Pleural effusion is a common finding, and exudative fluid raises the question of underlying malignancy; as many as 20% of cancers of thoracic origin present with pleural effusion, and malignant effusion is the second most common cause of exudative pleural effusions.1,2 In the in-hospital setting, 25% of all exudative pleural effusions are secondary to malignancy.3,4 When patients present with a potentially malignant effusion coupled with findings suspect for malignancy, a traditional approach has been to sample the fluid and then base subsequent work-up (retap, biopsy other areas, pleural biopsy, pleuroscopy) upon the results of cytologic evaluation of the initial fluid sample.2,5,6
The advent of an active interventional pulmonary group at our institution and of tools such as endobronchial and esophageal ultrasound (EBUS and EUS) has led to an alternative approach to the scenario of pleural effusion with suspect radiologic findings; to biopsy, when possible, all approachable suspect lesions at time of thoracentesis. The logic for this approach can be explained with the example of a patient with lung mass, adenopathy, and pleural effusion. It is true that if the pleural fluid comes back positive for malignancy, then the M1 status defined by that positivity dominates prognosis.7 The addition of EBUS/EUS nodal biopsy nevertheless confers several potential benefits: (1) A negative pleural cytology does not rule out malignancy or malignant pleural involvement. Pleural cytology has a reported sensitivity between 40% and 87% even when the suspicion for malignancy is high.2–4 A false negative in particular could delay diagnosis and care. (2) We have at our institution rapid on-site cytologic evaluation of needle aspiration biopsies, and the presence of malignancy can thus commonly be confirmed on site on the day of the biopsy, avoiding the delay involved in processing of pleural fluid and any diagnostic uncertaintly. (3) Positive (or negative) nodal biopsies contribute the N status in the TNM system. (4) With the advent of immunotherapy and mutation-specific therapies, it is now increasingly important to generate large cell blocks that are adequate for molecular testing.8 When positive, needle aspiration biopsies in our hands and those of some others have approximately 99% adequacy for molecular testing.9,10 Although newly developing techniques promise increased sensitivity and specificity for analysis of pleural fluid,11 one can at present expect that pleural fluid will be inadequate for testing in approximately 30% of cases.12,13
At our institution, IP practices as a subset of a larger pulmonary consulting service, and our database contains both (1) patients treated with the traditional “sequential approach,” with thoracentesis alone as the initial diagnostic procedure and (2) patients treated with the “concurrent approach,” in which patients with more than one suspect anatomic finding undergo two or more diagnostic procedures simultaneously. This dichotomy at our institution allowed us to review the database and perform an objective evaluation of the impact of the sequential versus the concurrent approach upon clinical course and outcomes. Our hypothesis was that the concurrent approach would lead to increased efficiency and decreased time to treatment.
PATIENTS AND METHODS
After Institutional Review Board approval at the University of Arkansas for Medical Sciences, a retrospective chart review was performed on all pleural fluid cytology studies performed at the University of Arkansas for Medical Sciences from May, 2014 to February, 2017. The University of Arkansas for Medical Sciences is a 437-bed tertiary care teaching hospital. It is the primary teaching hospital for the medical school and, as the only academic hospital in the state, has a broad referral base.
The study group was defined as all patients with previously undiagnosed exudative effusion, radiologic findings suspect for malignancy, and no prior diagnosis of primary lung malignancy. Both inpatients and outpatients were eligible for inclusion. Examples of radiologic findings suspect for malignancy included a chest mass, hilar/mediastinal/cervical adenopathy, evidence for distant metastasis, and pleural studding in addition to pleural fluid. We first identified 565 patients with exudative pleural fluid (per Light’s criteria)14 sampled during the 34-month study period. We then excluded 329 patients who did not have additional radiologic findings suspect for malignancy. An additional 191 patients were excluded because of a prior diagnosis of primary thoracic malignancy. The 45 patients that remained after these exclusions formed the study population. The sequential group was defined as patients for whom pleural fluid cytology was first obtained, with subsequent decisions based upon pleural fluid cytology results. The concurrent group was defined as those for whom additional suspect areas were biopsied on the same day as pleural fluid sampling. Some patients with pleural findings highly suspect for malignancy went directly to pleuroscopy in order to allow not only fluid sampling but simultaneous pleural biopsies and intervention and were included in the concurrent group.
Patient records were reviewed for demographic information, smoking history, family history of malignancy, clinical presentation, chest computed tomographic findings, pleural fluid results, pathology reports, and procedural notes. Charts were also reviewed for time to diagnosis, time to molecular studies (if any), time to consultation of hematology/oncology, time to therapeutic intervention (if any), and time to death (if available).
