Description:FIELD OF THE INVENTION
The present invention generally relates to obstructed lung diagnosis. More particularly, the present invention relates to a cost-effective reliable system for assessing pleural effusion by analysing chest X-ray image though deep learning models. The present invention aims to enhance cardiopulmonary care, bridge healthcare gaps, and optimize patient outcomes in both clinical and resource-constrained settings.

BACKGROUND OF THE INVENTION
Pleural effusion represents a pathological condition characterized by accumulation of fluid in the pleural space, an area between the tissues, which lines the lungs, and the chest wall. Pleural effusion primarily manifests as an abnormal buildup of fluid, often indicative of obstructed lungs. Pleural effusion has a broader spectrum of etiologies including but not limited to congestive heart failure, pneumonia, malignancies, pulmonary embolism, and liver or kidney diseases. Common symptoms of include dyspnea (shortness of breath), chest pain, cough, and reduced exercise tolerance. In some cases, pleural effusion may be asymptomatic and incidentally discovered during routine medical examinations or imaging studies.

Detecting pleural effusion is crucial for diagnosing and managing underlying medical conditions promptly. Pleural effusion may coexist with pleural thickening, but it can also occur independently. Therefore, accurate identification and differentiation of pleural effusion are essential for appropriate clinical management. Further, the patients having mild form of such condition or at its primitive stage very often go undiagnosed or unnoticed by inexperienced doctors.

Utilizing imaging modalities such as X-rays plays a pivotal role in diagnosing obstructed lungs. X-ray imaging offers several advantages, including non-invasiveness, cost-effectiveness, and widespread availability in healthcare settings. Despite the involvement of radiation, X-rays provide valuable diagnostic information while exposing patients to relatively low radiation doses compared to alternative modalities such as CT scans. Further, the X-ray checkup appears to be affordable and easily accessible.

With the recent development of machine learning technology, the object detection algorithms based on convolutional neural networks are being used in medical and diagnosis fields. However, all the existing machine leaning-based diagnostic techniques have several limitations in terms of real-time result delivery, implementation in low end computing device, computing speed, computing resource utilization, type of disease detection, diagnosis accuracy etc. Therefore, it is required come up with a cost-effective, non-invasive, and reliable approach to examine obstructed lungs, especially to assess severity/stages of pleural effusion.

A reference may be made to CA3140122A1 that discloses a method/system for identifying an anomaly in a chest X-ray, in which convolutional networks are trained to determine a probability of presence of broad range of pathologies: atelectasis, cardiomegaly, pleural effusion, infiltration, mass, nodule, pneumonia, pneumothorax, consolidation, edema, emphysema, fibrosis, pleural thickening, and diaphragmatic hernia. If any anomaly is detected, the system may generate a heat map to provide visual information about the regions in the image contributing to the identification of the pathology.

Another reference may be made to US10691980B1 that discloses a system and method for multi-abnormality classification based on chest X-ray images CNN and DBN models are deployed to predict abnormality classification scores for wide range abnormalities such as granuloma, infiltrate, nodule, scaring, effusion, atelectasis, bone/soft tissue lesion, fibrosis, cardiac abnormality, mass, pneumothorax, COPD, consolidation, pleural thickening, cardiomegaly, emphysema, edema, pneumonia, hilar abnormality, or hernias.

One more reference may be made to Indian patent application number 202223019813 that discloses a deep learning method to diagnose severity level of seventeen lung diseases using either X-ray image or CT scan images, wherein the method deploys combinative architecture of XChes13Net2.0 and YOLOV5, and uses RGB coloured heatmap and bounding box techniques.

Further reference may be made to US11436725B2 that discloses a self-supervised chest X-ray image analysis machine-learning model that utilizes transferable visual words (TransVW) for to reduce annotation effort in different pathologies.

