Pediatric tuberculosis detection using trained African giant pouched rats

Machine Learning for Prediction of Tuberculosis Detection: Case Study of Trained African Giant Pouched Rats

BACKGROUND

Technological advancement has led to the growth and rapid increase of tuberculosis (TB) medical data generated from different health care areas, including diagnosis. Prioritizing better adoption and acceptance of innovative diagnostic technology to reduce the spread of TB significantly benefits developing countries. Trained TB-detection rats are used in Tanzania and Ethiopia for operational research to complement other TB diagnostic tools. This technology has increased new TB case detection owing to its speed, cost-effectiveness, and sensitivity.

OBJECTIVE

During the TB detection process, rats produce vast amounts of data, providing an opportunity to identify interesting patterns that influence TB detection performance. This study aimed to develop models that predict if the rat will hit (indicate the presence of TB within) the sample or not using machine learning (ML) techniques. The goal was to improve the diagnostic accuracy and performance of TB detection involving rats.

METHODS

APOPO (Anti-Persoonsmijnen Ontmijnende Product Ontwikkeling) Center in Morogoro provided data for this study from 2012 to 2019, and 366,441 observations were used to build predictive models using ML techniques, including decision tree, random forest, naïve Bayes, support vector machine, and k-nearest neighbor, by incorporating a variety of variables, such as the diagnostic results from partner health clinics using methods endorsed by the World Health Organization (WHO).

RESULTS

The support vector machine technique yielded the highest accuracy of 83.39% for prediction compared to other ML techniques used. Furthermore, this study found that the inclusion of variables related to whether the sample contained TB or not increased the performance accuracy of the predictive model.

CONCLUSIONS

The inclusion of variables related to the diagnostic results of TB samples may improve the detection performance of the trained rats. The study results may be of importance to TB-detection rat trainers and TB decision-makers as the results may prompt them to take action to maintain the usefulness of the technology and increase the TB detection performance of trained rats.

Increased tuberculosis case detection in Tanzanian children and adults using African giant pouched rats

BACKGROUND

African giant pouched rats, trained by Anti-Persoonsmijnen Ontmijnende Product Ontwikkeling (APOPO), have demonstrated their ability to detect tuberculosis (TB) from sputum. We assessed rat-based case detection and compared the mycobacterium bacillary load (MTB-load) in children versus adults.

METHODS

From January-December 2022, samples were collected prospectively from 69 Directly Observed Therapy (DOT) facilities' presumed TB patients. Using an average of five rats, APOPO re-evaluated patients with bacteriologically negative (sputum-smear microscopy or Xpert MTB/RIF) results. Rat-positive samples were tested using concentrated smear light-emitting diode microscopy to confirm TB detection before treatment initiation. The rats' identification of pulmonary TB is based on smelling TB-specific volatile organic compounds (VOCs) in sputum. Using STATA, Chi-square for odds ratio and confidence interval was calculated and evaluated: (1) the yield of rat-based TB detection compared to that of the health facilities; (2) rat-based TB detection in children versus adults; and (3) rats' ability to detect TB across MTB-loads and between children and adults.

RESULTS

From 35,766 patients, 5.3% (1900/35,766) were smear-positive and 94.7% (33,866/35,766) were smear or Xpert-negatives at DOTS facility. Of those with negative results, 2029 TB cases were detected using rats, contributing to 52% (2029/3929 of total TB identified), which otherwise would have been missed. Compared to DOT facilities, rats were six-fold more likely to detect TB among Acid Fast Bacilli (AFB) 1+/scanty [90% (1829/2029) versus 60% (1139/1900), odds ratio, OR = 6.11, 95% confidence interval, CI: 5.14-7.26]; twice more likely to identify TB cases among children [71% (91/129) versus 51% (1795/3542), OR = 2.3, 95% CI: 1.59-3.42]; and twice more likely to identify TB cases among children with AFB 1+/scanty than adults with the same MTB-load [5% (86/1703) versus 3% (28/1067), OR = 2.0, 95% CI: 1.28-3.03].

CONCLUSIONS

Rats contributed over half of the TB cases identified in program settings, and children, especially those with a lower MTB-load, were more likely to be diagnosed with TB by rats. The chemical signatures, VOCs, were only available for adults, and further research describing the characteristics of VOCs in children versus adults may pave the way to enhance TB diagnosis in children.

Scent detection of Brucella abortus by African giant pouched rats (Cricetomys ansorgei)

BACKGROUND

Brucellosis is a contagious zoonosis caused by bacteria of the genus Brucella. While the disease has been eradicated in most developed countries, it remains endemic in sub-Saharan Africa where access to reliable diagnostics is limited. African giant pouched rats (Cricetomys ansorgei) have been trained to detect the scent of Mycobacterium tuberculosis to increase case detection in sub-Saharan Africa. Given the similar diagnostic challenges facing brucellosis and tuberculosis, we explored the feasibility of training African giant pouched rats to detect Brucella.

