis of one or more pathologies based on the analyzing of the at least one medical data;

generating, using the processing unit, at least one result based on the one or more outputs; and

storing, using a system memory, the at least one result and the at least one machine learning model.

2 . The method of claim 1 , wherein the at least one device comprises at least one medical imaging unit, wherein the at least one medical imaging unit is associated with at least one ultrasound imaging modality, wherein the at least one medical imaging unit is configured for generating the at least one medical data associated with the at least one user by imaging at least one portion of at least one body part of the at least one user, wherein the imaging comprises ultrasound imaging of the at least one portion of the at least one body part of the at least one user, wherein the at least one medical imaging unit comprises at least one transducer, wherein the at least one transducer generates and transmits ultrasound waves to the at least one portion of the at least one body part and receives echos of the ultrasound waves from the at least one portion of the at least one body part for the generating of the at least one medical data.

3 . The method of claim 1 , wherein the at least one medical data further comprises at least one echocardiography data corresponding to at least one study, wherein the analyzing of the at least one medical data further comprises analyzing the at least one echocardiography data using the at least one machine learning model, wherein the one or more outputs comprises one or more predictions of a presence of one or more conditions, one or more estimations of one or more measurements, one or more visualizations of the one or more measurements, and one or more conclusions of the diagnosis, wherein the obtaining of the one or more outputs comprises obtaining the one or more predictions of the presence of the one or more conditions, the one or more estimations of the one or more measurements, the one or more visualizations of the one or more measurements, and the one or more conclusions of the diagnosis from the at least one machine learning model, wherein the at least one result comprises an echocardiography report, wherein the generating of the at least one result comprises generating the echocardiography report by combining the one or more predictions of the presence of the one or more conditions, the one or more estimations of the one or more measurements, the one or more visualizations of the one or more measurements, and the one or more conclusions of the diagnosis.

4 . The method of claim 1 , wherein the at least one medical data further comprises at least one B mode image, wherein the analyzing of the at least one medical data further comprises analyzing the at least one B mode image using the at least one machine learning model, wherein the one or more outputs comprises one or more estimations of one or more Doppler echocardiographic measurements and one or more predictions of one or more conditions, wherein the obtaining of the one or more outputs comprises obtaining the one or more estimations of the one or more Doppler echocardiographic measurements and the one or more predictions of the one or more conditions from the at least one machine learning model, wherein the generating of the at least one result is further based on the one or more estimations of the one or more Doppler echocardiographic measurements and the one or more predictions of the one or more conditions.

5 . The method of claim 1 further comprising:

identifying, using the processing unit, a plurality of historical studies from at least one data source based on a time criterion, wherein each of the plurality of historical studies comprises at least one historical medical data and a historical result corresponding to the at least one historical medical data;

analyzing, using the processing unit, the plurality of studies using the at least one machine learning model;

generating, using the processing unit, a new result for each of the plurality of studies;

receiving, using the communication connection, at least one flagging criterion for flagging at least one of the plurality of studies from the at least one device;

identifying, using the processing unit, at least one historical study from the plurality of historical studies based on the at least one flagging criterion, wherein the new result of each of the at least one historical study disagrees with the historical result of each of the at least one historical study;

generating, using the processing unit, at least one warning based on the identifying of the at least one historical study, wherein the at least one warning indicates a misdiagnosis associated with the at least one historical study; and

transmitting, using the communication connection, the at least one warning to the at least one device.

6 . The method of claim 1 further comprising:

transmitting, using the communication connection, the at least one result to at least one clinician device associated with at least one clinician;

receiving, using the communication connection, at least one correction on the at least one result from the at least one clinician device;

updating, using the processing unit, the at least one result based on the at least one correction;

generating, using the processing unit, an updated result based on the updating;

retraining, using the processing unit, the at least one machine learning model based on the updated result; and

storing, using the system memory, the updated result.

