Survival and prognostic factor analyses in malignant giant cell tumour of bone

Background: The characteristics and survival of patients with malignant giant cell tumour of bone (GCTB) have not been investigated thoroughly due to the rarity of the disease. We evaluated these factors in a large cohort in the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) database. Methods: Data from patients who were diagnosed with malignant GCTB from 1975 to 2016 were extracted from the SEER database. The overall survival (OS) was calculated by Kaplan– Meier analysis, and intergroup differences were tested by the log-rank test. Univariate and multivariate Cox proportional hazard regression analyses were conducted to identify the independent survival factors. Results: A total of 325 patients with malignant GCTB were included. The overall 1-, 5, and 10-year survival rates were 94.3% (95% CI: 91.7-96.8), 82.3% (95% CI: 77.9-86.6), and 80.1% (95% CI: 75.4-84.7), respectively. A potential non-linear J-shaped dose–response relationship between the age or diagnosis year and survival. Multivariate Cox regression showed poor survival in patients with age from 35 to 60 years (hazard ratio (HR) =9.99, 95% CI: 1.34-74.80, P =0.025), age older than 60 years (HR=62.03, 95% CI: 7.94-484.38, P <0.001), with stage T2 disease (HR=4.85, 95% CI: 1.52-15.47, P =0.008), with stage T3 disease (HR=6.09, 95% CI: 1.03-36.23, P =0.047), and with distant tumours (HR=2.76, 95% CI: 1.14-6.65, P =0.024), and extraskeletal sites (HR=3.33, 95% CI: 1.02-10.85, P =0.046). Conclusions: This large population-based series described the clinical characteristics of malignant GCTB. Patients with age >34 years, stage T2/3

It is important to perform survival prediction to determine individual treatments. However, the issues have not been intensively studied due to the lack of clinical long-term follow-up data [9]. Itkin  This database has been extended, and some new factors have been added, including TNM stages, updated AJCC grade, insurance status, and records of specific metastatic organs. The updated data should be re-evaluated to provide further understanding of malignant GCTB. In the present study, based on the updated data from the database, we attempted to evaluate the survival of patients with malignant GCTB and to identify prognostic factors.

Data Source
Data from patients with malignant GCTB were extracted from the SEER database with SEER*Stat Software version 8. 3.5 (https://seer.cancer.gov/seerstat/) (Information Management Service, Inc. Calverton, MD, USA). According to the ICD-O-3/WHO2008 guidelines, data from patients who were diagnosed with malignant GCTB (code: 9250/3: giant cell tumour of bone, malignant) between 1975 and 2016 were collected. Patients were excluded if they were diagnosed at autopsy or via death certificate. Malignant GCTBs located in extraskeletal sites such as the kidney, lung, pancreas and thyroid were excluded.

Statistical analysis
From the records of patients in the SEER database, the following patient-related variables were selected for analysis. We investigated the relationships between age and diagnosis year as continuous measure and overall survival using non-linear restricted cubic splines (RCS) model with 3 knots. As reported by Harrell, the positions of knots were selected at 0.1, 0.5 and 0.9 quintiles of the distribution [12]. The age was divided into four sections (≤18, 19-34, 35-60 or >60 years), and the year interval of diagnosis was divided into 1975-1984, 1985-2004, 2005-2013, and 2014-2016. The other variables included sex (male or female); marital status (unmarried or married); race (white, black, Indian/Alaska Native (IA) or Asian or Pacific Islander (API)); primary site (long bones of lower limb with associated joints, long bones of upper limb with scapula and associated joints, pelvis and spine, irregular bone and other sites outside of bone); tumour grade (I, II, III, or IV); T stage (T1, T2, or T3); N stage (N0 or N1); M stage (M0 or M1); historic SEER stage (localized, regional, or distant); and surgical treatment of the primary site (none or yes). Patients without clear records were classified as unknown for the corresponding parameter.
Continuous data are described as the mean ± standard deviation. Differences between groups were analysed with univariate analysis. Categorical data are presented as the number and percentage, and the difference was evaluated by Pearson's chi-square test. The 1-, 5-, and 10-year overall survival (OS) values were calculated by the Kaplan-Meier method, and the log-rank test was used to evaluate the differences. The variables that had a P value <0.05 in the univariate Cox regression analysis and those that had a significant effect on the prognosis in the biologically based theoretical model were incorporated into the multivariate model. To evaluate the influence of extraskeletal sites on survival, a sensitivity analysis was performed by including or excluding cases.
All statistical analyses were performed using SPSS 23.0 (IBM Corporation, Armonk, NY, USA) and MedCalc 15.2.2. Two-sided P value <0.05 was considered statistically significant.