RESULTS

After 3 months of training, rats reliably identified cultured Brucella, achieving an average sensitivity of 93.56% (SD = 0.650) and specificity of 97.65% (SD = 0.016). Rats readily generalized to novel, younger Brucella cultures that presumably generated a weaker volatile signal and correctly identified at least one out of three fecal samples spiked with Brucella culture during a final test of feasibility.

DISCUSSION

To our knowledge, these experiments are the first to demonstrate Brucella emits a unique odor profile that scent detection animals can be trained to identify. Importantly, cultured E. coli samples were included throughout training and test to ensure the rats learned to specifically identify Brucella bacteria rather than any bacteria in comparison to bacteria-free culture medium. E. coli controls therefore served a crucial function in determining to what extent Brucella abortus emits a unique odor signature. Further research is needed to determine if a Brucella-specific volatile signature is present within clinical samples. If confirmed, the present results suggest trained rats could serve as a valuable, novel method for the detection of Brucella infection.

Pediatric Tuberculosis Diagnostics: Present and Future

The current diagnostic abilities for the detection of pediatric tuberculosis are suboptimal. Multiple factors contribute to the under-diagnosis of intrathoracic tuberculosis in children, namely the absence of pathognomonic features of the disease, low bacillary loads in respiratory specimens, challenges in sample collection, and inadequate access to diagnostic tools in high-burden settings. Nonetheless, the 2020s have witnessed encouraging progress in the area of novel diagnostics. Recent WHO-endorsed rapid molecular assays hold promise for use in service decentralization strategies, and new policy recommendations include stools as an alternative, child-friendly specimen for testing with the GeneXpert assay. The pipeline of promising assays in mid/late-stage development is expanding, and novel pediatric candidate biomarkers based on the host immune response are being identified for use in diagnostic and triage tests. For a new test to meet the pediatric target product profiles prioritized by the WHO, it is key that the peculiarities and needs of the hard-to-reach pediatric population are considered in the early planning phases of discovery, validation, and implementation studies.

Visual Analytics of Tuberculosis Detection Rat Performance

The diagnosis of tuberculosis (TB) disease remains a global challenge, and the need for innovative diagnostic approaches is inevitable. Trained African giant pouched rats are the scent TB detection technology for operational research. The adoption of this technology is beneficial to countries with a high TB burden due to its cost-effectiveness and speed than microscopy. However, rats with some factors perform better. Thus, more insights on factors that may affect performance is important to increase rats' TB detection performance. This paper intends to provide understanding on the factors that influence rats TB detection performance using visual analytics approach. Visual analytics provide insight of data through the combination of computational predictive models and interactive visualizations. Three algorithms such as Decision tree, Random Forest and Naive Bayes were used to predict the factors that influence rats TB detection performance. Hence, our study found that age is the most significant factor, and rats of ages between 3.1 to 6 years portrayed potentiality. The algorithms were validated using the same test data to check their prediction accuracy. The accuracy check showed that the random forest outperforms with an accuracy of 78.82% than the two. However, their accuracies difference is small. The study findings may help rats TB trainers, researchers in rats TB and Information systems, and decision makers to improve detection performance. This study recommends further research that incorporates gender factors and a large sample size.

Breath can discriminate tuberculosis from other lower respiratory illness in children

Pediatric tuberculosis (TB) remains a global health crisis. Despite progress, pediatric patients remain difficult to diagnose, with approximately half of all childhood TB patients lacking bacterial confirmation. In this pilot study (n = 31), we identify a 4-compound breathprint and subsequent machine learning model that accurately classifies children with confirmed TB (n = 10) from children with another lower respiratory tract infection (LRTI) (n = 10) with a sensitivity of 80% and specificity of 100% observed across cross validation folds. Importantly, we demonstrate that the breathprint identified an additional nine of eleven patients who had unconfirmed clinical TB and whose symptoms improved while treated for TB. While more work is necessary to validate the utility of using patient breath to diagnose pediatric TB, it shows promise as a triage instrument or paired as part of an aggregate diagnostic scheme.

