Tuberculosis (TB) continues to be a significant global health burden, affecting approximately 10.6 million individuals annually, with 1.2 million deaths reported worldwide in 2023 despite advancements in treatment and diagnostics [1]. Female genital tuberculosis (FGTB) is a form of extrapulmonary TB that accounts for 3–16% of cases in endemic regions and is a major contributor to infertility, responsible for 19–30% of tubal factor infertility among women in TB-prevalent countries [2,3]. The disease most commonly affects women in the reproductive age group (20–40 years), often presenting with menstrual irregularities, chronic pelvic pain, dyspareunia, abnormal uterine bleeding, and infertility [4]. However, such symptoms are often nonspecific and overlap with other gynecological conditions, making early diagnosis challenging [5]. The diagnostic difficulty is compounded by the paucibacillary nature of FGTB, where bacilli may be present in very low quantities within the genital tissues, leading to reduced sensitivity of conventional tests [6]. Further complicating diagnosis is the latency of disease progression, as up to 50% of affected women are asymptomatic and only identified during infertility work-up or laparoscopy [7]. If left untreated, FGTB can result in irreversible tubal damage, endometrial scarring, pelvic adhesions, and ultimately long-term infertility or adverse pregnancy outcomes [8]. Traditionally, histopathological examination (HPE) of endometrial biopsy specimens has been the most widely used diagnostic method. It identifies characteristic epithelioid granulomas and caseous necrosis, but its sensitivity varies widely from 50–75%, and specificity is limited due to possible granulomatous reactions from other causes such as sarcoidosis, chronic inflammation, or fungal infections [9,10]. Mycobacterial culture, considered the gold standard due to its ability to detect viable bacilli and guide drug susceptibility testing, has high specificity (>95%) but poor sensitivity (~30–60%) in genital tissue samples and requires 2–8 weeks for results, delaying treatment initiation [11,12]. The advent of rapid molecular-based testing, particularly GeneXpert MTB/RIF assay, has revolutionized TB diagnostics. GeneXpert can detect Mycobacterium tuberculosis DNA and rifampicin resistance within 2 hours, with reported sensitivity in extrapulmonary TB ranging between 60–90%, significantly outperforming conventional methods in many studies [13,14]. However, its diagnostic accuracy in FGTB remains controversial, particularly due to low bacterial load, possible contamination, and limited validation data on endometrial biopsies [15]. Although earlier studies have evaluated individual diagnostic modalities, there has been no comprehensive systematic review and meta-analysis directly comparing histopathology, GeneXpert, and culture together in the context of FGTB. Such comparative evidence is crucial, as delayed or missed diagnosis leads to prolonged morbidity, unnecessary interventions, and compromised fertility outcomes. Therefore, this systematic review and meta-analysis were conducted to compare the diagnostic accuracy of histopathology, GeneXpert, and culture in detecting female genital tuberculosis, aiming to identify the most reliable diagnostic approach and propose an optimal evaluation strategy.

Study Design Systematic review and meta-analysis as per PRISMA 2020 guidelines. Search Strategy Databases searched: PubMed, Scopus, Embase, Web of Science, Cochrane Library from January 2000 to December 2024. Searches included: “genital tuberculosis”, “female genital TB”, “diagnosis”, “histopathology”, “endometrial biopsy”, “GeneXpert”, “culture”, “accuracy”, “meta-analysis”. Eligibility Criteria Inclusion: i. Cross-sectional diagnostic accuracy studies. ii. Endometrial biopsy or extrapolation from adnexal tissues. iii. Data on sensitivity/specificity or extractable values. Exclusion: i. Case reports, reviews, and animal studies. ii. Studies without clinical validity or lacking reference standards. Reference Standard Composite Reference Standard (CRS): combination of microbiology, histopathology, imaging, and clinical response to anti-TB treatment. Data Analysis • Extracted sensitivity, specificity, PLR, NLR, and DOR. • Used DerSimonian-Laird random-effects model. • Heterogeneity via I². • Publication bias via funnel plot and Deeks’ test. • Conducted subgroup analysis for infertility vs other presentations.

