- Email: [email protected]
Postoperative staging of the neck dissection using extracapsular spread and lymph node ratio as prognostic factors in HPV-negative head and neck squamous cell carcinoma patients
Oral Oncology 77 (2018) 37–42
Contents lists available at ScienceDirect
Oral Oncology journal homepage: www.elsevier.com/locate/oraloncology
Postoperative staging of the neck dissection using extracapsular spread and lymph node ratio as prognostic factors in HPV-negative head and neck squamous cell carcinoma patients
T
Katarina Majercakovaa, Cristina Valerob, Montserrat Lópezb, Jacinto Garcíab, Nuria Farréa, ⁎ Miquel Querb, Xavier Leónb,c, a b c
Radiotherapy Oncology Department, Hospital de la Santa Creu i Sant Pau, Universitat Autònoma de Barcelona, Barcelona, Spain Otorhinolaryngology Department, Hospital de la Santa Creu i Sant Pau, Universitat Autònoma de Barcelona, Barcelona, Spain Centro de Investigación Biomédica en Red de Bioingeniería, Biomateriales y Nanomedicina (CIBER-BBN), Madrid, Spain
A R T I C L E I N F O
A B S T R A C T
Keywords: Extracapsular spread Lymph node ratio pN TNM classification Head and neck cancer HPV-negative
Objectives: The presence of nodes with extracapsular spread (ECS) and the lymph node ratio (LNR) have prognostic competence in the pathologic evaluation of patients with a head and neck squamous cell carcinoma (HNSCC) treated with a neck dissection. The purpose of this study is to assess the effect of ECS & LNR on prognosis of HPV negative HNSCC patients treated with neck dissection and to compare to 8th edition TNM/ AJCC classification. Materials and methods: We carried out a retrospective study of 1383 patients with HNSCC treated with a neck dissection between 1985 and 2013. We developed a classification of the patients according to the presence of nodes with ECS and the LNR value with a recursive partitioning analysis (RPA) model. Results: We obtained a classification tree with four terminal nodes: for patients without ECS (including patients pN0) the cut-off point for LNR was 1.6%, while for patients with lymph nodes with ECS it was 11.4%. The 5-year disease-specific survival for patients without ECS/LNR < 1.6% was 83.3%; for patients without ECS/ LNR ≥ 1.6% it was 61.5%; for patients with ECS/LNR < 11.4% it was 33.7%; and for patients with ECS/ LNR ≥ 11.4% it was 18.5%. The classification obtained with RPA had better discrimination between categories than the 8th edition of the TNM/AJCC classification. Conclusion: ECS status and LNR value proved high prognostic capacity in the pathological evaluation of the neck dissection. The combination of ECS and LNR improved the predictive capacity of the 8th edition of the TNM/ AJCC classification in HPV-negative HNSCC patients.
Introduction Lymph node status is one of the most important clinical predictors of survival for head and neck squamous cell carcinoma (HNSCC) patients. The standard pathological nodal staging (pN) of a neck dissection considers the number, size and location of positive lymph nodes. Several studies and meta-analysis show that the presence of lymph nodes with extracapsular spread (ECS), defined as extension of the tumor outside the lymph node capsule, negatively affects prognosis in HPV-negative HNSCC patients [1–4]. Interestingly, ECS did not affect survival in patients with HPV-positive oropharyngeal tumors [5,6]. These studies have led to the inclusion of the ECS into the pathological classification criteria in the 8th edition of the TNM/AJCC classification of HPV-negative patients [7,8]. The 8th edition TNM/AJCC
⁎
classification improves the differentiation in survival among the pN categories as well as the distribution of the number of patients per category in HNSCC HPV-negative patients when compared with the 7th edition TNM/AJCC [9]. To further improve the classification of nodal disease, several authors have analyzed the lymph node ratio (LNR). LNR is defined as the proportion of metastatic lymph nodes related to the total number of examined neck nodes. LNR has proved a very high prognostic capacity in neck dissection evaluation. High LNR values have consistently related to a worse overall and specific survival in most series (Supplementary material, Table 1). The objective of this study is to assess the prognostic competence of the pathological classification of the neck dissection obtained from evaluating ECS and LNR together and to compare this classification
Corresponding author at: Otorhinolaryngology Department, Hospital de la Santa Creu i Sant Pau, C/ Mas Casanovas, 90, 08041 Barcelona, Spain. E-mail address: [email protected] (X. León).
https://doi.org/10.1016/j.oraloncology.2017.12.010 Received 9 August 2017; Received in revised form 8 November 2017; Accepted 15 December 2017 1368-8375/ © 2017 Elsevier Ltd. All rights reserved.
Oral Oncology 77 (2018) 37–42
K. Majercakova et al.
study period. Patients with advanced tumor, either locally (pT3-T4) or regionally (pN2-N3), microscopically involved surgical margins, or ECS were considered candidates to adjuvant treatment. Postoperative radiotherapy was delivered in 2 Gy fractions to a total of 50 Gy in 5 weeks directed to both the primary site and the neck. In cases with ECS, a boost of up to 60–66 Gy was administered over the compromised areas. Cisplatin 100 mg/m2 was offered in selected cases with indication of postoperative radiotherapy from 2000 to present. The mean follow-up time was 5.7 years (standard deviation 5.4 years). During the follow-up period, 257 patients (18.6%) had local failure, 206 (14.9%) had regional failure, and 214 (15.15%) presented distant metastases. ECS was defined as any breach in the lymph node capsule caused by tumor cells. The pathological report of the neck dissections in our center did not include information about the microscopic or macroscopic character of the ECS. LNR was defined as the ratio between positive nodes and total number of removed nodes. The cut-off points for LNR were defined through recursive partitioning analysis (RPA) by the chi-square automatic interaction detection method, considering the specific survival as the dependent variable. Survival curves were estimated using the Kaplan-Meier actuarial method according to the presence of ECS and LNR. Differences in survival were compared using the log-rank test. We then performed a multivariate analysis (Cox proportional hazards model) including ECS and LNR as independent variables and disease-specific survival as the dependent variable. The RPA was repeated including both ECS and LNR in the model. Disease-specific survival was evaluated according to the categories obtained with this RPA. Finally, the prognostic capacity of the RPA model was compared with the 8th edition of the TNM/AJCC classification. We used measures of hazard discrimination and balance following the criteria defined by Groome et al. [53] to objectively compare the obtained RPA classification and the 8th edition of TNM/AJCC classification. Hazard discrimination measured how evenly the survival curves are spaced for each of the classification categories, and how large is the difference in survival between the best and the worst categories. The hazard discrimination ranges from zero to 1, where 1 represents an ideal classification with a full coverage of the survival area by evenly spaced curves. Balance quantifies the distribution of the number of patients for the classification scheme in each category. Balance ranges from zero to 1 and measures the adequacy to an ideal situation with the same number of patients in each category. The best classification system is the one with the highest balance. All procedures were reviewed by the Institutional Review Board at Santa Creu i Sant Pau Hospital. The study conforms to the principles outlined in the Declaration of Helsinki.
Table 1 Characteristics of the patients included in the study. Age (years)
Median 60.2/Standard deviation 11.1
Sex
Men Women
1251 (90.5%) 132 (9.5%)
Toxics consumption
Never Moderate Severe
85 (6.1%) 163 (11.8%) 1135 (82.1%)
Histological differentiation
Well Moderate Poor
135 (9.8%) 1074 (77.6%) 174 (12.6%)
Localization
Oral cavity Oropharynx Hypopharynx Supraglottic larynx Glottic larynx
316 250 214 375 228
(22.8%) (18.1%) (15.5%) (27.1%) (16.5%)
Local extension#
T1 T2 T3 T4
175 400 519 289
(12.7%) (28.9%) (37.5%) (20.9%)
Regional extension#
N0 N1 N2 N3
609 258 411 105
(44.0%) (18.7%) (29.7%) (7.6%)
#
According to 7th edition TNM classification.
with the 8th edition of the TNM/AJCC classification.
