Abstract
Acute respiratory failure (ARF) is a leading cause of unexpected admissions to the ICU in adults with active solid or haematologic cancer. Early trials highlighted some benefits of non-invasive ventilation (NIV) delivered with supplemental oxygen versus conventional oxygen. Whether this is still maintained in the high-flow nasal cannula (HFNC) era is doubtful. Therefore, a Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA)-guided, scenario-based systematic review was conducted to evaluate the effectiveness and harms of NIV versus oxygen/HFNC and to identify predictors and outcomes of NIV failure in adult cancer patients with ARF. Searches on PubMed from inception to 31 August 2025, Scopus, and the Cochrane Library from 2010 to 31 August 2025 were completed. In this review of trials and studies, primary outcomes were NIV failure, defined as intubation during the index episode, and in-hospital mortality. Secondary outcomes were length of stay, complications, and predictors of NIV failure. Two reviewers screened, extracted, and assessed risk of bias using Cochrane Risk of Bias 2 (RoB 2) and the Newcastle-Ottawa Scale. We did not perform a formal meta-analysis, and studies were compared qualitatively. Twenty-six studies met the inclusion criteria, including four randomised trials and 22 cohorts, with a pooled sample of approximately 12,000 patients. In de novo hypoxemic respiratory failure, the largest recent trial and adjusted cohorts showed no decline in intubation or mortality rates with early NIV versus oxygen or high-flow nasal cannula. In cancer-related acute respiratory distress syndrome (ARDS), a form of ARF, NIV failure was common at about 60% to 80% and greatly linked to death. In the case of NIV failure, independent predictors of failure included shock, lower arterial oxygen partial pressure to inspired oxygen fraction ratio (PaO₂/FiO₂), invasive fungal infection, higher severity scores, and undetermined aetiology. However, in cases of cancer and acute respiratory failure with cardiac dysfunction, observational data suggested lower ICU mortality with early NIV. Limitations of this review include heterogeneity, residual confounding, and variable protocols. To conclude, benefits of NIV in cancer-related ARF are context dependent. There is no clear advantage of NIV over HFNC/oxygen in de novo hypoxemia. In addition, high failure risk in ARDS with NIV mandates vigilant escalation. A context-specific NIV approach rather than a blanket one is highly recommended in the HFNC era.