Novel Image Processing Method for Detecting Strep Throat (Streptococcal Pharyngitis) Using Smartphone

Diagnostic utility of pharyngeal follicular structures in COVID-19: A large-scale cross-sectional study

OBJECTIVES

Pharyngeal follicles similar to those seen in influenza have been observed in patients with coronavirus disease 2019 (COVID-19), suggesting their potential as early-stage diagnostic markers. In this study, we examined the diagnostic potential of pharyngeal follicles for COVID-19, particularly the Omicron variant and its subtypes, to obtain basic data for AI-based diagnostic imaging tools.

METHODS

A cross-sectional study was conducted from July 21, 2022, to March 31, 2023, at the Tokyo Shinagawa Hospital's fever clinic. Participants aged ≥15 years who underwent real-time polymerase chain reaction testing for COVID-19 and pharyngeal examinations were included. Demographic details, symptom onset, throat pain, and vaccination status were also recorded. Pharyngeal structures were categorized into four groups: follicles, buds, mixed, or absent.

RESULTS

Of the 1223 participants, 829 (67.8%) tested positive for COVID-19. Among those who tested positive, 73.6% (95% CI: 70.6%-76.6%) had follicular structures, compared to 52.8% (95% CI: 47.9%-57.7%) of those who tested negative (P = 1.0 × 10-12). Overall, 818 participants exhibited follicular structures (439 with follicles, 281 with buds, and 98 with mixed structures), while 405 lacked any follicular structures. Regression analysis identified throat pain and follicular structures as significant COVID-19 predictors (95% confidence intervals: 2.49-4.85 and 1.43-2.59, respectively). Mixed follicles were identified as a potentially characteristic feature of COVID-19.

CONCLUSION

Pharyngeal follicular structures demonstrated high sensitivity for early COVID-19 diagnosis.

Harnessing AI for precision tonsillitis diagnosis: a revolutionary approach in endoscopic analysis

BACKGROUND

Diagnosing and treating tonsillitis pose no significant challenge for otolaryngologists; however, it can increase the infection risk for healthcare professionals amidst the coronavirus pandemic. In recent years, with the advancement of artificial intelligence (AI), its application in medical imaging has also thrived. This research is to identify the optimal convolutional neural network (CNN) algorithm for accurate diagnosis of tonsillitis and early precision treatment.

METHODS

Semi-supervised learning with pseudo-labels used for self-training was adopted to train our CNN, with the algorithm including UNet, PSPNet, and FPN. A total of 485 pharyngoscopic images from 485 participants were included, comprising healthy individuals (133 cases), patients with the common cold (295 cases), and patients with tonsillitis (57 cases). Both color and texture features from 485 images are extracted for analysis.

RESULTS

UNet outperformed PSPNet and FPN in accurately segmenting oropharyngeal anatomy automatically, with average Dice coefficient of 97.74% and a pixel accuracy of 98.12%, making it suitable for enhancing the diagnosis of tonsillitis. The normal tonsils generally have more uniform and smooth textures and have pinkish color, similar to the surrounding mucosal tissues, while tonsillitis, particularly the antibiotic-required type, shows white or yellowish pus-filled spots or patches, and shows more granular or lumpy texture in contrast, indicating inflammation and changes in tissue structure. After training with 485 cases, our algorithm with UNet achieved accuracy rates of 93.75%, 97.1%, and 91.67% in differentiating the three tonsil groups, demonstrating excellent results.

CONCLUSION

Our research highlights the potential of using UNet for fully automated semantic segmentation of oropharyngeal structures, which aids in subsequent feature extraction, machine learning, and enables accurate AI diagnosis of tonsillitis. This innovation shows promise for enhancing both the accuracy and speed of tonsillitis assessments.

Fifteen-minute consultation: Group A streptococcal pharyngitis, diagnosis and treatment in children

Group A streptococcus (GAS) is the most common bacterial cause of pharyngitis in children. GAS causes significant suppurative and non-suppurative complications including invasive GAS disease and acute rheumatic fever. This article describes the current epidemiology and clinical presentation of GAS pharyngitis and explores how diagnostic and treatment decisions differ globally. Several key decision support tools are discussed including international guidelines, clinical decision scores and laboratory tests along with the evidence for treatment choice and duration. With recent international reports describing an increase in GAS infections, clinicians should be familiar with their local GAS pharyngitis guidelines and the rationale for diagnosis and treatment of this common childhood illness.

