Description:FORM 2

THE PATENTS ACT, 1970
(39 of 1970)
&
THE PATENT RULES, 2003

COMPLETE SPECIFICATION
(See Section 10 and Rule 13)

Title of invention:
METHOD AND SYSTEM FOR DELIVERY OF OPTIMAL NASAL DRUG DEPOSITION IN A TARGET REGION OF NASAL CAVITY

Applicant:
Tata Consultancy Services Limited
A company Incorporated in India under the Companies Act, 1956
Having address:
Nirmal Building, 9th Floor,
Nariman Point, Mumbai 400021,
Maharashtra, India

The following specification particularly describes the invention and the manner in which it is to be performed.
TECHNICAL FIELD
[01] The disclosure herein generally relates to nasal drug, and more particularly, to method and system for delivery of optimal nasal drug in a target region of nasal cavity.
BACKGROUND
[02] The respiratory tract is a primary contact zone for the environment, it represents a gateway, not only for infectious particles such as bacteria and viruses, but also for potential treatments. Conventionally, nasal route is extensively used for delivery of drugs for treatment of local diseases such as nasal allergy, nasal congestion, and nasal infections. Recent developments have shown that the nasal route can be exploited for systemic delivery of drugs such as small molecular, weight polar drugs, peptides and proteins that are not easily administered via other routes than by injection, or where a rapid onset of action is required. Lately the use of nasal route for delivery of vaccines, especially against respiratory infections such as influenza has also attracted interest from the pharmaceutical companies specializing in vaccine delivery.
[03] It is crucial for Rhinologists to understand evaluation parameters, such as internal airflow velocity, wall shear stress, and pressure drop. While there are a number of approved drug formulations for local and systemic indications, the development of nasal drug formulations for nasal delivery is still a challenge. There is a need for devices that can deliver compounds to the upper nasal cavity for direct nose-to-brain delivery.
[04] Traditional nasal drug delivery devices do not adequately propel the drug from the device to the target location. To be effective, medication must be delivered at the target location, which is impractical for most subjects with existing drug delivery devices. Existing nasal sprays are typically ineffective because they are unable to reach interior region of the nasal cavity. Due to the variation in individual subject nasal anatomy which complicates drug delivery. Also, there is no method to directly track and monitor the nasal drug delivered in the target location. Thus, there is a need for an effective personalized drug delivery design and method to easily, effectively, and non-invasively deliver local medical drug to the target location of the nasal cavity of any subject as needed.
SUMMARY
[05] Embodiments of the present disclosure present technological improvements as solutions to one or more of the above-mentioned technical problems recognized by the inventors in conventional practice. For example, in one embodiment, a system for delivery of optimal nasal drug in a target region of nasal cavity is provided. The system includes receiving a scenario among a plurality of scenarios and corresponding input parameters, wherein the plurality of scenarios includes a first scenario, a second scenario and a third scenario.
[06] Scenario 1, in different clinical conditions ENT specialist prescribes a nasal drug to be sprayed on a specific region of the subject’s nasal cavity. It is important to position and orient the spray accurately to ensure optimum drug deposition at the intended region in the nasal cavity, in situations where optimum drug deposition is not possible by positioning and orienting the device a personalized device parameters needs to be estimated to make a personalized device. (ii) Scenario 2, A pharma researcher needs to fine tune the drug parameters such as drug density and drug viscosity these parameters affect the breakup of the drug into fine particles that travels in the nasal cavity when administered through a nasal spray and particle trajectory. Here, there is a need to design drug parameters to ensure targeted drug delivery. (iii) Scenario 3, Nasal spray designer designing a device needs to know the cone angle, particle size distribution, and nozzle parameters along with other details. Nasal device designers aim to design nasal devices for targeted nasal drug delivery in nasal cavity.
[07] Further, the input parameters are processed to select one of a machine learning model among an ensemble of machine learning models, wherein the ensemble of machine learning models includes a nasal model, and a drug-device model. Input parameters are analysed to trigger corresponding machine learning model, wherein the corresponding machine learning model is triggered to process the scenario.
[08] In one embodiment, if the received scenario corresponds to the first scenario the nasal model is triggered and a first set of inputs are obtained to determine personalized injection parameters or device parameters corresponding to optimum deposition at the target region, wherein the first set of inputs comprises a nasal CT scan of the subject, a breathing data of the subject, a target region of interest (ROI) within the nasal cavity, and a plurality of prescribed nasal drug parameters and a plurality of nasal spray device parameters.
[09] In another embodiment, if the received scenario corresponds to the second scenario the drug-device model is triggered and a second set of inputs are obtained to determine a plurality of drug parameters, wherein the second set of inputs comprises a breathing rate, nasal geometry, the target region of interest (ROI) within the nasal cavity, a plurality of device parameters and a plurality of injection parameters.
[010] In another embodiment, if the received scenario corresponds to the third scenario the drug-device model is triggered, and a third set of inputs are obtained to design a device, wherein the third set of inputs includes the breathing rate, the nasal geometry, the target region of interest (ROI) within the nasal cavity, the plurality of drug parameters, and the plurality of injection parameters.
[011] The breathing data includes a range of breathing rate, the plurality of drug parameters includes a drug viscosity and a drug density along with drug molecule, the plurality of injection parameters includes coordinate of tip position and orientation vector, and the plurality of device parameters includes a drug flow velocity, a nasal spray nozzle disc diameter, a cone angle, a drug particle diameter distribution, and a total number of drug particles.
[012] Furthermore, the scenario is executed by corresponding machine learning model to perform within the nasal cavity of the subject, wherein if the received scenario corresponds to the first scenario, personalized injection parameters or personalized device parameters are determined ensuring optimum deposition of nasal spray particles injected into the subject's nasal cavity at least one target clinical region of interest (ROI), by processing the first set of inputs and optimizing drug deposition at the region of interest estimated by the nasal model using the first set of processed inputs and optimization parameters such as the plurality of injection parameters or the plurality of device parameters.
[013] And, if the received scenario corresponds to the second scenario, the plurality of drug parameters are determined to ensure optimum drug deposition at the target region of interest (ROI) of the nasal cavity, by processing the second set of inputs and optimizing drug deposition at the region of interest estimated by the drug-device model using the second set of processed inputs.
[014] Further, if the received scenario corresponds to the third scenario, plurality of device parameters is determined to ensure optimum drug deposition at the target region of interest (ROI) of the nasal cavity, by processing the third set of inputs and optimizing drug deposition at the region of interest estimated by the drug-device model using the third set of processed inputs.
[015] Finally, an augmented reality visualization is created for the scenario where the nasal drug particle movement inside the nasal cavity of the subject is tracked by a drug particle tracking model.
[016] In another aspect, a method for delivery of optimal nasal drug in a target region of nasal cavity is provided. The method includes receiving a scenario among a plurality of scenarios and corresponding input parameters, wherein the plurality of scenarios includes a first scenario, a second scenario and a third scenario.
[017] Scenario 1, in different clinical scenario ENT specialist prescribes a nasal drug to be sprayed on a specific region of the subject’s nasal cavity. It is important to position and orient the spray accurately ensure optimum drug deposition at the intended region in the nasal cavity, in situation where optimum drug deposition is not possible by positioning and orienting the device a personalized device parameters needs to be estimated to make a personalized device. (ii) Scenario 2, A pharma researcher need to fine tune the drug parameters such as drug density and drug viscosity these parameters affect the breakup of the drug into fine particles that travels in the nasal cavity when administered through a nasal spray and particle trajectory. Here, there is a need to design drug parameters to ensure targeted drug delivery. (iii) Scenario 3, Nasal spray designer designing a device needs to know the cone angle, particle size distribution, and nozzle parameters along with other details. Nasal device designers aim to design nasal devices for targeted nasal drug delivery in nasal cavity.
[018] Further, the input parameters are processed to select one of a machine learning model among an ensemble of machine learning models, wherein the ensemble of machine learning models includes a nasal model, and a drug-device model. Input parameters are analysed to trigger corresponding machine learning model, wherein the corresponding machine learning model is triggered to process the scenario.
[019] In one embodiment, if the received scenario corresponds to the first scenario the nasal model is triggered and a first set of inputs are obtained to determine personalized injection parameters or device parameters corresponding to optimum deposition at the target region, wherein the first set of inputs comprises a nasal CT scan of the subject, a breathing data of the subject, a target region of interest (ROI) within the nasal cavity, and a plurality of prescribed nasal drug parameters and a plurality of nasal spray device parameters.
[020] In another embodiment, if the received scenario corresponds to the second scenario the drug-device model is triggered and a second set of inputs are obtained to determine a plurality of drug parameters, wherein the second set of inputs comprises a breathing rate, nasal geometry, the target region of interest (ROI) within the nasal cavity, a plurality of device parameters and a plurality of injection parameters.
[021] In another embodiment, if the received scenario corresponds to the third scenario the drug-device model is triggered, and a third set of inputs are obtained to design a device, wherein the third set of inputs includes the breathing rate, the nasal geometry, the target region of interest (ROI) within the nasal cavity, the plurality drug parameters, and the plurality of injection parameters.
[022] The breathing data includes a range of breathing rate, the plurality of drug parameters includes a drug viscosity and a drug density along with drug molecule, the plurality of injection parameters includes coordinate of tip position and orientation vector, and the plurality of device parameters includes a drug flow velocity, a nasal spray nozzle disc diameter, a cone angle, a drug particle diameter distribution, and a total number of drug particles.
[023] Furthermore, the scenario is executed by corresponding machine learning model to perform within the nasal cavity of the subject, wherein if the received scenario corresponds to the first scenario, personalized injection parameters or personalized device parameters are determined ensuring optimum deposition of nasal spray particles injected into the subject's nasal cavity at least one target clinical region of interest (ROI), by processing the first set of inputs and optimizing drug deposition at the region of interest estimated by the nasal model using the first set of processed inputs and optimization parameters such as the plurality of injection parameters or the plurality of device parameters.
[024] And, if the received scenario corresponds to the second scenario, the plurality of drug parameters are determined to ensure optimum drug deposition at the target region of interest (ROI) of the nasal cavity, by processing the second set of inputs and optimizing drug deposition at the region of interest estimated by the drug-device model using the second set of processed inputs.
[025] Further, if the received scenario corresponds to the third scenario, plurality of device parameters is determined to ensure optimum drug deposition at the target region of interest (ROI) of the nasal cavity, by processing the third set of inputs and optimizing drug deposition at the region of interest estimated by the drug-device model using the third set of processed inputs.
[026] Finally, an augmented reality visualization is created for the scenario where the nasal drug particle movement inside the nasal cavity of the subject is tracked by a drug particle tracking model.
[027] In yet another aspect, there are provided one or more non-transitory machine-readable information storage mediums comprising one or more instructions, which when executed by one or more hardware processors causes a method for delivery of optimal nasal drug in a target region of nasal cavity is provided. The method includes receiving a scenario among a plurality of scenarios and corresponding input parameters, wherein the plurality of scenarios includes a first scenario, a second scenario and a third scenario.
[028] Scenario 1, in different clinical scenario ENT specialist prescribes a nasal drug to be sprayed on a specific region of the subject’s nasal cavity. It is important to position and orient the spray accurately ensure optimum drug deposition at the intended region in the nasal cavity, in situation where optimum drug deposition is not possible by positioning and orienting the device a personalized device parameters needs to be estimated to make a personalized device. (ii) Scenario 2, A pharma researcher need to fine tune the drug parameters such as drug density and drug viscosity these parameters affect the breakup of the drug into fine particles that travels in the nasal cavity when administered through a nasal spray and particle trajectory. Here, there is a need to design drug parameters to ensure targeted drug delivery. (iii) Scenario 3, Nasal spray designer designing a device needs to know the cone angle, particle size distribution, and nozzle parameters along with other details. Nasal device designers aim to design nasal devices for targeted nasal drug delivery in nasal cavity.
[029] Further, the input parameters are processed to select one of a machine learning model among an ensemble of machine learning models, wherein the ensemble of machine learning models includes a nasal model, and a drug-device model. Input parameters are analysed to trigger corresponding machine learning model, wherein the corresponding machine learning model is triggered to process the scenario.
[030] In one embodiment, if the received scenario corresponds to the first scenario the nasal model is triggered and a first set of inputs are obtained to determine personalized injection parameters or device parameters corresponding to optimum deposition at the target region, wherein the first set of inputs comprises a nasal CT scan of the subject, a breathing data of the subject, a target region of interest (ROI) within the nasal cavity, and a plurality of prescribed nasal drug parameters and a plurality of nasal spray device parameters.
[031] In another embodiment, if the received scenario corresponds to the second scenario the drug-device model is triggered and a second set of inputs are obtained to determine a plurality of drug parameters, wherein the second set of inputs comprises a breathing rate, nasal geometry, the target region of interest (ROI) within the nasal cavity, a plurality of device parameters and a plurality of injection parameters.
[032] In another embodiment, if the received scenario corresponds to the third scenario the drug-device model is triggered, and a third set of inputs are obtained to design a device, wherein the third set of inputs includes the breathing rate, the nasal geometry, the target region of interest (ROI) within the nasal cavity, the plurality drug parameters, and the plurality of injection parameters.
[033] The breathing data includes a range of breathing rate, the plurality of drug parameters includes a drug viscosity and a drug density along with drug molecule, the plurality of injection parameters includes coordinate of tip position and orientation vector, and the plurality of device parameters includes a drug flow velocity, a nasal spray nozzle disc diameter, a cone angle, a drug particle diameter distribution, and a total number of drug particles.
[034] Furthermore, the scenario is executed by corresponding machine learning model to perform within the nasal cavity of the subject, wherein if the received scenario corresponds to the first scenario, personalized injection parameters or personalized device parameters are determined ensuring optimum deposition of nasal spray particles injected into the subject's nasal cavity at least one target clinical region of interest (ROI), by processing the first set of inputs and optimizing drug deposition at the region of interest estimated by the nasal model using the first set of processed inputs and optimization parameters such as the plurality of injection parameters or the plurality of device parameters.
[035] And, if the received scenario corresponds to the second scenario, the plurality of drug parameters are determined to ensure optimum drug deposition at the target region of interest (ROI) of the nasal cavity, by processing the second set of inputs and optimizing drug deposition at the region of interest estimated by the drug-device model using the second set of processed inputs.
[036] Further, if the received scenario corresponds to the third scenario, plurality of device parameters is determined to ensure optimum drug deposition at the target region of interest (ROI) of the nasal cavity, by processing the third set of inputs and optimizing drug deposition at the region of interest estimated by the drug-device model using the third set of processed inputs.
[037] Finally, an augmented reality visualization is created for the scenario where the nasal drug particle movement inside the nasal cavity of the subject is tracked by a drug particle tracking model.
[038] It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention, as claimed.