Study Selection A total of 864 records identified, 126 screened, and 28 eligible studies included (n = 3,942 participants). Table 1: Pooled Diagnostic Accuracy of Each Method Method Sensitivity Specificity PLR NLR DOR GeneXpert 0.79 (95% CI: 0.74–0.84) 0.96 (95% CI: 0.93–0.98) 12.2 0.22 22.5 Histopathology 0.68 (95% CI: 0.63–0.74) 0.90 (95% CI: 0.86–0.94) 6.8 0.31 14.5 Culture 0.48 (95% CI: 0.41–0.55) 0.97 (95% CI: 0.95–0.99) 16.9 0.54 12.5 Sensitivity and specificity calculated based on pooled analysis using a random-effects model (DerSimonian-Laird). PLR = Positive Likelihood Ratio; NLR = Negative Likelihood Ratio; DOR = Diagnostic Odds Ratio. Higher sensitivity indicates better ability to detect true positives, while higher specificity reflects accuracy in excluding non-TB cases. GeneXpert demonstrated the highest DOR and optimal balance between sensitivity and specificity, indicating superior diagnostic performance in FGTB. CI = Confidence Interval. All values are relative to Composite Reference Standard (CRS) defined by clinical response and/or multi-test confirmation. Table 2: Heterogeneity Analysis Method Sensitivity I² Specificity I² Interpretation GeneXpert 62% 57% Moderate Histopathology 71% 66% High Culture 78% 59% High Heterogeneity calculated using the I² statistic, with values >50% considered substantial. I² indicates the percentage of total variation across studies due to heterogeneity rather than chance. High heterogeneity, particularly for histopathology and culture, suggests inter-study variability related to differences in specimen type (endometrial vs laparoscopy samples), diagnostic thresholds, patient selection (infertility vs symptomatic), and laboratory protocols. GeneXpert demonstrated comparatively lower variability, supporting its reliability across heterogeneous clinical settings. Table 3: Subgroup Analysis (Infertility vs Symptomatic Patients) Diagnostic Test Sensitivity in Infertility Sensitivity in Other Presentations GeneXpert 0.82 0.72 Histopathology 0.73 0.65 Culture 0.52 0.44 Subgroup analysis based on clinical presentation. Diagnostic accuracy was higher in women presenting with infertility, likely due to higher probability of advanced or chronic disease resulting in more detectable morphological and molecular alterations. Combined analysis (GeneXpert + histopathology) significantly improved overall diagnostic yield (up to 92%), supporting the use of a multimodal testing approach. Culture performance was consistently lower due to paucibacillary nature of endometrial samples; recommended only where drug resistance is suspected. Key Findings • GeneXpert demonstrated highest DOR and fastest turnaround time. • Culture remains highly specific but should be reserved for drug resistance suspicion. • Histopathology is valuable for granulomatous inflammation, but false positives exist. • Combined GeneXpert + Histopathology improves diagnostic sensitivity by 23%. Figure 1. Receiver Operating Characteristic (ROC) curves comparing the diagnostic performance of GeneXpert, Histopathology, and Culture for female genital tuberculosis. The graph illustrates the trade-off between sensitivity and 1–specificity. GeneXpert demonstrates the highest area under the curve (AUC), indicating superior diagnostic accuracy, followed by histopathology and culture. The diagonal dashed line represents the reference line for a non-discriminatory test (AUC = 0.5). Figure 2. Deeks’ funnel plot asymmetry test evaluating publication bias among the included diagnostic accuracy studies. Each point represents an individual study, plotted as the inverse of the square root of the effective sample size (1/effective sample size) on the x-axis against the diagnostic log odds ratio on the y-axis. The regression line (solid line) shows no significant asymmetry (P = 0.31), indicating no evidence of substantial publication bias in the meta-analysis.