7 . The method of claim 1 further comprising:

receiving, using the communication connection, at least one additional medical data associated with the at least one user from the at least one device, wherein the at least one medical data is associated with at least one modality and the at least one additional medical data is associated with at least one additional modality; and

analyzing, using the processing unit, the at least one additional medical data using the at least one machine learning model, wherein the at least one machine learning model further comprises at least one additional artificial neural network, wherein the at least one artificial neural network is trained using a set of training medical data associated with the at least one modality and the at least one additional artificial neural network is trained using a set of additional training medical data associated with the at least one additional modality, wherein the set of additional training medical data corresponds to the set of training medical data, wherein the at least one artificial neural network and the at least one additional artificial neural network are jointly trained with a cross modality artificial neural network of the at least one machine learning model using a loss function created using outputs from the cross modality artificial neural network, the at least one artificial neural network, and the at least one additional artificial neural network, wherein the jointly training of the at least one artificial neural network and the at least one additional artificial neural network with the cross modality artificial neural network leverages at least one data fusion technique to enhance an accuracy of the at least one machine learning model by combining the at least one modality of the set of training medical data with the at least one additional modality of the set of training medical data, wherein the obtaining of the one or more outputs is further based on the analyzing of the at least one additional medical data.

8 . The method of claim 1 further comprising:

receiving, using the communication connection, at least one user input from at least one input device;

processing, using the processing unit, the at least one result and the at least one user input using a conversational model;

generating, using the processing unit, at least one response corresponding to the at least one user input based on the processing;

transmitting, using the communication connection, the at least one response to at least one output device; and

storing, using the system memory, the at least one user input and the at least one response.

9 . The method of claim 1 , wherein the at least one medical data further comprises a freeform text corresponding to a note, wherein the at least one artificial neural network comprises at least one additional output layer, wherein the at least one additional output layer is configured for predicting text tokens by decoding the lower dimensional abstract vector space representation, wherein the obtaining of the one or more outputs further comprises obtaining at least one text token corresponding to at least one additional assessment parameter for the diagnosis of the one or more pathologies, wherein the generating of the at least one result is further based on the at least one text token.

10 . A system for facilitating a diagnosis of pathologies using at least one machine learning model, the system comprising:

a communication connection configured for:

receiving at least one medical data associated with at least one user from at least one device; and

receiving at least one indication corresponding to at least one condition from the at least one device:

a processing unit communicatively coupled with the communication connection, wherein the processing unit is configured for:

identifying a plurality of clinician-annotated echocardiography reports and a plurality of echocardiography data associated with the plurality of clinician-annotated echocardiography reports from at least one data source based on the at least one indication;

extracting at least one category from the plurality of clinician-annotated echocardiography reports using at least one natural language processing model;

training one or more machine learning models using the at least one category and the plurality of echocardiography data;

selecting the at least one machine learning model from the one or more machine learning models based on the at least one medical data;

analyzing the at least one medical data using the at least one machine learning model, wherein the analyzing of the at least one medical data using the at least one machine learning model is further based on the selecting, wherein the at least one machine learning model comprises at least one artificial neural network comprising a plurality of layers, wherein the plurality of layers comprises at least one input layer, a plurality of output layers, and at least one middle layer between the at least one input layer and the plurality of output layers, wherein the at least one input layer is configured for taking a plurality of medical data as a plurality of inputs to the at least one input layer, wherein the at least one middle layer is configured for outputting a lower dimensional abstract vector space representation for each of the plurality of inputs by encoding the plurality of medical data to a lower dimensional abstract vector space, wherein the plurality of output layers is configured for classifying the lower dimensional abstract vector space representation to a plurality of outputs corresponding to a plurality of assessment parameters considered in the diagnosis of the pathologies, wherein each of the plurality of output layers is configured for communicating each of the plurality of outputs with at least one of the plurality of output layers, wherein each of the plurality of output layers is configured for adjusting each of the plurality of outputs based on the at least one of the plurality of outputs of at least one of the plurality of output layers, wherein the analyzing of the at least one medical data comprises inputting the at least one medical data to the at least one machine learning model;

obtaining one or more outputs corresponding to one or more assessment parameters from the at least one machine learning model for the diagnosis of one or more pathologies based on the analyzing of the at least one medical data; and

generating at least one result based on the one or more outputs; and

a system memory communicatively coupled with the processing unit, wherein the system memory is configured for storing the at least one result and the at least one machine learning model.

