Reginal Lymph Node Involvement is Associated With Poorer 1 Survivorship in Patients With Malignant Bone Tumor

of children, their parent or legal guardian. Background: Regional lymph node involvement is rare in patients with malignant bone 70 tumors. We aimed to analyze the prevalence and prognostic implications in the lymph- 71 involved patients with malignant bone tumors. 72 Materials and Methods: From 1988 to 2016, 9582 patients with primary malignant 73 bone tumors in the SEER database were enrolled. Overall survival (OS) was computed 74 using the Kaplan-Meier method. A multivariate analysis was performed using the Cox 75 regression. 76 Results: 346 (3.63%) patients exhibited regional lymph node involvement. Lymph- 77 involved patients had larger tumor size and more metastasis than patients without. 78 Ewing sarcoma most frequently developed lymph node involvement. Lymph-involved 79 patients (0.31 95% CI (0.26–0.37)) had lower 5-year OS rates than patients without 80 lymph node involvement (0.66 95% CI (0.65–0.68)) (p<0.001). 81 Conclusions: Lymph node involvement is rare in patients with malignant bone tumors. 82 They had a large tumor size, more distant metastasis and poor survival.


84
Primary malignant bone tumors constitute a heterogeneous group of rare solid 85 tumors, and the most prevalent types are osteosarcoma, chondrosarcoma and Ewing 86 sarcoma. The metastatic pattern in primary malignant bone tumor is typically 87 hematogenous, with the lung and bone as the most prevalent metastatic sites (1,2). 88 Although regional lymph node involvement is thought to be a relatively uncommon 89 5 event in patients with primary bone malignancies, it was reported to be associated with 90 a poor survival in several studies of patients with various histological subtypes of 91 tumors, such as osteosarcoma and chondrosarcoma (3-6). However, the overall 92 proportion of regional lymph node involvement and its effect on the survival of patients 93 with primary malignant bone tumors remain unclear. 94 Efforts to understand the percentage of patients with primary bone malignancies 95 who develop regional nodal involvement and the histological subtypes that are more or 96 less likely to undergo lymphatic spread may influence the decision to require the patient 97 to undergo further advanced imaging tests, such as PET-CT, or regional lymph node 98 biopsy. The status of the regional lymph nodes may alter the prognosis by substantially 99 guiding the overall treatment strategy and potentially altering the range of resection or 100 surgical method. However, currently, the process has not been systematically evaluated. 101 Challenges include the low disease incidence and heterogeneity of histological subtypes. 102 Given the paucity of data describing these patients, we completed the current 103 population-based study of this rare group of patients with primary malignant bone 104 tumors and regional lymph node involvement. We provide a description of the 105 prevalence, risk factors and effects on survival for this unique group of patients.  108 We accessed the US National Cancer Institute Surveillance Epidemiology and End 109 Results (SEER) database, which collects and publishes cancer incidence and survival 110 data (7), using the SEER*Stat version 8.3.5 (National Cancer Institute, Bethesda, MD, 111 6 USA) to obtain the data from patients with primary malignant bone tumors, including 112 clinical characteristics and outcomes and to answer the questions posed above.

114
The inclusion criteria were a histologically confirmed diagnosis of primary 115 malignant bone tumors from 1988 to 2016. The diagnosis was made when the patient 116 was alive, as we excluded a diagnosis determined at autopsy or a death certificate only.

