Abstract
Objective
Chondrosarcoma is a common bone malignancy, and the main treatment method is surgery. Different surgeries lead to different survival outcomes. The aim of this study was to construct a new clinical predictive tool to accurately predict the overall survival (OS) and cause specific survival (CSS) of patients with chondrosarcoma receiving different treatments.
Methods
A total of 620 patients with chondrosarcoma registered between January 1, 2000 and December 31, 2016 were recruited as study targets. The missing values are filled by multiple imputation. Two continuous variables, age and tumor size, were divided into binary variables based on Kaplan–Meier curve. Univariate and multivariate analyses were used to explore predictors and establish nomograms. Propensity score matching (PSM) analysis was used to reduce the impact of potential confounders to determine whether different surgical modalities had any survival benefits in subgroups.
Results
In a multivariate cox regression, age, grade, tumor size, radiotherapy, chemotherapy, and surgical methods were identified as independent prognostic factors for chondrosarcoma. To construct 1‐, 3‐, and 5‐year nomogram maps of OS and CSS with prognostic factors and verify the c index internally (OS, 0.807; CSS, 0.847) above American Joint Committee on Cancer (AJCC) (OS, 0.685; CSS, 0.732).
Conclusion
This study found that the 5 year overall survival rate of patients with non‐metastatic chondrosarcoma of the extremities was about 80%. Age, high malignancy, large tumor, prior chemoradiotherapy, and poor surgical selection were independent risk factors. Therefore, the nomogram established in this study will help to optimize clinicians' personalized decision making for patients.
Keywords: Chondrosarcoma, Nomogram, Non‐metastatic, Surgery methods
In this study, we investigated the risk factors of various surgical procedures for patients with different conditions of non‐metastatic chondrosarcoma of the extremities, and constructed a prognostic model for patients with chondrosarcoma. The established model helps clinicians to personalize treatment plans according to different patients.
Introduction
Chondrosarcoma is a group of heterogeneous tumors that co‐produce cartilage matrix. 1 Chondrosarcoma is the third most common primary bone malignancy, accounting for 20%–27% of all primary bone malignancies. 2 It is generally considered to be relatively resistant to chemotherapy and radiotherapy due to its low extracellular matrix, low proportion of dividing cells, and poor distribution of blood vessels. Extensive and total resection is the preferred surgical treatment for moderate and severe chondrosarcoma. For low‐grade chondrosarcomas limited to bone, extensive intrafocal curettage followed by local adjuvant therapy, cavity filling with bone graft, has good long‐term clinical results and satisfactory local control. 3 , 4 Advanced metastases suggest a poor prognosis, so identifying prognostic factors for premetastatic chondrosarcoma may help oncologists treat patients individually. 5 , 6
In recent years, with the study of the survival rate of various surgical methods for chondrosarcoma, people have gradually realized that a slightly smaller resection scope will not significantly reduce the survival rate, but also improve the quality of life of patients. 7 , 8 To date, direct comparison of prognostic and epidemiological data between surgical modalities for chondrosarcoma has not been reported.
The American Joint Committee on Cancer (AJCC) staging system is the most common standard for patients with chondrosarcoma. 9 Using the basic characteristics of the tumor, the AJCC staging system can make overall survival (OS) predictions. Failure to consider the effects of surgery can lead to limitations when predicting post‐operative OS. Song et al. studied prognostic variables in patients with chondrosarcoma and found that metastasis was associated with males, higher grade, larger tumor size, axial or craniofacial position, but did not further analyze 5‐ and 10‐year OS and cause specific survival (CSS) or establish prognostic models. 5 Using nomogram data to predict clinical outcomes in patients with metastatic chondrosarcoma, Wang et al. found that 5‐ and 10‐year OS rates for metastatic chondrosarcoma were 28.4% and 22.8%, and 5‐ and 10‐year CSS rates were 31.2% and 26.6%, respectively. 10 Among the prognostic factors, surgical removal of the primary tumor can significantly prolong the survival of patients with metastatic chondrosarcoma, but no specific classification of surgery has been made. 10 Jiang et al. used three types of surgery to construct a survival nomogram. 11 However, these studies did not include the complete type of surgery as a study variable, nor did they construct predictive models for post‐operative OS. 11 Therefore, it is extremely important to construct a reliable and accurate OS prediction model for patients with chondrosarcoma after surgery.
