Search

In Kilifi (Kenya), a pneumococcal conjugate vaccine (PCV10) was introduced in 2011 in infants (aged <1 year, 3 + 0 schedule) with a catch-up campaign in children aged 1–4 years. We aimed to measure the effect of PCV10 on population immunity. In this observational study, repeated cross-sectional serosurveys were conducted in independent random samples of 500 children younger than 15 years every 2 years between 2009 and 2017. During these surveys, blood samples were collected by venesection. Concentrations of anti-capsular IgGs against vaccine serotypes (VTs) 1, 4, 5, 6B, 7F, 9V, 14, 18C, 19F, and 23F, and against serotypes 6A and 19A, were assayed by ELISA. We plotted the geometric mean concentrations (GMCs) by birth year to visualise age-specific antibody profiles. In infants, IgG concentrations of 0·35 μg/mL or higher were considered protective. Of 3673 volunteers approached, 2152 submitted samples for analysis across the five surveys. Vaccine introduction resulted in an increase in the proportion of young children with protective IgG concentrations, compared with before vaccine introduction (from 0–33% of infants with VT-specific levels over the correlate of protection in 2009, to 60–94% of infants in 2011). However, among those vaccinated in infancy, GMCs of all ten VTs had waned rapidly by the age of 1, but rose again later in childhood. GMCs among children aged 10–14 years were consistently high over time (eg, the range of GMCs across survey rounds were between 0·45 μg/mL and 1·00 μg/mL for VT 23F and between 2·00 μg/mL and 3·11 μg/mL for VT 19F). PCV10 in a 3 + 0 schedule elicited protective IgG levels during infancy, when disease risk is high. The high antibody levels in children aged 10–14 years might indicate continued exposure to vaccine serotypes due to residual carriage or to memory responses to cross-reactive antigens. Despite rapid waning of IgG after vaccination, disease incidence among young children in this setting remains low, suggesting that lower thresholds of antibody, or other markers of immunity (eg, memory B cells), may be needed to assess population protection among children who have aged past infancy. Gavi, the Vaccine Alliance; Wellcome Trust. [ABSTRACT FROM AUTHOR]

Background: The global Immunisation Agenda 2030 highlights coverage and equity as a strategic priority goal to reach high equitable immunisation coverage at national levels and in all districts. We estimated inequities in full immunisation coverage associated with socioeconomic, geographic, maternal, child, and place of birth characteristics among children aged 12-23 months in Kenya.Methods: We analysed full immunisation coverage (1-dose BCG, 3-dose DTP-HepB-Hib (diphtheria, tetanus, pertussis, hepatitis B and Haemophilus influenzae type B), 3-dose polio, 1-dose measles, and 3-dose pneumococcal vaccines) of 3943 children aged 12-23 months from the 2014 Kenya Demographic and Health Survey. We disaggregated mean coverage by socioeconomic (household wealth, religion, ethnicity), geographic (place of residence, province), maternal (maternal age at birth, maternal education, maternal marital status, maternal household head status), child (sex of child, birth order), and place of birth characteristics, and estimated inequities in full immunisation coverage using bivariate and multivariate logistic regression.Results: Immunisation coverage ranged from 82% [81-84] for the third dose of polio to 97.4% [96.7-98.2] for the first dose of DTP-HepB-Hib, while full immunisation coverage was 68% [66-71] in 2014. After controlling for other background characteristics through multivariate logistic regression, children of mothers with primary school education or higher have at least 54% higher odds of being fully immunised compared to children of mothers with no education. Children born in clinical settings had 41% higher odds of being fully immunised compared to children born in home settings. Children in the Coast, Western, Central, and Eastern regions had at least 74% higher odds of being fully immunised compared to children in the North Eastern region, while children in urban areas had 26% lower odds of full immunisation compared to children in rural areas. Children in the middle and richer wealth quintile households were 43-57% more likely to have full immunisation coverage compared to children in the poorest wealth quintile households. Children who were sixth born or higher had 37% lower odds of full immunisation compared to first-born children.Conclusions: Children of mothers with no education, born in home settings, in regions with limited health infrastructure, living in poorer households, and of higher birth order are associated with lower rates of full immunisation. Targeted programmes to reach under-immunised children in these subpopulations will lower the inequities in childhood immunisation coverage in Kenya. [ABSTRACT FROM AUTHOR]