117
Patients who were not diagnosed with bone malignancies as the first primary 118 malignancy and patients without available information on the status of regional lymph 119 node involvement were also excluded ( Figure 1). who met the inclusion criteria were included in this study. Then, we separated all 138 patients into two groups according to the status of regional lymph node involvement.  two groups depending on the status of regional lymph node involvement. The chi-147 square test and Fisher's exact test were used to identify the differences in categorical 148 variables between the two groups. A Kaplan-Meier curve was constructed to estimate 149 the overall survival, and a log-rank test was applied to identify the significance of 150 differences in survival between groups. A Cox proportional hazard regression analysis 151 was performed to calculate the hazard ratios with 95% confidence intervals and to 152 determine the effect of regional lymph node involvement on the prognosis, while 153 controlling for the effects of age, sex, race, tumor size, tumor location, tumor grade, 154 distant metastasis. Statistical significance was determined using an α error of 0.05. patients were reported to have regional lymph node involvement, while 9236 (96.39%) 165 did not (Table 1). 166 3.2 Differences in characteristics between patients with and without regional 167 lymph node involvement 168 Briefly, patients with regional lymph node involvement were more likely to have  Table 2). 177 9 3.3 Association between histological subtypes and the prevalence of regional lymph 178 node involvement 179 The incidence of regional lymph node involvement varies among histological 180 subtypes. The prevalence of regional lymph node involvement was 8 with chordoma. Regional lymph node involvement was not noted in 13 subtypes of  (Table 3). 193 We observed a clear association between the histological subtype and the 194 prevalence of regional lymph node involvement. Compared with the overall proportion 195 of patients presenting with regional lymph node involvement (  3.4 Relationship between regional lymph node involvement and the prognosis 203 Overall survival was worse for patients with primary malignant bone tumors 204 presenting with regional lymph node involvement than patients without regional nodal 205 disease (p<0.001). The estimated 5-year overall survival rates of patients with and 206 without regional lymph node involvement were 31% (95% CI, 26-37%) and 66% (95% 207 CI, 65-68%), respectively (Figure 2A). 208 A subgroup analysis also confirmed these results. For patients without distant 209 metastasis, regional lymph node involvement was also associated with a worse 5-year 210 survival: 42% for patients with regional node involvement (95% CI, 34%-52%) and 72% 211 for patients without lymph node involvement (95% CI, 71-73%; p < 0.001) ( Figure   212 2B). For patients presenting with metastatic disease, regional lymph node involvement 213 was associated with a worse 5-year overall survival; the estimated rate was 20% for 214 patients with regional lymph node involvement (95% CI, 12-28%) and 29% for patients 215 without this presentation (95% CI, 25-32%; p=0.0061) ( Figure 2C). In patients with 216 osteosarcoma, the 5-year overall survival was 16% (95% CI, 9-32%) for patients with 217 regional lymph node involvement and 63% (95% CI, 58-63%; p < 0.001) for patients 218 without this presentation ( Figure 2D). The 5-year overall survival rate was 43% (95% 219 CI, 33%-55%) for patients with Ewing sarcoma presenting with regional lymph node 220 involvement and 65% (95% CI, 61%-68%; p < 0.001) for patients without this 221 11 presentation ( Figure 2E). In patients with malignant fibrous histiocytoma, the 5-year 222 overall survival rate was 22% (95% CI, 21-31%) for patients with regional lymph node 223 involvement and 60% (95% CI, 57%-61%; p < 0.001) for patients without regional 224 lymph node involvement ( Figure 2F). 225 Next, we built a multivariate Cox proportional hazard model to assess the 226 independent effect of regional node involvement on the overall survival. Based on the 227 results of the univariate analysis, covariates included in the Cox model were age, sex, 228 race, tumor size, primary tumor site, distant metastasis, grade and histological type. All 229 covariates, except race and primary tumor site, met the proportional hazard assumption.

230
The estimated hazard ratio (HR) of death in patients with regional lymph node 231 involvement was 1.74 (95% CI, 1.48-2.05; p < 0.001) compared with patients without 232 regional lymph node involvement (Figure 3). among others (8-12). Regional lymph node involvement is thought to be rare in patients 239 with primary malignant bone tumors, although the overall prevalence of nodal 240 involvement is poorly defined in the existing reports (13, 14). Likewise, our systematic 241 understanding of the effect of regional lymph node involvement on primary bone 242 malignancies is also limited, and we were only able to identify several limited studies 243 12 on the survival of patients with osteosarcoma, chondrosarcoma and Ewing sarcoma 244 presenting with and without lymph node metastasis (15-17). As shown in the current 245 study, 3.63% of all patients with primary malignant bone tumors presented with 246 regional lymph node involvement. Nodal involvement was associated with the 247 histological subtype. More importantly, regional lymph node involvement  Consistent with previous studies, our results reveal that regional lymph node 251 involvement is rare in patients with primary malignant bone tumors, with a prevalence 252 of 3.63%. Previous studies report a prevalence of nodal involvement in osteosarcoma 253 ranging from <1% to 10% (16). An analysis of chondrosarcoma based on the SEER 254 database reviewed a prevalence of 1.3% (15). The metastatic pattern of primary 255 malignant bone tumors is typically hematogenous, and the reason for the rarity of 256 regional lymph node involvement has not been clearly clarified to date. Some 257 researchers postulated that the paucity of lymphatic channels in bone may explain the 258 rare incidence of regional lymph node involvement (18).

