Machine learning in the loop for tuberculosis diagnosis support.

The use of machine learning (ML) for diagnosis support has advanced in the field of health. In the present paper, the results of studying ML techniques in a tuberculosis diagnosis loop in a scenario of limited resources are presented. Data are analyzed using a tuberculosis (TB) therapy program at a health institution in a main city of a developing country using five ML models. Logistic regression, classification trees, random forest, support vector machines, and artificial neural networks are trained under physician supervision following physicians' typical daily work. The models are trained on seven main variables collected when patients arrive at the facility. Additionally, the variables applied to train the models are analyzed, and the models' advantages and limitations are discussed in the context of the automated ML techniques. The results show that artificial neural networks obtain the best results in terms of accuracy, sensitivity, and area under the receiver operating curve. These results represent an improvement over smear microscopy, which is commonly used techniques to detect TB for special cases. Findings demonstrate that ML in the TB diagnosis loop can be reinforced with available data to serve as an alternative diagnosis tool based on data processing in places where the health infrastructure is limited.

Preliminary Text Analysis from Medical Records for TB Diagnosis Support.

Tuberculosis is an infectious disease that is spread through the air from one person to another and is one of the top ten causes of death in the world according to the World Health Organization. From biomedical engineering, decision support systems based on artificial intelligence have shown advantages for healthcare personnel in tasks such as diagnosis and screening. A specific area of the artificial intelligence is the natural language processing, however, most of these approaches are based on available data. This paper shows the construction of a dataset based on medical records of subjects suspected of tuberculosis. In addition, an initial exploration of the contents of the constructed dataset and how this approach can be followed by a natural language processing to support tuberculosis diagnosis in data demanding scenarios are presented.Clinical Relevance- In some developing countries as Colombia, it is difficult to develop systems based on artificial intelligence due to the availability of data. This proposal holds a strategy to build a dataset to train machine learning models, and to obtain support diagnosis tools, employing natural language from the medical scenario from text written by health professionals in the medical record. In this way, trained models based on this information available can be employed in places where medical infrastructure is precarious.

Diagnosis and treatment of latent tuberculosis in patients with multiple sclerosis, expert consensus. On behalf of the Colombian Association of Neurology, Committee of Multiple Sclerosis.

BACKGROUND: Multiple sclerosis is an inflammatory and neurodegenerative demyelinating disease. Current treatment of multiple sclerosis focuses on the use of immunomodulatory, immunosuppressant, and selective immunosuppressant agents. Some of these medications may result in high risk of opportunistic infections including tuberculosis. OBJECTIVE: The purpose of this study was to obtain consensus from a panel of neurologists, pulmonologists, infectious disease specialists, and epidemiology experts regarding the diagnosis, treatment, and monitoring of latent tuberculosis in patients with multiple sclerosis. METHODS: A panel of experts in multiple sclerosis and tuberculosis was established. The methodological process was performed in three phases: definition of questions, answer using Delphi methodology, and the discussion of questions not agreed. RESULTS: Tuberculosis screening is suggested when multiple sclerosis drugs are prescribed. The recommended tests for latent tuberculosis are tuberculin and interferon gamma release test. When an anti-tuberculosis treatment is indicated, monitoring should be performed to determine liver enzyme values with consideration of age as well as comorbid conditions such as a history of alcoholism, age, obesity, concomitant hepatotoxic drugs, and history of liver disease. CONCLUSION: Latent tuberculosis should be considered in patients with multiple sclerosis who are going to be treated with immunomodulatory and immunosuppressant medications. Transaminase level monitoring is required on a periodic basis depending on clinical and laboratory characteristics. In addition to the liver impairment, other side effects should be considered when Isoniazid is prescribed.

[Prevalence and risk factors associated to tuberculosis and non-tuberculous mycobacterial infections in HIV-positive patients in Bogotá].

