Deep Neural Networks Offer Morphologic Classification and Diagnosis of Bacterial Vaginosis

Artificial intelligence in the clinical laboratory

Laboratory medicine has become a highly automated medical discipline. Nowadays, artificial intelligence (AI) applied to laboratory medicine is also gaining more and more attention, which can optimize the entire laboratory workflow and even revolutionize laboratory medicine in the future. However, only a few commercially available AI models are currently approved for use in clinical laboratories and have drawbacks such as high cost, lack of accuracy, and the need for manual review of model results. Furthermore, there are a limited number of literature reviews that comprehensively address the research status, challenges, and future opportunities of AI applications in laboratory medicine. Our article begins with a brief introduction to AI and some of its subsets, then reviews some AI models that are currently being used in clinical laboratories or that have been described in emerging studies, and explains the existing challenges associated with their application and possible solutions, finally provides insights into the future opportunities of the field. We highlight the current status of implementation and potential applications of AI models in different stages of the clinical testing process.

A prospective observational multicentric clinical trial to evaluate microscopic examination of acid-fast bacilli in sputum by artificial intelligence-based microscopy system

Microscopy-based tuberculosis (TB) diagnosis i.e., Ziehl-Neelsen (ZN) stained smear screening still remains the primary diagnostic method in resource poor and high TB burden countries, however itrequires considerable experience and is bound to human errors. In remote areas, wherever expert microscopist is not available, timely diagnosis at initial level is not possible. Artificial intelligence (AI)-based microscopy may be a solution to this problem. A prospective observational multi-centric clinical trial to evaluate microscopic examination of acid-fast bacilli (AFB) in sputum by the AI based system was done in three hospitals in Northern India. Sputum samples from 400 clinically suspected cases of pulmonary tuberculosis were collected from three centres. Ziehl-Neelsen staining of smears was done. All the smears were observed by 3 microscopist and the AI based microscopy system. AI based microscopy was found to have a sensitivity, specificity, positive predictive value, negative predictive value and diagnostic accuracy of 89.25%, 92.15%, 75.45%, 96.94%, 91.53% respectively. AI based sputum microscopy has an acceptable degree of accuracy, PPV, NPV, specificity and sensitivity and thus may be used as a screening tool for the diagnosis of pulmonary tuberculosis.

The Use of Machine Learning for Image Analysis Artificial Intelligence in Clinical Microbiology

The growing transition to digital microbiology in clinical laboratories creates the opportunity to interpret images using software. Software analysis tools can be designed to use human-curated knowledge and expert rules, but more novel artificial intelligence (AI) approaches such as machine learning (ML) are being integrated into clinical microbiology practice. These image analysis AI (IAAI) tools are beginning to penetrate routine clinical microbiology practice, and their scope and impact on routine clinical microbiology practice will continue to grow. This review separates the IAAI applications into 2 broad classification categories: (i) rare event detection/classification or (ii) score-based/categorical classification. Rare event detection can be used for screening purposes or for final identification of a microbe including microscopic detection of mycobacteria in a primary specimen, detection of bacterial colonies growing on nutrient agar, or detection of parasites in a stool preparation or blood smear. Score-based image analysis can be applied to a scoring system that classifies images in toto as its output interpretation and examples include application of the Nugent score for diagnosing bacterial vaginosis and interpretation of urine cultures. The benefits, challenges, development, and implementation strategies of IAAI tools are explored. In conclusion, IAAI is beginning to impact the routine practice of clinical microbiology, and its use can enhance the efficiency and quality of clinical microbiology practice. Although the future of IAAI is promising, currently IAAI only augments human effort and is not a replacement for human expertise.

State of the Art for Diagnosis of Bacterial Vaginosis

Bacterial vaginosis (BV) is the most common cause of vaginal discharge among reproductive-age women. It is associated with multiple adverse health outcomes, including increased risk of acquisition of HIV and other sexually transmitted infections (STIs), in addition to adverse birth outcomes. While it is known that BV is a vaginal dysbiosis characterized by a shift in the vaginal microbiota from protective Lactobacillus species to an increase in facultative and strict anaerobic bacteria, its exact etiology remains unknown. The purpose of this minireview is to provide an updated overview of the range of tests currently used for the diagnosis of BV in both clinical and research settings. This article is divided into two primary sections: traditional BV diagnostics and molecular diagnostics. Molecular diagnostic assays, particularly 16S rRNA gene sequencing, shotgun metagenomic sequencing, and fluorescence in situ hybridization (FISH), are specifically highlighted, in addition to multiplex nucleic acid amplification tests (NAATs), given their increasing use in clinical practice (NAATs) and research studies (16S rRNA gene sequencing, shotgun metagenomic sequencing, and FISH) regarding the vaginal microbiota and BV pathogenesis. We also provide a discussion of the strengths and weaknesses of current BV diagnostic tests and discuss future challenges in this field of research.