259
Based on our findings, patients with regional lymph node involvement were more likely 260 to have a larger tumor size, a higher percentage of distant metastasis and to be diagnosed 261 with Ewing sarcoma than patients without regional lymph node involvement. The 262 involvement of regional lymph nodes is a signal of the extent of tumor invasion.

263
Therefore, as a marker of tumor invasion, the observation of a larger tumor size in 264 patients with regional nodal involvement is reasonable. Primary malignant bone tumors 265 13 typically display hematogenous metastasis. As a significant adverse factor, the 266 consistent appearance of distant metastasis with regional nodal involvement is rational. 267 We reported substantial variation in the risk of regional lymph node involvement among 268 patients with different histological subtypes of primary malignant bone tumors. Ewing 269 sarcoma (8.81%) was associated with a higher risk of regional lymph node involvement 270 than other subtypes of bone malignancies, followed by myxoid chondrosarcoma (4.49%)  Based on the results from the present study, patients with primary malignant bone 286 tumors and regional lymph node involvement have inferior overall survival outcomes. 287 14 Regional nodal involvement was previously reported to be an adverse prognostic factor 288 for several histological subtypes of primary bone malignancies, such as osteosarcoma 289 and chondrosarcoma (15, 16). Our study confirmed the adverse effect of regional lymph 290 node involvement patients with primary malignant bone tumors on overall survival, 291 including all common histological subtypes. Although the presence of regional nodal 292 involvement is rare, the effect on the prognosis of patients with all subtypes of bone 293 malignancies should not be ignored. Since an examination of regional lymph nodes is 294 not included in the routine evaluation of primary malignant bone tumors, we 295 recommend PET/CT imaging as an effective method to screening for lymph node 296 involvement (22, 23). Biopsy of suspicious regional lymph nodes may be needed for 297 confirmation when planning further treatment.

298
The analysis of data from the SEER database has several limitations. This study is 299 limited by its retrospective approach. Second, the SEER database did not provide 300 detailed information about the treatment of patients, including the management of 301 regional lymph node metastasis and surgical methods. Therefore, bias may exist in the 302 survival analysis. Additionally, a detailed analysis of recommendations for patients with 303 regional lymph node involvement was impossible. Moreover, the status of lymph node 304 involvement in patients included in the SEER database was determined by either 305 clinical, surgical or pathologic adjudication. Therefore, the actual prevalence of 306 regional lymph node involvement was unable to be determined. 307 We conclude that the prevalence of regional lymph node involvement in patients 308 with primary malignant bone tumors included in the SEER database was 3.63%.

15
Patients with nodal involvement were more likely to have a large tumor size and distant 310 metastasis. Ewing sarcoma was associated with a higher risk of regional lymph node 311 involvement than other subtypes of bone malignancies. Patients with regional nodal 312 involvement exhibit a poorer survival than patients without regional lymph node 313 involvement. The importance of a regional lymph node evaluation in patients with 314 primary malignant bone tumors might currently be underestimated. Based on the 315 independent association between regional lymph node involvement and poor survival, 316 we suggest a cautious assessment of the status of regional nodal involvement in patients 317 with primary bone malignancies. The mechanism of lymph node involvement, 318 association of this finding with a poor prognosis, and recommended management of 319 invaded lymph nodes require further investigation.