Level-specific analysis
Table 3 examines the frequency of nodal metastases in each neck level in specimens that were cN0 and cN+ve, respectively, at the corresponding level. ‘Occult’ nodes for each level were from here defined as those found on pathological staging in levels that had been cN0 at the corresponding level. It therefore demonstrates the concordance of ND specimens with preoperative clinical and radiological examination.
Two hundred individual neck level specimens were analysed (see Table 2) from the sixty-one patients. Seventy-seven neck level specimens had been ascribed cN+ve status at the corresponding level, of which 83.1% (64/77) had pathological evidence of metastatic disease. One hundred and twenty-three neck level specimens were cN0 in the corresponding level, of which occult disease was found in 13.0% (16/123). This occult disease was found in 58.3% (7/12) of cN0 level IIs, 13.3% (6/45) of cN0 level IIIs, 5.5% (3/61) of cN0 level IVs, and 0/5 cN0 level Vs. Overall, sixteen of sixty-one patients had pathological nodes in levels not established clinically. All four level I NDs were performed on necks cN+ve for level I, of which three were pN+ve in level I; two of the seven level V NDs were performed on necks cN+ve for level V, of which one was pN+ve in level V.
Table 4 presents the negative predictive values (NPV) and positive predictive values (PPV) for cN0 and cN+ve status, respectively, for each level. With regard to the distribution of lymph node metastasis at adjacent neck levels, it was found that patients who were cN+ve at level II had a PPV of 30.9% (95% CI 27.4% - 34.6%) for also being pN+ve at level III, and those who were cN0 at level II had a NPV of 87.5% (95% CI 48.1% - 98.2%) for being pN0 at level III. Patients who were cN+ve at level III had a PPV of 11.1% (95% CI 4.8% - 23.6%) for also being pN+ve in level IV, and a NPV of 97.8% (95% CI 89.8% - 99.6%) for the absence of pathological level IV nodes (data not shown).