11 . The system of claim 10 , wherein the at least one device comprises at least one medical imaging unit, wherein the at least one medical imaging unit is associated with at least one ultrasound imaging modality, wherein the at least one medical imaging unit is configured for generating the at least one medical data associated with the at least one user by imaging at least one portion of at least one body part of the at least one user, wherein the imaging comprises ultrasound imaging of the at least one portion of the at least one body part of the at least one user, wherein the at least one medical imaging unit comprises at least one transducer, wherein the at least one transducer generates and transmits ultrasound waves to the at least one portion of the at least one body part and receives echos of the ultrasound waves from the at least one portion of the at least one body part for the generating of the at least one medical data.

12 . The system of claim 10 , wherein the at least one medical data further comprises at least one echocardiography data corresponding to at least one study, wherein the analyzing of the at least one medical data further comprises analyzing the at least one echocardiography data using the at least one machine learning model, wherein the one or more outputs comprises one or more predictions of a presence of one or more conditions, one or more estimations of one or more measurements, one or more visualizations of the one or more measurements, and one or more conclusions of the diagnosis, wherein the obtaining of the one or more outputs comprises obtaining the one or more predictions of the presence of the one or more conditions, the one or more estimations of the one or more measurements, the one or more visualizations of the one or more measurements, and the one or more conclusions of the diagnosis from the at least one machine learning model, wherein the at least one result comprises an echocardiography report, wherein the generating of the at least one result comprises generating the echocardiography report by combining the one or more predictions of the presence of the one or more conditions, the one or more estimations of the one or more measurements, the one or more visualizations of the one or more measurements, and the one or more conclusions of the diagnosis.

13 . The system of claim 10 , wherein the at least one medical data further comprises at least one B mode image, wherein the analyzing of the at least one medical data further comprises analyzing the at least one B mode image using the at least one machine learning model, wherein the one or more outputs comprises one or more estimations of one or more Doppler echocardiographic measurements and one or more predictions of one or more conditions, wherein the obtaining of the one or more outputs comprises obtaining the one or more estimations of the one or more Doppler echocardiographic measurements and the one or more predictions of the one or more conditions from the at least one machine learning model, wherein the generating of the at least one result is further based on the one or more estimations of the one or more Doppler echocardiographic measurements and the one or more predictions of the one or more conditions.

14 . The system of claim 10 , wherein the processing unit is further configured for:

identifying a plurality of historical studies from at least one data source based on a time criterion, wherein each of the plurality of historical studies comprises at least one historical medical data and a historical result corresponding to the at least one historical medical data;

analyzing the plurality of studies using the at least one machine learning model;

generating a new result for each of the plurality of studies;

identifying at least one historical study from the plurality of historical studies based on at least one flagging criterion, wherein the new result of each of the at least one historical study disagrees with the historical result of each of the at least one historical study; and

generating at least one warning based on the identifying of the at least one historical study, wherein the at least one warning indicates a misdiagnosis associated with the at least one historical study, wherein the communication connection is further configured for:

receiving the at least one flagging criterion for flagging at least one of the plurality of studies from the at least one device; and

transmitting the at least one warning to the at least one device.

15 . The system of claim 10 , wherein the communication connection is further configured for:

transmitting the at least one result to at least one clinician device associated with at least one clinician; and

receiving at least one correction on the at least one result from the at least one clinician device, wherein the processing unit is further configured for:

updating the at least one result based on the at least one correction;

generating an updated result based on the updating; and

retraining the at least one machine learning model based on the updated result, wherein the system memory is further configured for storing the updated result.

16 . The system of claim 10 , wherein the communication connection is further configured for receiving at least one additional medical data associated with the at least one user from the at least one device, wherein the at least one medical data is associated with at least one modality and the at least one additional medical data is associated with at least one additional modality, wherein the processing unit is further configured for analyzing the at least one additional medical data using the at least one machine learning model, wherein the at least one machine learning model further comprises at least one additional artificial neural network, wherein the at least one artificial neural network is trained using a set of training medical data associated with the at least one modality and the at least one additional artificial neural network is trained using a set of additional training medical data associated with the at least one additional modality, wherein the set of additional training medical data corresponds to the set of training medical data, wherein the at least one artificial neural network and the at least one additional artificial neural network are jointly trained with a cross modality artificial neural network of the at least one machine learning model using a loss function created using outputs from the cross modality artificial neural network, the at least one artificial neural network, and the at least one additional artificial neural network, wherein the obtaining of the one or more outputs is further based on the analyzing of the at least one additional medical data.

