Abstract
Background: Lymph node ratio (LNR), log odds of positive lymph
nodes (LODDS), and the number of postoperative lymph node staging (pN)
are prognostic indicators of various cancers. However, comparison of the
prognostic values of these indicators remains unclear in hypopharyngeal
squamous cell carcinoma (HPSCC). This study
aims to compare the prognostic
values of LNR, LODDS and pN in stage Ⅲ-Ⅳ HPSCC.
Methods: We conducted a
retrospective study on 166 patients with stage Ⅲ-Ⅳ HPSCC. LNR and LODDS
were divided into two groups using X-tile version 3.6.1. Univariate and
multivariate analyses of the risk of overall survival (OS) and
disease-free survival (DFS) were performed using the log-rank
(Mantel-Cox) test and the Cox proportional hazards model, respectively.
We compared the prognostic value of LNR with that of LODDS and pN using
receiver operating characteristic (ROC) curves.
Results: According to
the X-tile, the cut-off values are 0.11 for LNR and -0.91 for LODDS.
LNR, LODDS, and pN were significantly correlated with DFS by univariate
analysis (P < 0.05). Multivariate analysis demonstrated that
LNR was an independent prognostic factor for DFS (P < 0.01).
Multivariate analysis also revealed that postoperative tumour staging
(pT) classification, LNR, and surgical margins were independent
prognostic factors for OS.
Compared with pN and LODDS, LNR
showed a stronger predictive power for DFS.
Conclusion: LNR may be a better predictor for DFS than pN and
LODDS in stage Ⅲ-Ⅳ HPSCC patients. LNR in the highest tertile (≥ 0.11)
may cause poor OS and DFS. LODDS
in the highest tertile (≥-0.91) may cause poor DFS.
Keywords: hypopharyngeal
squamous cell carcinoma, lymph node ratio, log odds of positive lymph
nodes, prognosis.
key points:
- We conducted a retrospective study on 166 patients with stage Ⅲ-Ⅳ
HPSCC.
- aims to compare the prognostic values of LNR, LODDS and pN in stage
Ⅲ-Ⅳ HPSCC.
- LNR in the highest tertile (≥ 0.11) may cause poor OS and DFS.
- LODDS in the highest tertile (≥-0.91) may cause poor DFS.
- Compared with pN and LODDS, LNR showed a stronger predictive power for
DFS.
Introduction
HPSCC is one of the foremost head and neck cancers. The overall 5-year
survival rate remains approximately 40%, with the occurrence of local
and distant recurrences after treatment in a few cases(1). Although the
tumor-node-metastasis (TNM)
classification is a widely used prognostic factor for many solid tumors,
the staging system has been reported to be associated with poor
prognosis of patients with HPSCC(2). Furthermore, N status in the TNM
system provides only limited prognostic information. In order to clarify
which patients are more suitable for aggressive treatment to reduce
mortality and recurrence rates. Improvements in the staging system and
patient risk stratification, particularly for post-operative patients,
are required..
Lymph node metastasis is a well-known prognostic factor for HPSCC. Most
recent studies on the parameters include the total number of harvested
lymph nodes (HLNs) and the number of positive lymph nodes (PLNs)(3, 4).
Low values of the parameters, which are affected by the cervical
lymphatic dissection technique, may result in misdiagnosis and
eventually lead to an inaccurate treatment. The LNR,
defined as PLNs /HLNs, is a
better prognostic indicator. It also shows a better predictive potential
because it combines the information on regional metastatic disease
burden and type of cervical lymphatic dissection; thus, it accommodates
the advantages of both parameters while bypassing their disadvantages
(5, 6). Another index, the LODDS, defined as log (PLNs +0.5)/((HLNs-
PLNs ) +0.5), is correlated with the prognosis of head and neck tumors
as well (7). SIRI has been proved to have independent prognostic value
in head and neck tumors(8). Our previous study also found that SIRI had
a better predictive value than neutrophil-to-lymphocyte ratio (NLR),
platelet-to-lymphocyte ratio (PLR), and lymphocyte-to-monocyte ratio
(LMR) in HPSCC patients.
Past research has found that LNR and LODDS staging may be superior to
PLNs staging in patients with incomplete lymph node clearance because
they are not significantly affected by the total number of LNs
examined(9-14). LODDS has been used to predict the prognosis of several
tumors, but there is little clinical evidence for the role of LODDS in
HPSCC. For the first time, this study compared the predictive values of
LNR, LODDS and PLNs for OS and DFS in stage Ⅲ-Ⅳ HPSCC patients and
revealed new parameters to improve the predictive power.
Materials and methods
Study design and patient selection
A total of 304 patients with complete medical records and pathologically
confirmed primary HPSCC without secondary primary and distant metastatic
lesions were recruited. All these patients were diagnosed and treated at
the Eye, Ear, Nose and Throat Hospital of Fudan University, China,
between January 1, 2003, and December 31, 2014.They were classified
according to the American Joint Committee on Cancer (AJCC)/TNM-Union for
International Cancer Control (UICC) Staging Classification, 7th edition.
Of these 304 patients, 176 patients with pathological stages III-IV were
selected. The exclusion criteria included : (1)patients who had
undergone cervical lymphatic dissection, chemotherapy, or radiation
therapy before surgical treatment; (2)ECOG score >1
;(3) a chronic inflammatory
condition, such as hepatitis B virus and hepatitis C virus, gastritis or
nephritis;(4) aggressive inflammatory disease or co-infection; (5) an
autoimmune disease or treatment with steroids; (6) hematological
disease. Ultimately, 166 patients were included in the study analysis.
Follow-up and clinical endpoints
A follow-up of disease progression and time to death was performed on
all patients by telephone and outpatient records every three months for
the first two years and every six months until the end. The last
follow-up visit was performed on January 10, 2018, and OS was recorded
from the date of surgery until death. DFS was defined as the time
between the start of therapy and tumor relapse (locoregional recurrence
and/or distant metastases) or non-cancer causes of death.
Statistical Analysis
The optimal cut-off points of LNR and LODDS were determined using X-tile
version 3.6.1 (Yale University), according to the highest Chi-squared
value and lowest P-value in DFS. LNR and LODDS were categorized into two
groups based on the results attained from X-tile. The SIRI was defined
as (N*M)/L, where N, M, and L were the absolute count of neutrophils,
monocytes, and lymphocytes obtained from the pretreatment blood test.
The median determined the cut-off value for the SIRI, MLN, and HLN. OS
and DFS curves were plotted according to the Kaplan–Meier method using
SPSS (version 25.0) and were built with Graphpad Prism (version 8.0).
Univariate and multivariate
analyses for the risk of OS and DFS were performed using the log-rank
(Mantel-Cox) test and the Cox proportional hazard models, respectively.
A P-value of 0.05 (two-tailed) was considered statistically significant.