Older cancer patients usually have relatively poorer survival rates than younger patients.25, 26 HNSCC patients diagnosed at ≥60 years of age have been reported to have increased cause-specific mortality.27 As it is known, advanced TNM stage and poor grade are usually associated with the poor survival of cancer patients. However, the association of age with T classification, grade or LNM remains controversial. Some reports stated that the biologic behavior of HNSCC in younger patients was more aggressive compared with that in elderly patients.28–31 However, other studies noted no significant difference in T stage or grade between young and old patients with HNSCC.28, 32 As to the association of LNM with age, previous studies also had inconsistent and controversial findings. A retrospective study based on a large patient cohort of tongue squamous cell carcinoma reported that the LNM rate in young adults (< 40y) was higher than that in middle-age adults (40y~70y).32 This conclusion was supported by several other studies.33–35 Meanwhile, Jun-Ook et al. noted no significant difference in N stage between different age groups.31 The inconsistency between these findings might be partially due to the differences in population size, inclusion criteria of patients, age range, and age grouping, etc. In this study, we were surprised to find that the LNM rate of young patients (<60y) compared to that of old patients (≥60y) showed the opposite tendency in the derivation and the validation cohort. In the derivation cohort, young patients had a slightly but not significantly higher LNM rate compared to old patients (p=0.0703). In contrast, the LNM rate of old patients was higher, though not significantly, than that of young patients (p=0.261) in the validation cohort. The latter may be ascribed to possible patient selection bias in the choice of neck dissection surgery. When formulating the surgical plan, we comprehensively considered the tumor malignancy, the patient's health status, and the patient's expectations of quality of life after surgery. For elderly patients, the choice of neck dissection would be made more rigorously, which might explain the higher LNM rate in our validation cohort than the actual rate in elderly patients with HNSCC. Obviously, more prospective validation research is required to support this assumption.
As LNM is major prognostic factor for HNSCC patients, figuring out its risk factors is of utmost importance.36, 37 Increasing evidence indicates that lymphatic metastasis is not entirely passively but also actively regulated by malignant cells.38–40 Thus, the characteristics of primary tumors may highly impact the outcome of LNM occurrence. Regarding the associations of LNM with tumor size, we found that greater tumors (T2~T4) had increased the LNM rate 1.607-fold compared to smaller tumors (T1) (95% CI:1.146~2.272, p<0.05). This finding is in good agreement with previous studies. Chung et al. reported that tumors with higher T stage (T3~T4) had higher LNM rate, though with no statistical significance, compared to those with lower T stage (T1) (OR:1.86, 95% CI: 0.95~3.77, p=0.07).41 Bataini et al. reported an increased clinical occurrence of LNM in patients with T4 lesions (70%) compared to those with T1 lesions (44%).42 Except for T stage, the histologic grading based on malignant histomorphology has been widely used to predict the outcomes of patients with various cancer types, including HNSCC.43–46 Kademani, et al. reported that a higher-grade of HNSCC tumors is associated with decreased survival rate.47 A few studies have proved that tumor grade, especially grading at the invasive front, is a predictor of LNM in HNSCC.48 Further reports identified the significant positive correlation of the degree of pathological grade of primary tumors and occurrence frequency of germinal center predominance pattern.49 Since HNSCC tumors with germinal center predominance pattern feature twice risk of LNM compared to those with lymphocyte predominance pattern,50, 51 this may partially explain why high-grade tumors metastasize to regional lymph nodes with higher frequency.
In addition to LNM status, many indicators related to metastatic lymph nodes have been documented to be significant predictors of HNSCC patient prognosis, including the number of metastatic nodes,52 lymph node density (number of positive lymph nodes/total number of excised lymph nodes, LND)9, 53, 54 and the presence of extra-nodal spread (ENS)7, etc. However, we could not make further associations because of the limitations of the original data.
To the best of our knowledge, this study is one of the first to identify the risk factors for LNM of HNSCC patients from the aspect of tumor pathology. We demonstrate that tumors with greater size and higher grade have a greater potential for lymph node metastasis. Based on this, the early screening and diagnosis of HNSCC is of critical importance for improving treatment efficacy and prolonging overall survival time. Despite its merits, the research has some limitations. Firstly, the present study had a retrospective design and relied upon evaluations of the pathology reports. The heterogeneity of surgeons performing the clinical practice was inevitable. Although patients in both the derivation cohort and the validation cohort received neck dissection, the decisions on procedures made by surgeons varied among individuals. Neck dissection is classified based on the dissected anatomic regions, including radical neck dissection, modified radical neck dissection, selective neck dissection, and extended neck dissection.55 The diagnosis of pathological LNM was basically dependent on the performance of neck dissection. The inconsistence of surgical practice might lead to underestimation of the incidence of pathological LNM and of the correlation of LNM with the OS of HNSCC patients, as well as the potential correlations of other variables with LNM. Secondly, the validation cohort was formed by patients from a single institution, therefore, a prospective multicenter randomized trial involving a larger number of patients is necessary to validate the findings.