Tonsils at Telethon: developing a standardised collection of tonsil photographs for group A streptococcal (GAS) research

Introduction

Group A streptococcus (GAS) infections, such as pharyngitis and impetigo, can lead to rheumatic fever and rheumatic heart disease (RHD). Australian Aboriginal and Torres Strait Islander populations experience high rates of RHD and GAS skin infection, yet rates of GAS pharyngitis are unclear. Anecdotally, clinical presentations of pharyngitis, including tonsillar hypertrophy and sore throat, are uncommon. This study aimed to develop a standardised set of tonsil photographs and determine tonsil size distribution from an urban paediatric population.

Methods

A prospective cohort of children aged 3-15 years were recruited at the public events "Discover Day" and "Telethon Weekend" (October 2017) in Perth, Western Australia, Australia. Tonsil photographs, symptomatology, and GAS rapid antigen detection tests (RADT) were collected. Tonsil size was graded from the photographs using the Brodsky Grading Scale of tonsillar hypertrophy (Brodsky) by two independent clinicians, and inter-rater reliability calculated. Pharyngitis symptoms and GAS RADT were correlated, and immediate results provided.

Results

Four hundred and twenty-six healthy children participated in the study over three days. The median age was seven years [interquartile range (IQR) 5.9-9.7 years]. Tonsil photographs were collected for 92% of participants, of which 62% were rated as good-quality photographs and 79% were deemed of adequate quality for assessment by both clinicians. When scored by two independent clinicians, 57% received the same grade. Average Brodsky grades (between clinicians) were 11%, 35%, 28%, 22% and 5% of grades 0,1,2,3 and 4, respectively. There was moderate agreement in grading using photographs, and minimal to weak agreement for signs of infection. Of 394 participants, 8% reported a sore throat. Of 334 GAS RADT performed, <1% were positive.

Discussion

We report the first standardised use of paediatric tonsil photographs to assess tonsil size in urban-living Australian children. This provides a proof of concept from an urban-living cohort that could be compared with children in other settings with high risk of GAS pharyngitis or rheumatic fever such as remote-living Australian Indigenous populations.

Deep Learning Model and its Application for the Diagnosis of Exudative Pharyngitis

OBJECTIVES

Telemedicine is firmly established in the healthcare landscape of many countries. Acute respiratory infections are the most common reason for telemedicine consultations. A throat examination is important for diagnosing bacterial pharyngitis, but this is challenging for doctors during a telemedicine consultation. A solution could be for patients to upload images of their throat to a web application. This study aimed to develop a deep learning model for the automated diagnosis of exudative pharyngitis. Thereafter, the model will be deployed online.

METHODS

We used 343 throat images (139 with exudative pharyngitis and 204 without pharyngitis) in the study. ImageDataGenerator was used to augment the training data. The convolutional neural network models of MobileNetV3, ResNet50, and EfficientNetB0 were implemented to train the dataset, with hyperparameter tuning.

RESULTS

All three models were trained successfully; with successive epochs, the loss and training loss decreased, and accuracy and training accuracy increased. The EfficientNetB0 model achieved the highest accuracy (95.5%), compared to MobileNetV3 (82.1%) and ResNet50 (88.1%). The EfficientNetB0 model also achieved high precision (1.00), recall (0.89) and F1-score (0.94).

CONCLUSIONS

We trained a deep learning model based on EfficientNetB0 that can diagnose exudative pharyngitis. Our model was able to achieve the highest accuracy, at 95.5%, out of all previous studies that used machine learning for the diagnosis of exudative pharyngitis. We have deployed the model on a web application that can be used to augment the doctor's diagnosis of exudative pharyngitis.

Point-wise spatial network for identifying carcinoma at the upper digestive and respiratory tract

PROBLEM

Artificial intelligence has been widely investigated for diagnosis and treatment strategy design, with some models proposed for detecting oral pharyngeal, nasopharyngeal, or laryngeal carcinoma. However, no comprehensive model has been established for these regions.

AIM

Our hypothesis was that a common pattern in the cancerous appearance of these regions could be recognized and integrated into a single model, thus improving the efficacy of deep learning models.

METHODS

We utilized a point-wise spatial attention network model to perform semantic segmentation in these regions.

RESULTS

Our study demonstrated an excellent outcome, with an average mIoU of 86.3%, and an average pixel accuracy of 96.3%.