Rapid TB diagnostic service and community action to FIND.TREAT.ALL#EndTB, Maputo, Mozambique, 2013-2018

Finding and treating all tuberculosis (TB) patients is crucial for ending TB. We investigated whether rapid diagnostic turnaround time (TAT) and patient tracking could increase TB treatment initiation in Maputo, Mozambique. Among 3329 TB patients newly diagnosed by the University Eduardo Mondlane-Anti-Persoonsmijnen Ontmijnende Product Ontwikkeling/Anti-Personnel Landmines Detection Product Development (APOPO) Laboratory between 2013 and 2018, on average 61% were verifiably linked to care. This proportion increased from 54% (first half 2013) to 79% (second half 2018) after introducing a 24-hour TAT in 2015 and patient tracking conducted by a community-based partner, Associação Kenguelekezé, in 2017. Rapid, well-connected TB diagnostic services can reduce pre-treatment loss to follow-up and support the joint initiative of WHO, Stop TB and Global Fund to 'FIND.TREAT.ALL.#EndTB'.

The Many Hosts of Mycobacteria 8 (MHM8): A conference report

Mycobacteria are important causes of disease in human and animal hosts. Diseases caused by mycobacteria include leprosy, tuberculosis (TB), nontuberculous mycobacteria (NTM) infections and Buruli Ulcer. To better understand and treat mycobacterial disease, clinicians, veterinarians and scientists use a range of discipline-specific approaches to conduct basic and applied research, including conducting epidemiological surveys, patient studies, wildlife sampling, animal models, genetic studies and computational simulations. To foster the exchange of knowledge and collaboration across disciplines, the Many Hosts of Mycobacteria (MHM) conference series brings together clinical, veterinary and basic scientists who are dedicated to advancing mycobacterial disease research. Started in 2007, the MHM series recently held its 8th conference at the Albert Einstein College of Medicine (Bronx, NY). Here, we review the diseases discussed at MHM8 and summarize the presentations on research advances in leprosy, NTM and Buruli Ulcer, human and animal TB, mycobacterial disease comorbidities, mycobacterial genetics and 'omics, and animal models. A mouse models workshop, which was held immediately after MHM8, is also summarized. In addition to being a resource for those who were unable to attend MHM8, we anticipate this review will provide a benchmark to gauge the progress of future research concerning mycobacteria and their many hosts.

Sniffing animals as a diagnostic tool in infectious diseases

BACKGROUND

Scents and odours characterize some microbes when grown in the laboratory, and experienced clinicians can diagnose patients with some infectious diseases based on their smell. Animal sniffing is an innate behaviour, and animals' olfactory acuity is used for detecting people, weapons, bombs, narcotics and food.

OBJECTIVES

We briefly summarized current knowledge regarding the use of sniffing animals to diagnose some infectious diseases and the potential use of scent-based diagnostic instruments in microbiology.

SOURCES

Information was sought through PubMed and extracted from peer-reviewed literature published between January 2000 and September 2019 and from reliable online news. The search terms 'odour', 'scent', 'bacteria', 'diagnostics', 'tuberculosis', 'malaria' and 'volatile compounds' were used.

CONTENT

Four major areas of using sniffing animals are summarized. Dogs have been used to reliably detect stool associated with toxigenic Clostridioides difficile and for surveillance. Dogs showed high sensitivity and moderate specificity for detecting urinary tract infections in comparison to culture, especially for Escherichia coli. African giant pouched rats showed superiority for diagnosing tuberculosis over microscopy, but inferiority to culture/molecular methods. Several approaches for detecting malaria by analysing host skin odour or exhaled breath have been explored successfully. Some microbial infections produce specific volatile organic compounds (VOCs), which can be analysed by spectrometry, metabolomics or other analytical approaches to replace animal sniffing.

IMPLICATIONS

The results of sniffing animal studies are fascinating, and animal sniffing can provide intermediate diagnostic solutions for some infectious diseases. Lack of reproducibility, and cost of animal training and housing are major drawbacks for wider implementation of sniffing animals. The ultimate goal is to understand the biological background of this animal ability and to characterize the specific VOCs that animals are recognizing. VOC identification, improvement of odour sampling methods and development of point-of-care instruments could allow implementation of scent-based tests for major human pathogens.

[Medical diagnosis by breath analysis: odor sensors]

An odor sensor is a device for detecting target odors within a mixture, used in many fields including medical diagnosis. Electronic noses are networks of olfactory sensors, consisting of a surface whose properties are modified in the presence of odors, coupled with a measurement system. Their olfactory signature is analyzed in comparison with databases. Such portable devices can monitor body odors, e.g. in the breath, so as to reliably diagnose various pathologies at an early stage and non-invasively. It is tempting to use the naturally optimized molecular recognition of odorants and intrinsic sensitivity of the animal olfactory system to detect and discriminate minute amounts of odorants. New bioelectronic hybrid devices or "bioelectronic noses" can be designed by replacing the artificial sensory elements of e-noses by proteins naturally binding odorants, particularly olfactory receptors. As in the animal olfactory system, the detection and discrimination of odorants require a network of olfactory receptors. Prototypes of such miniaturized bioelectronic noses yield promising results.