Patient characteristics
The selection process of the patients is illustrated in Figure 1. According to the inclusion criteria, 334 patients with malignant GCTB seen from 1975 to 2016 were initially selected. After excluding one case that was diagnosed via death certificate and eight patients with extraskeletal GCTB, 325 patients were finally included. Table 1 shows the detailed clinicopathological features of the patients.
The median age of the patients was 34 years, and a female predominance (52.3%) was observed.
Regarding historic SEER stage, the localized and regional stages were the most frequent, occurring in 61.2% of all the patients, while 13.5% of patients had disease at distant stage. Surgery was performed in 53.85% of the patients.
Eight patients were diagnosed with extraskeletal GCTB. Their data were recorded from 1987 to 2016, and five patients were female. Two of the patients were diagnosed at approximately 20 years old, while the others were older than 50 years. Regarding the primary site, three cases were found in the pancreas, one was found in the kidney, one was found in the thyroid, one was found in the mediastinum, one was found in the lung, and there was one subcutaneous case. There was no clear record of tumour grade in four patients, one of whom had grade I disease, two of whom had grade III disease, and one of whom had grade IV disease. Two patients had long survival (135 and 136 months), one died at 24 months after diagnosis, and five patients died within 7 months.

Survival analysis and prognostic factor identification
The overall survival of the entire cohort is illustrated in Figure 2. The overall 1-, 5-, and 10-year survival rates were 94.3% (95% CI: 91.7-96.8), 82.3% (95% CI: 77.9-86.6), and 80.1% (95% CI: 75.4-84.7), respectively. The detailed OS data for patients with malignant GCTB according to each characteristic are summarized in Table 2.We observed a potential non-linear J-shaped dose-response relationship between the age or diagnosis year and survival. As shown in the Figures 3A and 3B, three knots were applied to the data and stratified the age and diagnosis year into four groups.
The results of the univariate and multivariate Cox regression analyses are summarized in Table 3 [16,17]. In our study, 14 out of 89 cases extracted from 2010 to 2016 were recorded as M1 status, and the lung was the major metastatic site (10 out of 89 cases, 11.2%). The higher incidence of lung metastasis in this study than in the previous study may be a result of the introduction of regular screening. More lung metastases were reported to develop early after the initial diagnosis of GCTB and after recurrence than long after diagnosis [17,18]. Screening for possible metastasis should especially be performed in these situations.
TNM classification data have been added to the SEER database since 2004, and the results showed a trend for poor survival in patients with higher T stage or higher N stage. The multivariate Cox regression analysis suggested that higher T stage was a prognostic factor. Tumour size was previously found to be an independent risk factor for lung metastasis [19]. Based on the Campanacci classification, which has been widely accepted for evaluating giant cell tumours, grade III tumours are the aggressive type. A series of studies reported Campanacci grade III as a risk factor for tumour recurrence [20,21] and lung metastasis [22,23]. Recurrence was also reported to be a risk factor for metastasis [22,23]. Therefore, T stage and/or Campanacci classification played an important role in evaluating the survival of patients with giant cell tumours.
To our knowledge, this is the first study of extraskeletal malignant GCTB. Based on our results, an extraskeletal site was associated with poor survival. Only two patients lived longer than 11 years, while five patients died in less than seven months. Therefore, attention should be paid to patients with lesions in extraskeletal sites. Our study was limited by the few available cases, and further analysis revealing potential prognostic factors in a large population could not be conducted. More cases from multiple centres will be needed. Helsinki Declaration and its later amendments or comparable ethical standards.

Consent for publication
Not applicable.

Availability of data and material
Data from patients with malignant GCTB were extracted from the SEER database with SEER*Stat Software version 8.3.5 (https://seer.cancer.gov/seerstat/) (Information Management Service, Inc. Calverton, MD, USA).

Competing interests
All the authors report no conflicts of interest in this work.

Funding
The present study was sponsored by Natural Science     The overall survival of the patients with malignant giant cell tumor of bone from the SEER database.