CONCLUSION

The research confirmed that the mucosa of oral pharyngeal, nasopharyngeal, and laryngeal regions may share a common appearance, including the appearance of tumors, which can be recognized by a single artificial intelligence model. Therefore, a deep learning model could be constructed to effectively recognize these tumors.

Convolutional neural network-based vocal cord tumor classification technique for home-based self-prescreening purpose

BACKGROUND

In this study, we proposed a deep learning technique that can simultaneously detect suspicious positions of benign vocal cord tumors in laparoscopic images and classify the types of tumors into cysts, granulomas, leukoplakia, nodules and polyps. This technique is useful for simplified home-based self-prescreening purposes to detect the generation of tumors around the vocal cord early in the benign stage.

RESULTS

We implemented four convolutional neural network (CNN) models (two Mask R-CNNs, Yolo V4, and a single-shot detector) that were trained, validated and tested using 2183 laryngoscopic images. The experimental results demonstrated that among the four applied models, Yolo V4 showed the highest F1-score for all tumor types (0.7664, cyst; 0.9875, granuloma; 0.8214, leukoplakia; 0.8119, nodule; and 0.8271, polyp). The model with the lowest false-negative rate was different for each tumor type (Yolo V4 for cysts/granulomas and Mask R-CNN for leukoplakia/nodules/polyps). In addition, the embedded-operated Yolo V4 model showed an approximately equivalent F1-score (0.8529) to that of the computer-operated Yolo-4 model (0.8683).

CONCLUSIONS

Based on these results, we conclude that the proposed deep-learning-based home screening techniques have the potential to aid in the early detection of tumors around the vocal cord and can improve the long-term survival of patients with vocal cord tumors.

Unveiling the Enigmatic Adenoids and Tonsils: Exploring Immunology, Physiology, Microbiome Dynamics, and the Transformative Power of Surgery

Within the intricate realm of the mucosal immune system resides a captivating duo: the adenoids (or pharyngeal tonsils) and the tonsils (including palatine, tubal, and lingual variations), which harmoniously form the Waldeyer's ring. As they are strategically positioned at the crossroads of the respiratory and gastrointestinal systems, these exceptional structures fulfill a vital purpose. They function as formidable "gatekeepers" by screening microorganisms-both bacteria and viruses-with the mission to vanquish local pathogens via antibody production. However, under specific circumstances, their function can take an unsettling turn, inadvertently transforming them into reservoirs for pathogen incubation. In this review, we embark on a fascinating journey to illuminate the distinctive role of these entities, focusing on the local immune system inside their tissues. We delve into their behavior during inflammation processes, meticulously scrutinize the indications for surgical intervention, and investigate the metamorphosis of their microbiota in healthy and diseased states. We explore the alterations that occur prior to and following procedures like adenoidectomy, tonsillectomy, or their combined counterparts, particularly in pediatric patients. By comprehending a wealth of data, we may unlock the key to the enhanced management of patients with otorhinolaryngological disorders. Empowered with this knowledge, we can embrace improved therapeutic approaches and targeted interventions/surgeries guided by evidence-based guidelines and indications.

A pilot study to develop assessment tools for Group A Streptococcus surveillance studies

Introduction

Group A Streptococcus (GAS) causes pharyngitis (sore throat) and impetigo (skin sores) GAS pharyngitis triggers rheumatic fever (RF) with epidemiological evidence supporting that GAS impetigo may also trigger RF in Australian Aboriginal children. Understanding the concurrent burden of these superficial GAS infections is critical to RF prevention. This pilot study aimed to trial tools for concurrent surveillance of sore throats and skins sore for contemporary studies of RF pathogenesis including development of a sore throat checklist for Aboriginal families and pharynx photography.

Methods

Yarning circle conversations and semi-structured interviews were performed with Aboriginal caregivers and used to develop the language and composition of a sore throat checklist. The sore throat story checklist was combined with established methods of GAS pharyngitis and impetigo surveillance (examination, bacteriological culture, rapid antigen detection and serological tests) and new technologies (photography) and used for a pilot cross-sectional surveillance study of Aboriginal children attending their health clinic for a routine appointment. Feasibility, acceptability, and study costs were compiled.

Results

Ten Aboriginal caregivers participated in the sore-throat yarning circles; a checklist was derived from predominant symptoms and their common descriptors. Over two days, 21 Aboriginal children were approached for the pilot surveillance study, of whom 17 were recruited; median age was 9 years [IQR 5.5-13.5], 65% were female. One child declined throat swabbing and three declined finger pricks; all other surveillance elements were completed by each child indicating high acceptability of surveillance assessments. Mean time for screening assessment was 19 minutes per child. Transport of clinical specimens enabled gold standard microbiological and serological testing for GAS. Retrospective examination of sore throat photography concorded with assessments performed on the day.

