Try nucleic acid amplification methods along with sputum smear. Serological tests like IGRA are not reliable as in adults. Gastric lavage or good quality induced sputum should be the material. Beforehand nebulisation with salbutamol and 3%NaCl then chest physio may increase the yield of sputum. Do not go by clinical sense only. You will face MDR burden rapidly.
It is a combination of radiography of chest, gastric lavage for Mycobacterium tbc, and, if possible, flexible bronchoscopy with inspection, lavage and brush biopsy.
Diagnosis of pulmonary TB in kids is predominantly clinical & microbiologic confirmation occurs very infrequently (~20-50%). Culture is the reference standard, but clearly not a gold standard. PCR, usually via Gene Xpert, is WHO recommended. But a negative culture or Xpert should not delay initiation of ATT. Lots of good review articles on the challenges on pediatric TB diagnosis.