Material and methods We performed a retrospective study based on prospectively collected information of patients with HNSCC treated in our center [51]. A total of 1473 patients with HNSCC located in the oral cavity, oropharynx, hypopharynx, or larynx, diagnosed from 1985 until 2013, and treated with a unilateral or bilateral neck dissection were initially included in the study. HPV status in oropharyngeal tumors was analyzed retrospectively by HPV-DNA detection with SPF-10 real time PCR assay in combination with LiPA genotyping [52]. We excluded 24 patients with HPV-positive oropharyngeal cancer, 20 patients with selective neck dissection of only one lymph node area, 15 patients who lacked appropriate information about the pathological results of the neck dissection, and 31 patients who did not have a minimum follow-up of 2 years. Table 1 shows the characteristics of the 1383 patients finally included in the study. According to the interaction between tobacco and alcohol consumption, we created a combined variable of toxic consumption with three categories: no consumption; moderate consumption (< 20 cigarettes/day and/or < 80 g alcohol/day); and severe consumption (≥20 cigarettes/day and/or ≥80 g alcohol/day). We retrieved information concerning the type of neck dissection (unilateral or bilateral), the number of dissected nodes, the number of positive nodes, and the number of nodes with ECS for all patients. We performed 2140 neck dissections (469 radical neck dissections and 1671 selective neck dissections) in the patients included in the study. A total of 757 patients (54.7%) had bilateral neck dissections. In patients treated with a bilateral neck dissection, results were analyzed adding the neck nodes dissected on both sites of the neck. The mean number of lymph nodes studied per patient was 31.2 (standard deviation 19.0, range 7–118). In 201 cases (14.5%) we performed the neck dissections after a previous treatment with radiotherapy (n = 110) or chemoradiotherapy (n = 91). The interval between the radiotherapy or chemoradiotherapy and the neck dissection was 6–10 weeks (median, 8.5 weeks). A total of 751 patients (54.3%) had postoperative adjuvant treatment with radiotherapy (n = 677) or chemoradiotherapy (n = 74). The indications for adjuvant treatment were maintained throughout the
Results Five-year disease-specific survival for all the patients included in the study was 63.6% (95% CI: 60.9–66.3%). In 677 patients (49.0%) the pathological study of the neck dissection did not demonstrate the presence of lymph node metastases (pN0), 356 (25.7%) had lymph node metastases without extracapsular spread (pN+/ECS−) and 350 (25.3%) had lymph node metastases with extracapsular spread (pN +/ECS+). Fig. 1A shows the disease-specific survival curves related to the presence of lymph node metastases with ECS. The 5-year diseasespecific survival was 83.0% (95% CI: 80.1–85.9%) for pN0 patients, 62.3% (95% CI: 56.8–67.8%) for pN+/ECS− patients, and 25.7% (95% CI: 20.6–30.8%) for pN+/ECS+ patients. There were statistically significant differences in disease-specific survival regarding to the existence of lymph node metastases and ECS (P = .0001). The mean value of LNR of patients included in the study was 7.3% (standard deviation 13.6%, median 1.6%). In the subgroup of patients 38
Oral Oncology 77 (2018) 37–42
K. Majercakova et al.
survival was 83.3% (95% CI: 80.1–85.9%) for patients without ECS and LNR < 1.6%; 61.5% (95% CI: 56.0–67.0%) for patients without ECS and LNR ≥ 1.6%; 33.7% (95% CI: 25.7–41.7%) for patients with ECS and LNR < 11.4%; and 18.5% (95% CI: 12.2–24.8%) for patients with ECS and LNR ≥ 11.4%. These differences were statistically significant (P = .0001). Fig. 3B shows the disease-specific survival curves for pN according to the 8th edition of the TNM/AJCC classification. The 5-year diseasespecific survival figures read as follows: pN0 83.0% (95% CI: 80.1–85.9%), pN1 69.1% (95% CI: 61.5–76.7%), pN2 57.0% (95% CI: 49.7–64.3%), and pN3 22.2% (95% CI: 17.1–27.3%) (P = .0001). The comparison between both methods showed that the classification obtained with RPA had better capacity of discrimination between categories than the 8th edition of the TNM/AJCC classification. The value of hazard discrimination for the RPA was 0.868 compared to the value of 0.653 for the 8th edition of the TNM/AJCC classification (differences in hazard discrimination value of 0.215 in favor of the RPA classification). The distribution of patients per categories was very similar for both classifications. The balance value for RPA was 0.671 versus 0.680 for the 8ª edition of the TNM/AJCC classification (differences in balance value of 0.009 in favor of the TNM/AJCC classification). Table 2 shows 5-year local-, regional- and distant-recurrence-free survival for every category in both classifications. Additionally, it shows the values of hazard discrimination corresponding to every survival study. With the increasing burden of disease in our classification, a higher risk of local, regional or distant failure was observed. The hazard discrimination scores for regional and distant control were superior for the RPA classification, while the hazard discrimination score for local control favored the 8th edition of the TNM/AJCC classification. The supplementary material shows the local and regional recurrence-free survival and the distant metastases-free survival curves according to both classification methods.
Fig. 1. Disease-specific survival curves according to the presence of neck node metastases with extracapsular spread (ECS) (A) and to the lymph node ratio (LNR) category (B).
with pN+, the mean value of LNR was 13.3% (standard deviation 16.1%, median 8.7%). The RPA defined two cut-off points for LNR. The low cut-off point was set in 1.6% and the high cut-off point in 11.4%, defining three categories with significant differences in disease-specific survival. Fig. 1B shows disease-specific survival for the categories defined by these cut-off points. The 5-year disease-specific survival was 83.0% (95% CI: 80.1–85.9%) for patients with LNR < 1.6% (n = 692, 50%), 52.3% (95% CI: 47.0–57.6%) for patients with LNR 1.6–11.4% (n = 415, 30.0%), and 29.8% (95% CI: 23.9–35.7%) for patients with LNR > 11.4% (n = 276, 20.0%). These differences reached statistical significance (P = .0001). We carried out a multivariate study considering the disease-specific survival as dependent variable, including ECS and LNR category as independent variables. According to the results of this model, both variables showed prognostic capacity. Patients with one or more lymph nodes with ECS had 2.45 times (95% CI: 1.97–3.06) higher risk of disease-specific mortality compared to patients without evidence of ECS (P = .0001). Similarly, patients with intermediate LNR had 2.55 (95% CI: 1.97–3.31) higher risk than patients with low LNR, and patients with high LNR had 3.98 times (95% CI: 2.98–5.31) higher risk of disease-specific mortality than patients with low LNR. All these differences reached statistical significance (P = .0001). We performed a RPA considering disease-specific survival as dependent variable. ECS and LNR value were included as independent variables. A classification tree with four terminal nodes was obtained (Fig. 2). Interestingly, the same cut-off points for LNR delineated the terminal nodes. In patients without ECS (including pN0 patients) the corresponding cut-off point for LNR was 1.6%, while for patients with lymph nodes with ECS it was 11.4%. Fig. 3A shows the disease-specific survival curves for the terminal nodes obtained with the RPA model. The 5-year disease-specific
Discussion According to our results, the pathological classification of neck dissections in patients with HPV-negative HNSCC based in ECS and LNR value shows better prognostic capacity than the 8th edition of the TNM/AJCC classification [7,8]. Strong evidence in the literature supports the role of ECS as a negative prognostic factor for final survival. The results of a meta-analysis by Dünne et al. [3] showed that the presence of ECS in neck dissections had a negative impact on survival, with a summarized odds ratio of 2.7 (95% CI, 2.2–3.4). Additionally, a recent systematic review of the literature and meta-analysis carried out by Mermod et al. [4] confirmed the impact of ECS on loco-regional recurrence and distant metastases in HPV-negative HNSCC patients. Interestingly, the presence of ECS in patients with HPV-positive oropharyngeal tumors did not affect prognosis [5,6]. According to our results, there were statistically significant differences in disease-specific survival related to the presence of lymph node metastases with ECS in HPV-negative HNSCC patients, as shown in Fig. 1A. More recently, the LNR has emerged as a new an independent prognostic factor in the pathological evaluation of neck dissections. Early publications in 2009 by Shrime et al. [10,11] and Gil et al. [12] demonstrated the prognostic capacity of LNR in patients with oral cavity carcinomas. Following these data, multiple studies proved the relationship between the LNR and survival or disease control of HNSCC patients in different localizations (Supplementary material, Table 1). The LNR weighs three factors that can potentially influence nodal staging: (1) surgical factors: the comprehensiveness of the neck dissection measured as the total number of lymph nodes removed; (2) tumor factors: the number of positive neck nodes; and (3) staging factors: the completeness of the pathologic analysis. 39
Oral Oncology 77 (2018) 37–42
K. Majercakova et al.
Fig. 2. Recursive partitioning analysis classification according to the presence of lymph node metastases with extracapsular spread (ECS) and lymph node ratio (LNR) value (n, number of patients included in each node of the classification tree; the percentage in the pies indicates the proportion of patients who died as a consequence of the tumor).