The type of surgery is an important prognostic factor in patients with chondrosarcoma. Inappropriate types of surgery for patients of different ages with different physical conditions and disease stages can lead to a poorer prognosis and shorter survival. Little attention has been paid to the effect of specific types of surgery on the survival of patients with chondrosarcoma. We collected the clinical features and survival of patients with chondrosarcoma, analyzed independent risk factors, 5‐ and 10‐year OS and CSS, and established a nomogram to predict the probability of survival.
Method
Data Collection
We selected patients with a histological diagnosis of chondrosarcoma from our hospital database from 2000 to 2016. Inclusion criteria were: (i) upper and lower limb long bones; (ii) clear treatment options; and (iii) follow‐up of more than 5 years or death at follow‐up. According to the inclusion criteria, 935 patients were enrolled. Exclusion criteria were: (i) unproven surgery (n = 91); (ii) no definite staging (n = 122); and (iii) diagnosis of metastasis (n = 102). Ultimately, 620 patients were enrolled.
Data Extraction
Variables collected included baseline patient demographics (age, sex, and marital status), tumor characteristics (tumor size, primary site, lateral position, grade, American Joint Board on Cancer [AJCC] T, and AJCC stage), and treatment (surgical method, chemotherapy, and radiation). OS was regarded as the primary endpoint and was defined as the time in months from diagnosis to death due to any cause. CSS was regarded as the secondary endpoint and was defined as the time in months from diagnosis to death specific to chondrosarcoma. After checking data integrity and identifying missing values, the missing data is filled with the multiple imputation (MI) method. The following two continuous variables were transformed into binary variables: age and tumor size. In this study, we classified patients according to the following factors, such as age at diagnosis (≤65, >65 years), sex (female, male), marital status (married, single, widowed, other), primary site (upper limb, lower limb), grade (I, II, III, IV), laterality (left, right), tumor size (≤106, >106 mm), AJCC T stage (T1, T2, T3), AJCC stage (IA, IB, IIA, IIB, III), radiation (yes, none), chemotherapy (yes, none), and surgery method (no surgery, extensive resection, tumor segment resection, radical resection, amputation, tumor curette). Based on the surgical method received, patients were separated into six groups: no surgery, extensive resection, tumor segment resection, radical resection, amputation and tumor curette. The study was conducted in accordance with the principles of the Declaration of Helsinki, and the study protocol was approved by the ethics committee of Shanxi medical university (2023(YX)NO:115). Because of the retrospective nature of the study, patient consent for inclusion was waived.
Statistical Analysis
All data were analyzed using the SPSS version 27.0 (IBM, Armonk, NY, USA) and R software 4.05 (www.r-project.org). Two‐tailed p < 0.05 was the statistical difference. Optimal cutoff values were set for two continuous variables: age and tumor size, according to Kaplan–Meier curve. Chi‐square test was used to compare the clinicopathological data of each group. In addition, χ 2‐test was used to compare the OS and CSS of various prognostic factors in patients with chondrosarcoma. Different survival rates of variables were graphically evaluated by using the Kaplan–Meier method. Univariate and multivariate analyses of the variables in patients with chondrosarcoma were performed using the Cox proportional risk model. After univariate analysis, prognostic factors with significant influence on prognosis were included in multivariate Cox regression analysis to screen out independent prognostic factors. In order to avoid multicollinearity in multivariate analysis, we removed AJCC T and stage variables in the Cox proportional risk model. Based on clinically independent prognostic factors, nomogram model were constructed to predict 1‐, 3‐, and 5‐year survival probabilities of patients with chondrosarcoma. We calculate the C‐index, which quantifies the difference between observations and predictions and shows the predictive power of the model. The calibration diagram of the model shows the calibration between predicted and actual survival. Decision curve analysis (DCA) was used to evaluate the clinical effect and benefit of the prediction model. In addition, to compare the predictive ability of the AJCC staging systems and the nomograms, we applied the Kaplan–Meier method and calculated the C‐index of AJCC.
Propensity Score Matching
In retrospective cohort studies, treatment‐related selection bias resulting from an imbalance in the baseline characteristics is inevitable. 12 PSM can reduce the selection bias, offset differing clinical features among groups, and bolster the evidence of a retrospective cohort study. 12 , 13 The present study created a logistic regression model with propensity score to balance the baseline characteristics across surgical modalities. The PSM was performed in a 1:1 ratio using nearest neighbor matching with a caliper of 0.05. Chi‐square or Fisher's exact tests were used to compare baseline characteristics between groups.