BRIEF DESCRIPTION OF THE DRAWINGS
[039] The accompanying drawings, which are incorporated in and constitute a part of this disclosure, illustrate exemplary embodiments and, together with the description, serve to explain the disclosed principles:
[040] FIG.1 illustrates an illustrative system (alternatively referred as nasal targeted drug delivery system), in accordance with some embodiments of the present disclosure.
[041] FIG.2 illustrates a functional architectural view of the system for nasal targeted drug delivery using the system of FIG.1, in accordance with some embodiments of the present disclosure.
[042] FIG.3A illustrates an input processing unit to process the input corresponding to different scenario using the system of FIG.2, in accordance with some embodiments of the present disclosure.
[043] FIG.3B illustrates a 3D nasal geometry generated from the CT scan image of a subject using the system of FIG.2, in accordance with some embodiments of the present disclosure.
[044] FIG.4A illustrates an ML model selector triggering corresponding to different scenario for obtaining drug deposition using the system of FIG.2, in accordance with some embodiments of the present disclosure.
[045] FIG.4B illustrates optimizer generating optimum parameters for optimum drug delivery at target region corresponding to scenario using the system of FIG.2, in accordance with some embodiments of the present disclosure.
[046] FIG.4C illustrates optimization engine to work with selected ML model using the system of FIG.2, in accordance with some embodiments of the present disclosure.
[047] FIG.4D illustrates retraining of ML models implemented into the ML models engine using the system of FIG.2, in accordance with some embodiments of the present disclosure.
[048] FIG.5A illustrates method steps for training a nasal model implemented into the ML models engine using the system of FIG.2, in accordance with some embodiments of the present disclosure.
[049] FIG.5B illustrates method steps for training a drug-device model implemented into the ML models engine using the system of FIG.2, in accordance with some embodiments of the present disclosure.
[050] FIG.5C illustrates method steps for training a drug particle tracking model implemented into the ML models engine using the system of FIG.2, in accordance with some embodiments of the present disclosure.
[051] FIG.6A illustrates the nasal cavity with various clinical regions using the system of FIG.2, in accordance with some embodiments of the present disclosure.
[052] FIG.6B illustrates the nasal cavity segmented into various regions (x0 through x22) using the system of FIG.2, in accordance with some embodiments of the present disclosure.
[053] FIG.6C illustrates an example particle injection planes of the nasal cavity using the system of FIG.2, in accordance with some embodiments of the present disclosure.
[054] FIG.6D illustrates an example nasal drug particle deposited inside the nasal cavity using the system of FIG.2, in accordance with some embodiments of the present disclosure.
[055] FIG.6E depicts CFD simulation results of a nasal geometry injected with a nasal drug and its deposition percentage using the system of FIG.2, in accordance with some embodiments of the present disclosure.
[056] FIG.7 depicts a flow diagram of an example process for executing the scenario using the system of FIG.1, in accordance with some embodiments of the present disclosure.
DETAILED DESCRIPTION OF EMBODIMENTS
[057] Exemplary embodiments are described with reference to the accompanying drawings. In the figures, the left-most digit(s) of a reference number identifies the figure in which the reference number first appears. Wherever convenient, the same reference numbers are used throughout the drawings to refer to the same or like parts. While examples and features of disclosed principles are described herein, modifications, adaptations, and other implementations are possible without departing from the scope of the disclosed embodiments.
[058] Nasal cavity have important protective functions by filtering, warming, and humidifying the inhaled air before it reaches lower respiratory tract. The nasal cavity is divided into two regions, which reach from nostrils towards nasopharynx. They can be further separated into three regions: a vestibular region, a respiratory region, and an olfactory region. The respiratory region includes a superior turbinate, a middle turbinate, and an inferior turbinate. These regions play very important in conditioning inhaled air along with other functions.
[059] Nasal drug sprays have gained significant attention in recent years due to their non-invasive and effective approach for local respiration system to treat several conditions such as nasal obstruction, nasal polyps, and sinusitis. In addition, the non-invasive nature of nasal drug delivery through nasal spray makes it a promising option for long-term systemic administration, as it affords patient comfort and compliance often hindered by other drug therapy.
[060] Recent developments have highlighted the possibility of exploiting the nasal route for direct transport of drugs from nose to brain bypassing the blood brain barrier. Recently treatment of disease such as Parkinson, Alzheimer, and epilepsy have started through nasal spray utilizing this route.
[061] Delivery of nasal drug within a target region into the nasal cavity of the subject is challenging due to the structure of nostril lobe. Each subject may have structural change in the nostril regions where deposition of the nasal drug requires systematic approach for delivery of equivalent nasal drug. Existing methods are unable to provide systematic combinational approach as one system to address challenging problem of nasal drug delivery into the target region of interest, determining drug parameters to inject within the nasal cavity and tracking injected drug in augmented reality visualization.
[062] Further, to deliver the drug into the target region of interest in the nasal cavity administration of the drug requires a precise position and orientation of the nasal spray along with the plume angle and drug particle size. The nasal spray is meant to deliver a specific amount of the drug in each application. The drug coming out of the spray is in the form of a plume. The formation of the plume is controlled by a nozzle attached to the device. The drug inside the spray may be kept at a higher pressure. Drug while coming through the nozzle breaks into small droplets and spreads in a conical shape. The method of the present disclosure estimates drug parameters and device parameters that ensures optimum drug deposition over target region in the nasal cavity. These estimates may be useful in designing device and rapid development of drug at clinical trial phase.
[063] If the subject has deviation within the nasal cavity such as a deviated septum or dilated sinus, available devices may not be effective in achieving targeted drug delivery. Subject may require a specific position and orientation of the spray to ensure optimum deposition a specified target region. Moreover, a specialized personal spray device may also be required to optimal delivery to the target region. The method of the present disclosure provides injection parameters and personalized device parameters for subjects with nasal disorders or diseases.
[064] In addition, the method of the present disclosure is capable of providing augmented reality visualization of the nasal cavity. This helps the clinicians to visualize motion of the drug particle and its deposition in the nasal cavity.
[065] Embodiments herein provide a method and system delivery of optimal nasal drug deposition in a target region of nasal cavity. The system may be alternatively referred as a nasal targeted drug delivery system. The system is capable of achieving optimal nasal drug deposition in a target region within a nasal cavity of the subject in different scenario. In a first scenario, the system determine parameters to inject drug into the nasal cavity of a subject, design a personalized nasal spray corresponding to the device parameters based on the nasal anatomy of the subject. In a second scenario, system helps accelerate the design of a generic nasal spray by estimating device parameters for a given drug. And, finally in a third scenario, the system determine drug parameters for optimal nasal drug deposition percentage into the nasal cavity. Additionally, the system provides tracking of drug particles through augmented reality visualization.
[066] The system 100 receives scenario corresponding to one of the scenario and corresponding set of inputs comprising a CT scan of the subject nasal, a breathing rate of the subject, a device parameters, a drug parameters and name of the target region in the nasal cavity from at least one medical expert or a pharma expert via a graphical user interface (GUI) which further transmits the scenario to subsequent modules associated with the system. The set of inputs are preprocessed to create a 3D nasal geometry, obtain drug parameters and breathing parameters. The scenario is further analyzed to trigger corresponding ML model associated with a ML model selector to execute parameters. The selected ML model is then utilized by optimizer to obtain optimum parameters corresponding to the requested scenario. The system 100 also creates an augmented reality visualization of the nasal drug particle movement and deposition in the nasal cavity. In addition, the use of ML model for estimating percentage deposition and estimation of optimum parameters reduces computational time and makes the solution practical for real time use. The disclosed system is further explained with the method as described in conjunction with FIG.1 to FIG.7 below.
[067] Referring now to the drawings, and more particularly to FIG.1 through FIG.7, where similar reference characters denote corresponding features consistently throughout the figures, there are shown preferred embodiments, and these embodiments are described in the context of the following exemplary system and/or method.
[068] FIG.1 is an illustrative system (alternatively referred as nasal targeted drug delivery system), in accordance with some embodiments of the present disclosure. In an embodiment, the system 100 includes processor (s) 104, communication interface (s), alternatively referred as or input/output (I/O) interface(s) 106, and one or more data storage devices or memory 102 operatively coupled to the processor (s) 104. The system 100, with the processor(s) is configured to execute functions of one or more functional blocks of the system 100.
[069] Referring to the components of the system 100, in an embodiment, the processor (s) 104 can be one or more hardware processors 104. In an embodiment, the one or more hardware processors 104 can be implemented as one or more microprocessors, microcomputers, microcontrollers, digital signal processors, central processing units, state machines, logic circuitries, and/or any devices that manipulate signals based on operational instructions. Among other capabilities, the processor(s) 104 is configured to fetch and execute computer-readable instructions stored in the memory. In an embodiment, the system 100 can be implemented in a variety of computing systems, such as laptop computers, notebooks, hand-held devices, workstations, mainframe computers, servers, a network cloud, and the like.
[070] The I/O interface(s) 106 can include a variety of software and hardware interfaces, for example, a web interface, a graphical user interface, and the like and can facilitate multiple communications within a wide variety of networks N/W and protocol types, including wired networks, for example, LAN, cable, etc., and wireless networks, such as WLAN, cellular, or satellite. In an embodiment, the I/O interface (s) 106 can include one or more ports for connecting a number of devices (nodes) of the system 100 to one another or to another server.
[071] The memory 102 may include any computer-readable medium known in the art including, for example, volatile memory, such as static random-access memory (SRAM) and dynamic random-access memory (DRAM), and/or non-volatile memory, such as read only memory (ROM), erasable programmable ROM, flash memories, hard disks, optical disks, and magnetic tapes.
[072] In an embodiment, memory 102 includes a plurality of modules 108 can also include various sub-modules as depicted in FIG.2 such as an input processing unit 202, a ML models engine 204, a ML model selector 206, a target region optimizer 208, a drug particle tracking model 214, a drug particle motion visualizer 216, and a retraining unit 218. The plurality of modules 108 include programs or coded instructions that supplement applications or functions performed by the system 100 for executing different steps involved in the process of providing data privacy in service operations of the system 100. The plurality of modules 108, amongst other things, can include routines, programs, objects, components, and data structures, which perform particular tasks or implement particular abstract data types. The plurality of modules 110 may also be used as, signal processor(s), node machine(s), logic circuitries, and/or any other device or component that manipulates signals based on operational instructions. Further, the plurality of modules 108 can be used by hardware, by computer-readable instructions executed by the one or more hardware processors 104, or by a combination thereof.
[073] Memory 102 may comprise information pertaining to input(s)/output(s) of each step performed by the processor(s) 104 of the system 100 and methods of the present disclosure. Functions of the components of system 100, for delivery of optimal nasal drug in a target region of nasal cavity, are explained in conjunction with FIG.2, and FIG.3 providing flow diagram, architectural overviews, and performance analysis of the system 100.
[074] FIG.2 illustrates a functional architectural view of the system for nasal targeted drug delivery, in accordance with some embodiments of the present disclosure. The system 200 may be an example of the system 100 (FIG.1). In an example embodiment, the system 200 may be embodied in, or is in direct communication with the system, for example the system 100 (FIG.1).
