The present study included 101 cancer patients with symptoms andor radiological manifestations consistent with tuberculosis: Fourty threebronchogenic carcinoma, 37 hematologic malignancies and 21 solidtumors. All samples were subjected to Ziehl Neelsen (ZN) staining,culture on Lowenstein Jensen (LJ) and Mycobacterium growthindicator tube (MGIT) as well as molecular detection byMycobacterium tuberculosis Direct test( MTD) test. Only those samplesshowing positive culture or at least two positive tests were consideredtrue mycobacterial disease. Tuberculosis was detected in 32/101(31.6%) of patients with cancer. Sixteen out of 101(15.8%) cancerpatients had past history of TB and antituberculous treatment. Of these,only 6(37.5%) were found to have TB (reactivation) in association withpresent malignant condition. The overall sensitivity, specificity, PPVand NPV of ZN staining were 46.8%, 100%, 100%, 80.2%;respectively. The overall sensitivity, specificity, PPV and NPV of LJwere 62.5%, 100%, 100%, 85.1%; respectively. The overall sensitivity,specificity, PPV and NPV of MGIT were 56.2%, 100%, 100%, 84.1%;respectively. The overall sensitivity, specificity, PPV and NPV of MTDwere 56.2%, 82.6%, 60%, 80.2%; respectively. Combined sensitivity ofany two tests was 94%. Sputum was better than BAL for the recoveryof mycobacteria. Twelve cases were only MTD positive, wereconsidered negative but presumptive. It was concluded that:tuberculosis due to Mycobacterium tuberculosis and othermycobacteria should always be considered in the diagnostic approachto imunocompromised cancer patients. Utilization of combination of tests together with consideration of key clinical characteristics, couldimprove diagnostic accuracy.