The physician factor and anatomical site in 8846 consecutive mediastinal lymph node aspirations in a cross-sectional study

Abstract

Mediastinal lymph node fine needle aspiration (MLN-FNA) is a common procedure; however, the physician factor in pathological category, and anatomical site are not routinely assessed. Cytology reports for endobronchial ultrasound (EBUS)/endoscopic ultrasound (EUS) MLN-FNA specimens (8846) were retrieved for July 2012–Dec 2019, classified by hierarchical free text string match algorithm into 51 diagnostic categories, four mutually exclusive diagnostic groups (benign |suspicious |malignant |insufficient), and 24 anatomical sites. Pathologist and submitting physician/surgeon bias were assessed using logistic regression and funnel plots|control charts centered on the group median (diagnostic/capture) rate. Eleven pathologists and seven submitting physician/surgeon were involved in more than 250 specimens each. Overall, the MLN-FNAs were benign|suspicious|malignant|insufficient in 46%|4%|25%|24% of specimens. Percent malignant (number of samples) varied by station; 7| 4R| 4L| 2R| 10R| 11R| 11L were respectively 21%(3,101), 27%(2,453), 19%(1,289), 41%(435), 27%(497), 24%(357), 26%(229). The number of outlier (P < 0.05/P < 0.001) pathologists of 11 from the group median rate for benign|suspicious|malignant|insufficient was 0/0| 3/1| 0/0| 3/0 respectively. The outlier (P < 0.05/P < 0.001) submitting physicians/surgeons of 7 for benign|suspicious|malignant|insufficient was 3/2| 2/2| 3/2| 3/2 respectively. The physician and anatomical site are significant predictors of MLN-FNA pathology.