Results
Fill in Missing Values and Determine Cutoff Values for Continuous Variables
The missing value of each variable was identified, and the missing value was modified to NA. The data was filled with MI method to make the data table complete (Fig. S1). KM curve analysis showed that the optimal age threshold of chondrosarcoma patients was 65 years, so the patients were divided into two groups (≤65, >65 years) (Figs. S2 and S3); The optimal cutoff for tumor size was 106 mm, so tumor size was divided into two groups (≤106, >106 mm) (Figs. S4 and S5).
Patient Characteristics
The baseline characteristics of patients included in this study are presented in Table 1. Overall, 620 cases of chondrosarcoma were screened according to the screening criteria, wherein the patients who received no surgery, tumor segment resection, radical resection, amputation, extensive resection, and tumor curette were 36 (5.8%), 130 (21.0%), 360 (58.1%), 44 (7.1%), 44 (7.1%), 6 (1.0%), respectively. Age, primary site, tumor size, AJCC T, AJCC stage and grade were significantly different among the six groups (p < 0.05). Radical resection accounted for the highest proportion among all patients. Patients older than 65 were more likely to opt for no surgery or amputation. Tumors larger than 106 mm are more likely to result in amputation. This trend was also observed in grade III or even grade IV tumors, AJCC Stage T3, AJCC Stage IIB, and stage III.
TABLE 1.
Demographic data and tumor characteristics of chondrosarcoma patients
| Total | No surgery | Tumor segment resection | Radical resection | Amputation | Extensive resection | Tumor curette | Statistical value | p‐value (p < 0.05) | |
|---|---|---|---|---|---|---|---|---|---|
| Characteristic | 620 | 36 (5.8%) | 130 (21.0%) | 360 (58.1%) | 44 (7.1%) | 44 (7.1%) | 6 (1.0%) | ||
| Age | F = 4.166 | 0.001a | |||||||
| ≤65 | 512 (82.6) | 26 (72.2) | 117 (90.0) | 297 (82.5) | 28 (63.6) | 38 (86.4) | 6 (100.0) | ||
| >65 | 108 (17.4) | 10 (27.8) | 13 (10.0) | 63 (17.5) | 16 (36.4) | 6 (13.6) | 0 (0.0) | ||
| Sex | F = 1.587 | 0.161a | |||||||
| Male | 335 (54) | 17 (47.2) | 63 (48.5) | 200 (55.6) | 29 (65.9) | 21 (47.7) | 5 (83.3) | ||
| Female | 285 (46) | 19 (52.8) | 67 (51.5) | 160 (44.4) | 15 (34.1) | 23 (52.3) | 1 (16.7) | ||
| Marital status | F = 1.469 | 0.517a | |||||||
| Married | 365 (58.9) | 17 (47.2) | 80 (61.5) | 216 (60.0) | 22 (50.0) | 26 (59.1) | 4 (66.7) | ||
| Single | 164 (26.5) | 12 (33.3) | 35 (26.9) | 94 (26.1) | 10 (22.7) | 12 (27.3) | 1 (16.7) | ||
| Widowed | 33 (5.3) | 4 (11.1) | 5 (3.8) | 20 (5.6) | 3 (6.8) | 1 (2.3) | 0 (0.0) | ||
| Other | 58 (9.4) | 3 (8.3) | 10 (7.7) | 30 (8.3) | 9 (20.5) | 5 (11.4) | 1 (16.7) | ||
| Primary site | F = 2.533 | 0.028a | |||||||
| Upper limb | 253 (40.8) | 16 (44.4) | 61 (46.9) | 139 (38.6) | 10 (22.7) | 24 (54.5) | 3 (50.0) | ||
| Lower limb | 367 (59.2) | 20 (55.6) | 69 (53.1) | 221 (61.4) | 34 (77.3) | 20 (45.5) | 3 (50.0) | ||
| Grade | F = 7.741 | 0.001a | |||||||
| I | 284 (45.8) | 18 (50.0) | 77 (59.2) | 147 (40.8) | 9 (20.5) | 28 (63.6) | 5 (83.3) | ||
| II | 257 (41.5) | 16 (44.4) | 43 (33.1) | 156 (43.3) | 25 (56.8) | 16 (36.4) | 1 (16.7) | ||
| III | 59 (9.5) | 2 (5.6) | 8 (6.2) | 41 (11.4) | 8 (18.2) | 0 (0.0) | 0 (0.0) | ||
| IV | 20 (3.2) | 0 (0.0) | 2 (1.5) | 16 (4.4) | 2 (4.5) | 0 (0.0) | 0 (0.0) | ||
| Laterality | F = 0.871 | 0.498a | |||||||
| Left | 307 (49.5) | 17 (47.2) | 66 (50.8) | 169 (46.9) | 25 (56.8) | 26 (59.1) | 4 (66.7) | ||
