What is the minimum sample size, regardless of total number of population, that can be considered as reliable for performing advanced statistical tests?

I have 32 patients with tuberculous meningitis, if I determined the sensitivity and specificity of cerebrospinal fluid adenosine deaminase and I got significant results, does it mean the result can be generalized? Can I determine the predictors of mortality from this sample?