The spread of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in Africa is poorly described. The first case of SARS-CoV-2 in Kenya was reported on 12 March 2020, and an overwhelming number of cases and deaths were expected, but by 31 July 2020, there were only 20,636 cases and 341 deaths. However, the extent of SARS-CoV-2 exposure in the community remains unknown. We determined the prevalence of anti–SARS-CoV-2 immunoglobulin G among blood donors in Kenya in April–June 2020. Crude seroprevalence was 5.6% (174 of 3098). Population-weighted, testperformance-adjusted national seroprevalence was 4.3% (95% confidence interval, 2.9 to 5.8%) and was highest in urban counties Mombasa (8.0%), Nairobi (7.3%), and Kisumu (5.5%). SARS-CoV-2 exposure is more extensive than indicated by case-based surveillance, and these results will help guide the pandemic response in Kenya and across Africa. [ABSTRACT FROM AUTHOR]

This study examines the prevalence of SARS-CoV-2 antibodies among blood donors aged 16 to 64 years in Kenya from January to March 2021. [ABSTRACT FROM AUTHOR]

Objective To estimate the population prevalence of active pulmonary tuberculosis in Gambia. Methods Between December 2011 and January 2013, people aged ≥ 15 years participating in a nationwide, multistage cluster survey were screened for active pulmonary tuberculosis with chest radiography and for tuberculosis symptoms. For diagnostic confirmation, sputum samples were collected from those whose screening were positive and subjected to fluorescence microscopy and liquid tuberculosis cultures. Multiple imputation and inverse probability weighting were used to estimate tuberculosis prevalence. Findings Of 100 678 people enumerated, 55 832 were eligible to participate and 43 100 (77.2%) of those participated. A majority of participants (42 942; 99.6%) were successfully screened for symptoms and by chest X-ray. Only 5948 (13.8%) were eligible for sputum examination, yielding 43 bacteriologically confirmed, 28 definite smear-positive and six probable smear-positive tuberculosis cases. Chest X-ray identified more tuberculosis cases (58/69) than did symptoms alone (43/71). The estimated prevalence of smear-positive and bacteriologically confirmed pulmonary tuberculosis were 90 (95% confidence interval, CI: 53-127) and 212 (95% CI: 152-272) per 100 000 population, respectively. Tuberculosis prevalence was higher in males (333; 95% CI: 233-433) and in the 35-54 year age group (355; 95% CI: 219-490). Conclusion The burden of tuberculosis remains high in Gambia but lower than earlier estimates of 490 per 100 000 population in 2010. Less than half of all cases would have been identified based on smear microscopy results alone. Successful control efforts will require interventions targeting men, increased access to radiography and more accurate, rapid diagnostic tests. [ABSTRACT FROM AUTHOR]

Background: A Tuberculin skin test (TST) survey was conducted to assess the prevalence of latent TB Infection (LTBI) and to estimate the annual risk of M. tuberculosis infection (ARTI) in Gambian school children. The results are expected to contribute to understanding of Tuberculosis epidemiology in The Gambia. Methods: This was a nationwide, multi-cluster survey in children aged 6–11 years. Districts, 20 of 37, were selected by probability proportional to size and schools by simple random sampling. All TST were performed using the Mantoux method. Height and weight measurements were obtained for all participants. We calculated prevalence of LTBI using cut-off points of 10mm, the mirror and mixture modelling methods. Results: TST readings were completed 13,386 children with median age of 9 years (interquartile range [IQR] 8–10 years). Mixture analysis yielded a cut-off point of 12 mm, and LTBI prevalence of 6.9% [95%CI 6.47–7.37] and the ARTI was 0.75% [95%CI 0.60–0.91]. LTBI was associated gender and urban residence (p <0.01). Nutritional status was not associated with non-reactive TST or sizes of TST indurations. ARTI did not differ significantly by age, gender, BCG vaccination or residence. Conclusions: This estimates for LTBI prevalence and ARTI were low but this survey provides updated data. Malnutrition did not affect estimates of LTBI and ARTI. Given the low ARTI in this survey and the overlapping distribution of indurations with mixture modelling, further surveys may require complementary tests such as interferon gamma release assays or novel diagnostic tools. [ABSTRACT FROM AUTHOR]