| Right | 313 (50.5) | 19 (52.8) | 64 (49.2) | 191 (53.1) | 19 (43.2) | 18 (40.9) | 2 (33.3) | ||
| Tumor size | F = 12.888 | 0.001a | |||||||
| ≤106 mm | 465 (75) | 26 (72.2) | 122 (93.8) | 256 (71.1) | 18 (40.9) | 37 (84.1) | 6 (100.0) | ||
| >106 mm | 155 (25) | 10 (27.8) | 8 (6.2) | 104 (28.9) | 26 (59.1) | 7 (15.9) | 0 (0.0) | ||
| AJCC T | F = 7.897 | 0.001a | |||||||
| T1 | 391 (63.1) | 23 (63.9) | 105 (80.8) | 213 (59.2) | 13 (29.5) | 31 (70.5) | 6 (100.0) | ||
| T2 | 223 (36) | 12 (33.3) | 22 (16.9) | 146 (40.6) | 31 (70.5) | 12 (27.3) | 0 (0.0) | ||
| T3 | 6 (1) | 1 (2.8) | 3 (2.3) | 1 (0.3) | 0 (0.0) | 1 (2.3) | 0 (0.0) | ||
| AJCC stage | F = 6.774 | 0.001a | |||||||
| IA | 357 (57.6) | 22 (61.1) | 99 (76.2) | 189 (52.5) | 10 (22.7) | 31 (70.5) | 6 (100.0) | ||
| IB | 179 (28.9) | 11 (30.6) | 19 (14.6) | 113 (31.4) | 24 (54.5) | 12 (27.3) | 0 (0.0) | ||
| IIA | 34 (5.5) | 1 (2.8) | 6 (4.6) | 24 (6.7) | 3 (6.8) | 0 (0.0) | 0 (0.0) | ||
| IIB | 44 (7.1) | 1 (2.8) | 3 (2.3) | 33 (9.2) | 7 (15.9) | 0 (0.0) | 0 (0.0) | ||
| III | 6 (1) | 1 (2.8) | 3 (2.3) | 1 (0.3) | 0 (0.0) | 1 (2.3) | 0 (0.0) | ||
| Radiation | F = 1.080 | 0.368a | |||||||
| None | 601 (96.9) | 33 (91.7) | 125 (96.2) | 350 (97.2) | 44 (100.0) | 43 (97.7) | 6 (100.0) | ||
| Yes | 19 (3.1) | 3 (8.3) | 5 (3.8) | 10 (2.8) | 0 (0.0) | 1 (2.3) | 0 (0.0) | ||
| Chemotherapy | F = 0.535 | 0.748a | |||||||
| None | 604 (97.4) | 34 (94.4) | 128 (98.5) | 351 (97.5) | 42 (95.5) | 43 (97.7) | 6 (100.0) | ||
| Yes | 16 (2.6) | 2 (5.6) | 2 (1.5) | 9 (2.5) | 2 (4.5) | 1 (2.3) | 0 (0.0) |
Notes: Percentages may not total 100 because of rounding. aANOVA test
Abbreviations: Grade I, well differentiated; Grade II, moderately differentiated; Grade III, poorly differentiated; Grade IV, undifferentiated.
Univariate and multivariate cox regression analysis
We used the Cox proportional hazards regressions model to identify the potential risk factors correlating with OS and CSS in chondrosarcoma patients (Table 2). Kaplan–Meier survival curves were plotted for categorical variables of surgery method. The differences in the survival time distributions were examined using the log‐rank method. The differences in the OS time distributions between groups for these surgery methods were statistically significant (p < 0.05) (Fig. 1A). The differences in the CSS time distributions between groups for these surgery methods were no statistically significant (p > 0.05) (Fig. 1B). The age, sex, marital status, tumor size, primary site, grade, AJCC T, and AJCC stage surgical method, chemotherapy, and radiation in the univariate Cox regression that were statistically significant (p < 0.05) were incorporated into the multivariate analysis. In the multivariate analysis of OS, variables including age, grade, tumor size and radiation were all statistically significant. In the multivariate analysis of OS, variables including age, grade, tumor size and radiation were all statistically significant. According to multivariate analysis, the outcomes were improved in patients with younger age, well‐differentiated stage, lower tumor size, no radiation therapy. In the multivariate analysis of CSS, variables including age, grade and radiation were all statistically significant. According to multivariate analysis, the outcomes were improved in patients with younger age, well‐differentiated stage, localized growth, no radiation therapy.
