Abstract:

Objective To evaluate the value of GeneXpert MTB/RIF technology (GeneXpert) and BACTEC MGIT 960 liquid culture (MGIT 960) in detecting and diagnosing chest wall tuberculosis with different surgical specimens. Methods Two hundred and fourteen patients with suspected chest wall tuberculosis admitted from June 2016 to July 2019 in Xi’an Chest Hospital were recruited. Pus, caseous necrotic tissues and granulation tissues in the lesions were obtained through surgical operation. GeneXpert and MGIT 960 were used to detect the above specimens. One hundred and seventy-four cases of chest wall tuberculosis and 40 cases of non-tuberculous chest wall disease were confirmed by comprehensive clinical diagnosis and pathological examination. The sensitivity, specificity, positive predictive value, negative predictive value and Kappa value of GeneXpert and MGIT 960 in the diagnosis of chest wall tuberculosis were evaluated with the clinical diagnostic results as the reference standard. The results (positive rate difference) of GeneXpert and MGIT 960 were compared among the three specimens. Statistical analysis was performed by Chi-square test, and P value less than 0.05 was considered statistically significant. Results Among these 214 suspected chest wall tuberculosis patients,174 cases were confirmed by preoperative blood test, imaging data and postoperative pathological results, and 40 cases were non-tuberculous chest wall disease. There were 18 cases of pyogenic chest wall abscess, 9 cases of chest wall cyst with common bacterial infection, 8 cases of connective tissue tumors, 3 cases of rib osteomyelitis and 2 cases of chest wall actinomycosis. With the clinical diagnosis as the standard, the sensitivity of GeneXpert to identify tuberculosis was 82.18% (143/174), the specificity was 97.50% (39/40), the positive predictive value was 99.31% (143/144), the negative predictive value was 55.71% (39/70), and the Kappa value was 0.618; The sensitivity of MGIT 960 identification of tuberculosis was 52.30% (91/174), the specificity was 100.00% (40/40), the positive predictive value was 100.00% (91/91), the negative predictive value was 32.52% (40/123), and the Kappa value was 0.291. The positive test results of the two techniques for the three types of surgical specimens were: granulation tissue (81.9%, 113/138), caseous necrotic tissue (68.8%, 88/128), and abscess (44.7%, 68/152). The positive rates was compared between pus group and caseous necrotic tissue group, χ ²=16.23, P<0.001; the positive rates were compared between caseous necrotic tissue group and granulation tissue group, χ ²=4.19, P=0.013; the positive rates were compared between granulation tissue group and pus group, χ ²=42.54, P<0.001. Conclusion The sensitivity of chest wall tuberculosis identified by GeneXpert is significantly higher than MGIT 960, and the specificity is comparable to that of MGIT 960, which can be used for rapid detection of surgical specimens. GeneXpert and MGIT 960 combined detection of three types specimens of suspected chest wall tuberculosis patients, granulation tissue positive rate is higher.

Key words: Tuberculosis, Thoracic wall, Specimen handling, Laboratory techniques and procedures, Diagnosis, Evaluation studies