distribution [19], ROC analysis [20,21,31,34,35,45], the minimum p value from the log-rank test [17,23,25,27,28,39,41,43,44,46,48,49], principal component analysis [18,50], the maximally selected rank statistic method [10,11,24,37], a graphic determination from a logarithmic transformation of data [14], or previously published thresholds [22,42,47]. In the present study, we used a RPA method to select the cut-off points for our model. According the results of this model, the cut-off points defining populations with different prognosis were: LNR < 1.6% (including patients with pN0), LNR 1.6–11.4% and LNR > 11.4%. The studies that analyze both ECS and the LNR in multivariate analysis led to conflicting results. For some authors, the inclusion of LNR in the model leads to a loss of the independent prognostic capacity of ECS [10,12,14,15,27,31,32,34,43–46,48]. However, others consider that both ECS and LNR retain prognostic significance in multivariate studies [16,17,21,34,36,38,40,48]. According to our results, both ECS and LNR have an independent prognostic capacity in a multivariate study, which justifies their inclusion in the RPA model. The main advantage of RPA as a classification method is that it defines specific cut-off points that optimize the possibility of discrimination in different clinical situations. In our sample, the RPA algorithm defined a first level of classification based on the presence or the absence of neck nodes with ECS. The second level of classification defined a specific cut-off point for LNR for every ECS category. In patients without ECS the value of LNR with discrimination capacity in disease-specific survival was 1.6%, and in patients with ECS it corresponded to 11.4%. The comparison of the RPA classification with the 8th edition of the TNM/AJCC classification yielded satisfactory results in the homogeneous distribution of the differences in survival between categories (hazard discrimination) in favor of the RPA. The distribution of the number of patients per category was very similar in both classifications. One of the main goals of a pathological classification of the neck dissection is to provide prognostic data on the risk of regional recurrence. This level of risk has a direct impact in the planning of adjuvant treatment and its intensity. The classification based on RPA showed higher values of hazard discrimination in regional and distant-metastases-free survival than the 8th edition pathological TNM/AJCC in HPV-negative HNSCC patients. These results indicate that this method has better prognostic capacity evaluating the regional and distant disease control. One of the limitations of our study is that we do not have
Fig. 3. Disease-specific survival according to the recursive partitioning analysis classification (A) and to pN category as defined by the 8th edition TNM/AJCC classification (B).
The cut-off points for LNR values oscillate between 2.5% and 20% in different studies (Supplementary material, Table 1). These differences are due to the heterogeneity of patients between studies. While some series analyze HNSCC with a specific location, others include tumors from all head and neck locations. Some series only include patients with pathological lymph node metastases (pN+) while others include also patients without nodal disease (pN0). Besides, the methods of determination of cut-off points are different between studies. Some studies uses the median value of the LNR [12,13,15,26,29,30,32,33], quartile 40
Oral Oncology 77 (2018) 37–42
K. Majercakova et al.
Table 2 Five-year local (rT), regional (rN), and distant (rM) recurrence-free survival and hazard discrimination values according to the recursive partitioning analysis (RPA) classification and the 8th edition of the TNM classification.
RPA
ECS−/LNR ECS−/LNR ECS+/LNR ECS+/LNR
< > < >
1.6% 1.6% 11.4% 11.4%
Hazard discrimination 8th TNM
rT (CI 95%)
rN (CI 95%)
rM (CI 95%)
87.3% 73.7% 62.8% 59.0%
93.6% 85.1% 65.2% 48.6%
92.6% 81.9% 55.3% 47.5%
(84.8–89.8%) (68.7–78.7%) (52.9–72.7%) (48.9–69.1%)
0.517 pN0 pN1 pN2 pN3
Hazard discrimination
87.2% 77.5% 71.0% 59.9%
(91.7–95.5%) (80.8–89.4%) (56.6–73.8%) (39.2–58.0%)
0.647 (84.5–89.9%) (70.5–84.5%) (64.4–77.6%) (52.5–67.3%)
0.653
93.3% 88.2% 83.8% 53.4% 0.208
(90.5–94.7%) (77.4–86.4%) (45.4–65.2%) (37.6–57.4%)
0.662 (91.4–95.2%) (83.0–87.4%) (78.2–89.4%) (46.6–30.2%)
92.5% 87.2% 79.2% 46.1%
(90.4–94.6%) (81.8–92.6%) (73.0–85.4%) (38.5–53.7%)
0.328
Abbreviations: ECS, extracapsular spread; LNR, lymph node ratio.
information about the microscopic or macroscopic character of the ECS. Besides, this study is a retrospective evaluation of patients treated at a single institution. Our population of study is mainly composed of heavy smokers and drinkers (82% of our patients), with a high proportion of laryngeal tumors (43.6% of patients). This limits the extension of our conclusions and its utility for generalized use across all tumor types and patient populations, and calls for further studies to validate the prognostic capacity of the joint assessment of the presence of nodes with ECS and the LNR as a pathological criterion in the evaluation of neck dissections in HNSCC patients.
[5] Haughey BH, Sinha P. Prognostic factors and survival unique to surgically treated p16+ oropharyngeal cancer. Laryngoscope 2012;122(Suppl 2):S13–33. [6] Kharytaniuk N, Molony P, Boyle S, O'Leary G, Werner R, Heffron C, et al. Association of extracapsular spread with survival according to human papillomavirus status in oropharynx squamous cell carcinoma and carcinoma of unknown primary site. JAMA Otolaryngol Head Neck Surg 2016;142:683–90. [7] UICC International Union Against Cancer. TNM classification of malignant tumors. In: Brierley JD, Gospodarowicz MK, Wittekind Ch, editors. 8th ed., Wiley-Blackwell; 2016. [8] AJCC Cancer Staging Manual. In: Amin MB, Edge S, Greene F, Byrd DR, Brookland RK, Washington MK, et al., editors. 8th ed. Springer International Publishing; 2017. [9] García J, López M, López L, Bagué S, Granell E, Quer M, et al. Validation of the pathological classification of lymph node metastasis for head and neck tumors according to the 8th edition of the TNM Classification of Malignant Tumors. Oral Oncol 2017;70:29–33. [10] Shrime MG, Bachar G, Lea J, Volling C, Ma C, Gullane PJ, et al. Nodal ratio as an independent predictor of survival in squamous cell carcinoma of the oral cavity. Head Neck 2009;31:1482–8. [11] Shrime MG, Ma C, Gullane PJ, Gilbert RW, Irish JC, Brown DH, et al. Impact of nodal ratio on survival in squamous cell carcinoma of the oral cavity. Head Neck 2009;31:1129–36. [12] Gil Z, Carlson DL, Boyle JO, Kraus DH, Shah JP, Shaha AR, et al. Lymph node density is a significant predictor of outcome in patients with oral cancer. Cancer 2009;115:5700–10. [13] Süslü N, Hoşal AS, Sözeri B. Prognostic value of metastatic lymph node ratio in node-positive head and neck carcinomas. Am J Otolaryngol 2010;31:315–9. [14] Ebrahimi A, Clark JR, Zhang WJ, Elliott MS, Gao K, Milross CG, et al. Lymph node ratio as an independent prognostic factor in oral squamous cell carcinoma. Head Neck 2011;33:1245–51. [15] Kim SY, Nam SY, Choi SH, Cho KJ, Roh JL. Prognostic value of lymph node density in node-positive patients with oral squamous cell carcinoma. Ann Surg Oncol 2011;18:2310–7. [16] Lanzer M, Kruse A, Lübbers HT, Zemann W, Reinisch S. Lymph node ratio and capsule penetration as independent risk factors in head and neck squamous cell carcinoma. Head Neck Oncol 2012;4:89. [17] Liao CT, Hsueh C, Lee LY, Lin CY, Fan KH, Wang HM, et al. Neck dissection field and lymph node density predict prognosis in patients with oral cavity cancer and pathological node metastases treated with adjuvant therapy. Oral Oncol 2012;48:329–36. [18] Kim KY, Cha IH. Risk stratification of oral cancer patients using a combined prognostic factor including lymph node density and biomarker. J Cancer Res Clin Oncol 2012;138:483–90. [19] Yu Y, Wang XL, Xu ZG, Fan CC, Li Q. Prognostic value of lymph node ratio in hypopharyngeal squamous cell carcinoma after chemoradiotherapy. Chin Med J (Engl) 2013;126:4139–44. [20] Patel SG, Amit M, Yen TC, Liao CT, Chaturvedi P, Agarwal JP, et al. Lymph node density in oral cavity cancer: results of the International Consortium for Outcomes Research. Br J Cancer 2013;15(109):2087–95. [21] Sayed SI, Sharma S, Rane P, Vaishampayan S, Talole S, Chaturvedi P, et al. Can metastatic lymph node ratio (LNR) predict survival in oral cavity cancer patients? J Surg Oncol 2013;108:256–63. [22] Urban D, Gluck I, Pfeffer MR, Symon Z, Lawrence YR. Lymph node ratio predicts the benefit of post-operative radiotherapy in oral cavity cancer. Radiother Oncol 2013;106:74–9. [23] Mizrachi A, Hadar T, Rabinovics N, Shpitzer T, Guttman D, Feinmesser R, et al. Prognostic significance of nodal ratio in cutaneous squamous cell carcinoma of the head and neck. Eur Arch Otorhinolaryngol 2013;270:647–53. [24] Reinisch S, Kruse A, Bredell M, Lübbers HT, Gander T, Lanzer M. Is lymph-node ratio a superior predictor than lymph node status for recurrence-free and overall survival in patients with head and neck squamous cell carcinoma? Ann Surg Oncol 2014;21:1912–8. [25] Rudra S, Spiotto MT, Witt ME, Blair EA, Stenson K, Haraf DJ. Lymph node density–prognostic value in head and neck cancer. Head Neck 2014;36:266–72.