Background: The identification of Mycobacterium-tuberculosis (Mtb) infected individuals remains a challenge due to an insufficient understanding of immune responses detected with the current diagnostic tests for latent tuberculosis i.e. the tuberculin skin test (TST) or IFN–γ release assays (IGRAs) and an inability to distinguish infection stages with current immunologic assays. Further classification based on markers other than IFN–γ may help to define markers of early Mtb infection. Methods: We assessed the TST status of Mtb-exposed household contacts at baseline and at 6 months. Contacts were classified into those with initial positive TST (TST+); those with baseline negative TST but TST conversion at 6 months (TST converters, TSTC) and those with persistently negative TST (PTST−). We assessed their short- and long-term immune responses to PPD and ESAT–6/CFP–10 (EC) via IFN–γ ELISPOT and a multiplex cytokine array in relation to TST status and compared them to those of TB cases to identify immune profiles associated with a spectrum of infection stages. Results: After 1 and 6 days stimulation with EC, 12 cytokines (IFN–γ, IL–2, IP–10, TNF–α, IL–13, IL–17, IL–10, GMCSF, MIP–1β, MCP–3, IL–2RA and IL–1A) were not different in TSTC compared to TST+ suggesting that robust adaptive Mtb-specific immune responses precede TST conversion. Stratifying contacts by baseline IFN–γ ELISPOT to EC in combination with TST results revealed that IP–10 and IL–17 were highest in the group of TST converters with positive baseline ELISPOT, suggesting they might be markers for recent infection. Conclusion: We describe a detailed analysis of Mtb-specific biomarker profiles in exposed household contacts in a TB endemic area that provides insights into the dynamic immune responses to Mtb infection and may help to identify biomarkers for ‘at-risk’ populations beyond TST and IGRA. [ABSTRACT FROM AUTHOR]

Background: Introduction of pneumococcal vaccines in Nigeria is a priority as part of the Accelerated Vaccine Introduction Initiative (AVI) of the Global Alliance for Vaccines and Immunisation (GAVI). However, country data on the burden of pneumococcal disease (IPD) is limited and coverage by available conjugate vaccines is unknown. This study was carried out to describe the pre vaccination epidemiology and population biology of pneumococcal carriage in Nigeria. Methods: This was a cross sectional survey. Nasopharyngeal swabs (NPS) were obtained from a population sample in 14 contiguous peri-urban Nigerian communities. Data on demographic characteristics and risk factor for carriage were obtained from all study participants. Pneumococci isolated from NPS were characterised by serotyping, antimicrobial susceptibility and Multi Locus Sequencing Typing (MLST). Results: The prevalence of pneumococcal carriage was 52.5%. Carriage was higher in children compared to adults (67.4% vs. 26%), highest (≈90%) in infants aged ,9 months and reduced significantly with increasing age (P<0.001). Serotypes 19F (18.6%) and 6A (14.4%) were most predominant. Potential vaccine coverage was 43.8%, 45.0% and 62% for PCV-7, PCV-10 and PCV-13 respectively. There were 16 novel alleles, 72 different sequence types (STs) from the isolates and 3 Sequence Types (280, 310 and 5543) were associated with isolates of more than one serotype indicative of serotype switching. Antimicrobial resistance was high for cotrimoxazole (93%) and tetracycline (84%), a third of isolates had intermediate resistance to penicillin. Young age was the only risk factor significantly associated with carriage. Conclusions: Pneumococcal carriage and serotype diversity is highly prevalent in Nigeria especially in infants. Based on the coverage of serotypes in this study, PCV-13 is the obvious choice to reduce disease burden and prevalence of drug resistant pneumococci. However, its use will require careful monitoring. Our findings provide sound baseline data for impact assessment following vaccine introduction in Nigeria. [ABSTRACT FROM AUTHOR]

Background: Qualitative and quantitative changes in IGRA response offer promise as biomarkers to monitor Tuberculosis (TB) drug therapy, and for the comparison of new interventions. We studied the decay kinetics of TB-specific antigen T-cell responses measured with an in-house ELISPOT assay during the course of therapy. Methods: Newly diagnosed sputum smear positive TB cases with typical TB chest radiographs were recruited. All patients were given standard anti-TB treatment. Each subject was followed up for 6 months and treatment outcomes were documented. Blood samples were obtained for the ESAT-6 and CFP-10 (EC) ELISPOT at diagnosis, 1-, 2-, 4- and 6-months. Qualitative and quantitative reversion of the ELISPOT results were assessed with McNemar test, conditional logistic regression and mixed-effects hierarchical Poisson models. Results: A total of 116 cases were recruited and EC ELISPOT was positive for 87% (95 of 109) at recruitment. There was a significant decrease in the proportion of EC ELISPOT positive cases over the treatment period (p<0.001). Most of the reversion occurred between the start and first month of treatment and at completion at 6 months. ESAT-6 had higher median counts compared to CFP-10 at all time points. Counts for each antigen declined significantly with therapy (p<0.001). Reverters had lower median SFUs at the start of treatment compared to non-Reverters for both antigens. Apart from the higher median counts for non-Reverters, no other risk factors for non-reversion were found. Conclusions: TB treatment induces qualitative and quantitative reversion of a positive in-house IGRA in newly diagnosed cases of active TB disease. As this does not occur reliably in the majority of cured individuals, qualitative and quantitative reversion of an IGRA ELISPOT has limited clinical utility as a surrogate marker of treatment efficacy. [ABSTRACT FROM AUTHOR]