All the existing convolutional neural networks as employed in X-ray image analysis are primarily focused on heatmap technique to detect too broad range of abnormality patterns; therefore, a further precise analysis is required on chest X-ray images to diagnose a specific life-threatening disease, particularly assessing pleural effusion even at very primitive stage that cannot be visually noticeable by naked eyes in X-ray images. Moreover, there are some critical anatomical characteristics such as gap between the ribs and the lungs, and marks/patches on the lungs, which need to be examined very cautiously and meticulously to show results of plural thickening location in an online user interface instantly. Therefore, it is required to devise a user-friendly, low-cost, and reliable system for assessing pleural effusion through advanced deep learning-based image segmentation and annotation application on chest X-ray images, which includes all the advantages of the conventional/existing techniques/methodologies and overcomes the deficiencies of such techniques/methodologies.

OBJECT OF THE INVENTION
It is an object of the present invention to empower healthcare providers with a reliable and efficient tool for early diagnosis obstructed lungs without much complicated tests, facilitating timely interventions and improving patient outcomes.

It is another object of the present invention to automate identification of visually unnoticeable pathological conditions of the obstructed lungs which could be indications pleural effusion at very primitive stage.

It is one more object of the present invention to develop novel advanced deep learning models for precise diagnosis of pleural effusion using frontal (posteroanterior) chest X-ray images.

It is a further object of the present invention is to devise a user-friendly, cost-effective, and reliable computing (online) system for assessing pleural effusion from chest X-ray images.

SUMMARY OF THE INVENTION
In one aspect, the present invention provides a user-friendly, reliable, and cost-effective system for assessing pleural effusion by analysing chest X-ray image though deep learning models such as YoloV8. The system comprises an comprising an image upload field for uploading chest X-ray images, and a result display field for displaying indication of the pleural effusion in real-time; and a server being in communication with the user interface via a wireless network. The server is configured to: deploy a neural network model on the uploaded X-ray image to locate thereon a first set of coordinates (three key points on left and right side on chest X-ray image) along a region between lower corner edge of lungs and adjacent upper edge of diaphragm, and a second set of coordinates coordinates (three key points on left and right side on chest X-ray image) along a region between lower side edge of heart and adjacent upper edge of diaphragm; execute angle measurement codes to measure costophrenic angle using the first set of coordinates, and cardiophrenic angle using the second set of coordinates; and feed the measured angle values into a threshold checking module to determine whether the X-ray image content corresponds to healthy or unhealthy stages based on comparison of the measured angle values with threshold values as set/stored in the threshold checking module. The healthy stage indicates absence of pleural effusion, and the unhealthy stage indicates presence of mild pleural effusion or severe pleural effusion.

Other aspects, advantages, and salient features of the present invention will become apparent to those skilled in the art from the following detailed description, which delineate the present invention in different embodiments.

BRIEF DESCRIPTION OF DRAWINGS
These and other features, aspects, and advantages of the present invention will become better understood when the following detailed description is read with reference to the accompanying figures.

Fig. 1 illustrates different key points used in measuring costophrenic angle (L) (Fig. 1a) and cardiophrenic angle (H) (Fig. 1b), and both angles marked in frontal (posteroanterior) chest X-ray image (Fig. 1c) with corresponding coordinates (Fig. 1d).

Fig. 2 illustrates different hardware components of the system for assessing pleural effusion, in accordance with an embodiment of the present invention.

Fig. 3 illustrates X-ray image analysis operational steps for assessing pleural effusion, in accordance with an embodiment of the present invention.

Fig. 4 illustrates Yolo-V8 model architecture, in accordance with an exemplary embodiment of the present invention.

List of reference numerals
100 user interface
102 image upload field
104 result display field
200 server (processing unit)
202 neural network model
204 measurement codes
206 threshold checking module
L costophrenic angle
H cardiophrenic angle

DETAILED DESCRIPTION OF THE INVENTION
Various embodiments described herein are intended only for illustrative purposes and subject to many variations. It is understood that various omissions and substitutions of equivalents are contemplated as circumstances may suggest or render expedient, but are intended to cover the application or implementation without departing from the scope of the present invention. Also, it is to be understood that the phraseology and terminology used herein is for the purpose of description and should not be regarded as limiting.