Conclusion Besides the classical pathological criteria of nodal metastases, ECS status and LNR value proved elevated prognostic capacity in the evaluation of patients with HPV-negative HNSCC treated with neck dissection. The combination of ECS and LNR improved the predictive capacity of the 8th edition of the TNM/AJCC classification. Conflict of interest statement All the authors certify that there is no conflict of interest with any financial organization regarding the material discussed in the manuscript. Acknowledgements This work was supported by a grant from the Instituto de Salud Carlos III – Spain (FIS PI14/01918). The study was cofunded by Fondo Europeo de Desarrollo Regional (FEDER), A Way to Build Europe. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.oraloncology.2017.12. 010. References [1] de Juan J, García J, López M, Orús C, Esteller E, Quer M, et al. Inclusion of extracapsular spread in the pTNM classification system: a proposal for patients with head and neck carcinoma. JAMA Otolaryngol Head Neck Surg 2013;139:483–8. [2] Wreesmann VB, Katabi N, Palmer FL, Montero PH, Migliacci JC, Gönen M, et al. Influence of extracapsular nodal spread extent on prognosis of oral squamous cell carcinoma. Head Neck 2016;38:1192–9. [3] Dünne AA, Müller HH, Eisele DW, Kessel K, Moll R, Werner JA. Meta-analysis of the prognostic significance of perinodal spread in head and neck squamous cell carcinomas (HNSCC) patients. Eur J Cancer 2006;42:1863–8. [4] Mermod M, Tolstonog G, Simon C, Monnier Y. Extracapsular spread in head and neck squamous cell carcinoma: a systematic review and meta-analysis. Oral Oncol 2016;62:60–71.
41
Oral Oncology 77 (2018) 37–42
K. Majercakova et al.
Arch 2016;469:635–41. [41] Yildiz MM, Petersen I, Eigendorff E, Schlattmann P, Guntinas-Lichius O. Which is the most suitable lymph node predictor for overall survival after primary surgery of head and neck cancer: pN, the number or the ratio of positive lymph nodes, or log odds? J Cancer Res Clin Oncol 2016;142:885–93. [42] Roberts TJ, Colevas AD, Hara W, Holsinger FC, Oakley-Girvan I, Divi V. Number of positive nodes is superior to the lymph node ratio and American Joint Committee on Cancer N staging for the prognosis of surgically treated head and neck squamous cell carcinomas. Cancer 2016;122:1388–97. [43] Suzuki H, Matoba T, Hanai N, Nishikawa D, Fukuda Y, Koide Y, et al. Lymph node ratio predicts survival in hypopharyngeal cancer with positive lymph node metastasis. Eur Arch Otorhinolaryngol 2016;273:4595–600. [44] Imre A, Pinar E, Dincer E, Ozkul Y, Aslan H, Songu M, et al. Lymph node density in node-positive laryngeal carcinoma: analysis of prognostic value for survival. Otolaryngol Head Neck Surg 2016;155:797–804. [45] Park JO, Joo YH, Cho KJ, Kim MS. Lymph node density as an independent prognostic factor in node-positive patients with tonsillar cancer. Head Neck 2016;38(Suppl 1):E705–11. [46] Lieng H, Gebski VJ, Morgan GJ, Veness MJ. Important prognostic significance of lymph node density in patients with node positive oral tongue cancer. ANZ J Surg 2016;86:681–6. [47] Ong W, Zhao R, Lui B, Tan W, Ebrahimi A, Clark JR, et al. Prognostic significance of lymph node density in squamous cell carcinoma of the tongue. Head Neck 2016;38(Suppl 1):E859–66. [48] Suzuki H, Beppu S, Hanai N, Hirakawa H, Hasegawa Y. Lymph node density predicts lung metastases in oral squamous cell carcinoma. Br J Oral Maxillofac Surg 2016;54:213–8. [49] Tseros EA, Gebski V, Morgan GJ, Veness MJ. Prognostic significance of lymph node ratio in metastatic cutaneous squamous cell carcinoma of the head and neck. Ann Surg Oncol 2016;23:1693–8. [50] Kim KY, Zhang X, Kim SM, Lee BD, Cha IH. A combined prognostic factor for improved risk stratification of patients with oral cancer. Oral Dis 2017;23:91–6. [51] León X, Orús C, Quer M. Design, maintenance, and exploitation of an oncologic database for patients with malignant tumors of the head and neck. Acta Otorrinolaringol Esp 2002;53:185–90. [52] Castellsagué X, Alemany L, Quer M, Halec G, Quirós B, Tous S, et al. HPV involvement in head and neck cancers: comprehensive assessment of biomarkers in 3680 patients. J Natl Cancer Inst 2016;108:djv403. [53] Groome PA, Schulze K, Boysen M, Hall SF, Mackillop WJ. A comparison of published head and neck stage groupings in carcinomas of the oral cavity. Head Neck 2001;23:613–24.
[26] Ampil FL, Caldito G, Ghali GE. Can the lymph node ratio predict outcome in head and neck cancer with single metastasis positive-node? Oral Oncol 2014;50:e18–20. [27] Prabhu RS, Hanasoge S, Magliocca KR, Hall WA, Chen SA, Higgins KA, et al. Lymph node ratio influence on risk of head and neck cancer locoregional recurrence after initial surgical resection: implications for adjuvant therapy. Head Neck 2015;37:777–82. [28] Wang YL, Li DS, Wang Y, Wang ZY, Ji QH. Lymph node ratio for postoperative staging of laryngeal squamous cell carcinoma with lymph node metastasis. PLoS ONE 2014;9:e87037. [29] Künzel J, Mantsopoulos K, Psychogios G, Grundtner P, Koch M, Iro H. Lymph node ratio as a valuable additional predictor of outcome in selected patients with oral cavity cancer. Oral Surg Oral Med Oral Pathol Oral Radiol 2014;117:677–84. [30] Künzel J, Psychogios G, Mantsopoulos K, Grundtner P, Waldfahrer F, Iro H. Lymph node ratio as a predictor of outcome in patients with oropharyngeal cancer. Eur Arch Otorhinolaryngol 2014;271:1171–80. [31] Park GC, Jung JH, Roh JL, Lee JH, Cho KJ, Choi SH, et al. Prognostic value of metastatic nodal volume and lymph node ratio in patients with cervical lymph node metastases from an unknown primary tumor. Oncology 2014;86:170–6. [32] Chen CC, Lin JC, Chen KW. Lymph node ratio as a prognostic factor in head and neck cancer patients. Radiat Oncol 2015;10:181. [33] Künzel J, Mantsopoulos K, Psychogios G, Agaimy A, Grundtner P, Koch M, et al. Lymph node ratio is of limited value for the decision-making process in the treatment of patients with laryngeal cancer. Eur Arch Otorhinolaryngol 2015;272:453–61. [34] Ryu IS, Roh JL, Cho KJ, Choi SH, Nam SY, Kim SY. Lymph node density as an independent predictor of cancer-specific mortality in patients with lymph nodepositive laryngeal squamous cell carcinoma after laryngectomy. Head Neck 2015;37:1319–25. [35] Joo YH, Cho KJ, Kim SY, Kim MS. Prognostic significance of lymph node density in patients with hypopharyngeal squamous cell carcinoma. Ann Surg Oncol 2015;22(Suppl 3):S1014–9. [36] Hua YH, Hu QY, Piao YF, Tang Q, Fu ZF. Effect of number and ratio of positive lymph nodes in hypopharyngeal cancer. Head Neck 2015;37:111–6. [37] Chan RC, Chan JY. Impact of nodal ratio on survival in recurrent nasopharyngeal carcinoma. Head Neck 2015;37:12–7. [38] Hong HR, Roh JL, Cho KJ, Choi SH, Nam SY, Kim SY. Prognostic value of lymph node density in high-grade salivary gland cancers. J Surg Oncol 2015;111:784–9. [39] Suzuki H, Hanai N, Hirakawa H, Nishikawa D, Hasegawa Y. Lymph node density is a prognostic factor in patients with major salivary gland carcinoma. Oncol Lett 2015;10:3523–8. [40] de Ridder M, Marres CC, Smeele LE, van den Brekel MW, Hauptmann M, Balm AJ, et al. A critical evaluation of lymph node ratio in head and neck cancer. Virchows
42
Contents lists available at ScienceDirect
Oral Oncology journal homepage: www.elsevier.com/locate/oraloncology
Postoperative staging of the neck dissection using extracapsular spread and lymph node ratio as prognostic factors in HPV-negative head and neck squamous cell carcinoma patients
T
Katarina Majercakovaa, Cristina Valerob, Montserrat Lópezb, Jacinto Garcíab, Nuria Farréa, ⁎ Miquel Querb, Xavier Leónb,c, a b c
