High variability in lymph node counts among an international cohort of pathologists: Questioning the scientific validity of node counts

Abstract

Background: The enumeration of lymph nodes (LNs) from surgical specimens plays a critical role in the staging of patients with cancer. LN count (LNC) can affect prognosis, staging, adequacy of resection, and/or eligibility for clinical trials. However, there is no standard method for counting LNs. Most studies in the literature site the pathology report as the source of LN data, without discussion of the counting criteria. Patients and Methods: Four microscopic slides from separate pelvic LN dissections were digitally scanned and uploaded with their gross descriptions to an online library and an online survey. Respondents were asked how many LNs they would count per slide as part of a staging procedure. The survey was distributed to an international cohort of pathologists. Results: A total of 122 surveys were returned: 79 from practicing pathologists and 43 from residents/fellows. There was no statistical difference between the groups. All slides showed significant individual variability. The LNC range for each slide was as follows: slide 1, 1-3; slide 2, 0-13; slide 3, 1-8; slide 4, 1-11. The intraclass correlation (ICC) for all responders was 0.26 (95% CI, 0.05