INTRODUCTION: Tuberculosis is one of the most widely distributed infectious diseases worldwide. It is the most common cause of mortality among AIDS patients. In Colombia, 12,918 tuberculosis cases were notified, and 926 deaths were reported in 2015. OBJECTIVE: To determine the prevalence and risk factors associated to mycobacterial infections in HIVpositive patients in two public hospitals from Bogotá. MATERIALS AND METHODS: A prospective and descriptive study was carried out by an active search for tuberculosis cases and non-tuberculous mycobacterial infections in HIV-positive patients. We considered demographic, social, clinical, and personal habits as variables. Statistical analyses were done using Stata 13™ software. RESULTS: Three hundred and fifty six patients were included, 81.2% were men and 18.8% were women; the mean age was 36.5 years. Tuberculosis infection had a frequency of 19.9% (95% CI: 15.9-24.5%) and non-tuberculous mycobacterial infection had a 3.9% frequency (95% CI: 2.16-6.5%). Bivariate analysis showed a statistically significant association between tuberculosis infection and CD4+ T cell counts (p=0.003), viral load (p=0.008), antiretroviral therapy (p=0.014), and body mass index (BMI) <18 kg/m2 (p=0.000). In non-tuberculous mycobacterial infections there was a statistically significantassociation with BMI (p=0.027) and CD4+ T cell counts (p=0.045). CONCLUSION: Factors associated with an impaired immune system caused by HIV infection are an important risk factor for developing tuberculosis. The lack of antiretroviral therapy and the BMI were also important risk factors for tuberculosis.

La desnutrición, factor de riesgo para complicaciones en pacientes con tuberculosis en Bogotá / Malnutrition, a risk factor for complications in patients with tuberculosis in Bogotá

En la Revista Colombiana de Neumología, número dos, volumen 34 del año 2022, está publicado el artículo original de los doctores Daniel Adolfo Suárez, Andrea Carolina Córdoba y Oscar Alberto Sáenz titulado "Factores de riesgo para complicaciones en pacientes con tuberculosis en una institución de tercer nivel de la ciudad de Bogotá". Es un estudio observacional retrospectivo de la cohorte de 130 pacientes con tuberculosis pulmonar diagnosticados entre los años 2017 y 2018, cuyo objetivo fue el de identificar posibles factores de riesgo asociados al desarrollo de complicaciones por la tuberculosis. Los autores presentan la revisión de la literatura con relación a las complicaciones de la tuberculosis. En general, las complicaciones de la tuberculosis se han dividido según el sitio anatómico comprometido, esto es, pulmonar o respiratorio, neurológico, cardiovascular, gastrointestinal u otros y han sido atribuidas al efecto patológico de la micobacteria o al efecto inflamatorio desencadenado por la respuesta inmune del huésped.

In the Colombian Journal of Pneumology, number two, volume 34 of the year 2022, the original article by doctors Daniel Adolfo Suárez, Andrea Carolina Córdoba and Oscar Alberto Sáenz entitled "Risk factors for complications in patients with tuberculosis in a hospital third level of the city of Bogotá". It is a retrospective observational study of the cohort of 130 patients with pulmonary tuberculosis diagnosed between 2017 and 2018, whose objective was to identify possible risk factors associated with the development of complications from tuberculosis. The authors present a review of the literature regarding the complications of tuberculosis. In general, the complications of tuberculosis have been divided according to the anatomical site involved, that is, pulmonary or respiratory, neurological, cardiovascular, gastrointestinal or others, and have been attributed to the pathological effect of the mycobacteria or to the inflammatory effect triggered by the immune response. of the host.

High levels of anti-tuberculin (IgG) antibodies correlate with the blocking of T-cell proliferation in individuals with high exposure to Mycobacterium tuberculosis.

OBJECTIVES: To determine the effect of anti-tuberculin antibodies in the T-cell proliferation in response to tuberculin and Candida antigens in individuals with different levels of tuberculosis (TB) risk. METHODS: Sixteen high-risk TB individuals, 30 with an intermediate TB risk (group A), and 45 with a low TB risk (group B), as well as 49 control individuals, were studied. Tuberculin skin test (TST) results were analyzed and serum levels of antibodies (IgG and IgM) against purified protein derivative (PPD) were measured by ELISA. Tuberculin and Candida antigens were used to stimulate T-cell proliferation in the presence of human AB serum or autologous serum. RESULTS: High levels of anti-tuberculin IgG antibodies were found to be significantly associated with the blocking of T-cell proliferation responses in cultures stimulated with tuberculin but not with Candida antigens in the presence of autologous serum. This phenomenon was particularly frequent in high-risk individuals with high levels of anti-tuberculin IgG antibodies in the autologous serum when compared to the other risk groups, which exhibited lower levels of anti-tuberculin antibodies. CONCLUSIONS: Although cellular immunity plays a central role in the protection against TB, humoral immunity is critical in the control of Mycobacterium tuberculosis infection in high-risk individuals with latent TB infection.