Radiotherapy Oncology Department, Hospital de la Santa Creu i Sant Pau, Universitat Autònoma de Barcelona, Barcelona, Spain Otorhinolaryngology Department, Hospital de la Santa Creu i Sant Pau, Universitat Autònoma de Barcelona, Barcelona, Spain Centro de Investigación Biomédica en Red de Bioingeniería, Biomateriales y Nanomedicina (CIBER-BBN), Madrid, Spain
A R T I C L E I N F O
A B S T R A C T
Keywords: Extracapsular spread Lymph node ratio pN TNM classification Head and neck cancer HPV-negative
Objectives: The presence of nodes with extracapsular spread (ECS) and the lymph node ratio (LNR) have prognostic competence in the pathologic evaluation of patients with a head and neck squamous cell carcinoma (HNSCC) treated with a neck dissection. The purpose of this study is to assess the effect of ECS & LNR on prognosis of HPV negative HNSCC patients treated with neck dissection and to compare to 8th edition TNM/ AJCC classification. Materials and methods: We carried out a retrospective study of 1383 patients with HNSCC treated with a neck dissection between 1985 and 2013. We developed a classification of the patients according to the presence of nodes with ECS and the LNR value with a recursive partitioning analysis (RPA) model. Results: We obtained a classification tree with four terminal nodes: for patients without ECS (including patients pN0) the cut-off point for LNR was 1.6%, while for patients with lymph nodes with ECS it was 11.4%. The 5-year disease-specific survival for patients without ECS/LNR < 1.6% was 83.3%; for patients without ECS/ LNR ≥ 1.6% it was 61.5%; for patients with ECS/LNR < 11.4% it was 33.7%; and for patients with ECS/ LNR ≥ 11.4% it was 18.5%. The classification obtained with RPA had better discrimination between categories than the 8th edition of the TNM/AJCC classification. Conclusion: ECS status and LNR value proved high prognostic capacity in the pathological evaluation of the neck dissection. The combination of ECS and LNR improved the predictive capacity of the 8th edition of the TNM/ AJCC classification in HPV-negative HNSCC patients.
Introduction Lymph node status is one of the most important clinical predictors of survival for head and neck squamous cell carcinoma (HNSCC) patients. The standard pathological nodal staging (pN) of a neck dissection considers the number, size and location of positive lymph nodes. Several studies and meta-analysis show that the presence of lymph nodes with extracapsular spread (ECS), defined as extension of the tumor outside the lymph node capsule, negatively affects prognosis in HPV-negative HNSCC patients [1–4]. Interestingly, ECS did not affect survival in patients with HPV-positive oropharyngeal tumors [5,6]. These studies have led to the inclusion of the ECS into the pathological classification criteria in the 8th edition of the TNM/AJCC classification of HPV-negative patients [7,8]. The 8th edition TNM/AJCC
⁎
classification improves the differentiation in survival among the pN categories as well as the distribution of the number of patients per category in HNSCC HPV-negative patients when compared with the 7th edition TNM/AJCC [9]. To further improve the classification of nodal disease, several authors have analyzed the lymph node ratio (LNR). LNR is defined as the proportion of metastatic lymph nodes related to the total number of examined neck nodes. LNR has proved a very high prognostic capacity in neck dissection evaluation. High LNR values have consistently related to a worse overall and specific survival in most series (Supplementary material, Table 1). The objective of this study is to assess the prognostic competence of the pathological classification of the neck dissection obtained from evaluating ECS and LNR together and to compare this classification
Corresponding author at: Otorhinolaryngology Department, Hospital de la Santa Creu i Sant Pau, C/ Mas Casanovas, 90, 08041 Barcelona, Spain. E-mail address: [email protected] (X. León).
https://doi.org/10.1016/j.oraloncology.2017.12.010 Received 9 August 2017; Received in revised form 8 November 2017; Accepted 15 December 2017 1368-8375/ © 2017 Elsevier Ltd. All rights reserved.
Oral Oncology 77 (2018) 37–42
K. Majercakova et al.
study period. Patients with advanced tumor, either locally (pT3-T4) or regionally (pN2-N3), microscopically involved surgical margins, or ECS were considered candidates to adjuvant treatment. Postoperative radiotherapy was delivered in 2 Gy fractions to a total of 50 Gy in 5 weeks directed to both the primary site and the neck. In cases with ECS, a boost of up to 60–66 Gy was administered over the compromised areas. Cisplatin 100 mg/m2 was offered in selected cases with indication of postoperative radiotherapy from 2000 to present. The mean follow-up time was 5.7 years (standard deviation 5.4 years). During the follow-up period, 257 patients (18.6%) had local failure, 206 (14.9%) had regional failure, and 214 (15.15%) presented distant metastases. ECS was defined as any breach in the lymph node capsule caused by tumor cells. The pathological report of the neck dissections in our center did not include information about the microscopic or macroscopic character of the ECS. LNR was defined as the ratio between positive nodes and total number of removed nodes. The cut-off points for LNR were defined through recursive partitioning analysis (RPA) by the chi-square automatic interaction detection method, considering the specific survival as the dependent variable. Survival curves were estimated using the Kaplan-Meier actuarial method according to the presence of ECS and LNR. Differences in survival were compared using the log-rank test. We then performed a multivariate analysis (Cox proportional hazards model) including ECS and LNR as independent variables and disease-specific survival as the dependent variable. The RPA was repeated including both ECS and LNR in the model. Disease-specific survival was evaluated according to the categories obtained with this RPA. Finally, the prognostic capacity of the RPA model was compared with the 8th edition of the TNM/AJCC classification. We used measures of hazard discrimination and balance following the criteria defined by Groome et al. [53] to objectively compare the obtained RPA classification and the 8th edition of TNM/AJCC classification. Hazard discrimination measured how evenly the survival curves are spaced for each of the classification categories, and how large is the difference in survival between the best and the worst categories. The hazard discrimination ranges from zero to 1, where 1 represents an ideal classification with a full coverage of the survival area by evenly spaced curves. Balance quantifies the distribution of the number of patients for the classification scheme in each category. Balance ranges from zero to 1 and measures the adequacy to an ideal situation with the same number of patients in each category. The best classification system is the one with the highest balance. All procedures were reviewed by the Institutional Review Board at Santa Creu i Sant Pau Hospital. The study conforms to the principles outlined in the Declaration of Helsinki.
Table 1 Characteristics of the patients included in the study. Age (years)
Median 60.2/Standard deviation 11.1
Sex
Men Women
1251 (90.5%) 132 (9.5%)
Toxics consumption
Never Moderate Severe
85 (6.1%) 163 (11.8%) 1135 (82.1%)
Histological differentiation
Well Moderate Poor
135 (9.8%) 1074 (77.6%) 174 (12.6%)
Localization
Oral cavity Oropharynx Hypopharynx Supraglottic larynx Glottic larynx
316 250 214 375 228
(22.8%) (18.1%) (15.5%) (27.1%) (16.5%)
Local extension#
T1 T2 T3 T4
175 400 519 289
(12.7%) (28.9%) (37.5%) (20.9%)
Regional extension#
N0 N1 N2 N3
609 258 411 105
(44.0%) (18.7%) (29.7%) (7.6%)
#
According to 7th edition TNM classification.
with the 8th edition of the TNM/AJCC classification.