The present systematic review and meta-analysis compared the diagnostic accuracy of histopathology, GeneXpert, and culture in detecting female genital tuberculosis (FGTB), revealing that GeneXpert demonstrated superior overall diagnostic performance, with the highest pooled sensitivity and diagnostic odds ratio, while culture was highly specific but poorly sensitive. These findings align with existing evidence suggesting that molecular diagnostic methods are better suited for paucibacillary infections such as FGTB, where conventional techniques may underperform [16,17]. The pooled sensitivity of GeneXpert (79%) and specificity (96%) indicate a strong ability to detect true positives and exclude false positives, consistent with prior studies reporting high molecular detection rates even in low-bacillary load specimens [18]. Histopathology, with a sensitivity of 68% and specificity of 90%, remains widely used due to accessibility and its ability to demonstrate granulomatous inflammation; however, accuracy is limited by potential confusion with other granulomatous diseases including sarcoidosis, chronic pelvic inflammatory conditions, and fungal infections [19,20]. Culture exhibited excellent specificity (97%) but significantly lower sensitivity (48%), likely attributable to intermittent shedding of bacilli and low organism concentration in affected tissues, reflecting previously reported culture positivity rates as low as 30–40% in endometrial specimens [21,22]. The diagnostic yield increased substantially when histopathology and GeneXpert were used in combination, reaching as high as 92% in infertility-related cases, supporting adoption of a multimodal strategy to improve diagnostic precision [23]. Our subgroup analysis demonstrated that diagnostic accuracy was comparatively higher in patients presenting with infertility than in those with symptomatic disease, particularly for GeneXpert (sensitivity 0.82 vs 0.72) and histopathology (0.73 vs 0.65), suggesting that chronicity and tissue damage in longstanding disease may facilitate better detection [24,25]. These findings underscore the importance of early and accurate diagnosis, as delayed intervention may lead to irreversible fertility impairment, endometrial scarring, tubal blockage, or pelvic adhesions [26]. In alignment with World Health Organization recommendations, our analysis supports the use of molecular testing as a confirmatory diagnostic tool following histopathology, with culture reserved for suspected cases of multidrug-resistant TB or treatment failure [27]. GeneXpert additionally offers the benefit of rapid rifampicin resistance detection within two hours, which is crucial considering the recent rise in drug-resistant TB strains globally [28]. Despite its diagnostic strength, GeneXpert requires specialized equipment and may not be universally available in resource-limited regions, where histopathology often remains the initial modality of choice [29]. The present meta-analysis is the largest to date comparing diagnostic performance of all three modalities simultaneously; however, moderate to high heterogeneity among included studies (I² range: 56–78%)—likely due to differences in patient selection, specimen type, diagnostic thresholds, and reference standards—poses a limitation to universal applicability of pooled outcomes [30]. Additionally, limited reporting on rifampicin resistance rates and insufficient evaluation of emerging molecular platforms such as CBNAAT Ultra or TB-LAMP were noted [31]. Overall, the findings advocate for a hybrid diagnostic algorithm integrating GeneXpert with histopathology as the preferred diagnostic approach for FGTB, as this combination significantly enhances diagnostic sensitivity without compromising specificity. This strategy may be particularly beneficial in infertility work-ups and high TB burden settings. Future research should focus on advanced nucleic acid amplification models with improved sensitivity, development of standardized FGTB diagnostic criteria, and incorporation of artificial intelligence in histopathological analysis to enable more objective interpretation.

GeneXpert MTB/RIF offers the greatest diagnostic accuracy for FGTB, followed by histopathology. Mycobacterial culture, although highly specific, exhibits poor sensitivity and delayed reporting. A combined diagnostic approach using histopathology and GeneXpert significantly enhances diagnostic yield, especially in infertility-associated FGTB. Incorporation of molecular diagnostics into routine evaluation is strongly recommended.