Conclusion

Yarning circle conversations were effective in deriving culturally appropriate sore throat questionnaires for GAS pharyngitis surveillance. New and established tools were feasible, practical and acceptable to participants and enable surveillance to determine the burden of superficial GAS infections in communities at high risk of RF. Surveillance of GAS pharyngitis and impetgio in remote Australia informs primary RF prevention with potential global translation.

Recent Advances in the Rheumatic Fever and Rheumatic Heart Disease Continuum

Nearly a century after rheumatic fever (RF) and rheumatic heart disease (RHD) was eradicated from the developed world, the disease remains endemic in many low- and middle-income countries (LMICs), with grim health and socioeconomic impacts. The neglect of RHD which persisted for a semi-centennial was further driven by competing infectious diseases, particularly the human immunodeficiency virus (HIV) pandemic. However, over the last two-decades, slowly at first but with building momentum, there has been a resurgence of interest in RF/RHD. In this narrative review, we present the advances that have been made in the RF/RHD continuum over the past two decades since the re-awakening of interest, with a more concise focus on the last decade's achievements. Such primary advances include understanding the genetic predisposition to RHD, group A Streptococcus (GAS) vaccine development, and improved diagnostic strategies for GAS pharyngitis. Echocardiographic screening for RHD has been a major advance which has unearthed the prevailing high burden of RHD and the recent demonstration of benefit of secondary antibiotic prophylaxis on halting progression of latent RHD is a major step forward. Multiple befitting advances in tertiary management of RHD have also been realized. Finally, we summarize the research gaps and provide illumination on profitable future directions towards global eradication of RHD.

Smartphone-based imaging systems for medical applications: a critical review

SIGNIFICANCE

Smartphones come with an enormous array of functionality and are being more widely utilized with specialized attachments in a range of healthcare applications. A review of key developments and uses, with an assessment of strengths/limitations in various clinical workflows, was completed.

AIM

Our review studies how smartphone-based imaging (SBI) systems are designed and tested for specialized applications in medicine and healthcare. An evaluation of current research studies is used to provide guidelines for improving the impact of these research advances.

APPROACH

First, the established and emerging smartphone capabilities that can be leveraged for biomedical imaging are detailed. Then, methods and materials for fabrication of optical, mechanical, and electrical interface components are summarized. Recent systems were categorized into four groups based on their intended application and clinical workflow: ex vivo diagnostic, in vivo diagnostic, monitoring, and treatment guidance. Lastly, strengths and limitations of current SBI systems within these various applications are discussed.

RESULTS

The native smartphone capabilities for biomedical imaging applications include cameras, touchscreens, networking, computation, 3D sensing, audio, and motion, in addition to commercial wearable peripheral devices. Through user-centered design of custom hardware and software interfaces, these capabilities have the potential to enable portable, easy-to-use, point-of-care biomedical imaging systems. However, due to barriers in programming of custom software and on-board image analysis pipelines, many research prototypes fail to achieve a prospective clinical evaluation as intended. Effective clinical use cases appear to be those in which handheld, noninvasive image guidance is needed and accommodated by the clinical workflow. Handheld systems for in vivo, multispectral, and quantitative fluorescence imaging are a promising development for diagnostic and treatment guidance applications.

CONCLUSIONS

A holistic assessment of SBI systems must include interpretation of their value for intended clinical settings and how their implementations enable better workflow. A set of six guidelines are proposed to evaluate appropriateness of smartphone utilization in terms of clinical context, completeness, compactness, connectivity, cost, and claims. Ongoing work should prioritize realistic clinical assessments with quantitative and qualitative comparison to non-smartphone systems to clearly demonstrate the value of smartphone-based systems. Improved hardware design to accommodate the rapidly changing smartphone ecosystem, creation of open-source image acquisition and analysis pipelines, and adoption of robust calibration techniques to address phone-to-phone variability are three high priority areas to move SBI research forward.