Diagnostic performance of an automated microscopy and pH test for diagnosis of vaginitis

Vaginitis is a common gynecological problem, nevertheless, its clinical evaluation is often insufficient. This study evaluated the performance of an automated microscope for the diagnosis of vaginitis, by comparison of the investigated test results to a composite reference standard (CRS) of wet mount microscopy performed by a specialist in vulvovaginal disorders, and related laboratory tests. During this single-site cross-sectional prospective study, 226 women reporting vaginitis symptoms were recruited, of which 192 samples were found interpretable and were assessed by the automated microscopy system. Results showed sensitivity between 84.1% (95%CI: 73.67-90.86%) for Candida albicans and 90.9% (95%CI: 76.43-96.86%) for bacterial vaginosis and specificity between 65.9% (95%CI: 57.11-73.64%) for Candida albicans and 99.4% (95%CI: 96.89-99.90%) for cytolytic vaginosis. These findings demonstrate the marked potential of machine learning-based automated microscopy and an automated pH test of vaginal swabs as a basis for a computer-aided suggested diagnosis, for improving the first-line evaluation of five different types of infectious and non-infectious vaginal disorders (vaginal atrophy, bacterial vaginosis, Candida albicans vaginitis, cytolytic vaginosis, and aerobic vaginitis/desquamative inflammatory vaginitis). Using such a tool will hopefully lead to better treatment, decrease healthcare costs, and improve patients' quality of life.

Morphologic identification of clinically encountered moulds using a residual neural network

The use of morphology to diagnose invasive mould infections in China still faces substantial challenges, which often leads to delayed diagnosis or misdiagnosis. We developed a model called XMVision Fungus AI to identify mould infections by training, testing, and evaluating a ResNet-50 model. Our research achieved the rapid identification of nine common clinical moulds: Aspergillus fumigatus complex, Aspergillus flavus complex, Aspergillus niger complex, Aspergillus terreus complex, Aspergillus nidulans, Aspergillus sydowii/Aspergillus versicolor, Syncephalastrum racemosum, Fusarium spp., and Penicillium spp. In our study, the adaptive image contrast enhancement enabling XMVision Fungus AI as a promising module by effectively improve the identification performance. The overall identification accuracy of XMVision Fungus AI was up to 93.00% (279/300), which was higher than that of human readers. XMVision Fungus AI shows intrinsic advantages in the identification of clinical moulds and can be applied to improve human identification efficiency through training. Moreover, it has great potential for clinical application because of its convenient operation and lower cost. This system will be suitable for primary hospitals in China and developing countries.

End-to-End Deep Learning-Based Cells Detection in Microscopic Leucorrhea Images

Vaginitis is a prevalent gynecologic disease that threatens millions of women’s health. Although microscopic examination of vaginal discharge is an effective method to identify vaginal infections, manual analysis of microscopic leucorrhea images is extremely time-consuming and labor-intensive. To automate the detection and identification of visible components in microscopic leucorrhea images for early-stage diagnosis of vaginitis, we propose a novel end-to-end deep learning-based cells detection framework using attention-based detection with transformers (DETR) architecture. The transfer learning was applied to speed up the network convergence while maintaining the lowest annotation cost. To address the issue of detection performance degradation caused by class imbalance, the weighted sampler with on-the-fly data augmentation module was integrated into the detection pipeline. Additionally, the multi-head attention mechanism and the bipartite matching loss system of the DETR model perform well in identifying partially overlapping cells in real-time. With our proposed method, the pipeline achieved a mean average precision (mAP) of 86.00% and the average precision (AP) of epithelium, leukocyte, pyocyte, mildew, and erythrocyte was 96.76, 83.50, 74.20, 89.66, and 88.80%, respectively. The average test time for a microscopic leucorrhea image is approximately 72.3 ms. Currently, this cell detection method represents state-of-the-art performance.

A Data-Efficient Framework for the Identification of Vaginitis Based on Deep Learning

Vaginitis is a gynecological disease affecting the health of millions of women all over the world. The traditional diagnosis of vaginitis is based on manual microscopy, which is time-consuming and tedious. The deep learning method offers a fast and reliable solution for an automatic early diagnosis of vaginitis. However, deep neural networks require massive well-annotated data. Manual annotation of microscopic images is highly cost extensive because it not only is a time-consuming process but also needs highly trained people (doctors, pathologists, or technicians). Most existing active learning approaches are not applicable in microscopic images due to the nature of complex backgrounds and numerous formed elements. To address the problem of high cost of labeling microscopic images, we present a data-efficient framework for the identification of vaginitis based on transfer learning and active learning strategies. The proposed informative sample selection strategy selected the minimal training subset, and then the pretrained convolutional neural network (CNN) was fine-tuned on the selected subset. The experiment results show that the proposed pipeline can save 37.5% annotation cost while maintaining competitive performance. The proposed promising novel framework can significantly save the annotation cost and has the potential of extending widely to other microscopic imaging applications, such as blood microscopic image analysis.

