Background: Prophylactic central neck dissection (PCND) for papillary thyroid cancer (PTC) is controversial. Recent publications suggest that the number and size of nodes and the presence of extranodal extension (ENE) are important features for risk stratification of lymph node metastases. We analyzed these features in clinically unapparent nodes that would not otherwise be removed. We also investigated the impact of surgeon experience on the ability to detect metastatic lymph nodes intraoperatively. Methods: Forty-seven patients with well-differentiated PTC, with no preoperative evidence of central metastases, were included in this study. Intraoperatively, clinically apparent disease was determined by inspection and palpation by the senior surgeon and a fellow/senior resident, and recorded in a blinded fashion. Rate of occult metastases based on intraoperative evaluation were tabulated for each group of surgeons. Histopathologic features of occult nodes were analyzed to determine what clinicians would be missing by foregoing a PCND, and how that would have impacted the patient management. Results: The rate of occult metastases, based on senior surgeon assessment, was 26%, and did not differ significantly from fellow/senior resident assessment. The level of agreement between these two surgeon groups was moderate (k=0.665). Analysis of the false negative cases revealed that the size of the largest undetected node ranged from 0.1 to 1.3cm; 36% of patients with occult metastases demonstrated five or more positive nodes, and 27% showed ENE. Discussion: Clinical assessment based on intraoperative inspection and palpation had poor sensitivity and specificity in identifying metastatic central nodes, regardless of the level of experience of the surgeon. There was moderate agreement between surgeons of different experience levels. Sensitivity improved significantly with larger size of positive nodes, but not with the presence of multiple positive nodes or presence of ENE. In foregoing PCND in this patient population, our results suggest that treating clinicians miss potentially virulent disease with a large number of occult positive central nodes and occult nodes with ENE. This is the first report to address the pathologic features of clinically nonevident central nodes showing a high incidence of clinically relevant, adverse histologic features, as well as the impact of surgeon experience in performing the important intraoperative determination of whether there are clinically evident nodes that require removal.