Most of the studies that address the use of NIV in cancer patients focus on patients with hematological malignancies (Table 67.2). These studies have shown improved outcome of patients with ARF managed by NIV compared with invasive mechanical ventilation. In one study of 1,302 patients with hematologic malignancies admitted to the ICU with ARF, NIV was attempted in 21 % of these patients and 46 % of them later required invasive mechanical ventilation [7]. Favorable outcomes were encountered in NIV patients, with mortality of 46 % compared with 69 % in patients who were initially put on invasive mechanical ventilation and 77 % in patients who failed NIV and later required.

Studies that specifically evaluate the outcome of patients with solid malignancies and ARF treated by NIV are sparse. Two major studies have reported on the use of NIV in patients with solid malignancies. The most recent study by Azoulay et al. [21] in 2014 was a retrospective analysis of data from six previously published retrospective and prospective studies on patients with malignancy admitted to the ICU in 14 university and university-affiliated hospitals in France and Belgium. In this review, there were 147 (14.6 %) patients with solid malignancy, in whom breast and lung cancers were the most commonly encountered. NIV was used in 387 patients, among whom 174 survived and 213 did not survive, with a 64 % overall mortality. Among the 387 patients initiated on NIV, 276 subsequently required mechanical ventilation. NIV failure was strongly associated with the severity of ARDS, with patients failing in the moderate to severe categories by the Berlin definition. Following a multivariate analysis, patients with solid tumors were independently associated with lower mortality, OR 0.51 (0.34–0.77) p = 0.002. As illustrated in Fig. 67.1, mortality from solid tumors also decreased over time compared with hematologic malignancies.

The other major study that reported on the use of NIV in solid malignancies was a prospective study by Azevedo et al. [22], which was a multicenter study carried out in 28 ICUs in Brazil. In their patient population, 227 (86 %) patients had solid malignancies and only 36 (14 %) patients had hematologic malignancies. The most frequently encountered malignancies were lung, breast, and lower gastrointestinal cancers. NIV was initially used in 85 patients (32.3 %), and 45 (47.9 %) of these patients required mechanical ventilation. Patients who ended up only using NIV had a mortality of 40 %, whereas those who required mechanical ventilation had a mortality rate of 68.9 %. Reasons for NIV failure were not evaluated in this study. Multivariate analysis to identify independent risk factors for mortality did not reveal that the type of tumor was a risk factor, but significant risk factors for mortality included medical admission, cancer status, tumor as a reason for ventilator support, poor performance status, NIV followed by mechanical ventilation, use of mechanical ventilation only, and higher SOFA (sequential organ failure assessment) scores.

Major factors found to be associated with NIV failure following multivariate analysis include respiratory rate > 20 breaths per minute, organ failure, particularly renal insufficiency requiring renal replacement therapy, hemodynamic instability requiring vasopressors, ARDS, PaO₂/FiO₂ < 146 1 h after NIV initiation, persistent organ failure over the first few ICU days, pneumonia, pH <7.25, excessive air leak, lack of tolerance, agitation during NIV, high severity of illness scores (APACHE II ≥29 or SAPS II ≥35), and longer duration of NIV dependency [5, 13, 19–21, 23] (Table 67.3). The above predictors should serve as a guideline when making decisions on the initial respiratory support for patients with hematologic malignancy in ARF.

It is also imperative to understand the characteristics of patients with malignancies who have had favorable outcomes with the use of NIV. Following multivariate logistic regression, patient factors that have been associated with favorable outcomes include good performance status, especially if patient is ambulatory, age less than 40, no recurrence of cancer, and absence of airway involvement [13]. Patients with solid tumors were better candidates than those with hematologic malignancy (OR 0.51 (0.34–0.77) p = 0.002). Patients with ARDS from primary malignancy had better outcomes than those with ARDS from secondary or undermined causes (OR 0.41 (0.2–0.88) p = 0.02) [21, 22].