A Vaginitis Classification Method Based on Multi-Spectral Image Feature Fusion

Vaginitis is one of the commonly encountered diseases of female reproductive tract infections. The clinical diagnosis mainly relies on manual observation under a microscope. There has been some investigation on the classification of vaginitis diseases based on computer-aided diagnosis to reduce the workload of clinical laboratory staff. However, the studies only using RGB images limit the development of vaginitis diagnosis. Through multi-spectral technology, we propose a vaginitis classification algorithm based on multi-spectral image feature layer fusion. Compared with the traditional RGB image, our approach improves the classification accuracy by 11.39%, precision by 15.82%, and recall by 27.25%. Meanwhile, we prove that the level of influence of each spectrum on the disease is distinctive, and the subdivided spectral image is more conducive to the image analysis of vaginitis disease.

Deep Convolutional Neural Networks Implementation for the Analysis of Urine Culture

BACKGROUND

Urine culture images collected using bacteriology automation are currently interpreted by technologists during routine standard-of-care workflows. Machine learning may be able to improve the harmonization of and assist with these interpretations.

METHODS

A deep learning model, BacterioSight, was developed, trained, and tested on standard BD-Kiestra images of routine blood agar urine cultures from 2 different medical centers.

RESULTS

BacterioSight displayed performance on par with standard-of-care-trained technologist interpretations. BacterioSight accuracy ranged from 97% when compared to standard-of-care (single technologist) and reached 100% when compared to a consensus reached by a group of technologists (gold standard in this study). Variability in image interpretation by trained technologists was identified and annotation "fuzziness" was quantified and found to correlate with reduced confidence in BacterioSight interpretation. Intra-testing (training and testing performed within the same institution) performed well giving Area Under the Curve (AUC) ≥0.98 for negative and positive plates, whereas, cross-testing on images (trained on one institution's images and tested on images from another institution) showed decreased performance with AUC ≥0.90 for negative and positive plates.

CONCLUSIONS

Our study provides a roadmap on how BacterioSight or similar deep learning prototypes may be implemented to screen for microbial growth, flag difficult cases for multi-personnel review, or auto-verify a subset of cultures with high confidence. In addition, our results highlight image interpretation variability by trained technologist within an institution and globally across institutions. We propose a model in which deep learning can enhance patient care by identifying inherent sample annotation variability and improving personnel training.

Bacterial Vaginosis: What Do We Currently Know?

The vaginal microbiome is a well-defined compartment of the human microbiome. It has unique conditions, characterized by the dominance of one bacterial species, the Lactobacilli. This microbiota manifests itself by a low degree of diversity and by a strong dynamic of change in its composition under the influence of various exogenous and endogenous factors. The increase in diversity may paradoxically be associated with dysbiosis, such as bacterial vaginosis (BV). BV is the result of a disturbance in the vaginal ecosystem; i.e., a sudden replacement of Lactobacilli by anaerobic bacteria such as Gardnerella vaginalis, Atopobium vaginae, Ureaplasma urealyticum, Mycoplasma hominis, and others. It is the most common cause of vaginal discharge in women of childbearing age, approximately 30% of all causes. The etiology of this dysbiosis remains unknown, but its health consequences are significant, including obstetrical complications, increased risk of sexually transmitted infections and urogenital infections. Its diagnosis is based on Amsel’s clinical criteria and/or a gram stain based on the Nugent score. While both of these methods have been widely applied worldwide for approximately three decades, Nugent score are still considered the “gold standard” of BV diagnostic tools. Given the limitations of these tools, methods based on molecular biology have been developed as alternative rational strategies for the diagnosis of BV. The treatment of BV aims at restoring the balance of the vaginal flora to stop the proliferation of harmful microorganisms. Prescription of antibiotics such as metronidazole, clindamycin, etc. is recommended. Faced with the considerable uncertainty about the cause of BV, the high rate of recurrence, the unacceptable treatment options, and clinical management which is often insensitive and inconsistent, research on this topic is intensifying. Knowledge of its composition and its associated variations represents the key element in improving the therapeutic management of patients with the most suitable treatments possible.