The use of terms “comprises/comprising”, ‘includes/including’ or “having/have/has” and variations thereof herein are meant to encompass the items listed thereafter and equivalents thereof as well as additional items. Further, the terms, “an” and “a” herein do not denote a limitation of quantity, but rather denote the presence of at least one of the referenced items.

According to an embodiment of the present invention, as shown in Fig. 1, the present invention selects two major anatomical (morphological) factors of cardio-pulmonary regions such as blunting of costophrenic angle (L) (Fig. 1a) and cardiophrenic angle (H) (Fig. 1b) for precise analysis of obstructed lung images. The costophrenic angle (L) is formed between lower corner edges of lungs and adjacent upper edges of diaphragm. The cardiophrenic angle (H) is formed between lower side edges of heart and adjacent upper edge of diaphragm. Fig, 1c shows a sample of frontal (posteroanterior) chest X-ray image in which both costophrenic angle (L) and cardiophrenic angle (H) are annotated. At the very early/primitive stage, the blunting of the said angles in the X-ray images very often may not be noticed by the radiologist. Further, frontline medical staffs including nurses and in-experienced doctors find difficulty in distinguishing pleural effusion condition from other obstructed lung condition. Therefore, the present invention makes assessment of pleural effusion fully automated, user-friendly, and simplified. Further, the present invention provides an online (website) or a mobile app-based system through which the healthcare providers just upload the raw chest frontal (posteroanterior) X-ray images, and the accurate results are displayed in real-time. The results include whether the patient has any indication of pleural effusion or not, and severity/stage of pleural effusion, if detected.

According to an embodiment of the present invention, as shown in Fig. 1, the system for assessing pleural effusion is depicted. The system comprises a user interface (100), and a server (processing unit) (200) being in communication with the user interface (100) via a wireless network. The user interface (100) is web-based or a mobile app interface. The user interface (100) comprises an image upload section/field (102) for uploading chest X-ray images, and a result display section/field (104) for displaying indication of the pleural effusion in real-time. The server (200) comprises a memory and a processor, where the memory stores a set of processor executable codes/modules (software/algorithm) to carry out pleural effusion assessment operation.

According to an embodiment of the present invention, the server (200) has embedded therein a neural network model (202); angle measurement codes (204); a threshold checking module (206).

According to an embodiment of the present invention, the neural network model (202) is deployed on the uploaded X-ray image to locate thereon a first set of coordinates along a region between lower corner edge of lungs and adjacent upper edge of diaphragm, and a second set of coordinates along a region between lower side edge of heart and adjacent upper edge of diaphragm. For example, as shown in Fig. 1c - 1d, the model is trained to identify three key points (pixels) on each of the target regions in chest X-ray image, and corresponding coordinates of the identified key pixels [such as (x1, y1) for 1st key pixel point, (x2, y2) for 2nd key pixel point, (x3, y3) for 3rd key pixel point].

According to an embodiment of the present invention, the angle measurement codes (204) are to measure costophrenic angle (L) using the first set of coordinates, and cardiophrenic angle (H) using the second set of coordinates. For scripting the codes, Python 3.10.12 is used along with ultralytics library. For example, based on coordinate values of 3 key points both the angles are computed using equation 1.
secϴ=((|BA) ̅||(BC) ̅|)/((BA.) ̅(BC) ̅ )
(BA) ̅=A-B=(x1,y1)-(x2,y2)
(BC) ̅=(x3,y3)-(x2,y2)
|(BC) ̅ |= √((x3-x2)^2+(y3-y2)^2 )
|(BA) ̅ |= √(〖(x2-x1)〗^2-〖(y2-y1)〗^2 )
L or H = ϴ = 〖sec〗^(-1) [1/(((BA) ̅.(BC) ̅)/(|(BA) ̅ ||(BC) ̅|))] equation 1