TABLE 2.
Cox proportional hazards regression model for overall survival and cause‐specific survival in patients with non‐metastatic conventional chondrosarcoma of the extremities
| OS | CSS | |||||||
|---|---|---|---|---|---|---|---|---|
| Univariate analysis | p‐value | Multivariate analysis | p‐value | Univariate analysis | p‐value | Multivariate analysis | p‐value | |
| Characteristic | Hazard ratio (95% CI) | Hazard ratio (95% CI) | <0.05 | Hazard ratio (95% CI) | Hazard ratio (95% CI) | <0.05 | ||
| Age | ||||||||
| ≤65 | Reference | Reference | Reference | Reference | ||||
| >65 | 5.52 (3.86–7.88) | <0.01 | 3.53 (2.32–5.38) | <0.01 | 4.77 (3.01–7.56) | <0.01 | 2.76 (1.61–4.76) | <0.01 |
| Sex | ||||||||
| Male | Reference | Reference | Reference | |||||
| Female | 0.68 (0.47–0.97) | 0.04 | 0.77 (0.52–1.13) | 0.19 | 0.76 (0.48–1.20) | 0.24 | ||
| Marital status | ||||||||
| Married | Reference | Reference | Reference | Reference | ||||
| Single | 1.03 (0.66–1.62) | 0.89 | 1.31 (0.81–2.10) | 0.27 | 1.22 (0.70–2.16) | 0.48 | 1.58 (0.86–2.89) | 0.14 |
| Widowed | 5.72 (3.45–9.48) | <0.01 | 1.75 (0.97–3.17) | 0.06 | 5.86 (3.04–11.31) | <0.01 | 2.00 (0.96–4.20) | 0.07 |
| Other | 1.17 (0.63–2.17) | 0.62 | 1.20 (0.62–2.33) | 0.58 | 1.58 (0.76–3.29) | 0.22 | 1.34 (0.61–2.90) | 0.47 |
| Primary Site | ||||||||
| Upper limb | Reference | Reference | Reference | |||||
| Lower limb | 1.63 (1.11–2.40) | 0.01 | 1.34 (0.89–2.01) | 0.16 | 1.49 (0.92–2.43) | 0.11 | ||
| Grade | ||||||||
| I | Reference | Reference | Reference | Reference | ||||
| II | 2.58 (1.66–4.01) | <0.01 | 1.47 (0.89–2.45) | 0.14 | 3.82 (1.94–7.53) | <0.01 | 2.32 (1.12–4.83) | 0.02 |
| III | 4.34 (2.51–7.50) | <0.01 | 1.84 (0.97–3.46) | 0.11 | 9.05 (4.28–19.17) | <0.01 | 4.16 (1.76–9.86) | 0.00 |
| IV | 9.67 (4.80–19.48) | <0.01 | 3.61 (1.55–8.38) | 0.05 | 23.50 (10.13–54.55) | <0.01 | 11.12 (3.99–30.99) | <0.01 |
| Laterality | ||||||||
| Left | Reference | Reference | ||||||
| Right | 1.13 (0.80–1.61) | 0.49 | 1.20 (0.76–1.89) | 0.44 | ||||
| Tumor size | ||||||||
| ≤106 mm | Reference | Reference | Reference | Reference | ||||
| >106 mm | 3.52 (2.47–5.01) | <0.01 | 1.76 (1.18–2.63) | 0.00 | 3.46 (2.20–5.45) | <0.01 | 1.60 (0.97–2.64) | 0.07 |
| Radiation | ||||||||
| None | Reference | Reference | Reference | Reference | ||||
| Yes | 6.08 (3.41–10.84) | <0.01 | 2.55 (1.31–4.94) | 0.00 | 8.61 (4.52–16.40) | <0.01 | 2.94 (1.39–6.23) | 0.01 |
| Chemotherapy | ||||||||
| None | Reference | Reference | Reference | Reference | ||||
| Yes | 5.64 (3.03–10.51) | <0.01 | 2.22 (1.08–4.54) | 0.03 | 8.67 (4.44–16.95) | <0.01 | 1.82 (0.81–4.06) | 0.15 |
| Surgery | ||||||||
| No surgery | Reference | Reference | Reference | Reference | ||||
| Tumor segment resection | 0.25 (0.12–0.52) | <0.01 | 0.50 (0.23–1.12) | 0.09 | 0.25 (0.09–0.70) | 0.01 | 0.49 (0.17–1.45) | 0.20 |
| Radical resection | 0.43 (0.24–0.78) | 0.01 | 0.45 (0.23–0.87) | 0.02 | 0.49 (0.22–1.10) | 0.08 | 0.53 (0.21–1.31) | 0.17 |
| Amputation | 1.33 (0.67–2.66) | 0.42 | 1.02 (0.48–2.18) | 0.95 | 1.67 (0.68–4.15) | 0.27 | 1.50 (0.55–4.14) | 0.43 |