Background: T cells producing multiple factors have been shown to be required for protection from disease progression in HIV but we have recently shown this not to be the case in TB. Subjects with active disease had a greater proportion of polyfunctional cells responding to ESAT-6/CFP-10 stimulation than their infected but non-diseased household contacts (HHC). We therefore wanted to assess this profile in subjects who had successfully completed standard TB chemotherapy. Methods: We performed a cross-sectional study using PBMC from TB cases (pre- and post-treatment) and HHC. Samples were stimulated overnight with TB antigens (ESAT-6/CFP-10 and PPD) and their CD4+ and CD8+ T cells were assessed for production of CD107a, IFN-γ, IL-2 and TNF-α and the complexity of the responses was determined using SPICE and PESTLE software. Results and Conclusions: We found that an increase in complexity (i.e., production of more than 1 factor simultaneously) of the T cell profile was associated with TB disease and that this was significantly reduced following TB treatment. This implies that T cells are able to respond adequately to TB antigens with active disease (at least initially) but the ability of this response to protect the host from disease progression is hampered, presumably due to immune evasion strategies by the bacteria. These findings have implications for the development of new diagnostics and vaccine strategies. [ABSTRACT FROM AUTHOR]

Interferon gamma remains a key effector molecule that is still widely used as the most informative biomarker for screening human immune responses against tuberculosis, particularly in ELISPOT assays. We investigated the participation of CD4+ and CD8+ T lymphocytes in the PBMC responses to Mycobacterium tuberculosis (Mtb) specific antigens in 33 TB cases and 49 contacts. Responses to ESAT-6 were higher than CFP-10. There was no significant difference in responses to both Mtb antigens between cases and contacts. PBMCs response to ESAT-6 but not CFP-10 in cases was significantly reduced by depletion of CD4+ cells whereas CD8+ cell depletion had no impact. In conclusion, ESAT-6 is a more recognized antigen in this population, and CD4+ lymphocytes are the main participants in IFN-γ response by ELISPOT. Thus, a decline of CD4+ T lymphocytes below a critical level might affect the sensitivity of IFN-γ release assays for detecting Mtb infection. [ABSTRACT FROM AUTHOR]

Background: IFN-γ Release Assays (IGRAs) have been licensed for the diagnosis of latent Mycobacterium tuberculosis infection (LTBI). Their performance may depend on assay format and may vary across populations and settings. We compared the diagnostic performance of an in-house T -cell and commercial whole blood-based IGRAs for the diagnosis of LTBI and TB disease in The Gambia. Methods: Newly diagnosed sputum smear positive cases and their household contacts were recruited. Cases and contacts were bled for IGRA and contacts had a Mantoux skin test. We assessed agreement and discordance between the tests and categorized a contact's level of M. tuberculosis exposure according to where s/he slept relative to a case: the same room, same house or a different house. We assessed the relationship between exposure and test results by multiple logistic regression. Results: In 80 newly diagnosed TB cases, the sensitivity of ELISPOT was 78.7% and for QFT-GIT was 64.0% (p = 0.047). Of 194 household contacts 57.1% and 58.8% were positive for ELISPOT and QFT-GIT respectively. The overall agreement between both IGRAs for LTBI in contacts was 71.4% and there was no significant discordance (p = 0.29). There was significant discordance between the IGRAs and TST. Neither IGRA nor TST had evidence of false positive results because of Bacille Calmette Gu&erin;rin (BCG) vaccination. However, agreement between QFT-GIT and TST as well as discordance between both IGRAs and TST were associated with BCG vaccination. Both IGRAs responded to the M. tuberculosis exposure gradient and were positively associated with increasing TST induration (p = 0.003 for ELISPOT and p = 0.001 for QFT-GIT). Conclusion: The ELISPOT test is more sensitive than the QFT-GIT for diagnosing TB disease. The two tests perform similarly in the diagnosis of LTBI in TB contacts. Significant discordance between the two IGRAs and between each and the TST remain largely unexplained. [ABSTRACT FROM AUTHOR]