Where (BA) ̅ and (BC) ̅ represent vectors indicating positions of two corresponding points
|(BC) ̅ | and |(BA) ̅ | represent magnitudes of said vectors

According to an embodiment of the present invention, the measured angle values (L, H)are fed into the threshold checking module (206) to determine whether the X-ray image content corresponds to healthy or unhealthy stages based on comparison of the measured angle values (L, H) with threshold values as set in the threshold checking module (206). The healthy stage indicates absence of pleural effusion, and the unhealthy stage indicates presence of mild pleural effusion or severe pleural effusion. The threshold angle parameters (values) are set/stored/defined in the server, and the measured angle values (L, H, ) are compared with the threshold angle parameters as shown in Table 1 to indicate appropriate stages of the pleural effusion in form of text message to be displayed in the result display section (104).

Table 1
Threshold Angle Parameters Indication (Stage Classification)
Costophrenic Angle = 30-45 degrees Normal (absence of pleural effusion)
Costophrenic Angle > 45 degrees Mild Pleural Effusion
Cardiophrenic Angle = 80-110 degrees Normal (healthy lung)
Costophrenic Angle >45 degrees
Cardiophrenic Angle >110 degrees Severe Pleural Effusion
Invisible Angle Regions
(i.e., Half/Full lung region invisible) Severe Pleural Effusion

In an exemplary embodiment, the threshold checking module (206) defines a state possessing the costophrenic angle (L) of 30-45 degrees and/or the cardiophrenic angle (H) of 80-110 degrees as an indicative of the healthy stage (the absence of pleural effusion), and it is displayed on the result display field (104).

In an exemplary embodiment, the threshold checking module (206) defines a state possessing the costophrenic angle (L) greater than 45 degrees and/or the cardiophrenic angle (H) of 80-110 degrees as an indicative of the mild pleural effusion stage, and it is displayed on the result display field (104).

In an exemplary embodiment, the threshold checking module (206) defines a state possessing the costophrenic angle (L) greater than 45 degrees and the cardiophrenic angle (H) greater than 110 degrees as an indicative of the severe pleural effusion stage, and it is displayed on the result display field (104).

In an exemplary embodiment, the threshold checking module (206) defines a state where any region of the angles (L, H) is invisible as an indicative of the severe pleural effusion stage, and it is displayed on the result display field (104).

According to an embodiment of the present invention, as shown in Fig. 4, the frontal (posteroanterior) chest X-ray image analysis operation for detecting pleural thickening is depicted. The image analysis operation includes steps of: uploading (S1) chest X-ray images through an image upload field of a user interface; preprocessing (S2) the images by converting their size into defined dimensions; locating/identifying (S3) on the images a first set of coordinates (3 key points on each side) along a region between lower corner edge of lungs and adjacent upper edge of diaphragm, and a second set of coordinates (3 key points on each side) along a region between lower side edge of heart and adjacent upper edge of diaphragm; measuring (S4) costophrenic angle using the first set of coordinates, and cardiophrenic angle using the second set of coordinates; determining (S5) whether the image content corresponds to healthy or unhealthy stages based on comparison of the measured angle values with threshold values; and displaying (S6) stages of the pleural effusion in the result display field of the user interface.

According to an embodiment of the present invention, as shown in Fig. 5, the neural network model (202) is trained through YoloV8 (You Only Look Once) architecture. A dataset is prepared by collecting frontal (posteroanterior) chest X-ray images of healthy unhealthy patients from various diagnostic centres, followed by preprocessing of the said images in term of resizing each image into 1024 x 1024 dimensions, and annotating different pixel points and/or contour corresponding to various anatomical parts of lungs, heart, diaphragm, ribs, chest wall which appear in the X-ray images. There are marked three key points on each image, and the Yolo V8- pose model is trained for generating three key points. Once these points are generated then the costophrenic/cardiophrenic angles angle are calculated using equation 1. Preferably, the plural lines/layers/space variations with respect to changes in the costophrenic/cardiophrenic angles are considered as learning factors to differentiate the obstructed lungs (plural effusion condition) from healthy lungs. The model is trained, optimized, and validated. Finally, the results are analyzed for insights and improvements. The YOLO V8 architecture selected for the proposed model has 1 input units/layers, 53 hidden units/layers (DarkNet 53), and 1 output unit/layer. The different hyperparameters used in the YOLO V8 are stochastic gradient descent optimizer, with a momentum of 0.937 and trained for 100 epochs in one A100 GPU with each image size of 1024 x 1024.

