The majority of children and adolescents diagnosed with cancer will experience one or more episodes of fever or infection during their course of therapy. The most common microbiologically documented infection is bloodstream infection (BSI), which can be associated with severe sepsis or death. Current methods of diagnosis require a significant load of live bacteria in the blood making early detection difficult. Delayed diagnosis and delayed optimal therapy of BSIs are associated with increased morbidity and mortality. This study seeks to identify whether next generation sequencing (NGS) of pathogens can identify patients with impending bloodstream infection. This would enable preemptive targeted therapy to replace antibacterial prophylaxis which often leads ot high-density broad-spectrum antibiotic exposure and contributes to subsequent development of antibiotic resistance. PRIMARY OBJECTIVE: * To estimate the sensitivity and specificity of next generation pathogen sequencing for prediction of bloodstream infection in children with cancer at high risk of infection.