Population-level health and mortality data are crucial for evidence-informed policy but scarce in Nigeria. To fill this gap, we undertook a comprehensive assessment of the burden of disease in Nigeria and compared outcomes to other west African countries. In this systematic analysis, using data and results of the Global Burden of Diseases, Injuries, and Risk Factors Study 2019, we analysed patterns of mortality, years of life lost (YLLs), years lived with disability (YLDs), life expectancy, healthy life expectancy (HALE), and health system coverage for Nigeria and 15 other west African countries by gender in 1998 and 2019. Estimates of all-age and age-standardised disability-adjusted life-years for 369 diseases and injuries and 87 risk factors are presented for Nigeria. Health expenditure per person and gross domestic product were extracted from the World Bank repository. Between 1998 and 2019, life expectancy and HALE increased in Nigeria by 18% to 64·3 years (95% uncertainty interval [UI] 62·2–66·6), mortality reduced for all age groups for both male and female individuals, and health expenditure per person increased from the 11th to third highest in west Africa by 2018 (US$18·6 in 2001 to $83·75 in 2018). Nonetheless, relative outcomes remained poor; Nigeria ranked sixth in west Africa for age-standardised mortality, seventh for HALE, tenth for YLLs, 12th for health system coverage, and 14th for YLDs in 2019. Malaria (5176·3 YLLs per 100 000 people, 95% UI 2464·0–9591·1) and neonatal disorders (4818·8 YLLs per 100 000, 3865·9–6064·2) were the leading causes of YLLs in Nigeria in 2019. Nigeria had the fourth-highest under-five mortality rate for male individuals (2491·8 deaths per 100 000, 95% UI 1986·1–3140·1) and female individuals (2117·7 deaths per 100 000, 1756·7–2569·1), but among the lowest mortality for men older than 55 years. There was evidence of a growing non-communicable disease burden facing older Nigerians. Health outcomes remain poor in Nigeria despite higher expenditure since 2001. Better outcomes in countries with equivalent or lower health expenditure suggest health system strengthening and targeted intervention to address unsafe water sources, poor sanitation, malnutrition, and exposure to air pollution could substantially improve population health. The Bill & Melinda Gates Foundation. [ABSTRACT FROM AUTHOR]

Background Vaccination against coronavirus disease 2019 (COVID-19) is a cost-effective mitigation strategy against the pandemic. As the COVID-19 vaccine becomes more available, low uptake is now a global threat and understanding the underpinnings in local contexts is a priority for intervention development. We aimed to evaluate behavioural determinants of COVID-19 vaccine acceptance that could inform engagement strategies to improve vaccine uptake in Makoko, an urban slum in Lagos, Nigeria. Methods A population-based case–control study utilized the barrier analysis (BA) approach to evaluate the beliefs and behaviours of 45 'doers' and 45 'non-doers'. The standardized BA tabulation sheet was used to assess differences in the proportions between the two groups to identify significant factors that could be addressed through a behaviour change strategy. Results Perceived social norms (family, friend, healthcare workers) that approve the vaccine and expected vaccine protection against diseases among doers were determinants of behaviour. Perceived poor accessibility, safety concerns, lack of trust, low vaccine efficacy and low susceptibility to the infection were the most important determinants of behaviour among non-doers. Conclusions Measures to improve COVID-19 vaccine acceptance in Makoko should include improvement in accessibility and exposing myths and misinformation through clear, concise and evidence-based community education delivered by trusted persons such as healthcare workers and religious leaders. [ABSTRACT FROM AUTHOR]