Further, it is observed that the proposed model demonstrates robust performance in detecting pleural effusion, achieving an accuracy of 96.54%. With a precision of 0.9537, it showcases high correctness in identifying pleural effusion cases accurately. Additionally, the model exhibits a sensitivity of 97.02%, effectively capturing a significant portion of true positive cases, while maintaining a specificity of 95.22%, thereby minimizing the occurrence of false positives and false negatives in the diagnosis of pleural effusion

The foregoing descriptions of exemplary embodiments of the present invention have been presented for purposes of illustration and description. They are not intended to be exhaustive or to limit the invention to the precise forms disclosed, and obviously many modifications and variations are possible in light of the above teaching. The exemplary embodiment was chosen and described in order to best explain the principles of the invention and its practical application, to thereby enable the persons skilled in the art to best utilize the invention and various embodiments with various modifications as are suited to the particular use contemplated. It is understood that various omissions, substitutions of equivalents are contemplated as circumstance may suggest or render expedient, but is intended to cover the application or implementation without departing from the scope of the claims of the present invention. , Claims:We Claim:

1. A system for assessing pleural effusion, the system comprises:
a user interface (100) comprising an image upload field (102) for uploading chest X-ray images, and a result display field (104) for displaying indication of the pleural effusion in real-time; and
a server (200) being in communication with the user interface (100) via a wireless network, wherein the server (200) is configured to:
deploy a neural network model (202) on the uploaded X-ray image to locate thereon a first set of coordinates along a region between lower corner edge of lungs and adjacent upper edge of diaphragm, and a second set of coordinates along a region between lower side edge of heart and adjacent upper edge of diaphragm;
execute angle measurement codes (204) to measure costophrenic angle (L) using the first set of coordinates, and cardiophrenic angle (H) using the second set of coordinates; and
feed the measured angle values (L, H) into a threshold checking module (206) to determine whether the X-ray image content corresponds to healthy or unhealthy stages based on comparison of the measured angle values (L, H) with threshold values as set in the threshold checking module (206), wherein the healthy stage indicates absence of pleural effusion, and the unhealthy stage indicates presence of mild pleural effusion or severe pleural effusion.

2. The system as claimed in claim 1, wherein the neural network model (202) is trained through YoloV8 (You Only Look Once) architecture that employs stochastic gradient descent optimizer with momentum of 0.937.

3. The system as claimed in claim 1, wherein the user interface (100) is web-based or smartphone installable application interface.

4. The system as claimed in claim 1, wherein the threshold checking module (206) defines a state possessing the costophrenic angle (L) of 30-45 degrees and the cardiophrenic angle (H) of 80-110 degrees as an indicative of the healthy stage (the absence of pleural effusion) that is displayed in the result display field (104).

5. The system as claimed in claim 1, wherein the threshold checking module (206) defines a state possessing the costophrenic angle (L) greater than 45 degrees and the cardiophrenic angle (H) of 80-110 degrees as an indicative of the mild pleural effusion stage that is displayed in the result display field (104).

6. The system as claimed in claim 1, wherein the threshold checking module (206) defines a state possessing the costophrenic angle (L) greater than 45 degrees and the cardiophrenic angle (H) greater than 110 degrees as an indicative of the severe pleural effusion stage that is displayed in the result display field (104).

7. The system as claimed in claim 1, wherein the threshold checking module (206) defines a state where any region of the angles (L, H) is invisible as an indicative of the severe pleural effusion stage that is displayed in the result display field (104).