Discussion
This study examines the distribution of clinically known and occult nodal disease in neck dissection specimens of patients with p16+ve OPSCC who have undergone primary treatment with TORS/TOLM and ND. The data reaffirms previously reported data on the importance of including levels II, III and IV in ND for any cN+ve p16+ve OPSCC, although the volume of previously published literature is low 9-11. The frequency of occult node metastasis in patients staged overall as cN0 (four of nine NDs, 44%) underlines the importance of sound oncological surgical technique when undertaking selective neck dissection in this cohort. There is no place for limiting the neck dissection to levels II and III; levels II, III and IV is the minimum neck dissection patients should be offered regardless of the absence of clinical disease in prior echelon levels.
Three other studies previously published have examined the nodal distribution specific to p16+ve disease 9, 10, 12, with just one comparing the breakdown of clinically known and occult nodes 10. Previous studies had examined the distribution pattern of known and occult OPSCC nodal disease irrespective of p16 status,4, 13-17 with HPV only relatively recently demarcated as a disease subtype with a specific natural history and transmission.3 A recent Zenga et al. study examined a cohort of 324 pN+ve p16+ve OPSCC patients; however this study did not assess the distribution of LNs by neck level.11
The data we report here is consistent with studies reporting the frequency of occult lymph node metastasis in equivalent patient groups in level II (24-80%), level III (0-60%), and level IV nodes (0-27%).4, 10, 15, 17 The reported frequency of all LN metastasis are 76-90% for level II, 22-50% for level III and 9-14% for level IV which are comparable with our data. 4, 9, 10, 12, 13, 17. The insight gained from this study is similar to Amsbaugh et al who compared nodal distribution frequency in p16+ve and p16-ve OPSCC,12 but which only examined the distribution of cN+ve LNs. This found that the distribution of nodes in p16+ve disease was not significantly different to an historical Lindbergh study on overall nodal distributions in OPSCC conducted prior to establishment of the HPV aetiology.13 However, the Amsbaugh et al study had a 67% rate of smokers, introducing a confounding risk factor compared to our cohort with only 39% of patients reporting as current or significant smokers in the past.
We acknowledge that there is a weakness in our reporting of level IIA and IIB disease as this was only available for a small subset of patients (table 2), however with the limited data available we have shown that the presence of level IIB disease was always associated with level IIA disease. Generally, the senior authors’ policy is to dissect level IIA and IIB as standard in this patient group, but these specimens have, for the majority, been sent as a single level II specimen without sub-division. A larger study found isolated IIB disease in 2.2% without level IIA disease, although none in cN0 necks.4Recommendations in the United Kingdom remain that IIB can be left intact in T1-2 tumours if clinical disease in IIA is absent,3, 4 with heightened suspicion in tonsillar tumours.4
As with this study, previous reports on occult nodes for NDs reported uncommon involvement of levels I and V, with frequencies of 0-9% for level I and of 0-7% for level V.4, 10, 14, 15Dissection of level I is not considered routine and is advised in the presence of anterior tumour expansion into the oral cavity or in the presence of cN+ve level I lymph nodes.3