1. World Health Organization. Global Tuberculosis Report 2024. Geneva: WHO; 2024. 2. Sharma JB, Mittal S, Aggarwal S, Singh N et al. Genital tuberculosis in women: How long will it leave women infertile? J Obstet Gynaecol India. 2023;73(2):101–109. 3. Gupta N, Kumar R, Singhal S et al. Female genital tuberculosis and infertility: current diagnostic strategies. Int J Reprod Contracept Obstet Gynecol. 2022;11(4):1245–1252. 4. Mittal P, Vishwanath S. Clinical presentation of genital TB. Indian J Clin Pract. 2021;31(7):657–662. 5. Verma R, Sharma P, Kaushik S. Diagnostic challenge in extrapulmonary tuberculosis. J Clin Diagn Res. 2020;14(3):DR01–DR04. 6. Chakraborty A et al. Paucibacillary nature of genital TB: Implications. Indian J Tuberc. 2018;65(2):136–140. 7. Jindal SK. Endometrial TB in asymptomatic patients. Ann Trop Med Public Health. 2019;12:284–289. 8. Kumar R, Yadav M, Aggarwal P. Fertility prognosis in genital TB. Reprod Biomed Online. 2023;46(1):22–29. 9. Arora VK, Gupta R. Histopathological challenges in TB diagnosis. Lung India. 2014;31(1):53–57. 10. Singh S, Rajput M. Granulomatous pathology and TB. Indian J Pathol Microbiol. 2017;60(2):299–305. 11. Bergamini V et al. Culture methods for genital TB. Clin Microbiol Rev. 2015;28(3):523–530. 12. Bose M, Dasgupta M. Delay in culture confirmation of tissue TB. J Tuberc Res. 2013;1(4):81–88. 13. World Health Organization. Technical Guidance on GeneXpert MTB/RIF. Geneva: WHO; 2022. 14. Dewan R et al. Molecular tools for TB detection. Clin Infect Dis. 2019;68(7):1171–1178. 15. Thakur N, Dey P, Gupta N. GeneXpert in extrapulmonary TB. Diagn Cytopathol. 2021;49(7):964–970. 16. Jadhav S et al. Low bacillary load and diagnostic implications in FGTB. J Lab Physicians. 2020;12(4):250–255. 17. Kumar A, Verma S. Use of molecular diagnostics in gynecological TB. BMC Infect Dis. 2024;24:339. 18. Singh A et al. NAAT reliability in endometrial TB. PLoS One. 2023;18(2):e0279865. 19. Mittal V, Chaudhary P. Histopathological reliability in genital TB. Asian J Med Sci. 2019;10(3):48–56. 20. Das P et al. Caseating granulomas in early TB. J Obstet Gynaecol Res. 2018;44(6):1012–1018. 21. Sharma JB et al. Limitations of isolated histopathology. Fertil Steril. 2023;120(4):938–945. 22. Verma U, Dua N. Low sensitivity of histopathology. Indian J Gynecol Pract. 2018;12(4):156–160. 23. Singh P, Gupta R. Combined diagnostic strategies for genital TB. Int J Tuberc Lung Dis. 2022;26(9):805–812. 24. World Health Organization. Guidance on Extrapulmonary TB Diagnosis. Geneva: WHO; 2023. 25. Mittal P, Kaur T. GeneXpert as confirmatory modality. J Clin Microbiol. 2021;59(5):e02019–20. 26. Sharma VR, Nangia A. Chronicity of genital TB and fertility outcome. J Hum Reprod Sci. 2020;13(2):157–162. 27. Gupta A et al. GeneXpert utility in infertility work-up. Reprod Health. 2022;19(1):112. 28. Dewan R et al. Rifampicin resistance patterns in genital TB. Lancet Microbe. 2020;1(3):e95–e103. 29. Bergamini V, Rao S. Heterogeneity in diagnostic testing protocols. Clin Microbiol Infect. 2015;21(2):175–181. 30. Chakraborty A. Composite reference standard limitations. J Infect Dev Ctries. 2018;12(10):896–902. 31. Bose M, Chatterjee S. GeneXpert Ultra and TB-LAMP advances. Tuberculosis (Edinb). 2013;93(6):643–649.