Material and methods We performed a retrospective study based on prospectively collected information of patients with HNSCC treated in our center [51]. A total of 1473 patients with HNSCC located in the oral cavity, oropharynx, hypopharynx, or larynx, diagnosed from 1985 until 2013, and treated with a unilateral or bilateral neck dissection were initially included in the study. HPV status in oropharyngeal tumors was analyzed retrospectively by HPV-DNA detection with SPF-10 real time PCR assay in combination with LiPA genotyping [52]. We excluded 24 patients with HPV-positive oropharyngeal cancer, 20 patients with selective neck dissection of only one lymph node area, 15 patients who lacked appropriate information about the pathological results of the neck dissection, and 31 patients who did not have a minimum follow-up of 2 years. Table 1 shows the characteristics of the 1383 patients finally included in the study. According to the interaction between tobacco and alcohol consumption, we created a combined variable of toxic consumption with three categories: no consumption; moderate consumption (< 20 cigarettes/day and/or < 80 g alcohol/day); and severe consumption (≥20 cigarettes/day and/or ≥80 g alcohol/day). We retrieved information concerning the type of neck dissection (unilateral or bilateral), the number of dissected nodes, the number of positive nodes, and the number of nodes with ECS for all patients. We performed 2140 neck dissections (469 radical neck dissections and 1671 selective neck dissections) in the patients included in the study. A total of 757 patients (54.7%) had bilateral neck dissections. In patients treated with a bilateral neck dissection, results were analyzed adding the neck nodes dissected on both sites of the neck. The mean number of lymph nodes studied per patient was 31.2 (standard deviation 19.0, range 7–118). In 201 cases (14.5%) we performed the neck dissections after a previous treatment with radiotherapy (n = 110) or chemoradiotherapy (n = 91). The interval between the radiotherapy or chemoradiotherapy and the neck dissection was 6–10 weeks (median, 8.5 weeks). A total of 751 patients (54.3%) had postoperative adjuvant treatment with radiotherapy (n = 677) or chemoradiotherapy (n = 74). The indications for adjuvant treatment were maintained throughout the
Results Five-year disease-specific survival for all the patients included in the study was 63.6% (95% CI: 60.9–66.3%). In 677 patients (49.0%) the pathological study of the neck dissection did not demonstrate the presence of lymph node metastases (pN0), 356 (25.7%) had lymph node metastases without extracapsular spread (pN+/ECS−) and 350 (25.3%) had lymph node metastases with extracapsular spread (pN +/ECS+). Fig. 1A shows the disease-specific survival curves related to the presence of lymph node metastases with ECS. The 5-year diseasespecific survival was 83.0% (95% CI: 80.1–85.9%) for pN0 patients, 62.3% (95% CI: 56.8–67.8%) for pN+/ECS− patients, and 25.7% (95% CI: 20.6–30.8%) for pN+/ECS+ patients. There were statistically significant differences in disease-specific survival regarding to the existence of lymph node metastases and ECS (P = .0001). The mean value of LNR of patients included in the study was 7.3% (standard deviation 13.6%, median 1.6%). In the subgroup of patients 38
Oral Oncology 77 (2018) 37–42
K. Majercakova et al.
survival was 83.3% (95% CI: 80.1–85.9%) for patients without ECS and LNR < 1.6%; 61.5% (95% CI: 56.0–67.0%) for patients without ECS and LNR ≥ 1.6%; 33.7% (95% CI: 25.7–41.7%) for patients with ECS and LNR < 11.4%; and 18.5% (95% CI: 12.2–24.8%) for patients with ECS and LNR ≥ 11.4%. These differences were statistically significant (P = .0001). Fig. 3B shows the disease-specific survival curves for pN according to the 8th edition of the TNM/AJCC classification. The 5-year diseasespecific survival figures read as follows: pN0 83.0% (95% CI: 80.1–85.9%), pN1 69.1% (95% CI: 61.5–76.7%), pN2 57.0% (95% CI: 49.7–64.3%), and pN3 22.2% (95% CI: 17.1–27.3%) (P = .0001). The comparison between both methods showed that the classification obtained with RPA had better capacity of discrimination between categories than the 8th edition of the TNM/AJCC classification. The value of hazard discrimination for the RPA was 0.868 compared to the value of 0.653 for the 8th edition of the TNM/AJCC classification (differences in hazard discrimination value of 0.215 in favor of the RPA classification). The distribution of patients per categories was very similar for both classifications. The balance value for RPA was 0.671 versus 0.680 for the 8ª edition of the TNM/AJCC classification (differences in balance value of 0.009 in favor of the TNM/AJCC classification). Table 2 shows 5-year local-, regional- and distant-recurrence-free survival for every category in both classifications. Additionally, it shows the values of hazard discrimination corresponding to every survival study. With the increasing burden of disease in our classification, a higher risk of local, regional or distant failure was observed. The hazard discrimination scores for regional and distant control were superior for the RPA classification, while the hazard discrimination score for local control favored the 8th edition of the TNM/AJCC classification. The supplementary material shows the local and regional recurrence-free survival and the distant metastases-free survival curves according to both classification methods.
Fig. 1. Disease-specific survival curves according to the presence of neck node metastases with extracapsular spread (ECS) (A) and to the lymph node ratio (LNR) category (B).
with pN+, the mean value of LNR was 13.3% (standard deviation 16.1%, median 8.7%). The RPA defined two cut-off points for LNR. The low cut-off point was set in 1.6% and the high cut-off point in 11.4%, defining three categories with significant differences in disease-specific survival. Fig. 1B shows disease-specific survival for the categories defined by these cut-off points. The 5-year disease-specific survival was 83.0% (95% CI: 80.1–85.9%) for patients with LNR < 1.6% (n = 692, 50%), 52.3% (95% CI: 47.0–57.6%) for patients with LNR 1.6–11.4% (n = 415, 30.0%), and 29.8% (95% CI: 23.9–35.7%) for patients with LNR > 11.4% (n = 276, 20.0%). These differences reached statistical significance (P = .0001). We carried out a multivariate study considering the disease-specific survival as dependent variable, including ECS and LNR category as independent variables. According to the results of this model, both variables showed prognostic capacity. Patients with one or more lymph nodes with ECS had 2.45 times (95% CI: 1.97–3.06) higher risk of disease-specific mortality compared to patients without evidence of ECS (P = .0001). Similarly, patients with intermediate LNR had 2.55 (95% CI: 1.97–3.31) higher risk than patients with low LNR, and patients with high LNR had 3.98 times (95% CI: 2.98–5.31) higher risk of disease-specific mortality than patients with low LNR. All these differences reached statistical significance (P = .0001). We performed a RPA considering disease-specific survival as dependent variable. ECS and LNR value were included as independent variables. A classification tree with four terminal nodes was obtained (Fig. 2). Interestingly, the same cut-off points for LNR delineated the terminal nodes. In patients without ECS (including pN0 patients) the corresponding cut-off point for LNR was 1.6%, while for patients with lymph nodes with ECS it was 11.4%. Fig. 3A shows the disease-specific survival curves for the terminal nodes obtained with the RPA model. The 5-year disease-specific
Discussion According to our results, the pathological classification of neck dissections in patients with HPV-negative HNSCC based in ECS and LNR value shows better prognostic capacity than the 8th edition of the TNM/AJCC classification [7,8]. Strong evidence in the literature supports the role of ECS as a negative prognostic factor for final survival. The results of a meta-analysis by Dünne et al. [3] showed that the presence of ECS in neck dissections had a negative impact on survival, with a summarized odds ratio of 2.7 (95% CI, 2.2–3.4). Additionally, a recent systematic review of the literature and meta-analysis carried out by Mermod et al. [4] confirmed the impact of ECS on loco-regional recurrence and distant metastases in HPV-negative HNSCC patients. Interestingly, the presence of ECS in patients with HPV-positive oropharyngeal tumors did not affect prognosis [5,6]. According to our results, there were statistically significant differences in disease-specific survival related to the presence of lymph node metastases with ECS in HPV-negative HNSCC patients, as shown in Fig. 1A. More recently, the LNR has emerged as a new an independent prognostic factor in the pathological evaluation of neck dissections. Early publications in 2009 by Shrime et al. [10,11] and Gil et al. [12] demonstrated the prognostic capacity of LNR in patients with oral cavity carcinomas. Following these data, multiple studies proved the relationship between the LNR and survival or disease control of HNSCC patients in different localizations (Supplementary material, Table 1). The LNR weighs three factors that can potentially influence nodal staging: (1) surgical factors: the comprehensiveness of the neck dissection measured as the total number of lymph nodes removed; (2) tumor factors: the number of positive neck nodes; and (3) staging factors: the completeness of the pathologic analysis. 39
Oral Oncology 77 (2018) 37–42
K. Majercakova et al.
Fig. 2. Recursive partitioning analysis classification according to the presence of lymph node metastases with extracapsular spread (ECS) and lymph node ratio (LNR) value (n, number of patients included in each node of the classification tree; the percentage in the pies indicates the proportion of patients who died as a consequence of the tumor).