The SPSS statistics software (IBM Corp., Armonk, NY) package was used to perform data analysis. The χ2 test or Fisher exact test was used as appropriate to compare nominal variables between the 2 groups. Since the data were not normally distributed, the Mann-Whitney test was used to compare the medians of continuous variables between the 2 groups.15
Seventeen of the 45 study patients (38%) had undergone sequential testing and 28 (62%) had undergone concurrent testing. Table 1 summarizes the baseline clinical and demographic characteristics of the 2 groups. The mean age was 64.5 years, 80% of the patient population was white, and 71% was male. On comparing baseline demographic and clinical characteristics between the 2 groups, patients in the sequential group were more likely to be male (88% vs. 61%).
Tables 2 and 3 summarize presentation, biopsies performed, and findings and will not be reiterated in writing. Staging designations are per the most recent staging system, with malignant pleural effusion listed as stage IV disease. There were no significant differences between study groups in radiologic findings or method of biopsy. Combining both groups, a total of 41 biopsies were performed, 36 (80%) by interventional pulmonologists and 5 (20%) by interventional radiology. (At our institution, EUS-FNA for the diagnosis of thoracic diseases is not a gastroenterology procedure; it is routinely performed by pulmonology using the convex curvilinear ultrasound bronchoscope.) Five of the 45 study patients (11%) were ultimately found not to have cancer. For those with cancer, there was no significant difference between the 2 groups with respect to tissue histology or cancer stage. The patients tolerated the procedures well, with only 1 case of pneumothorax after thoracentesis in the sequential group that resolved without intervention.
Of patients identified with cancer, 6 were lost to follow-up, 5 (29%) in the sequential and 1 (4%) in the concurrent group (P=0.023). Of the 5 lost to follow-up in the sequential group, one was lost to follow-up before he could be informed of his diagnosis and the remaining 4 were lost before a therapeutic plan had been formulated.
Sequential timing for all patients is documented on Table 4, with time zero considered to be the initial evaluation of CT findings. The median lag time to thoracentesis was significantly shorter for the sequential group (1 d) than for the concurrent group (3 d) (P=0.002). The median lag time from discovery of pleural effusion to biopsy was significantly shorter for the concurrent group (the same 3 d) than for the sequential group (9 d; P=0.006). The median time from day of thoracentesis to diagnosis was significantly shorter for the concurrent group (2 d) than for the sequential group (6.5 d; P<0.001). (Although on-site analysis often confirmed a diagnosis of malignancy on the day of biopsy, this statistical analysis did not include on-site diagnoses; only final cytology reports were counted.) For those with cancer, the median lag time to consult of the hematology-oncology service was significantly shorter for the concurrent group (7 d) than for the sequential group (16 d, P=0.039). Molecular study results were available in a median of 13.1 days for concurrent group versus 31.5 days for sequential approach (P<0.0001). For those patients whose diagnosis led to an intervention such as radiation or surgery, there was no statistically significant difference between groups in the median lag time to intervention (13 d for the concurrent group vs. 20 d for the sequential group, P=0.204). For those with available data, there was no significant difference in the median time to death between the 2 groups (66 d for the sequential group vs. 65 d for the concurrent group, P=0.407).
Reports and guidelines have recommended that an initial thoracentesis be followed by invasive investigations for tissue sampling if nondiagnostic.2,5,6 These guidelines were formulated before the advances in interventional pulmonary noted above. Increased expertise and availability of tools/procedures such as regular ultrasound, EBUS, EUS, pleuroscopy, and radial ultrasound have markedly increased our capacity to access thoracic abnormalities. Advances in therapy have led to a need for cell blocks large enough to allow subtyping for actionable biomarkers. We hypothesized that if a patient with undiagnosed pleural effusion has additional findings suggestive of malignancy that are amenable to diagnostic sampling, then concomitant diagnostic sampling would increase efficiency and accelerate care for those with cancer. Although we suspect that there are other centers that sample concurrently, to our knowledge this is the only study that has attempted to compare concurrent and sequential approaches. This retrospective study supports the hypothesis with objective data; for patients who underwent concurrent thoracentesis and biopsy, there was increased patient retention, decreased time to diagnosis, decreased time to availability of results of molecular testing, and decreased time to oncologic consultation. Mortality was not impacted.