Prevalencia y factores asociados a la tuberculosis y las micobacteriosis en pacientes positivos para HIV en Bogotá / Prevalence and risk factors associated to tuberculosis and non-tuberculous mycobacterial infections in HIV-positive patients in Bogotá

Resumen Introducción. La tuberculosis es una de las enfermedades infecciosas de más amplia distribución en el mundo y constituye una de las primeras causas de muerte en pacientes con sida. En Colombia, en el 2015, se notificaron 12.918 casos de tuberculosis y 926 muertes. Objetivo. Determinar la prevalencia y los factores asociados a infecciones micobacterianas en pacientes infectados con el virus de inmunodeficiencia humana (HIV) en dos hospitales públicos de Bogotá. Materiales y métodos. Se hizo un estudio descriptivo de corte transversal con búsqueda activa de casos de tuberculosis y micobacteriosis en pacientes positivos para HIV. Se estudiaron variables demográficas, sociales, clínicas y de hábitos personales. Los análisis estadísticos se hicieron con el programa Stata 13TM. Resultados. Se incluyeron en el estudio 356 pacientes: 81,2 % hombres y 18,8 %, mujeres, con una media de edad de 36,5 años. La frecuencia de la tuberculosis fue de 19,9 % (IC95% 15,9-24,5 %) y la de infecciones por micobacterias no tuberculosas, de 3,9 % (IC95% 2,16-6,5 %). El análisis bivariado evidenció una asociación estadísticamente significativa entre la tuberculosis y el conteo de linfocitos TCD4+ (p=0,003), la carga viral (p=0,0008), el tratamiento antirretroviral (p=0,017) y un índice de masa corporal (IMC) menor de 18 kg/m2 (p=0,000). En las micobacteriosis solamente se presentó asociación estadísticamente significativa con el IMC (p=0,017) y con el conteo de linfocitos TCD4+ (p=0,045). Conclusión. Los factores asociados al deterioro del sistema inmunitario causados por el HIV, así como el no administrar el tratamiento antirretroviral de gran actividad y el IMC, constituyeron factores de riesgo para desarrollar la tuberculosis.

Abstract Introduction. Tuberculosis is one of the most widely distributed infectious diseases worldwide. It is the most common cause of mortality among AIDS patients. In Colombia, 12,918 tuberculosis cases were notified, and 926 deaths were reported in 2015. Objective. To determine the prevalence and risk factors associated to mycobacterial infections in HIV-positive patients in two public hospitals from Bogotá. Materials and methods. A prospective and descriptive study was carried out by an active search for tuberculosis cases and non-tuberculous mycobacterial infections in HIV-positive patients. We considered demographic, social, clinical, and personal habits as variables. Statistical analyses were done using Stata 13TM software. Results. Three hundred and fifty six patients were included, 81.2% were men and 18.8% were women; the mean age was 36.5 years. Tuberculosis infection had a frequency of 19.9% (95% CI: 15.9-24.5%) and non-tuberculous mycobacterial infection had a 3.9% frequency (95% CI: 2.16-6.5%). Bivariate analysis showed a statistically significant association between tuberculosis infection and CD4+ T cell counts (p=0.003), viral load (p=0.008), antiretroviral therapy (p=0.014), and body mass index (BMI) <18 kg/m2 (p=0.000). In non-tuberculous mycobacterial infections there was a statistically significant association with BMI (p=0.027) and CD4+ T cell counts (p=0.045). Conclusion. Factors associated with an impaired immune system caused by HIV infection are an important risk factor for developing tuberculosis. The lack of antiretroviral therapy and the BMI were also important risk factors for tuberculosis.

Humoral immunity in tuberculin skin test anergy and its role in high-risk persons exposed to active tuberculosis.

The most common test to identify latent tuberculosis is the tuberculin skin test that detects T cell responses of delayed type hypersensitivity type IV. Since it produces false negative reactions in active tuberculosis or in high-risk persons exposed to tuberculosis patients as shown in this report, we studied antibody profiles to explain the anergy of such responses in high-risk individuals without active infection. Our results showed that humoral immunity against tuberculin, regardless of the result of the tuberculin skin test is important for protection from active tuberculosis and that the presence of high antibody titers is a more reliable indicator of infection latency suggesting that latency can be based on the levels of antibodies together with in vitro proliferation of peripheral blood mononuclear cells in the presence of the purified protein derivative. Importantly, anti-tuberculin IgG antibody levels mediate the anergy described herein, which could also prevent reactivation of disease in high-risk individuals with high antibody titers. Such anti-tuberculin IgG antibodies were also found associated with blocking and/or stimulation of in vitro cultures of PBMC with tuberculin. In this regard, future studies need to establish if immune responses to Mycobacterium tuberculosis can generate a broad spectrum of reactions either toward Th1 responses favoring stimulation by cytokines or by antibodies and those toward diminished responses by Th2 cytokines or blocking by antibodies; possibly involving mechanisms of antibody dependent protection from Mtb by different subclasses of IgG.

