Pulmonary complications are common events that result in significant morbidity and mortality for cancer patients of all ages (1–4). The presence of pulmonary infiltrates in cancer patients presents a diagnostic dilemma. The differential diagnosis includes malignant disease, inflammatory response, and infectious disease, the last of which could have a bacterial, viral, or fungal etiology. Noninvasive imaging is of limited utility in differentiating among these diagnoses. Thus, clinicians must often choose among invasive procedures, including flexible bronchoscopy (FB) with bronchoalveolar lavage (BAL), endobronchial ultrasound (EBUS) (5), radial EBUS, robotic bronchoscopy and computed tomography (CT) guided navigation (6), needle aspiration, and open lung biopsy, to yield a diagnosis. Of these options, FB with BAL is uniquely useful as a diagnostic and therapeutic lung examination procedure. Moreover, the BAL fluid obtained during FB provides a sample of the alveolar microenvironment, reflecting the greater respiratory system. This fluid may yield microbes such as bacteria, inflammatory cells such as macrophages, or malignant cells such as leukemic cells, which, if diagnostic, can guide the management of cancer patients with pulmonary infiltrates (7, 8).

In the present study, our primary objective was to describe the diagnostic yield of FB with BAL in pediatric (9) cancer patients. Our secondary objectives were to identify patient demographics that influence the diagnostic yield; determine the safety of FB and describe its short-, medium-, and long-term outcomes; and describe the rate of patients in whom FB results changed clinical management.

In the current study, we describe the utility of FB with BAL among 196 pediatric cancer patients. In this, single-center review, we found that an increased total fluid volume instilled during BAL was significantly associated with increased probabilities of positive BAL culture, positive bacterial BAL culture, and positive viral BAL culture in pediatric cancer patients. Furthermore, we found that older patients were more likely to receive a large BAL instillation volume and thus had increased probabilities of a positive BAL culture and a positive BAL culture with multiple organisms. Consensus on fluid instillation, acquisition and processing in children was lacking prior to Ratjen’s (17) and Riedler’s studies (18). In 2000, The European Respiratory Society Task Force on bronchoalveolar lavage in children described a variety of protocols for BAL fluid instillation in pediatric patients (19). One approach is to instill 3 mL/kg divided into three equal aliquots for children weighing less than 20 kg and to instill 3 mL/kg in 20-mL portions for children weighing more than 20 kg (18, 20). Another approach is to adjust the BAL instillation volume according to the child’s functional residual capacity (21). Others use the method applied in adults, which is to instill 2-4 aliquots of the same volume (generally 10-20 mL), regardless of the patient’s body weight and age (22, 23). In our study, children had BAL with 3mL/kg in 10-20mL aliquots while adolescents had 2-5 aliquots of the same volume (20 mL) following best practice guidelines as referenced.

We also found a trend for both neutropenia and monocytopenia to be associated with an increased probability of positive fungal BAL culture. This is consistent with the findings of a prospective multicenter study by Villarroel et al., who found that an ANC ≤500 neutrophils/mL and AMC ≤100 monocytes/mL were each independently associated with an increased probability of invasive fungal disease (24).

In addition, we also found that the type of HSCT was associated with the probability of a positive BAL viral culture; among HSCT patients, those who underwent allogeneic unrelated-donor HSCT were the most likely to have a positive culture. This association could be explained in part by the myeloablative conditioning regimens patients receive in preparation for HSCT, whereas the predominance of allogeneic transplants may be explained by the use of immunomodulators to prevent graft-versus-host disease, each resulting in a relatively greater degree of immunosuppression (25). Our study also showed that longer post-HSCT duration was significantly associated with increased probabilities of both positive BAL culture and positive BAL culture with multiple organisms. This may be explained by the fact that after engraftment, B-cell and CD4 T-cell recover slowly; however, HSCT patients are still at risk for viral and fungal infections as well as infections caused by encapsulated bacteria (26). In addition, our findings support those of prior studies showing that BAL results influence treatment decisions (1, 27). Antimicrobial treatment was altered after more than half (69.1%) of the FB cases in the present study, with treatment narrowed in 27% of cases and treatment expanded in 42.1% of cases. In the cases where antibiotics were expanded, the most common class switch were done in patients with positive BAL cultures, worsening respiratory symptomatology or as a result of overall clinical deterioration. These included class switches from cephalosporins to carbapenems, addition of gram-positive coverage, anaerobic, antiviral, PJP and or atypical organisms’ coverage respectively. In cases where antibiotics were narrowed, the most common class switches were from carbapenems to cephalosporins or as a result of de-escalation of antiviral, PJP, antifungal coverage for patients with negative cultures, improved clinical condition or recommendation from the infectious disease consultant. Although we did not report sensitivities in the study, the data we report reflects subset of oncology and stem cell transplant patients who developed infections following standard prophylaxis strategies. While antimicrobial prophylaxis and therapies against bacterial, viral and fungal infections have improved over time, pediatric cancer and stem cell transplant patients, and perhaps, those in particular with lower respiratory tract infections may not exhibit robust responses to anti-microbial therapies in the absence of host immune defense systems (26, 28, 29).