Introduction: The high proportion of SARS-CoV-2 infections that have remained undetected presents a challenge to tracking the progress of the pandemic and estimating the extent of population immunity. Methods: We used residual blood samples from women attending antenatal care services at three hospitals in Kenya between August 2020 and October 2021and a validated IgG ELISA for SARS-Cov-2 spike protein and adjusted the results for assay sensitivity and specificity. We fitted a two-component mixture model as an alternative to the threshold analysis to estimate of the proportion of individuals with past SARS-CoV-2 infection. Results: We estimated seroprevalence in 2,981 women; 706 in Nairobi, 567 in Busia and 1,708 in Kilifi. By October 2021, 13% of participants were vaccinated (at least one dose) in Nairobi, 2% in Busia. Adjusted seroprevalence rose in all sites; from 50% (95%CI 42–58) in August 2020, to 85% (95%CI 78–92) in October 2021 in Nairobi; from 31% (95%CI 25–37) in May 2021 to 71% (95%CI 64–77) in October 2021 in Busia; and from 1% (95% CI 0–3) in September 2020 to 63% (95% CI 56–69) in October 2021 in Kilifi. Mixture modelling, suggests adjusted cross-sectional prevalence estimates are underestimates; seroprevalence in October 2021 could be 74% in Busia and 72% in Kilifi. Conclusions: There has been substantial, unobserved transmission of SARS-CoV-2 in Nairobi, Busia and Kilifi Counties. Due to the length of time since the beginning of the pandemic, repeated cross-sectional surveys are now difficult to interpret without the use of models to account for antibody waning. [ABSTRACT FROM AUTHOR]

OBJECTIVE. To compare the enzyme-linked immunospot (ELISPOT) assay with the tuberculin skin test (TST) in children for the diagnosis of Mycobacterium tuberculosis infection in the Gambia. METHODS. We divided child contacts of sputum smear-positive tuberculosis cases into 3 age categories (<5, 5-9, and 10-14 years) and assessed agreement between the 2 tests plus their relationship to prior Bacille Calmette-Guerin (BCG) vaccination. We categorized a child's level of M tuberculosis exposure according to where he/she slept relative to a case: the same room, same house, or a different house. The relationship between exposure and test result was assessed by multiple logistic regression. RESULTS. In child contacts of 287 cases, 225 (32.5%) of 693 were positive by TST and 232 (32.3%) of 718 by ELISPOT. The overall agreement between tests was 83% and the discordance was not significant. Both tests responded to the M tuberculosis exposure gradient in each age category. The percentage of those who were TST positive/ELISPOT negative increased with increasing exposure. At the lowest exposure level, the percentage of ELISPOT-positive children who were TST negative was increased compared with the highest exposure level. Neither test had evidence of false positive results because of BCG. CONCLUSIONS. In Gambian children, the ELISPOT is slightly less sensitive than the TST in the diagnosis of M tuberculosis infection from recent exposure, and neither test is confounded by prior BCG vaccination. Evidence of reduced TST sensitivity in subjects with the lowest known recent M tuberculosis exposure suggests that, when maximal sensitivity is important, the 2 tests may lie best used together. [ABSTRACT FROM AUTHOR]

Policy decisions on COVID-19 interventions should be informed by a local, regional and national understanding of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) transmission. Epidemic waves may result when restrictions are lifted or poorly adhered to, variants with new phenotypic properties successfully invade, or infection spreads to susceptible subpopulations. Three COVID-19 epidemic waves have been observed in Kenya. Using a mechanistic mathematical model, we explain the first two distinct waves by differences in contact rates in high and low social-economic groups, and the third wave by the introduction of higher-transmissibility variants. Reopening schools led to a minor increase in transmission between the second and third waves. Socioeconomic and urbanÐrural population structure are critical determinants of viral transmission in Kenya. [ABSTRACT FROM AUTHOR]

To reach performance levels comparable to human experts, computer-aided detection (CAD) systems are typically optimized following a supervised learning approach that relies on large training databases comprising manually annotated lesions. However, manually outlining those lesions constitutes a difficult and time-consuming process that renders detailedly annotated data difficult to obtain. In this paper, we investigate an alternative approach, namely multiple-instance learning (MIL), that does not require detailed information for optimization. We have applied MIL to a CAD system for tuberculosis detection. Only the case condition (normal or abnormal) was required during training. Based upon the well-known miSVM technique, we propose an improved algorithm that overcomes miSVM’s drawbacks related to positive instance underestimation and costly iteration. To show the advantages of our MIL-based approach as compared with a traditional supervised one, experiments with three X-ray databases were conducted. The area under the receiver operating characteristic curve was utilized as a performance measure. With the first database, for which training lesion annotations were available, our MIL-based method was comparable to the supervised system (0.86 versus 0.88). When evaluating the remaining databases, given their large difference with the previous image set, the most appealing strategy was to retrain the CAD systems. However, since only the case condition was available, only the MIL-based system could be retrained. This scenario, which is common in real-world applications, demonstrates the better adaptation capabilities of the proposed approach. After retraining, our MIL-based system significantly outperformed the supervised one (0.86 versus 0.79 and 0.91 versus 0.85, p<0.0001 and p=0.0002, respectively). [ABSTRACT FROM AUTHOR]