This study has limitations. It is a retrospective chart review study, which makes it subject to observer bias and inaccuracy of chart documentation and data collection. There is the possibility that the groups were not completely similar in their clinical characteristics and presentation, a potential unrecognized factor in the differences between groups. The study was performed at a single academic center, which could limit its generalizability. The concurrent approach was initiated by interventional pulmonary, and most of the concurrent procedures were performed by interventional pulmonologists; at some institutions the concurrent approach may be less feasible. For those with cancer, our pleural cytology yield was 40% (at the lower end of the reported spectrum of positivity). This could have impacted our final results. Eighty percent of the study population was white, which may make it less generalizable to populations with different racial proportions. Finally, the sample size is small and does decrease the power of the study. (In contrast, if there were only a small difference between the 2 groups the small number of study patients would have been inadequate for detection of statistical significance.)
There have been dramatic increases both in the capacity of pulmonologists to access and sample thoracic abnormalities and in the need for molecular testing. We hypothesized that when a patient with a pleural effusion of unknown etiology has other findings suspect for malignancy that are amenable to biopsy, then concomitant thoracentesis and biopsy should be performed. The clinical data from this study support that hypothesis. We would recommend that guidelines for undiagnosed pleural effusion be tailored to reflect the changing landscape of cancer diagnosis.
1. Martin Diaz E, Arnau Obrer A, Martorell Cebollada M, et al. Thoracocentesis for the assessment of lung cancer
with pleural effusion
. Arch Bronconeumol. 2002;38:479–484.
2. Light RW. Pleural effusions. Med Clin North Am. 2011;95:1055–1070.
3. Bartter T, Santarelli R, Akers SM, et al. The evaluation of pleural effusion
. Chest. 1994;106:1209–1214.
4. Leff A, Hopewell PC, Costello J. Pleural effusion
from malignancy. Ann Intern Med. 1978;88:532–537.
5. Marel M, Štastny B, Melinová L, et al. Diagnosis of pleural effusions. Chest. 1995;107:1598–1603.
6. Hooper C, Lee YCG, Maskell N. Investigation of a unilateral pleural effusion
in adults: British Thoracic Society pleural disease guideline 2010. Thorax. 2010;65 (suppl 2):ii4–ii17.
7. Detterbeck FC, Boffa DJ, Kim AW, et al. The eighth edition lung cancer
stage classification. Chest. 2017;151:193–203.
8. Cagle PT, Raparia K, Portier BP. Emerging Biomarkers in Personalized Therapy of Lung Cancer
. Springer, Cham. 2016:25–36.
9. Jeyabalan A, Bhatt N, Plummeridge M, et al. Adequacy of endobronchial ultrasound‑guided transbronchial needle aspiration samples processed as histopathological samples for genetic mutation analysis in lung adenocarcinoma. Mol Clin Oncol. 2015;4:119–125.
10. Hasan Z, Fedda F, Meena N, et al. On-site adequacy in pulmonologist-performed EUS-B-FNA of subdiaphragmatic lesions—an institutional experience. J Am Soc Cytopathol. 2017;6:S40–S41.
11. Liu X, Lu Y, Zhu G, et al. The diagnostic accuracy of pleural effusion
and plasma samples versus tumour tissue for detection of EGFR mutation in patients with advanced non-small cell lung cancer
: comparison of methodologies. J Clin Pathol. 2013;66:1065–1069.
12. Hassan H, Lourdes Y, Samjot D, et al. Adequacy of malignant pleural effusion
for epidermal growth factor receptor mutation analysis using the pyrosequencing method. Pleura. 2015;2:1–6.
13. Demaio A, Dash R, Sebastian S. JC-AJRC, 2017 U. Comprehensive Genetic Mutation Analysis On Pleural Effusions Related To Lung Adenocarcinoma. Am Thorac Soc
. 2017;195:A7554. Available at: http://www.atsjournals.org/doi/pdf/10.1164/ajrccm-conference.2017.195.1_MeetingAbstracts.A7554
. Accessed March 1, 2018.
14. Light RW, Macgregor MI, Luchsinger PC, et al. Pleural effusions: the diagnostic separation of transudates and exudates. Ann Intern Med [Internet]. 1972;77:507–513; Available at: http://www.ncbi.nlm.nih.gov/pubmed/4642731
. Accessed March 3, 2018.
15. Mann HB, Whitney DR. On a Test of Whether one of Two Random Variables is Stochastically Larger than the Other. Vol. 18, The Annals of Mathematical Statistics. Durham, NC: Institute of Mathematical Statistics; 1947:50–60.