[075] Nasal drug delivery is advantageous when compared with other routes of drug delivery. Clinical medical experts widely use nasal drug delivery as an efficient route for topical treatment of local diseases in nose and paranasal sinuses such as allergic and non-allergic rhinitis, sinusitis, and the like. Clinical medical experts may be an ENT specialist, a pharma specialist, a nasal device designer and the like.
[076] Each subject may have a different medical history and may require different types of medical treatment. Clinicians diagnose disease based on symptoms and health history of the subject. The below listed examples provides illustrations or practical implications in various scenarios,
[077] 1. Scenario 1 or first scenario, in different clinical conditions ENT specialist prescribes a nasal drug to be sprayed on a specific region of the subject’s nasal cavity. It is important to position and orient the spray accurately ensure optimum drug deposition at the intended region in the nasal cavity. Failing that will result in reduced effectiveness of the drug, associated side effects and wastage of the drug. Position and orientation for each target region is different, a standard position and orientation will not ensure optimum deposition over the target region. Therefore, a solution needs to be there to quickly estimate the position and orientation of the nasal drug for the subject to ensure optimum drug delivery.
[078] Moreover, subjects may have different nasal geometry, generalized nasal devices may fail to deliver drug at the intended region with all possible position and orientation of the nasal spray. A personalized device parameters need to be available to create a personalized device for the subject. Therefore, a system needs to be there to quickly replicate subject’s anatomy in silico and come up with personalized device and injection parameters to ensure optimum drug delivery at the target region in the subject’s nasal cavity.
[079] Scenario 2 or second scenario, A pharma researcher needs to fine tune the drug parameters such as drug density and drug viscosity these parameters affect the breakup of the drug into fine particles that travels in the nasal cavity when administered through a nasal spray. The size of these particles highly affects the percentage of deposition of the drug. Therefore, a precise value of the density and viscosity is highly important. An in-silico system that can quickly calculate these parameters for an available device will speed up the drug development process to a great extent.
[080] Scenario 3 or third scenario, A designer designing a device needs to know the cone angle, particle size distribution, and nozzle parameters along with other details. These parameters are crucial in achieving a precise target drug delivery of the nasal cavity.
[081] An in-silico system that can quickly calculate these parameters, will speed up the device development process to a great extent. This will reduce the cost of developing the spray and reduce the time to market.
[082] The system 200 addresses different scenario 1 through 3. In one embodiment, system 200 may have different role-based login to handle user of each scenario. The system 200 may alternatively have a scenario selection post login. GUI application window routes to corresponding functional unit based on the scenario received.
[083] In one embodiment, the system 200 includes an input processing unit 202, a ML models engine 204, a ML model selector 206, a target region optimizer 208, a drug particle tracking model 214, a drug particle motion visualizer 216, and a retraining unit 218. The input processing unit 202 process the input received from the user along with various scenarios selected by the user or from the user role. The input processing unit 202 may also trigger the re-training unit 218 if required. The processed inputs are then passed to the ML model selector 206 to select corresponding ML model based on the inputs and scenario. The selected ML model is then passed to the target region optimizer 208 to optimize required output parameters that ensure optimum drug delivery at the specified target region in the nasal cavity. The target region optimizer 208 estimates optimum parameters which is displayed to the user through GUI. The target region optimizer 208 also provides values of the percentage deposition at target regions. This data is also made available at the GUI for user reference. The optimum parameters are then passed to the drug particle tracking model 214 to create position and velocity data at different time steps of the drug particle cloud injected from the device. This drug particle cloud is then passed to drug particle motion visualizer 216 to create a virtual reality animation of the drug particle cloud motion in the nasal cavity of the subject. This virtual reality visualization is available for the user through GUI.
[084] Input processing unit 202 - Referring now to FIG.3A, users of the system 200 provides the inputs in terms of different scenario or objective and corresponding details comprising a combination of nasal cavity CT scan of a subject, a breathing data of the subject, a nasal spray along with name of the target region within the nasal cavity depending on the scenario 1 through 3. The input processing unit 202 receives these inputs from the user through graphical user interface (GUI) and processes them. The nasal cavity CT scan of the subject is processed to construct a 3D nasal geometry 202a. Further, the 3D nasal geometry 202a is used to determine parametric geometry 202b. The determination of the parametric geometry may be obtained from techniques such as proper orthogonal decomposition, singular value decomposition and the like. Further, the breathing data of the subject is analyzed to select subject specific breathing rate at which subject would be breathing while taking the nasal drug through the nasal spray. In one example the breathing rate could be in the range 5-12 liters per min. The drug and device parameters are obtained from the pre-defined library based on the prescribed nasal spray by the ENT specialist through the GUI. The plurality of drug parameters include density of the drug and viscosity of the drug along with formulation and active molecule details. Whereas plurality of device parameters include the size of the drug particle, plume cone angel, amount of drug, drug flow rate and particle velocities along with device pressure, and nozzle details. Further, the name of the target region of interest (ROI) is specified on the GUI where the drug is intended to be injected. The name of the region is marked on the nasal geometry and used to calculate optimum percentage deposition.
[085] Referring now to FIG.3B, in one embodiment, the nasal CT scan of the subject is utilized to construct the 3D nasal geometry using software or codes developed to construct 3D geometry from CT data such as 3D Slicer, Athena DICOM Viewer, 3D-DOCTOR, Vesalius3D or any other computer program or code by performing the steps:
a. Cropping is performed over the nasal CT scan of the subject using a ‘crop volume’ nasal cavity where a nasal airway passage sub-volume is cropped. This cropped sub-volume is further used for segmentation.
b. Segmentation is performed using a ‘segment editor’ where a new segment is added and a threshold tool is applied by setting a threshold range of about 1024 to 299 HU (Hounsfield Units) obtained by trial and error, the CT volume was extracted.
c. Extraction is performed to obtain the nasal airway passage.
d. Smoothing The 3D nasal geometry is generated to verify, and smoothing is performed to remove small holes or spikes using a “Gaussian smoothing filter’ or any other suitable mathematical or machine learning based filter, and finally, the 3D nasal geometry is exported as a Standard Tessellation Language (STL) file.
The generated 3D nasal geometry is further splitted into different regions of clinical interest namely Vestibule, Olfactory, Superior Turbinate, Middle Turbinate, Middle Turbinate, and Nasopharynx. One of these regions are selected as regions of interest for delivery of the nasal drug.
[086] ML models engine 204 includes ensemble of trained ML models comprising a nasal model, a drug-device model and drug particle model. The nasal model is developed to identify injection location and orientation of the nasal spray for subject. The model also estimates device parameters for subjects having deviations in the nasal cavity to obtain optimum drug delivery at the target region. This ML model is triggered in scenario 1 by ENT specialist while prescribing the drug to a patient.
[087] The drug-device model calculates the plurality of drug parameters for a nasal cavity given a spray device and the target region, additionally the drug-device ML model calculates device parameters for a given drug, nasal cavity, and target region. This model is triggered in scenario 2 and scenario 3 by a pharma expert and a designer.
[088] ML model selector 206 and Target region optimizer 208 - Referring now to FIG.4A and FIG 4B, the ML model selector 206 selects specific ML model to estimate percentage deposition over the nasal regions as per the scenario through the user login or through selection at the GUI. The target region optimizer 208 estimates optimum parameters. In scenario 1 the user, ENT specialist, wants to find out drug injection parameter in terms of positional x, y, z coordinate and orientation of the device in terms of angel from the vertical and side, for a subject with no deviation or obstruction in the nasal cavity. The user, ENT specialist, provides subject CT scan, specifies target region and prescribed spray through GUI. These inputs are processed through input processing unit 202 and produces geometrical parameters, drug parameters, and device parameters. The Model selector 206 selects nasal ML model to estimate percentage of drug deposition at the nasal regions for the subject for a given injection parameters. Optimization for target region 404 estimates injection parameters corresponding to the optimum drug deposition at the target region.
[089] In situations of subjects having deviation or obstruction in nasal and just changing injection parameter optimum deposition is not achieved through the system 200 achieves optimum deposition through optimization for target region 404 by calculating personalized device parameters such as nasal parameters and injection parameters. The personalized device may be 3D printed using calculated parameters by the system 200.
[090] In scenario 2, the user is a pharma specialist who wants to fine-tune drug parameters to optimize deposition of the drug at a target region in the nasal cavity. As the user changes the parameters of the drug, the user wants to estimate the drug parameter that ensures optimum deposition through the system 200. The scenario is captured in the system though the login of the type of the user or selection at the GUI. The user selects, spray device, nasal geometry, target region, and injection parameters. With these parameters, the ML model selector 206 selects drug-device ML model which quickly estimates percentage deposition of the drug at the target region. This eliminates the need for carrying out any physical experiments for estimating deposition. The optimum drug parameters such as density and viscosity for the selected inputs are estimated by optimizer 406 of the system 200.
[091] In scenario 3, the user is a CAD designer who wants to design a drug spray. The designer needs to know device parameters that ensure optimum drug delivery at the target region in the nasal cavity. As the user changes the design parameters of the device the percentage deposition at the target region changes. The user wants to estimate the device parameters through system 200. The scenario is captured in the system though the login of the type of the user or selection at the GUI. The user selects the drug, nasal geometry, target region, and injection parameters. With these inputs, processed by the input processing unit 202 the ML model selector 206 selects drug-device ML model which quickly estimates percentage deposition of the drug at the target region. This eliminates the need for manufacturing and testing the device for its effectiveness. System 200 further estimates the device parameters to obtain optimum drug delivery at the specified target region by the optimizer 408. The device parameters are drug flow rate, drug particle velocity, drug plume cone angle, and particle size. These parameters are then used to calculate nozzle parameters and pressure inside the canister to be used in design of the spray device.
[092] Target region optimizer 208 - In one embodiment, referring now FIG.4C, optimizers (404, 406 and 408) in FIG.4B to carry out optimization are collectively written as optimizer 402. The optimizer 402 is a multi-class optimization engine. The optimizer 402 includes an acquisition function to estimate next inputs, and a surrogate model. The optimizer may be Genetic Algorithm (GA), Bayesian Algorithm (BA), Simulated Annealing (SA), Particle Swarm Optimization (PSO). The selected ML model by ML model selector 206 is used to evaluate the percentage deposition at all the regions of the nasal cavity corresponding to input parameters. Based on the scenario from 1 to 3 received in the system 200 the target region optimizer 208 initially creates a DOE table using an injection parameters range, a drug parameters range, and a device parameters range Further, each set of DOE table is passed to the ML model to determine percentage deposition which are further used to create a surrogate model. For every iteration acquisition, function and optimization evaluates the surrogate model to determine the next level of inputs. The optimization loop executes for a predefined number of iterations and terminates when optimum parameters are determined.