distribution [19], ROC analysis [20,21,31,34,35,45], the minimum p value from the log-rank test [17,23,25,27,28,39,41,43,44,46,48,49], principal component analysis [18,50], the maximally selected rank statistic method [10,11,24,37], a graphic determination from a logarithmic transformation of data [14], or previously published thresholds [22,42,47]. In the present study, we used a RPA method to select the cut-off points for our model. According the results of this model, the cut-off points defining populations with different prognosis were: LNR < 1.6% (including patients with pN0), LNR 1.6–11.4% and LNR > 11.4%. The studies that analyze both ECS and the LNR in multivariate analysis led to conflicting results. For some authors, the inclusion of LNR in the model leads to a loss of the independent prognostic capacity of ECS [10,12,14,15,27,31,32,34,43–46,48]. However, others consider that both ECS and LNR retain prognostic significance in multivariate studies [16,17,21,34,36,38,40,48]. According to our results, both ECS and LNR have an independent prognostic capacity in a multivariate study, which justifies their inclusion in the RPA model. The main advantage of RPA as a classification method is that it defines specific cut-off points that optimize the possibility of discrimination in different clinical situations. In our sample, the RPA algorithm defined a first level of classification based on the presence or the absence of neck nodes with ECS. The second level of classification defined a specific cut-off point for LNR for every ECS category. In patients without ECS the value of LNR with discrimination capacity in disease-specific survival was 1.6%, and in patients with ECS it corresponded to 11.4%. The comparison of the RPA classification with the 8th edition of the TNM/AJCC classification yielded satisfactory results in the homogeneous distribution of the differences in survival between categories (hazard discrimination) in favor of the RPA. The distribution of the number of patients per category was very similar in both classifications. One of the main goals of a pathological classification of the neck dissection is to provide prognostic data on the risk of regional recurrence. This level of risk has a direct impact in the planning of adjuvant treatment and its intensity. The classification based on RPA showed higher values of hazard discrimination in regional and distant-metastases-free survival than the 8th edition pathological TNM/AJCC in HPV-negative HNSCC patients. These results indicate that this method has better prognostic capacity evaluating the regional and distant disease control. One of the limitations of our study is that we do not have
Fig. 3. Disease-specific survival according to the recursive partitioning analysis classification (A) and to pN category as defined by the 8th edition TNM/AJCC classification (B).
The cut-off points for LNR values oscillate between 2.5% and 20% in different studies (Supplementary material, Table 1). These differences are due to the heterogeneity of patients between studies. While some series analyze HNSCC with a specific location, others include tumors from all head and neck locations. Some series only include patients with pathological lymph node metastases (pN+) while others include also patients without nodal disease (pN0). Besides, the methods of determination of cut-off points are different between studies. Some studies uses the median value of the LNR [12,13,15,26,29,30,32,33], quartile 40
Oral Oncology 77 (2018) 37–42
K. Majercakova et al.
Table 2 Five-year local (rT), regional (rN), and distant (rM) recurrence-free survival and hazard discrimination values according to the recursive partitioning analysis (RPA) classification and the 8th edition of the TNM classification.
RPA
ECS−/LNR ECS−/LNR ECS+/LNR ECS+/LNR
< > < >
1.6% 1.6% 11.4% 11.4%
Hazard discrimination 8th TNM
rT (CI 95%)
rN (CI 95%)
rM (CI 95%)
87.3% 73.7% 62.8% 59.0%
93.6% 85.1% 65.2% 48.6%
92.6% 81.9% 55.3% 47.5%
(84.8–89.8%) (68.7–78.7%) (52.9–72.7%) (48.9–69.1%)
0.517 pN0 pN1 pN2 pN3
Hazard discrimination
87.2% 77.5% 71.0% 59.9%
(91.7–95.5%) (80.8–89.4%) (56.6–73.8%) (39.2–58.0%)
0.647 (84.5–89.9%) (70.5–84.5%) (64.4–77.6%) (52.5–67.3%)
0.653
93.3% 88.2% 83.8% 53.4% 0.208
(90.5–94.7%) (77.4–86.4%) (45.4–65.2%) (37.6–57.4%)
0.662 (91.4–95.2%) (83.0–87.4%) (78.2–89.4%) (46.6–30.2%)
92.5% 87.2% 79.2% 46.1%
(90.4–94.6%) (81.8–92.6%) (73.0–85.4%) (38.5–53.7%)
0.328
Abbreviations: ECS, extracapsular spread; LNR, lymph node ratio.
information about the microscopic or macroscopic character of the ECS. Besides, this study is a retrospective evaluation of patients treated at a single institution. Our population of study is mainly composed of heavy smokers and drinkers (82% of our patients), with a high proportion of laryngeal tumors (43.6% of patients). This limits the extension of our conclusions and its utility for generalized use across all tumor types and patient populations, and calls for further studies to validate the prognostic capacity of the joint assessment of the presence of nodes with ECS and the LNR as a pathological criterion in the evaluation of neck dissections in HNSCC patients.
[5] Haughey BH, Sinha P. Prognostic factors and survival unique to surgically treated p16+ oropharyngeal cancer. Laryngoscope 2012;122(Suppl 2):S13–33. [6] Kharytaniuk N, Molony P, Boyle S, O'Leary G, Werner R, Heffron C, et al. Association of extracapsular spread with survival according to human papillomavirus status in oropharynx squamous cell carcinoma and carcinoma of unknown primary site. JAMA Otolaryngol Head Neck Surg 2016;142:683–90. [7] UICC International Union Against Cancer. TNM classification of malignant tumors. In: Brierley JD, Gospodarowicz MK, Wittekind Ch, editors. 8th ed., Wiley-Blackwell; 2016. [8] AJCC Cancer Staging Manual. In: Amin MB, Edge S, Greene F, Byrd DR, Brookland RK, Washington MK, et al., editors. 8th ed. Springer International Publishing; 2017. [9] García J, López M, López L, Bagué S, Granell E, Quer M, et al. Validation of the pathological classification of lymph node metastasis for head and neck tumors according to the 8th edition of the TNM Classification of Malignant Tumors. Oral Oncol 2017;70:29–33. [10] Shrime MG, Bachar G, Lea J, Volling C, Ma C, Gullane PJ, et al. Nodal ratio as an independent predictor of survival in squamous cell carcinoma of the oral cavity. Head Neck 2009;31:1482–8. [11] Shrime MG, Ma C, Gullane PJ, Gilbert RW, Irish JC, Brown DH, et al. Impact of nodal ratio on survival in squamous cell carcinoma of the oral cavity. Head Neck 2009;31:1129–36. [12] Gil Z, Carlson DL, Boyle JO, Kraus DH, Shah JP, Shaha AR, et al. Lymph node density is a significant predictor of outcome in patients with oral cancer. Cancer 2009;115:5700–10. [13] Süslü N, Hoşal AS, Sözeri B. Prognostic value of metastatic lymph node ratio in node-positive head and neck carcinomas. Am J Otolaryngol 2010;31:315–9. [14] Ebrahimi A, Clark JR, Zhang WJ, Elliott MS, Gao K, Milross CG, et al. Lymph node ratio as an independent prognostic factor in oral squamous cell carcinoma. Head Neck 2011;33:1245–51. [15] Kim SY, Nam SY, Choi SH, Cho KJ, Roh JL. Prognostic value of lymph node density in node-positive patients with oral squamous cell carcinoma. Ann Surg Oncol 2011;18:2310–7. [16] Lanzer M, Kruse A, Lübbers HT, Zemann W, Reinisch S. Lymph node ratio and capsule penetration as independent risk factors in head and neck squamous cell carcinoma. Head Neck Oncol 2012;4:89. [17] Liao CT, Hsueh C, Lee LY, Lin CY, Fan KH, Wang HM, et al. Neck dissection field and lymph node density predict prognosis in patients with oral cavity cancer and pathological node metastases treated with adjuvant therapy. Oral Oncol 2012;48:329–36. [18] Kim KY, Cha IH. Risk stratification of oral cancer patients using a combined prognostic factor including lymph node density and biomarker. J Cancer Res Clin Oncol 2012;138:483–90. [19] Yu Y, Wang XL, Xu ZG, Fan CC, Li Q. Prognostic value of lymph node ratio in hypopharyngeal squamous cell carcinoma after chemoradiotherapy. Chin Med J (Engl) 2013;126:4139–44. [20] Patel SG, Amit M, Yen TC, Liao CT, Chaturvedi P, Agarwal JP, et al. Lymph node density in oral cavity cancer: results of the International Consortium for Outcomes Research. Br J Cancer 2013;15(109):2087–95. [21] Sayed SI, Sharma S, Rane P, Vaishampayan S, Talole S, Chaturvedi P, et al. Can metastatic lymph node ratio (LNR) predict survival in oral cavity cancer patients? J Surg Oncol 2013;108:256–63. [22] Urban D, Gluck I, Pfeffer MR, Symon Z, Lawrence YR. Lymph node ratio predicts the benefit of post-operative radiotherapy in oral cavity cancer. Radiother Oncol 2013;106:74–9. [23] Mizrachi A, Hadar T, Rabinovics N, Shpitzer T, Guttman D, Feinmesser R, et al. Prognostic significance of nodal ratio in cutaneous squamous cell carcinoma of the head and neck. Eur Arch Otorhinolaryngol 2013;270:647–53. [24] Reinisch S, Kruse A, Bredell M, Lübbers HT, Gander T, Lanzer M. Is lymph-node ratio a superior predictor than lymph node status for recurrence-free and overall survival in patients with head and neck squamous cell carcinoma? Ann Surg Oncol 2014;21:1912–8. [25] Rudra S, Spiotto MT, Witt ME, Blair EA, Stenson K, Haraf DJ. Lymph node density–prognostic value in head and neck cancer. Head Neck 2014;36:266–72.