Aerobic Vaginitis Diagnosis Criteria Combining Gram Stain with Clinical Features: An Establishment and Prospective Validation Study

Wet-mount microscopy aerobic vaginitis (AV) diagnostic criteria need phase-contrast microscopy and keen microscopists, and the preservation of saline smears is less common in clinical practice. This research work developed new AV diagnostic criteria that combine Gram stain with clinical features. We enrolled 325 AV patients and 325 controls as a study population to develop new AV diagnostic criteria. Then, an independent group, which included 500 women, was used as a validation population. AV-related microscopic findings on Gram-stained and wet-mount smears from the same participants were compared. The accuracy of bacterial indicators from the two methods was verified by bacterial 16S rRNA V4 sequencing (n = 240). Logistic regression was used to analyse AV-related clinical features. The screened clinical features were combined with Gram-stain microscopic indicators to establish new AV diagnostic criteria. There were no significant differences in the leukocyte counts or the parabasal epitheliocytes (PBC) proportion between the Gram-stain and wet-mount methods (400×). Gram stain (1000×) satisfied the ability to identify bacteria as verified by 16S rRNA sequencing but failed to identify toxic leukocytes. The new criteria included: Lactobacillary grades (LBG) and background flora (Gram stain, 1000×), leukocytes count and PBC proportion (Gram stain, 400×), and clinical features (vaginal pH > 4.5, vagina hyperemia, and yellow discharge). These criteria satisfied the accuracy and reliability for AV diagnosis (Se = 86.79%, Sp = 95.97%, and Kendall’s W value = 0.899) in perspective validation. In summary, we proposed an alternative and valuable AV diagnostic criteria based on the Gram stain, which can make it possible to diagnose common vaginitis like AV, BV, VVC, and mixed infections on the same smear and can be available for artificial intelligence diagnosis in the future.

Contribution of Lactobacillus iners to Vaginal Health and Diseases: A Systematic Review

Lactobacillus iners, first described in 1999, is a prevalent bacterial species of the vaginal microbiome. As L. iners does not easily grow on de Man-Rogosa-Sharpe agar, but can grow anaerobically on blood agar, it has been initially overlooked by traditional culture methods. It was not until the wide application of molecular biology techniques that the function of L. iners in the vaginal microbiome was carefully explored. L. iners has the smallest genome among known Lactobacilli and it has many probiotic characteristics, but is partly different from other major vaginal Lactobacillus species, such as L. crispatus, in contributing to the maintenance of a healthy vaginal microbiome. It is not only commonly present in the healthy vagina but quite often recovered in high numbers in bacterial vaginosis (BV). Increasing evidence suggests that L. iners is a transitional species that colonizes after the vaginal environment is disturbed and offers overall less protection against vaginal dysbiosis and, subsequently, leads to BV, sexually transmitted infections, and adverse pregnancy outcomes. Accordingly, under certain conditions, L. iners is a genuine vaginal symbiont, but it also seems to be an opportunistic pathogen. Further studies are necessary to identify the exact role of this intriguing species in vaginal health and diseases.

Automated interpretation and analysis of bronchoalveolar lavage fluid

BACKGROUND

The cytological analysis of bronchoalveolar lavage fluid (BALF) plays an essential role in the differential diagnosis of respiratory diseases. In recent years, deep learning has demonstrated excellent performance in image processing and object recognition.

OBJECTIVES

We aim to apply deep learning to the automated interpretation and analysis of BALF.

METHOD

Visual images were acquired using an automated biological microscopy platform. We propose a three-step algorithm to automatically interpret BALF cytology based on a convolutional neural network (CNN). The clinical value was evaluated at the patient level.

RESULTS

Our model successfully detected most cells in BALF specimens and achieved a sensitivity, precision, and F1 score of over 0.9 for most cell types. In two tests in the clinical context, the algorithm outperformed experienced practitioners.

CONCLUSION

The program can automatically provide the cytological background of BALF and augment clinical decision-making for clinicians.

Artificial Intelligence and Mapping a New Direction in Laboratory Medicine: A Review

BACKGROUND

Modern artificial intelligence (AI) and machine learning (ML) methods are now capable of completing tasks with performance characteristics that are comparable to those of expert human operators. As a result, many areas throughout healthcare are incorporating these technologies, including in vitro diagnostics and, more broadly, laboratory medicine. However, there are limited literature reviews of the landscape, likely future, and challenges of the application of AI/ML in laboratory medicine.

CONTENT

In this review, we begin with a brief introduction to AI and its subfield of ML. The ensuing sections describe ML systems that are currently in clinical laboratory practice or are being proposed for such use in recent literature, ML systems that use laboratory data outside the clinical laboratory, challenges to the adoption of ML, and future opportunities for ML in laboratory medicine.

SUMMARY

AI and ML have and will continue to influence the practice and scope of laboratory medicine dramatically. This has been made possible by advancements in modern computing and the widespread digitization of health information. These technologies are being rapidly developed and described, but in comparison, their implementation thus far has been modest. To spur the implementation of reliable and sophisticated ML-based technologies, we need to establish best practices further and improve our information system and communication infrastructure. The participation of the clinical laboratory community is essential to ensure that laboratory data are sufficiently available and incorporated conscientiously into robust, safe, and clinically effective ML-supported clinical diagnostics.