[093] In one example, the x, y, z coordinate of the injection location may vary from 0.00215m to 0.06620m, - 0.08856m to -0.07841, and 0.21117m to 0.21959m. Similarly, injection velocity may vary 15m/s to 20 m/s and drug plume cone angle may vary from 35 degrees to 45 degrees and particle size may vary from 5 micrometers to 40 micrometers. A DOE table created using these ranges and percentage deposition of drug is estimated using selected ML model and surrogate model is created for optimization.
[094] Drug particle tracking model 214 –This model obtains optimum parameter as input from the optimizer 402 . The drug particle tracking model 214 uses these inputs and calculates the position and velocity of the drug particles as they come out of the nasal spray. The plume of particles coming out of the nozzle has particles in the range from 50,000 particles to 100,000 particles. The position of the particles is randomly distributed in the disc of the nozzle opening and their velocity vectors are distributed along the cone of the plume estimated by the system 200. The position and velocity of all these particles are estimated after a time interval ranging from few microseconds to milliseconds by the drug particle tracking model. These new positions and velocity of the particle are then taken as input for the estimation of position and velocity of the particles after next the time step. The position and velocity of all the particles are calculated in the same fashion till they reach the walls of the nasal cavity and deposit there.
[095] Drug particle motion visualizer 216 – Series of position and velocity data corresponding to optimized parameters for all the drug particles calculated by drug particle tracking model 214 is fed to drug particle motion visualizer 216. This module constructs 3D augmented reality visualization of intranasal drug particle movement inside the nasal cavity for better insight of the drug delivery over the target region.
[096] Retraining unit 218 – Referring now to FIG.4D, the retraining unit obtains inputs from the input processing unit 202 to retrain the ensemble of ML models of the ML model engine 204 to ensure continuous learning and improve quality of the ML models to process new set of inputs and scenarios. The inputs from the user through GUI are processed through input processing unit 202. These inputs are then compared with the injection parameters range, the drug parameters range, and the device parameters range for which ensemble of ML models are trained. Further, the retraining unit 218 determines whether processed inputs are within the range or outside of the range. Retraining unit 218 is triggered when the processed input parameters deviate from the range to retrain the ensemble of ML models of the ML model engine 204. The ensemble of ML model re-trained will include Nasal Model, Drug-Device Model, and Particle tracking Model. The procedure of training these models is same as discussed in FIG.5A.
[097] Training Nasal ML Model - FIG.5A illustrates method steps for training a nasal model implemented into the ML models engine 204 using the system of FIG.2, in accordance with some embodiments of the present disclosure. In one embodiment, training steps performed by the system 200 are listed below.
[098] A dataset of nasal spray and subject CT is collected to create training data for the nasal model. Ranges of the drug parameter and device parameter are obtained from nasal spray data. 3D STL geometry of the nasal is created from the CT data and segmented with clinical regions. The range of injection parameters are obtained from the nostril volumes of the 3D nasal cavity and verified by the ENT specialist. Further the 3D nasal geometry is decomposed into parameters using a suitable technique such as proper orthogonal decomposition (POD).
[099] A first DOE table is created using ranges of parametric nasal geometry, drug parameters, device parameters and injection parameters. The corresponding 3D nasal geometry is meshed to create a CFD model using drug, device, and injection parameters from the first DOE table.
[0100] The CFD model is simulated using the first DOE table and the CFD model to determine corresponding drug deposition percentage over the regions of the nasal cavity. Further, a first set of ensemble models are trained using combination of elements in the first DOE table and corresponding drug deposition percentage. For example, the ensemble of ML models trained may include Logistic Regression, Naive Bayes, K Neighbors Classifier, SVM - Linear Kernel, Decision Tree Classifier, Random Forest Classifier, Gradient Boosting Classifier, Light Gradient Boosting Machine, Linear Discriminant Analysis, Quadratic Discriminant Analysis, and the like. Each machine learning model is ranked in decreasing order of their prediction accuracy. Further, an optimal nasal model is selected based on ranking. Finally, the hyper parameters of the selected models are tuned to further improve the accuracy of the selected ML model. In case of re-training set of new rows are added to the DOE table corresponding to new range of inputs and training of ensemble of ML model and truing of hyper parameters are carried out. The trained nasal model is further utilized to determine drug deposition percentage within the nasal cavity in real time with good accuracy.
[0101] Training drug-device ML Model - FIG.5B illustrates method steps for training a drug-device ML model implemented into the ML models engine 204 of FIG.2, in accordance with some embodiments of the present disclosure. The drug-device model associated with the ML models engine 204 is trained for a generic nasal cavity. Ranges of the drug parameter and device parameter ranges are obtained from nasal spray data. The range of injection parameters are obtained from the nostril volumes of the 3D nasal cavity and verified by the ENT specialist. A second DOE table is created using ranges of drug parameters, device parameters and injection parameters. The generic 3D nasal geometry is meshed to create a CFD model using drug, device, and injection parameters from the second DOE table.
[0102] The CFD model is simulated using the first DOE table and the CFD model to determine corresponding drug deposition percentage over the regions of the nasal cavity. Further, a second set of ensemble models are trained using combination of elements in the second DOE table and corresponding drug deposition percentage. For example, the ensemble of ML models trained may include Logistic Regression, Naive Bayes, K Neighbors Classifier, SVM - Linear Kernel, Decision Tree Classifier, Random Forest Classifier, Gradient Boosting Classifier, Light Gradient Boosting Machine, Linear Discriminant Analysis, Quadratic Discriminant Analysis, and the like. Each machine learning model in ranked in decreasing order of their prediction accuracy. Further, an optimal nasal model is selected based on ranking. Finally, the hyper parameters of the selected models are tuned to further improve the accuracy of the selected ML model. In case of re-training new row is added to the DOE table and training of ensemble of ML model and truing of hyper parameters are carried out. The trained drug-device model is further utilized to determine drug deposition percentage within the nasal cavity in real time with good accuracy.
[0103] Training drug particle tracking model 214 - FIG.5C illustrates method steps for training a drug particle tracking model implemented into the ML models engine using the system of FIG.2, in accordance with some embodiments of the present disclosure. The drug particle tracking model is trained with position and velocity components of each drug particle emanating from the nasal spray at different time steps. The CFD simulation corresponding to the first DOE table of FIG.5A and second DOE table of FIG.5B creates data of drug particle location and velocity at each time step.
[0104] A training and testing data set consisting of input position, velocity, particle size and time step and output position and velocity is created from CFD simulated data. Further, a third set of ensemble models are trained on this data. For example, the ensemble of ML models trained may include Logistic Regression, Naive Bayes, K Neighbors Classifier, SVM - Linear Kernel, Decision Tree Classifier, Random Forest Classifier, Gradient Boosting Classifier, Light Gradient Boosting Machine, Linear Discriminant Analysis, Quadratic Discriminant Analysis, and the like. Each machine learning model in ranked in decreasing order of their prediction accuracy. Further, an optimal nasal model is selected based on ranking. Finally, the hyper parameters of the selected models are tuned to further improve the accuracy of the selected ML model. In case of re-training new row is added to the DOE table and training of ensemble of ML model and truing of hyper parameters are carried out. Further, the trained drug particle tracking model predicts position and velocity of the drug particle for a given drug, device, and injection parameter in real time. These data are then used to create augmented reality visualization of the nasal drug particle movement inside the nasal cavity of the subject.
[0105] Drug Deposition Modeling using CFD – Results of CFD modeling and simulation are accurate enough to estimate the drug deposition in the nasal cavity. The CFD modeling the drug deposition requires meshed 3D geometry of the nasal cavity. The CFD model is created using the mesh and boundary conditions along with suitable numerical settings. To model drug deposition using CFD motion of each drug particle needs to be modeled. CFD software such as Ansys Fluent, Star CCM, Ansys CFX, OpenFOAM, or a code may be used to carry out CFD modeling. These software models breathing of the subject along with the flow of the drug particles.
[0106] Referring now FIG.6A, illustrating the nasal cavity of the subject with various nasal regions. The 3D nasal cavity is obtained by constructing CT scan of the nose region using one of conventional tools for example 3D Slicer, Athena DICOM Viewer, 3D-DOCTOR, Vesalius3D or any other computer program or code To study the drug deposition, the nasal cavity is divided into different regions, these regions were further segmented into small region. FIG.6A represents the nasal cavity which starts from nostrils and ends towards nasopharynx. The nasal cavity is further divided into three regions the vestibular region, the respiratory region, and the olfactory region. The respiratory region includes the superior turbinate, the middle turbinate, and the inferior turbinate.
[0107] Referring now to FIG.6B illustrates a segmented regions (x0 through x22) of the nasal cavity with various nostril regions of FIG.6A. FIG.6B illustrates the nasal cavity segmented into regions from x0-x22. The injected drug gets deposited in at least one of the regions from x0-x22 within the nasal cavity. The nasal cavity is segmented into regions as depicted in Table 1 and FIG.6A. These regions help optimizer to estimate precise parameters.
Table 1 – Nasal cavity regions and labels
Nasal section Region
Vestibule x0
Olfactory x1-x4
Superior Turbinate X5-x9
Middle Turbinate X10-x15
Middle Turbinate X16-x21
Nasophraynx X22
[0108] In one embodiment, breathing of the subject is first modelled by applying breathing rate boundary conditions at one of the nostrils of the subject as inlet boundary and nasal cavity as the wall boundary while another end of the nasal cavity as outlet boundary. Drug particles are modeled as spherical particles over the flow field obtained.
[0109] Referring now FIG.6C, in one embodiment, the nasal geometry boundary conditions used for simulation does not include the device or bottle. Instead, the drug particles are modeled from a point in either of the nostril in the form of a solid cone. The tip of the nozzle is assumed to be 10mm inside from the nostril. In this simulation, boundary conditions assigned to the drug particles are allowed to stick to the wall and escape from the outlet. The walls of the nasal regions are set to trap, while the outlet is set to escape. Table 2 provides the particle injection parameters used in simulation. Once the simulation is completed, the obtained results are post-processed using Paraview, ANSYS CFD Post or any suitable post processing software for analysis and visualization of the simulation results. The obtained simulated results are validated with experimental results through existing prior arts for flow and drug deposition. However, the device is not modeled in the simulation.
Table 2 – Particle injection parameters
Parameter Value
Particle Diameter 5-50 µm
Particle density 1000 kg/m3
Particle velocity 15-20 m/s
Particle injection type solid cone with disc injection
Particle injection angle 35°- 45°
Number of particles 50,000-100,000
[0110] FIG.6D illustrates an example nasal drug particle deposited inside the nasal cavity. The nasal drug is delivered through the nasal cavity using CFD simulations to investigate the flow dynamics and drug deposition patterns. The device delivers a fine mist of drug particles into the nasal cavity of the subject. FIG.6D have a release point where the drug is injected towards the tip inside the nasal cavity. The olfactory deposited particles entered the nose through the inner superior corner of the nostril. The middle meatus deposited particles entered the nose through the top third of the nostril. The inferior deposited particles entered via the bottom floor regions of the nostril.
[0111] Now referring to FIG.6E represents segmented regions of the nasal cavity within x0-x22. The particle deposited region percentage is represented with values and non-deposited particle deposited region percentage is represented with zero values. The particle deposited region percentage is represented in the regions shown in Table 3. In the current embodiment the target region namely middle turbinate receives maximum deposition.
Table 3 – Region wise particle deposition
Region % Deposition Region % Deposition
X0 0.0 X12 9.5
X1 0.0 X13 34.9
X2 0.0 X14 0.6
X3 0.0 X15 0.0
X4 0.0 X16 0.0
X5 0.0 X17 0.0
X6 0.0 X18 0.0
X7 0.1 X19 25.2
X8 0.0 X20 1.8
X9 0.0 X21 0.2
X10 0.0 X22 0.0
X11 0.0