Conclusion Besides the classical pathological criteria of nodal metastases, ECS status and LNR value proved elevated prognostic capacity in the evaluation of patients with HPV-negative HNSCC treated with neck dissection. The combination of ECS and LNR improved the predictive capacity of the 8th edition of the TNM/AJCC classification. Conflict of interest statement All the authors certify that there is no conflict of interest with any financial organization regarding the material discussed in the manuscript. Acknowledgements This work was supported by a grant from the Instituto de Salud Carlos III – Spain (FIS PI14/01918). The study was cofunded by Fondo Europeo de Desarrollo Regional (FEDER), A Way to Build Europe. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.oraloncology.2017.12. 010. References [1] de Juan J, García J, López M, Orús C, Esteller E, Quer M, et al. Inclusion of extracapsular spread in the pTNM classification system: a proposal for patients with head and neck carcinoma. JAMA Otolaryngol Head Neck Surg 2013;139:483–8. [2] Wreesmann VB, Katabi N, Palmer FL, Montero PH, Migliacci JC, Gönen M, et al. Influence of extracapsular nodal spread extent on prognosis of oral squamous cell carcinoma. Head Neck 2016;38:1192–9. [3] Dünne AA, Müller HH, Eisele DW, Kessel K, Moll R, Werner JA. Meta-analysis of the prognostic significance of perinodal spread in head and neck squamous cell carcinomas (HNSCC) patients. Eur J Cancer 2006;42:1863–8. [4] Mermod M, Tolstonog G, Simon C, Monnier Y. Extracapsular spread in head and neck squamous cell carcinoma: a systematic review and meta-analysis. Oral Oncol 2016;62:60–71.
41
Oral Oncology 77 (2018) 37–42
K. Majercakova et al.
Arch 2016;469:635–41. [41] Yildiz MM, Petersen I, Eigendorff E, Schlattmann P, Guntinas-Lichius O. Which is the most suitable lymph node predictor for overall survival after primary surgery of head and neck cancer: pN, the number or the ratio of positive lymph nodes, or log odds? J Cancer Res Clin Oncol 2016;142:885–93. [42] Roberts TJ, Colevas AD, Hara W, Holsinger FC, Oakley-Girvan I, Divi V. Number of positive nodes is superior to the lymph node ratio and American Joint Committee on Cancer N staging for the prognosis of surgically treated head and neck squamous cell carcinomas. Cancer 2016;122:1388–97. [43] Suzuki H, Matoba T, Hanai N, Nishikawa D, Fukuda Y, Koide Y, et al. Lymph node ratio predicts survival in hypopharyngeal cancer with positive lymph node metastasis. Eur Arch Otorhinolaryngol 2016;273:4595–600. [44] Imre A, Pinar E, Dincer E, Ozkul Y, Aslan H, Songu M, et al. Lymph node density in node-positive laryngeal carcinoma: analysis of prognostic value for survival. Otolaryngol Head Neck Surg 2016;155:797–804. [45] Park JO, Joo YH, Cho KJ, Kim MS. Lymph node density as an independent prognostic factor in node-positive patients with tonsillar cancer. Head Neck 2016;38(Suppl 1):E705–11. [46] Lieng H, Gebski VJ, Morgan GJ, Veness MJ. Important prognostic significance of lymph node density in patients with node positive oral tongue cancer. ANZ J Surg 2016;86:681–6. [47] Ong W, Zhao R, Lui B, Tan W, Ebrahimi A, Clark JR, et al. Prognostic significance of lymph node density in squamous cell carcinoma of the tongue. Head Neck 2016;38(Suppl 1):E859–66. [48] Suzuki H, Beppu S, Hanai N, Hirakawa H, Hasegawa Y. Lymph node density predicts lung metastases in oral squamous cell carcinoma. Br J Oral Maxillofac Surg 2016;54:213–8. [49] Tseros EA, Gebski V, Morgan GJ, Veness MJ. Prognostic significance of lymph node ratio in metastatic cutaneous squamous cell carcinoma of the head and neck. Ann Surg Oncol 2016;23:1693–8. [50] Kim KY, Zhang X, Kim SM, Lee BD, Cha IH. A combined prognostic factor for improved risk stratification of patients with oral cancer. Oral Dis 2017;23:91–6. [51] León X, Orús C, Quer M. Design, maintenance, and exploitation of an oncologic database for patients with malignant tumors of the head and neck. Acta Otorrinolaringol Esp 2002;53:185–90. [52] Castellsagué X, Alemany L, Quer M, Halec G, Quirós B, Tous S, et al. HPV involvement in head and neck cancers: comprehensive assessment of biomarkers in 3680 patients. J Natl Cancer Inst 2016;108:djv403. [53] Groome PA, Schulze K, Boysen M, Hall SF, Mackillop WJ. A comparison of published head and neck stage groupings in carcinomas of the oral cavity. Head Neck 2001;23:613–24.
[26] Ampil FL, Caldito G, Ghali GE. Can the lymph node ratio predict outcome in head and neck cancer with single metastasis positive-node? Oral Oncol 2014;50:e18–20. [27] Prabhu RS, Hanasoge S, Magliocca KR, Hall WA, Chen SA, Higgins KA, et al. Lymph node ratio influence on risk of head and neck cancer locoregional recurrence after initial surgical resection: implications for adjuvant therapy. Head Neck 2015;37:777–82. [28] Wang YL, Li DS, Wang Y, Wang ZY, Ji QH. Lymph node ratio for postoperative staging of laryngeal squamous cell carcinoma with lymph node metastasis. PLoS ONE 2014;9:e87037. [29] Künzel J, Mantsopoulos K, Psychogios G, Grundtner P, Koch M, Iro H. Lymph node ratio as a valuable additional predictor of outcome in selected patients with oral cavity cancer. Oral Surg Oral Med Oral Pathol Oral Radiol 2014;117:677–84. [30] Künzel J, Psychogios G, Mantsopoulos K, Grundtner P, Waldfahrer F, Iro H. Lymph node ratio as a predictor of outcome in patients with oropharyngeal cancer. Eur Arch Otorhinolaryngol 2014;271:1171–80. [31] Park GC, Jung JH, Roh JL, Lee JH, Cho KJ, Choi SH, et al. Prognostic value of metastatic nodal volume and lymph node ratio in patients with cervical lymph node metastases from an unknown primary tumor. Oncology 2014;86:170–6. [32] Chen CC, Lin JC, Chen KW. Lymph node ratio as a prognostic factor in head and neck cancer patients. Radiat Oncol 2015;10:181. [33] Künzel J, Mantsopoulos K, Psychogios G, Agaimy A, Grundtner P, Koch M, et al. Lymph node ratio is of limited value for the decision-making process in the treatment of patients with laryngeal cancer. Eur Arch Otorhinolaryngol 2015;272:453–61. [34] Ryu IS, Roh JL, Cho KJ, Choi SH, Nam SY, Kim SY. Lymph node density as an independent predictor of cancer-specific mortality in patients with lymph nodepositive laryngeal squamous cell carcinoma after laryngectomy. Head Neck 2015;37:1319–25. [35] Joo YH, Cho KJ, Kim SY, Kim MS. Prognostic significance of lymph node density in patients with hypopharyngeal squamous cell carcinoma. Ann Surg Oncol 2015;22(Suppl 3):S1014–9. [36] Hua YH, Hu QY, Piao YF, Tang Q, Fu ZF. Effect of number and ratio of positive lymph nodes in hypopharyngeal cancer. Head Neck 2015;37:111–6. [37] Chan RC, Chan JY. Impact of nodal ratio on survival in recurrent nasopharyngeal carcinoma. Head Neck 2015;37:12–7. [38] Hong HR, Roh JL, Cho KJ, Choi SH, Nam SY, Kim SY. Prognostic value of lymph node density in high-grade salivary gland cancers. J Surg Oncol 2015;111:784–9. [39] Suzuki H, Hanai N, Hirakawa H, Nishikawa D, Hasegawa Y. Lymph node density is a prognostic factor in patients with major salivary gland carcinoma. Oncol Lett 2015;10:3523–8. [40] de Ridder M, Marres CC, Smeele LE, van den Brekel MW, Hauptmann M, Balm AJ, et al. A critical evaluation of lymph node ratio in head and neck cancer. Virchows
42