Molecular signatures distinguishing active from latent tuberculosis in peripheral blood mononuclear cells, after in vitro antigenic stimulation with purified protein derivative of tuberculin (PPD) or Candida: a preliminary report.

Purified protein derivative (PPD) or tuberculin skin testing is used to identify infected individuals with Mycobacterium tuberculosis (Mtb) and to assess cell-mediated immunity to Mtb. In the present study, we compared PBMC cultures in the presence of tuberculin or Candida antigens using cytokine bead arrays and RNA microarrays. Measurements of different cytokines and chemokines in supernatants of PMBC cultures in the presence of PPD showed increased levels of interferon (IFN)-gamma in active tuberculosis infection (ATBI) and latent TB infected (LTBI) compared to controls, and increased levels of TNF-alpha in ATBI compared with LTBI. Also, we found increase of IL-6 in cultures of PPD positive and controls but not in the cultures with Candida. We also report the molecular signature of tuberculosis infection, in ATBI patients, the following genes were found to be up-regulated and absent in LTBI individuals: two kinases (JAK3 and p38MAPK), four interleukins (IL-7, IL-2, IL-6, and IFNbeta1), a chemokine (HCC-4) a chemokine receptor (CxCR5), two interleukin receptors (IL-1R2 and IL-18R1), and three additional ones (TRAF5, Smad2, CIITA, and NOS2A). By contrast, IL-17 and IGFBP3 were significantly up-regulated in LTBI. And, STAT4, GATA3, Fra-1, and ICOS were down-regulated in ATBI but absent in LTBI. Conversely, TLR-10, IL-15, DORA, and IKK-beta were down-regulated in LTBI but not in ATBI. Interestingly, the majority of the up-regulated genes found in ATBI were found in cultures stimulated with tuberculin (PPD) or Candida antigens, suggesting that these pathogens stimulate similar immunological pathways. We believe that the molecular signature distinguishing active from latent tuberculosis infection may require using cytokine bead arrays along with RNA microarrays testing cell cultures at different times following in vitro proliferation assays using several bacterial antigens and PPD.

Actividad de renina plasmática en enfermedad pulmonar obstructiva crónica

Introducción: En estados avanzados de la EPOC pueden demostrarse alteraciones neurohormonales significativas, que se correlacionan con la aparición de hipercapnia y retención de sodio y agua. No se conoce si hay activación del eje renina-angiotensina-aldosterona en etapas tempranas de la enfermedad, como se ha detectado que ocurre en la enfermedad ventricular izquierda. Objetivo: evaluar la actividad de renina plasmática, un índice de la función del eje renina-angiotensión-aldosterona, en pacientes con EPOC que no han presenta edemas ni descompensación clínica. Pacientes y Métodos: Diez y seis pacientes fueron estudiados con pruebas de función pulmonar (espirometría, curvo flujo-volumen), gasimetría arterial, radiografía del tórax y ecocardiograma transtorácico. En sangre venosa y en condiciones de reposo se determinó actividad de renina plasmática por radioinmunoensayo. Resultados: Se estudiaron 16 pacientes, 14 hombres y 2 mujeres, edad promedio de 61 años con EPOC moderado a severo (promedio del VEF 1 predicho 50 por ciento, relación VEF1/CVF 51 por ciento), PaO2 promedio 52 mm Hg y PaCO2 promedio 34,3 mm Hg, con hipertensión pulmonar moderada a severa (presión sistólica pulmonar promedio 61 mmHg y media pulmonar 41,7) y alteración de la función sistólica del ventriculo derecho. La función ventrícular izquierda presentaba alteración diastólica en 87 por ciento de los sujetos, mientras que la función sistólica era completamente normal, con una fracción de eyección normal (promedio 59,7 por ciento) todos los sujetos estudiados tuvieron actividad de renina plasmática normal. Conclusión: No se detectó alteración en la actividad de renina plasmática en este grupo de pacientes con EPOC estable y sin antecedentes de edemas o disfunción ventricular izquierda. Para alcanzar el verdadero papel de las alteraciones neurohormonales en la fisiopatología de la EPOC es necesario estudiar sujetos con mayor grado de hipercapnia o con diferente función ventricular izquierda.