Abstract

In lung carcinomas, the most important factor predicting the prognosis is the stage of the tumour. Infiltration of lymph nodes by tumour cells is important in staging and histopathological examination gives the real stage.

In this study, the relation between lymph node size detected on postoperative specimens and histopathological type of the tumour and involvement of lymph nodes by tumour cells were examined. A retrospective evaluation was made on 261 cases with NSLSC to whom medi - astinal curettage or mediastinal sampling was applied. Totally 867 nodes detected on N1-N2 locations, 137 with and 730 without tumour infiltration. A statistically significant relation was found between tumour infiltration and long and short axis of lymph nodes. For hilar nodes when the criterion was 10 mm for long axis of lymph nodes, sensitivity was 97.74% and specificity 26.0%, when it was 15 mm 78.9% and 64.20% and when it was 20 mm 36.84% and 79.0% respectively. When the criterion was 10 mm for the short axis, the sensitivity was 36.36, and the specificity was 87.8%; for 15 mm 27.0% and 91.0% respectively. For mediastinal lymph nodes, if the criterion is 10 mm; the rates were 92.31%, 33.41%, for 15 mm 74.36%, 62.9%, for 20 mm 58.97%, 80.09%, respectively. When the criterion is 10 mm for the short axis; sensitivity was 92.59%, and spesificity was 52.23%.

When histopathological types and involvement of lymph nodes were considered together; nodal infiltration was highest in squamous cell carcinoma, but it was not statistically significant. In conclusion, as dimensions of nodes increase, the risk of involvement by tumour increase. If 15 mm in long axis and 10 mm in short axis were taken as criteria in hilar and mediastinal nodes; sensitivity and specificity approached nearly to optimal values. But when total accuracy rates were considered, no optimal values for both dimensions could be determined.