Table 2– Details and performance characteristics of chest radiograph scoring systems
StudyFeatureTest detailsRuleSensitivity % (95% CI)#Specificity % (95% CI)#Area under the curve (SE)
Studies that included all patients with possible PTB
Bock et al . [20], 1996Logistic regression DOR (95%CI)Any of 1–3 or 4 (in the absence of 5) considered test positive81
1) CXR with upper lobe infiltrate5 (2.38–10.51)
2) CXR with cavity3.93 (1.06–14.62)
3) Knew someone with active TB2.42 (1.1–5.32)
4) Self-report of positive tuberculin skin test in past5.67 (1.57–22.01)
5) Self-report of isoniazid preventive therapy in the past0.18 (0.04–0.82)
El-solh et al . [21], 1997Upper zone disease
Weight loss
Diabetes mellitus
Used classification and regression tree analysisUpper zone disease absent and fever absent = test negative
Upper zone disease absent, but fever present = test negative if no weight loss and CD4+ ≥200.
Everyone else to be considered test positive
100 (78–100)50 (44–57)0.878 (0.029)
El-solh et al . [22], 1999Age, CD4+ counts, diabetes mellitus, HIV, tuberculin skin test positivityUsed general regression neural network100 (91–100)72 (65–77)0.947 (0.028)
Chest pain, weight loss, cough, night sweats, fever, shortness of breath
Upper lobe infiltrate, lower lobe infiltrate, upper lobe cavity, lower lobe cavity, adenopathy, unilateral pleural effusion, bilateral pleural effusion, pleural thickening, miliary pattern, normal
Moran et al . [23], 20091) Apical infiltrateUsed Chi-squared recursive partitioningAny of 1–7 present = test positive9649
2) Cavitation
3) Immigrant
4) Weight loss
5) Positive TB history
6) Homeless
7) Incarcerated
Mylotte et al . [24], 1997DOR (95% CI)ScoreScore ≥3 = test positive8863.20.86
(0.04)
1) AFB positive smear5.8 (3–11)3
2) Localised CXR change2.5 (1.3–4.9)2
3) Correctional facility residence2.3 (1.2–4.4)2
4) History of weight loss1.8 (1–3.2)1
DOR (95% CI)Score
Solari et al . [25], 20081) Age <35 years0.97 (0.96– 0.99)0Score ≥3 = test positive93420.809
2) Age 35–60 years-1(0.05)
3) Age ≥61 years-2
4) Weight loss2.79 (1.51–5.18)5
5) History of PTB0.51 (0.28–0.95)-3
6) Miliary pattern8.04 (2.79–23.16)10
7) Cavity2.54 (1.4–4.62)5
8) Upper lobe infiltrate5.64 (3.2–9.93)9
Studies that only included smear-negative patients with possible PTB
Lagrange-Xélot et al . [26], 2010Score≥1 = test positive96.221.3
TB risk factors or chronic symptoms4
Self-report of positive tuberculin skin test in past5
Shortness of breath-3
Temperature <38.5°C0
Temperature 38.5 –39°C3
Temperature >39°C6
Crackles on physical examination-3
Upper lobe disease on CXR6
Soto et al . [27], 2008DOR (95% CI)Score<0 = low probability
>4 = high probability
At a cut-off<0At a cut-off of >2
0.83 (0.07)
Haemoptysis3.24 (1.11– 9.22)29350
Weight loss2.35 (0.86–6.43)1At a cut-off >4
Age >45 years2.01 (1.01–3.01)-192
Expectoration0.35 (0.14–0.9)-1
Apical infiltrate4.29 (1.7–10.86)3
Miliary infiltrate9.31 (2.21–39.24)4
Soto et al . [28], 2011Same as above – validation study≥5 = high probabilityAt a cut-off<0
97.8 (94.5–99.4)14 (11–17.4)
At a cut-off >5
23.9 (17.9–30.7)93.1 (90.5–95.2)
Wisnivesky et al . [30], 2000DOR (95% CI)Score>1 = test positive98 (95–100)46
(33–59)
TB risk factors or chronic symptoms7.9 (4.4–24.2)4
Self-report of positive tuberculin skin test in past13.2 (4.4–40.7)5
Shortness of breath0.2 (0.1–0.5)-3
Temperature<38.5°C0.3 (0.1–0.5)0
Temperature 38.5–39°C3
Temperature >39°C6
Crackles on physical examination0.3 (0.1–0.5)-3
Upper lobe disease on CXR14.6 (3.7–57.5)6
Wisnivesky et al. [29], 2005Same as above95 (74–100)35 (31–40)
Study that excluded people living with HIV
Rakoczy et al . [31], 2008DOR (95% CI)Score>4 = test positive9742
Chronic symptoms10.21 (2.95–35.4)6
Immunosuppression+
(other than HIV)
8.14 (2.08–31.8)4
Foreign birth7.01 (2.1–23.8)2
CXR upper zone findings5.28 (1.6–17.2)2
Shortness of breath0.13 (0.04–0.45)-2
Validation of the score developed by Wisnivesky9618
  • CXR: chest radiograph; TB: tuberculosis; DOR: diagnostic odds ratio; AFB: acid-fast bacilli; PTB: pulmonary TB. #: 95% CIs are included when the published study either provided these, or the data were sufficient for the CIs to be calculated; : validation of the score developed by Wisnivesky et al. [30]; +: Calculated assuming a prevalence of 2%.