Purpose of Review <p>Despite a reasonable initial workup, a significant proportion of pleural effusions (PE) remain undiagnosed. We review here the management of the undiagnosed PE with a focus on an evidence-based evaluation of persistent unexplained PEs.</p> Recent Findings <p>The pooled sensitivity of pleural fluid cytology for malignancy is low (58.2%) and varies considerably based on tumor type without significant increase upon re-testing. However, pleural tissue biopsy via thoracoscopy carries a much higher diagnostic yield for maligancy, particularly mesothelioma, with a low complication rate. Importantly, the combination of the thoracoscopic visual assessment of the pleura, pleural touch preparations, and the presence of pleural nodules on computed tomography carries a high predictive accuracy for malignant PE (94%). In tuberculosis-endemic countries, thoracoscopy may not be necessary since closed needle pleural biopsy’s diagnostic yield is high. Also, pleural fluid adenosine deaminase carries a good sensitivity (87.5%) and specificity (87.82%). More recently, anti-inflammatory therapy with intravenous dexamethasone for parapneumonic effusions did not improve serum inflammatory markers compared to placebo. The combination of intrapleural tissue plasminogen activator and DNase reduced the frequency of referral for surgical treatment.</p> Summary <p>Work up of the undiagnosed PE requires a thorough and systematic approach. Management relies on patient-centered outcomes. Artificial intelligence and machine learning may help provide a new venue in the future to differentiate malignant from non-malignant PEs.</p>