Our findings did not demonstrate a relationship between timing of FB and diagnostic yield for infectious pathogens or between timing of FB and FB complications. This may be due to the fact that in 88.3% FBs, our subjects were receiving empiric antibiotics at the time of the FB. This is contrary to others findings, which found that as the duration of empiric antimicrobials increased, the BAL yield for positive cultures decreased (30–35).

The complication rate of FB in our study was 9.8%, with the most common complication being hypoxia (5.3%). This complication rate is lower than that described by Efrati et al. (30.6%) (36), but higher than in other studies (7, 37). A proposed pathophysiologic mechanism is that as the pathologic process involving the lung evolves, the parenchyma may become more inflamed, thereby increasing the propensity for complications such as bleeding and pneumothorax (31, 32).

Furthermore, in our study, the 3-day survival rate (99.0%) was higher than the 28-day and 180 day survival rates (86.4% and 60.1%, respectively), which suggests that patients were more likely to die from their disease than from complications related to FB.

Prior studies have also shown that FB with BAL is useful in the evaluation of pulmonary infiltrates in pediatric cancer patients. In a prospective trial of 14 pediatric patients with histories as varied as lymphoid and solid organ malignancies, solid organ transplant, and systemic lupus erythematosus, Pattishall et al. found that FB had a diagnostic yield of 71%, with the most common organisms being PJP, cytomegalovirus (CMV), Candida, and Aspergillus (27). In a retrospective analysis, Efrati et al. compared a cohort of children with cancer or primary immune deficiency (85% of whom previously received broad-spectrum antibiotics) with a control cohort of patients without a malignant diagnosis who underwent FB with BAL during the same period. A diagnosis was obtained in more than 80% of patients in both groups, and a definitive organism was detected in 53% of the patients in the cancer cohort and 18% of the patients in the control group, resulting in treatment alteration rates of 33% and 21%, respectively. Approximately 30% of the patients in the cancer cohort experienced complications, most of which were related to minor desaturations, whereas only 15% of the patients in the control group experienced complications. One cancer patient required mechanical ventilation following the procedure (36). In another retrospective study of FB with BAL in 33 patients with leukemia and pulmonary infiltrates, tracheobronchitis was identified in 51% of patients, endobronchial hypersecretion in 18%, and pulmonary hemorrhage in 21%. BAL cultures were positive in 63% of patients, revealing Candida in 39% of patients and Aspergillus in 9% of patients. There was a trend for early bronchoscopy to be associated with improved overall survival (35). In a recent review of 64 children diagnosed with leukemia, FB with BAL revealed that 56% of patients had positive microbiology results, including bacterial, viral, and fungal infections. Antimicrobial coverage was changed in more than 83% of patients. Approximately 38% of patients who underwent BAL had complications, the most common of which was hypoxia (38). Prior studies report changes in antimicrobial management after BAL (37–39) but won’t comment on prognosis and survival except for three studies in which positive and negative BAL results prompted antimicrobial changes (1, 40) but failed to show improvement in survival (36). Most of these studies, whose numbers of pediatric cancer patients ranged from 14 to 117, were relatively small compared to adult cohort studies (1, 35, 37, 41).

Our study had limitations. First, because this study was an exploratory, retrospective, single-center review, its findings may reflect technical aspects of the performance of FB and/or laboratory analysis that are unique to our institution. Second, while the data obtained via FB led to alteration of antimicrobial management in most cases, it is not clear whether those alterations impacted patient outcomes. The retrospective design of our study did not allow for this analysis. Lastly, the statistical analyses involved multiple testing of associations among the many variables, and no adjustment was made for multiple testing. A larger, multicenter prospective observational trial would attenuate institutional clinical practice influences and have the statistical power necessary to definitively show significant associations.

While acknowledging these limitations, our study also has several strengths. To our knowledge, this is the largest study of BAL in the pediatric oncology and HSCT population and the only BAL study in pediatric cancer patients that reports on total fluid volume instillation having an impact on BAL culture results. Additionally, our observations regarding BAL instillation volume and BAL culture positivity may help inform clinicians’ practice while performing the BAL procedure. Furthermore, the fact that BAL results altered antimicrobial treatment in more than half of the FB cases we observed also informs clinicians of the diagnostic utility of BAL when presented with pneumonias of unclear etiology in pediatric oncology and HSCT patients.

In summary, our study suggests that in pediatric oncology and HSCT patients, FB is safe, provides diagnostic and/or therapeutic benefits, and has implications for treatment decisions. Larger multicenter study evaluating the role of FB in pediatric cancer patients may further elucidate its utility for clinical care in this population.

The raw data supporting the conclusions of this article will be made available by the authors, upon request, without undue reservation.

The studies involving human participants were reviewed and approved by University of Texas MD Anderson Cancer Center Institutional Review Board. Written informed consent from the participants’ legal guardian/next of kin was not required to participate in this study in accordance with the national legislation and the institutional requirements.

RM: Fiberoptic bronchoscopies, study conception, data acquisition, analysis, and manuscript writing. AA: Fiberoptic bronchoscopy assistance, data acquisition, analysis, and manuscript writing. BB: Data acquisition, analysis, and manuscript writing. CA: Statistical modeling. JC, SR, MG, and LE: Fiberoptic bronchoscopy assistance. DP and KM: Analysis. All authors contributed to the article and approved the submitted version.

This study was supported in part by the NIH through MD Anderson’s Cancer Center Support Grant (CA16672).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.