The purpose is to determine the optimal pleural fluid volume needed to diagnose malignancy using pleural fluid cytology. Median survival rates after malignant pleural effusion (MPE) diagnosis are approximately 4-6 months, highlighting the importance of early diagnosis to guide therapy. Past studies suggest a wide range of volumes, from 10-300mL, as sufficient to make a diagnosis. However, the sensitivity for MPE diagnosis are low. Pleural fluid cells are heterogeneously distributed within the pleural cavity and larger volume samples may be more likely to recover malignant cells. There is no literature comparing the diagnostic yield of high volume pleural fluid samples (>1000mL) to low volume samples. Retrospective analysis of 233 thoracenteses from 2013 to 2017 was performed. Inclusion criteria were cases with known or suspected MPE with fluid sent for cytologic analysis by combined direct smear/cytospin and cell block preparations. Demographic, clinical, and procedural data were obtained from review of the medical records. Fluid samples were defined as large volume if greater than 1000mL and low volume if between 1mL and 1000mL. Statistical significance of outcomes between these volumes was measured using Pearson’s chi-squared test. 109 thoracentesis samples from 96 (42 female and 54 male) patients were sent for cytologic analysis in cases of known or suspected malignancy. 26 cases (24%) were positive for malignancy. Average volume for large volume samples was 1200mL ± 245 (range, 1000 to 1750mL) and for low volume samples was 63mL ± 135 (range, 2 to 800mL). 64% of large volume samples were positive for malignancy compared to 19% positivity in low volume samples. The rate of cytological detection of malignancy was significantly higher with large volume thoracentesis (p = 0.001). Large volume thoracentesis analysis allows for a significant improvement in cytological detection of malignancy when compared to low volume samples. Often, in practice a small sample is obtained for analysis and the remaining extracted fluid is discarded. This study emphasizes the need to implement analysis of larger volume samples (>1000mL) to improve the diagnostic yield for MPE.