B80-4-34 Pixels Predict Positivity: Pleural Fluid Hounsfield Units and Cytology Yield in Malignant Effusions

Rationale Cytology sensitivity in malignant pleural effusions (MPE) is variable, and repeat sampling decisions are often empirical. Because pleural fluid attenuation on CT (Hounsfield units, HU) reflects cellular/proteinaceous content, we evaluated whether higher HU is associated with cytology positivity in MPE and estimated the probability of a positive tap across HU values. Methods We conducted a single-center retrospective study at Wyckoff Heights Medical Center, Brooklyn, NY) on all adults who underwent thoracentesis from January 2022 though December 2024 with pre-procedure CT chest and complete fluid/serum studies. We looked at all effusions that were identified as malignant. Among MPEs, cytology outcome was coded from “Cytology - Malignant.” Pleural fluid attenuation was measured as the mean of four intrafluid ROIs (HU average). Primary comparison of HU between cytology-positive vs cytology-negative MPE used Mann-Whitney U (pre-specified) with Welch’s t-test as sensitivity. Effect size was Hedges’ g. A logistic regression modeled cytology positivity as a function of HU (odds ratio [OR] per 1 HU), adjusting for age and sex when available. We plotted predicted probability of positivity across the observed HU range and summarized positivity by HU bands (≤8, 8-12, >12). Results Among malignant effusions with complete data (n = 35), 12 (34.3%) were cytology-positive and 23 (65.7%) cytology-negative. HU tended to be higher in cytology-positive MPE (mean 15.72±5.93; median 17.27) than cytology-negative MPE (mean 12.93±7.32; median 11.95); Mann-Whitney p = 0.079; Welch p = 0.235. The effect size was moderate (Hedges’ g = 0.40). In the logistic model, each +1 HU was associated with 6.5% higher odds of cytology positivity (OR 1.07; 95% CI 0.96-1.19; p = 0.25). Predicted probability increased across the HU range (figure provided). By HU bands, observed cytology positivity was higher when HU was >12 than when ≤8 (illustrative PPV bar chart provided), but confidence intervals were wide due to small sample size. Conclusions In this single-center cohort of malignant effusions, pleural fluid attenuation shows a trend toward higher HU in cytology-positive taps with a moderate effect size, but the association did not reach statistical significance and logistic modeling CIs crossed unity. Clinically, HU may serve as a supportive heuristic—very low HU suggests lower cytology yield, and higher HU suggests greater yield—but HU alone should not drive the decision to repeat or forego cytology. A pragmatic role is HU-informed triage alongside tumor burden, imaging, and prior cytology results. Larger, multi-center cohorts with tumor subtype/stage and time-to-processing data are needed to determine whether HU can meaningfully refine cytology strategies. This abstract is funded by: None