[0112] FIG.7 depicts a flow diagram of an example process for executing the scenario using the system of FIG.1, in accordance with some embodiments of the present disclosure. In an embodiment, the system 200 comprises one or more data storage devices or the memory 102 operatively coupled to the processor(s) 104 and is configured to store instructions for execution of steps of the method 700 by the processor(s) or one or more hardware processors 104. The steps of the method 700 of the present disclosure will now be explained with reference to the components or blocks of the system 200 as depicted in FIG.1 through FIG.2, and the steps of flow diagram as depicted in FIG.7. Although process steps, method steps, techniques or the like may be described in a sequential order, such processes, methods, and techniques may be configured to work in alternate orders. In other words, any sequence or order of steps that may be described does not necessarily indicate a requirement that the steps to be performed in that order. The steps of processes described herein may be performed in any practical order. Further, some steps may be performed simultaneously.
[0113] Referring to the steps of the method 700, at step 702 a one or more hardware processor to receive a scenario among a plurality of scenarios and corresponding input parameters, wherein the plurality of scenarios includes a first scenario, a second scenario and a third scenario.
[0114] The system 200 may be deployed in clinical medical environment for practical applications. Each subject may have a different medical history and may require different types of medical treatment. Clinicians diagnose disease based on symptoms and health history of the subject. Referring to the clinical medical experts may be the ENT specialist. Alternatively other expert using the system 200 may be the pharma specialist, the nasal device designer, and the like. The scenario may correspond to scenarios 1-3 to be discussed in the embodiments as discussed in FIG.2.
[0115] Clinical medical expert may provide inputs such as the nasal cavity CT scan of the subject, the breathing data of the subject, prescribe nasal spray, at least one target region of interest (ROI) and the scenario via the graphical user interface (GUI). The scenario may include illustrative examples (Scenario 1-3) as depicted above in description corresponding to FIG.2 for various practical application scenarios. Reiterating the above Scenarios again,
[0116] (i) Scenario 1, in different clinical scenario ENT specialist prescribes a nasal drug to be sprayed on a specific region of the subject’s nasal cavity. It is important to position and orient the spray accurately ensure optimum drug deposition at the intended region in the nasal cavity, in situation where optimum drug deposition is not possible by positioning and orienting the device a personalized device parameters needs to be estimated to make a personalized device. (ii) Scenario 2, A pharma researcher need to fine tune the drug parameters such as drug density and drug viscosity these parameters affect the breakup of the drug into fine particles that travels in the nasal cavity when administered through a nasal spray and particle trajectory. Here, there is a need to design drug parameters to ensure targeted drug delivery. (iii). Scenario 3, nasal spray designer designing a device needs to know the cone angle, particle size distribution, and nozzle parameters along with other details. Nasal device designers aim to design nasal devices for targeted nasal drug delivery in nasal cavity.
[0117] To verify the rate of drug deposition in the target region(s) based on the diagnostic profile, pharmaceutical specialist can virtualize the rate of drug depiction in augmented reality. This analysis helps in in understanding flow and deposition of the drug particles in the nasal cavity after injection.
[0118] Referring to the steps of the method 700, at step 704 via the one or more hardware processors to process the input parameters, to select one of a machine learning model among an ensemble of machine learning models, wherein the ensemble of machine learning models includes a nasal model, and a drug-device model.
[0119] The input processing unit 206 as discussed above in FIG.2 and FIG.3A processes the inputs received at step 702, where the nasal cavity CT scan of the subject is utilized to construct a 3D nasal geometry. The 3D nasal geometry is used to determine the parametric geometry. Further, the breathing data of the subject is analyzed to select subject specific breathing rate, drug and device parameters are selected. The 3D nasal geometry is constructed by performing the steps as described above in FIG.2 and FIG.3B.
[0120] Once the inputs are processed by the input processing unit 202, the ML model selector 206 receives the scenario to select corresponding ML model from the ensemble of trained ML models from the ML models engine 204 (referring to FIG.2).
[0121] Referring to the steps of method 700, at step 706 via the one or more hardware processors to analyse the input parameters to trigger corresponding machine learning model, wherein the corresponding machine learning model is triggered to process the scenario:
[0122] if the received scenario corresponds to the first scenario the nasal model is triggered and a first set of inputs are obtained to determine personalized injection parameters or personalized device parameters corresponding to optimum deposition at the target region, wherein the first set of inputs comprises a nasal CT scan of the subject, a breathing data of the subject, a target region of interest (ROI) within the nasal cavity, and a plurality of prescribed nasal drug parameters and a plurality of nasal spray device parameters.
[0123] if the received scenario corresponds to the second scenario the drug-device model is triggered and a second set of inputs are obtained to determine a plurality of drug parameters, wherein the second set of inputs comprises a breathing rate, nasal geometry, the target region of interest (ROI) within the nasal cavity, a plurality of device parameters and a plurality of injection parameters.
[0124] if the received scenario corresponds to the third scenario the drug-device model is triggered, and a third set of inputs are obtained to design a device, wherein the third set of inputs includes the breathing rate, the nasal geometry, the target region of interest (ROI) within the nasal cavity, the plurality drug parameters, and the plurality of injection parameters.
[0125] The breathing data includes a range of breathing rate, the plurality of drug parameters includes a drug viscosity and a drug density along with drug molecule, the plurality of injection parameters includes coordinate of tip position and orientation vector, and the plurality of device parameters includes a drug flow velocity, a nasal spray nozzle disc diameter, a cone angle, a drug particle diameter distribution, and a total number of drug particles.
[0126] The trained nasal model (204) is used to process the received scenario corresponds to the first scenario (Referring to FIG.5A) to determine the plurality of injection parameters or personalized device parameters.
[0127] In another embodiment, referring to Scenario 2, if received scenario corresponds to the second scenario the drug-device model is triggered, and a second set of inputs are obtained to determine a plurality of drug parameters to ensure optimum deposition at the target region. The second set of inputs comprises a nasal geometry of the subject, the target region of interest (ROI) within the nasal cavity, a plurality of device parameters and a plurality of injection parameters.
[0128] The trained drug-device model (204) is used to process the received scenario corresponds to the second scenario (Referring to FIG.5B).
[0129] The drug-device model generates a plurality of drug parameters to ensure optimum deposition at the target region of interest (ROI).
[0130] The plurality of drug parameters includes drug viscosity, drug density and drug molecule.
[0131] In another embodiment, referring to the Scenario 3, if the received scenario corresponds to the third scenario the drug-device model is triggered, and a third set of inputs are obtained to design a device carrying the drug particle to transport the drug particle at target region. The third set of inputs includes the nasal geometry, the target region of interest (ROI) within the nasal cavity, a plurality drug parameter and a plurality of injection parameters.
[0132] The trained drug-device model (204) is used to process the received scenario corresponds to the third scenario (Referring to FIG.5B) to determine the plurality of device parameters.
[0133] Referring to the steps of method 700, at step 708 via the one or more hardware processors to execute the scenario by corresponding machine learning model to perform within the nasal cavity of the subject, wherein, if the received scenario corresponds to the first scenario, personalized injection parameters or personalized device parameters are determined ensuring optimum deposition of nasal spray particles injected into the subject's nasal cavity at least one target clinical region of interest (ROI), by processing the first set of inputs and optimizing drug deposition at the region of interest estimated by the nasal model using the first set of processed inputs and optimization parameters such as the plurality of injection parameters or the plurality of device parameters.
[0134] Here, the target region optimizer 208 (referring to FIG.2) an optimizer of the nasal model obtains the first set of optimized parameters comprising the plurality of injection parameters or personalized device parameters. The optimizer enables delivery of optimum amount of drug deposited in the target ROI.
[0135] Further, if the received scenario corresponds to the second scenario, the plurality of drug parameters are determined to ensure optimum drug deposition at the target region of interest (ROI) of the nasal cavity, by processing the second set of inputs and optimizing drug deposition at the region of interest estimated by the drug-device model using the second set of processed inputs.
[0136] Further, referring to the target region optimizer 208 (referring to FIG.2) the optimizer of the drug-device model determines the second set of optimized parameters comprising the plurality of drug parameters.
[0137] Further, if the received scenario corresponds to the third scenario, plurality of device parameters are determined to ensure optimum drug deposition at the target region of interest (ROI) of the nasal cavity, by processing the third set of inputs and optimizing drug deposition at the region of interest estimated by the drug-device model using the third set of processed inputs.
[0138] Further, referring to the target region optimizer 208 (referring to FIG.2) the optimizer the optimizer of the drug-device model determines the third set of optimized parameters comprising the plurality of device parameters.
[0139] In one embodiment, the optimum drug deposition over the region of interest (ROI) is determined by initially creating a DOE table from a range of optimization parameters for the selected scenario among the plurality of scenarios. Further, the drug deposition percentages are obtained using the DOE table, processed inputs corresponding to the scenario and the selected model. Then, a surrogate model is created to obtain next set of optimization parameters from an acquisition function. Finally, the above steps are repeated for a predefined number of time until optimized parameters are obtained.
[0140] Referring to the steps of the method 700, at step 710 creating for the scenario via the one or more hardware processors, an augmented reality visualization of the nasal drug particle movement inside the nasal cavity of the subject by a drug particle tracking model.
[0141] Referring to FIG.2, the trained drug particle tracking model is utilized to create the augmented reality visualization of the drug transportation inside the nasal cavity using the plurality of injected parameters.
[0142] Additionally, the system 200 enables to retrain the ensemble of ML models associated with the ML models engine 204 to ensure continuous learning and improve quality of the ML models to process new set of inputs and scenario.
[0143] The written description describes the subject matter herein to enable any person skilled in the art to make and use the embodiments. The scope of the subject matter embodiments is defined by the claims and may include other modifications that occur to those skilled in the art. Such other modifications are intended to be within the scope of the claims if they have similar elements that do not differ from the literal language of the claims or if they include equivalent elements with insubstantial differences from the literal language of the claims.
[0144] The embodiments of present disclosure herein addresses unresolved problem of nasal drug. The embodiment, thus provides method and system for delivery of optimal nasal drug deposition in a target region of nasal cavity. Moreover, the embodiments herein further provides an efficient method to determine drug deposition pattern. The method of the present disclosure demonstrates the on-shelf drug delivered through the nasal cavity using the efficient system. The system optimizes the drug parameters and the device design which can help in improving drug delivery and therapeutic efficacy.
[0145] It is to be understood that the scope of the protection is extended to such a program and in addition to a computer-readable means having a message therein; such computer-readable storage means contain program-code means for implementation of one or more steps of the method, when the program runs on a server or mobile device or any suitable programmable device. The hardware device can be any kind of device which can be programmed including e.g., any kind of computer like a server or a personal computer, or the like, or any combination thereof. The device may also include means which could be e.g., hardware means like e.g., an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or a combination of hardware and software means, e.g., an ASIC and an FPGA, or at least one microprocessor and at least one memory with software processing components located therein. Thus, the means can include both hardware means, and software means. The method embodiments described herein could be implemented in hardware and software. The device may also include software means. Alternatively, the embodiments may be implemented on different hardware devices, e.g., using a plurality of CPUs.
[0146] The embodiments herein can comprise hardware and software elements. The embodiments that are implemented in software include but are not limited to, firmware, resident software, microcode, etc. The functions performed by various components described herein may be implemented in other components or combinations of other components. For the purposes of this description, a computer-usable or computer readable medium can be any apparatus that can comprise, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device.
[0147] The illustrated steps are set out to explain the exemplary embodiments shown, and it should be anticipated that ongoing technological development will change the manner in which particular functions are performed. These examples are presented herein for purposes of illustration, and not limitation. Further, the boundaries of the functional building blocks have been arbitrarily defined herein for the convenience of the description. Alternative boundaries can be defined so long as the specified functions and relationships thereof are appropriately performed. Alternatives (including equivalents, extensions, variations, deviations, etc., of those described herein) will be apparent to persons skilled in the relevant art(s) based on the teachings contained herein. Such alternatives fall within the scope of the disclosed embodiments. Also, the words “comprising,” “having,” “containing,” and “including,” and other similar forms are intended to be equivalent in meaning and be open ended in that an item or items following any one of these words is not meant to be an exhaustive listing of such item or items or meant to be limited to only the listed item or items. It must also be noted that as used herein and in the appended claims, the singular forms “a,” “an,” and “the” include plural references unless the context clearly dictates otherwise.
[0148] Furthermore, one or more computer-readable storage media may be utilized in implementing embodiments consistent with the present disclosure. A computer-readable storage medium refers to any type of physical memory on which information or data readable by a processor may be stored. Thus, a computer-readable storage medium may store instructions for execution by one or more processors, including instructions for causing the processor(s) to perform steps or stages consistent with the embodiments described herein. The term “computer-readable medium” should be understood to include tangible items and exclude carrier waves and transient signals, i.e., be non-transitory. Examples include random access memory (RAM), read-only memory (ROM), volatile memory, nonvolatile memory, hard drives, CD ROMs, DVDs, flash drives, disks, and any other known physical storage media.
[0149] It is intended that the disclosure and examples be considered as exemplary only, with a true scope of disclosed embodiments being indicated by the following claims.
, Claims:
1. A processor implemented method (700) for delivery of optimal nasal drug deposition in a target region of nasal cavity, the method comprising:
receiving (702) via a one or more hardware processor, a scenario among a plurality of scenarios and corresponding input parameters, wherein the plurality of scenarios includes a first scenario, a second scenario and a third scenario;
processing (704) the input parameters via the one or more hardware processors, to select one of a machine learning model among an ensemble of machine learning models, wherein the ensemble of machine learning models includes a nasal model, and a drug-device model;
analysing (706) via the one or more hardware processors, the input parameters to trigger corresponding machine learning model, wherein the corresponding machine learning model is triggered to process the scenario:
(i) if the received scenario corresponds to the first scenario the nasal model is triggered and a first set of inputs are obtained to determine personalized injection parameters or device parameters corresponding to optimum deposition at the target region, wherein the first set of inputs comprises a nasal CT scan of the subject, a breathing data of the subject, a target region of interest (ROI) within the nasal cavity, and a plurality of prescribed nasal drug parameters and a plurality of nasal spray device parameters,
(ii) if the received scenario corresponds to the second scenario the drug-device model is triggered and a second set of inputs are obtained to determine a plurality of drug parameters, wherein the second set of inputs comprises a breathing rate, nasal geometry, the target region of interest (ROI) within the nasal cavity, a plurality of device parameters and a plurality of injection parameters and
(iii) if the received scenario corresponds to the third scenario the drug-device model is triggered, and a third set of inputs are obtained to design a device, wherein the third set of inputs includes the breathing rate, the nasal geometry, the target region of interest (ROI) within the nasal cavity, the plurality drug parameters, and the plurality of injection parameters,
wherein the breathing data includes a range of breathing rate, the plurality of drug parameters includes a drug viscosity and a drug density along with drug molecule, the plurality of injection parameters includes coordinate of tip position and orientation vector, and the plurality of device parameters includes a drug flow velocity, a nasal spray nozzle disc diameter, a cone angle, a drug particle diameter distribution, and a total number of drug particles;
executing (708) via the one or more hardware processors the scenario by corresponding machine learning model to perform within the nasal cavity of the subject, wherein,
(i) if the received scenario corresponds to the first scenario, personalized injection parameters or personalized device parameters are determined ensuring optimum deposition of nasal spray particles injected into the subject's nasal cavity at least one target clinical region of interest (ROI), by processing the first set of inputs and optimizing drug deposition at the region of interest estimated by the nasal model using the first set of processed inputs and optimization parameters such as the plurality of injection parameters or the plurality of device parameters,
(ii) if the received scenario corresponds to the second scenario, the plurality of drug parameters are determined to ensure optimum drug deposition at the target region of interest (ROI) of the nasal cavity, by processing the second set of inputs and optimizing drug deposition at the region of interest estimated by the drug-device model using the second set of processed inputs, and
(iii) if the received scenario corresponds to the third scenario, plurality of device parameters are determined to ensure optimum drug deposition at the target region of interest (ROI) of the nasal cavity, by processing the third set of inputs and optimizing drug deposition at the region of interest estimated by the drug-device model using the third set of processed inputs; and
creating (710) for the scenario via the one or more hardware processors, an augmented reality visualization of the nasal drug particle movement inside the nasal cavity of the subject by a drug particle tracking model.