Diagnostic Methods, Clinical Guidelines, and Antibiotic Treatment for Group A Streptococcal Pharyngitis: A Narrative Review

The most common bacterial cause of pharyngitis is infection by Group A β-hemolytic streptococcus (GABHS), commonly known as strep throat. 5-15% of adults and 15-35% of children in the United States with pharyngitis have a GABHS infection. The symptoms of GABHS overlap with non-GABHS and viral causes of acute pharyngitis, complicating the problem of diagnosis. A careful physical examination and patient history is the starting point for diagnosing GABHS. After a physical examination and patient history is completed, five types of diagnostic methods can be used to ascertain the presence of a GABHS infection: clinical scoring systems, rapid antigen detection tests, throat culture, nucleic acid amplification tests, and machine learning and artificial intelligence. Clinical guidelines developed by professional associations can help medical professionals choose among available techniques to diagnose strep throat. However, guidelines for diagnosing GABHS created by the American and European professional associations vary significantly, and there is substantial evidence that most physicians do not follow any published guidelines. Treatment for GABHS using analgesics, antipyretics, and antibiotics seeks to provide symptom relief, shorten the duration of illness, prevent nonsuppurative and suppurative complications, and decrease the risk of contagion, while minimizing the unnecessary use of antibiotics. There is broad agreement that antibiotics with narrow spectrums of activity are appropriate for treating strep throat. But whether and when patients should be treated with antibiotics for GABHS remains a controversial question. There is no clearly superior management strategy for strep throat, as significant controversy exists regarding the best methods to diagnose GABHS and under what conditions antibiotics should be prescribed.

Toward automated severe pharyngitis detection with smartphone camera using deep learning networks

PURPOSE

Severe pharyngitis is frequently associated with inflammations caused by streptococcal pharyngitis, which can cause immune-mediated and post-infectious complications. The recent global pandemic of coronavirus disease (COVID-19) encourages the use of telemedicine for patients with respiratory symptoms. This study, therefore, purposes automated detection of severe pharyngitis using a deep learning framework with self-taken throat images.

METHODS

A dataset composed of two classes of 131 throat images with pharyngitis and 208 normal throat images was collected. Before the training classifier, we constructed a cycle consistency generative adversarial network (CycleGAN) to augment the training dataset. The ResNet50, Inception-v3, and MobileNet-v2 architectures were trained with transfer learning and validated using a randomly selected test dataset. The performance of the models was evaluated based on the accuracy and area under the receiver operating characteristic curve (ROC-AUC).

RESULTS

The CycleGAN-based synthetic images reflected the pragmatic characteristic features of pharyngitis. Using the synthetic throat images, the deep learning model demonstrated a significant improvement in the accuracy of the pharyngitis diagnosis. ResNet50 with GAN-based augmentation showed the best ROC-AUC of 0.988 for pharyngitis detection in the test dataset. In the 4-fold cross-validation using the ResNet50, the highest detection accuracy and ROC-AUC achieved were 95.3% and 0.992, respectively.

CONCLUSION

The deep learning model for smartphone-based pharyngitis screening allows fast identification of severe pharyngitis with a potential of the timely diagnosis of pharyngitis. In the recent pandemic of COVID-19, this framework will help patients with upper respiratory symptoms to improve convenience in diagnosis and reduce transmission.

An IoT-based framework for early identification and monitoring of COVID-19 cases

The world has been facing the challenge of COVID-19 since the end of 2019. It is expected that the world will need to battle the COVID-19 pandemic with precautious measures, until an effective vaccine is developed. This paper proposes a real-time COVID-19 detection and monitoring system. The proposed system would employ an Internet of Things (IoTs) framework to collect real-time symptom data from users to early identify suspected coronaviruses cases, to monitor the treatment response of those who have already recovered from the virus, and to understand the nature of the virus by collecting and analyzing relevant data. The framework consists of five main components: Symptom Data Collection and Uploading (using wearable sensors), Quarantine/Isolation Center, Data Analysis Center (that uses machine learning algorithms), Health Physicians, and Cloud Infrastructure. To quickly identify potential coronaviruses cases from this real-time symptom data, this work proposes eight machine learning algorithms, namely Support Vector Machine (SVM), Neural Network, Naïve Bayes, K-Nearest Neighbor (K-NN), Decision Table, Decision Stump, OneR, and ZeroR. An experiment was conducted to test these eight algorithms on a real COVID-19 symptom dataset, after selecting the relevant symptoms. The results show that five of these eight algorithms achieved an accuracy of more than 90 %. Based on these results we believe that real-time symptom data would allow these five algorithms to provide effective and accurate identification of potential cases of COVID-19, and the framework would then document the treatment response for each patient who has contracted the virus.