17 . The system of claim 10 , wherein the communication connection is further configured for:

receiving at least one user input from at least one input device; and

transmitting at least one response to at least one output device, wherein the processing unit is further configured for:

processing the at least one result and the at least one user input using a conversation al model; and

generating the at least one response corresponding to the at least one user input based on the processing, wherein the system memory is further configured for storing the at least one user input and the at least one response.

18 . The system of claim 10 , wherein the at least one medical data further comprises a freeform text corresponding to a note, wherein the at least one artificial neural network comprises at least one additional output layer, wherein the at least one additional output layer is configured for predicting text tokens by decoding the lower dimensional abstract vector space representation, wherein the obtaining of the one or more outputs further comprises obtaining at least one text token corresponding to at least one additional assessment parameter for the diagnosis of the one or more pathologies, wherein the generating of the at least one result is further based on the at least one text token.

19 . A device for facilitating a diagnosis of pathologies using at least one machine learning model, the device comprising:

at least one medical imaging unit configured for generating at least one medical data associated with at least one user by imaging at least one portion of at least one body part of the at least one user, wherein the at least one medical imaging unit is associated with at least one ultrasound imaging modality, wherein the imaging comprises ultrasound imaging of the at least one portion of the at least one body part of the at least one user, wherein the at least one medical imaging unit comprises at least one transducer, wherein the at least one transducer generates and transmits ultrasound waves to the at least one portion of the at least one body part and receives echos of the ultrasound waves from the at least one portion of the at least one body part for the generating of the at least one medical data;

a processing unit communicatively coupled with the at least one medical imaging unit, wherein the processing unit is configured for:

identifying a plurality of clinician-annotated echocardiography reports and a plurality of echocardiography data associated with the plurality of clinician-annotated echocardiography reports from at least one data source based on at least one indication corresponding to at least one condition;

extracting at least one category from the plurality of clinician-annotated echocardiography reports using at least one natural language processing model;

training one or more machine learning models using the at least one category and the plurality of echocardiography data;

selecting the at least one machine learning model from the one or more machine learning models based on the at least one medical data;

analyzing the at least one medical data using the at least one machine learning model, wherein the analyzing of the at least one medical data using the at least one machine learning model is further based on the selecting, wherein the at least one machine learning model comprises at least one artificial neural network comprising a plurality of layers, wherein the plurality of layers comprises at least one input layer, a plurality of output layers, and at least one middle layer between the at least one input layer and the plurality of output layers, wherein the at least one input layer is configured for taking a plurality of medical data as a plurality of inputs to the at least one input layer, wherein the at least one middle layer is configured for outputting a lower dimensional abstract vector space representation for each of the plurality of inputs by encoding the plurality of medical data to a lower dimensional abstract vector space, wherein the plurality of output layers is configured for classifying the lower dimensional abstract vector space representation to a plurality of outputs corresponding to a plurality of assessment parameters considered in the diagnosis of the pathologies, wherein each of the plurality of output layers is configured for communicating each of the plurality of outputs with at least one of the plurality of output layers, wherein each of the plurality of output layers is configured for adjusting each of the plurality of outputs based on the at least one of the plurality of outputs of at least one of the plurality of output layers, wherein the analyzing of the at least one medical data comprises inputting the at least one medical data to the at least one machine learning model;

obtaining one or more outputs corresponding to one or more assessment parameters from the at least one machine learning model for the diagnosis of one or more pathologies based on the analyzing of the at least one medical data; and

generating at least one result based on the one or more outputs; and

a system memory communicatively coupled with the processing unit, wherein the system memory is configured for storing the at least one result and the at least one machine learning model.

20 . The device of claim 19 , wherein the at least one medical data further comprises at least one B mode image, wherein the analyzing of the at least one medical data further comprises analyzing the at least one B mode image using the at least one machine learning model, wherein the one or more outputs comprises one or more estimations of one or more Doppler echocardiographic measurements and one or more predictions of one or more conditions, wherein the obtaining of the one or more outputs comprises obtaining the one or more estimations of the one or more Doppler echocardiographic measurements and the one or more predictions of the one or more conditions from the at least one machine learning model, wherein the generating of the at least one result is further based on the one or more estimations of the one or more Doppler echocardiographic measurements and the one or more predictions of the one or more conditions.