2. The processor implemented method as claimed in claim 1, wherein the optimum drug deposition over the region of interest (ROI) is determined by performing the steps of:
creating a DOE table from a range of optimization parameters for the selected scenario among the plurality of scenarios;
obtaining drug deposition percentage using the DOE table, processed inputs corresponding to the scenario and the selected model;
creating a surrogate model and obtaining next set of optimization parameters from an acquisition function; and
repeating the steps for a predefined number of time until optimized parameters are obtained.

3. The processor implemented method as claimed in claim 1, wherein the nasal model is trained to determine the nasal drug particle deposition percentage within the nasal cavity of the subject, the steps of training comprises:
creating a first design of experiments (DOE) table with at least one of the plurality of drug parameters, the plurality of device parameters, the plurality of injection parameters and the subject nasal geometry in parametric form;
generating a training dataset and a testing dataset using a drug deposition rate at each region of interest by performing computational fluid dynamics (CFD) simulation corresponding to each combination of the DOE elements;
training and testing the ensemble of machine learning models using the training dataset and the testing dataset, and ranking each machine learning model in decreasing order to test accuracy;
selecting an optimal nasal model based on ranking to predict optimal nasal geometry after training; and
tuning a plurality of hyper parameters of the selected model to obtain the optimized nasal model.

4. The processor implemented method as claimed in claim 1, wherein the drug-device model is trained to determine the nasal drug particle deposition percentage within the nasal cavity, the step of training comprises:
creating a second design of experiments (DOE) table with at least of the plurality of devices parameters, the plurality of drug parameters, and the plurality of injection parameters;
performing the computational fluid dynamics (CFD) simulation corresponding to each combination of the second design of experiments (DOE) table and obtaining a drug deposition percentage at each region of the nasal cavity;
generating a test dataset and a training dataset by splitting each combination from the second DOE table to obtain the drug deposition rate at each region of interest;
training and testing ensemble of machine learning models with the training dataset and the testing dataset and ranking each model in decreasing order to test accuracy;
selecting an optimal model to estimate of drug parameters deposition percentage at each region after training; and
tuning the plurality of hyper parameters to obtain the optimized drug-device model.

5. The processor implemented method as claimed in claim 1, wherein the drug particle tracking model is trained to obtain position and velocity data of drug particle movement inside the nasal cavity at each time step, the step of training comprises:
creating a third design of experiments (DOE) table using the plurality of drug parameters, the plurality of device parameters and the nasal geometry in parametric form;
performing the computational fluid dynamics (CFD) simulation corresponding to each of the third DOE elements and obtaining a position and a velocity of each drug particle at different time intervals;
receiving the position and the velocity of each drug particle along with the particle size as input parameter and position and velocity of the drug particle as output parameter;
segmenting the data set into the test data set and the training dataset;
training and testing the ensemble of machine learning models with the training dataset and the testing dataset and ranking each model in decreasing order to test accuracy;
selecting the drug particle tracking model to predict an optimum velocity of each drug particle after training; and
tuning the plurality of hyper parameters to obtain optimized drug particle tracking model.

6. The processor implemented method as claimed in claim 3, wherein the CFD model is utilized to determine the nasal drug particle deposition percentage of at least one target clinical region of interest within the nasal cavity by,
obtaining the 3D nasal geometry from the nasal CT scan;
segmenting the 3D nasal cavity into at least one clinical region of interest (ROI);
generating a computational mesh from the 3D nasal model;
constructing and simulating a CFD model to estimate the drug particle deposition over the plurality of clinical regions of interest (ROI) based on the breathing rate, and the injected drug; and
calculating the drug particle deposition percentage at every clinical region among the plurality of clinical regions of interest (ROI) within the nasal cavity.

7. A system (100) for delivery of optimal nasal drug deposition in a target region of nasal cavity comprising:
a memory (102) storing instructions;
one or more communication interfaces (106); and
one or more hardware processors (104) coupled to the memory (102) via the one or more communication interfaces (106), wherein the one or more hardware processors (104) are configured by the instructions to:
receive a scenario among a plurality of scenarios and corresponding input parameters, wherein the plurality of scenarios includes a first scenario, a second scenario and a third scenario;
process the input parameters, to select one of a machine learning model among an ensemble of machine learning models, wherein the ensemble of machine learning models includes a nasal model, and a drug-device model;
analyse the input parameters to trigger corresponding machine learning model, wherein the corresponding machine learning model is triggered to process the scenario:
(i) if the received scenario corresponds to the first scenario the nasal model is triggered and a first set of inputs are obtained to determine personalized injection parameters or device parameters corresponding to optimum deposition at the target region, wherein the first set of inputs comprises a nasal CT scan of the subject, a breathing data of the subject, a target region of interest (ROI) within the nasal cavity, and a plurality of prescribed nasal drug parameters and a plurality of nasal spray device parameters,
(ii) if the received scenario corresponds to the second scenario the drug-device model is triggered and a second set of inputs are obtained to determine a plurality of drug parameters, wherein the second set of inputs comprises a breathing rate, nasal geometry, the target region of interest (ROI) within the nasal cavity, a plurality of device parameters and a plurality of injection parameters and
(iii) if the received scenario corresponds to the third scenario the drug-device model is triggered, and a third set of inputs are obtained to design a device, wherein the third set of inputs includes the breathing rate, the nasal geometry, the target region of interest (ROI) within the nasal cavity, the plurality drug parameters, and the plurality of injection parameters,
wherein the breathing data includes a range of breathing rate, the plurality of drug parameters includes a drug viscosity and a drug density along with drug molecule, the plurality of injection parameters includes coordinate of tip position and orientation vector, and the plurality of device parameters includes a drug flow velocity, a nasal spray nozzle disc diameter, a cone angle, a drug particle diameter distribution, and a total number of drug particles;
execute the scenario by corresponding machine learning model to perform within the nasal cavity of the subject, wherein,
(i) if the received scenario corresponds to the first scenario, personalized injection parameters or personalized device parameters are determined ensuring optimum deposition of nasal spray particles injected into the subject's nasal cavity at least one target clinical region of interest (ROI), by processing the first set of inputs and optimizing drug deposition at the region of interest estimated by the nasal model using the first set of processed inputs and optimization parameters such as the plurality of injection parameters or the plurality of device parameters,
(ii) if the received scenario corresponds to the second scenario, the plurality of drug parameters are determined to ensure optimum drug deposition at the target region of interest (ROI) of the nasal cavity, by processing the second set of inputs and optimizing drug deposition at the region of interest estimated by the drug-device model using the second set of processed inputs, and
(iii) if the received scenario corresponds to the third scenario, plurality of device parameters are determined to ensure optimum drug deposition at the target region of interest (ROI) of the nasal cavity, by processing the third set of inputs and optimizing drug deposition at the region of interest estimated by the drug-device model using the third set of processed inputs; and
create for the scenario an augmented reality visualization of the nasal drug particle movement inside the nasal cavity of the subject by a drug particle tracking model.

8. The system of claim 7, wherein the optimum drug deposition over the region of interest (ROI) is determined by performing the steps of:
creating a DOE table from a range of optimization parameters for the selected scenario among the plurality of scenarios;
obtaining drug deposition percentage using the DOE table, processed inputs corresponding to the scenario and the selected model;
creating a surrogate model and obtaining next set of optimization parameters from an acquisition function; and
repeating the steps for a predefined number of time until optimized parameters are obtained.

9. The system of claim 7, wherein the nasal model is trained to determine the nasal drug particle deposition percentage within the nasal cavity of the subject, the step of training comprises:
creating a first design of experiments (DOE) table with at least one of the plurality of drug parameters, the plurality of device parameters, the plurality of injection parameters and the subject nasal geometry in parametric form;
generating a training dataset and a testing dataset using a drug deposition rate at each region of interest by performing computational fluid dynamics (CFD) simulation corresponding to each combination of the DOE elements;
training and testing the ensemble of machine learning models using the training dataset and the testing dataset, and ranking each machine learning model in decreasing order to test accuracy;
selecting an optimal nasal model based on ranking to predict optimal nasal geometry after training; and
tuning a plurality of hyper parameters of the selected model to obtain the optimized nasal model.

10. The system of claim 7, wherein the drug-device model is trained to determine the nasal drug particle deposition percentage within the nasal cavity t, the step of training comprises:
creating a second design of experiments (DOE) table with at least of the plurality of devices parameters, the plurality of drug parameters, and the plurality of injection parameters;
performing the computational fluid dynamics (CFD) simulation corresponding to each combination of the second design of experiments (DOE) table and obtaining a drug deposition percentage at each region of the nasal cavity;
generating a test dataset and a training dataset by splitting each combination from the second DOE table to obtain the drug deposition rate at each region of interest;
training and testing ensemble of machine learning models with the training dataset and the testing dataset and ranking each model in decreasing order to test accuracy;
selecting an optimal model to estimate of drug parameters deposition percentage at each region after training; and
tuning the plurality of hyper parameters to obtain the optimized drug-device model.

11. The system of claim 7, wherein the drug particle tracking model is trained to obtain position and velocity data of drug particle movement inside the nasal cavity at each time step, the step of training comprises:
creating a third design of experiments (DOE) table using the plurality of drug parameters, the plurality of device parameters and the nasal geometry in parametric form;
performing the computational fluid dynamics (CFD) simulation corresponding to each of the third DOE elements and obtaining a position and a velocity of each drug particle at different time intervals;
receiving the position and the velocity of each drug particle along with the particle size as input parameter and position and velocity of the drug particle as output parameter;
segmenting the data set into the test data set and the training dataset;
training and testing the ensemble of machine learning models with the training dataset and the testing dataset and ranking each model in decreasing order to test accuracy;
selecting the drug particle tracking model to predict an optimum velocity of each drug particle after training; and
tuning the plurality of hyper parameters to obtain optimized drug particle tracking model.

12. The system of claim 9, wherein the CFD model is utilized to determine the nasal drug particle deposition percentage of at least one target clinical region of interest within the nasal cavity by,
obtaining the 3D nasal geometry from the nasal CT scan;
segmenting the 3D nasal cavity into at least one clinical region of interest (ROI);
generating a computational mesh from the 3D nasal model;
constructing and simulating a CFD model to estimate the drug particle deposition over the plurality of clinical regions of interest (ROI) based on the breathing rate, and the injected drug; and
calculating the drug particle deposition percentage at every clinical region among the plurality of clinical regions of interest (ROI) within the nasal cavity.