| Extensive resection | 0.17 (0.06–0.53) | 0.00 | 0.30 (0.09–0.96) | 0.04 | 0.25 (0.07–0.97) | 0.05 | 0.54 (0.13–2.19) | 0.39 |
| Tumor curette | 0.38 (0.05–2.89) | 0.35 | 1.50 (0.18–12.18) | 0.70 | 0.71 (0.09–5.74) | 0.75 | 4.47 (0.48–40.82) | 0.19 |
Abbreviations: CSS, cause‐specific survival; OS, overall survival.
FIG. 1.
Kaplan–Meier survival analysis of overall survival (A) and cancer‐specific survival (B) in all patients with chondrosarcoma.
Nomogram and Validation
Nomograms based on final Cox models were constructed to evaluate predictive ability of the 1‐, 3‐, and 5‐year OS and CSS with predictors in Fig. 2A,B. Then, to validate nomograms internally, we applied bootstrapped validation internally with the dataset. The C‐index for internal validation of OS and CSS was 0.807 and 0.847, respectively. Meanwhile, we used the calibration plots (Fig. 2C,D) to validate the concordance of nomograms by comparing predicted values with the actual endpoints. Finally, we set up a DCA diagram for the model to observe the estimated net benefits, DCA results showed that the model provided a good net benefit for chondrosarcoma patients at 1‐, 3‐, and 5‐year survival times (Fig. 2E,F). To compare nomograms with traditional staging systems, we obtained the C‐index of AJCC (OS, 0.685; CSS, 0.732) (Fig. 3A,B).
FIG. 2.
Nomogram analysis of chondrosarcoma. (A) A nomogram for predicting 1‐, 3‐, and 5‐year OS of patients. (B) A nomogram for predicting 1‐, 3‐ and 5‐year CSS of patients. (C) Calibration curve of the nomogram predicting 1‐, 3‐, and 5‐year OS. (D) Calibration curve of the nomogram predicting 1‐, 3‐ and 5‐year CSS. (E) DCA of the nomogram for predicting 1‐, 3‐, and 5‐year OS rates in patients with chondrosarcoma. (F) DCA of the nomogram for predicting 1‐, 3‐ and 5‐year CSS rates in patients with chondrosarcoma. CSS, cancer‐specific survival; DCA, Decision curve analysis; OS, overall survival.
FIG. 3.
Kaplan–Meier curves of overall survival (OS) (A) and cause‐specific survival (CSS) (B) for the American Joint Committee on Cancer (AJCC) staging system.
Subgroup Analysis after PSM
To reduce bias between groups, we used PSM to accurately match patients in the following groups: no surgery versus surgery, no surgery versus radical resection, tumor segment resection versus radical resection, and radical resection versus amputation. After PSM, there was a significant difference in OS between only tumor segment resection and radical resection (HR: 1.74 [1.08–2.8] p = 0.049) (Table 3, Figs. S1 and S2). Radical resection has a better OS than tumor segment resection (Figs. 4, 5, 6 and S6–S8).
TABLE 3.
Comparison of baseline variables between Tumor segment resection and Radical resection in original and matched datasets of patients with chondrosarcoma
| Before PSM | After PSM | |||||||
|---|---|---|---|---|---|---|---|---|
| Tumor segment resection | Radical resection | Statistical value | p‐value | Tumor segment resection | Radical resection | Statistical value | p‐value | |
| Characteristic | (N = 129) | (N = 359) | <0.05 | (N = 121) | (N = 121) | <0.05 | ||
| Age | χ 2 = 3.547 | 0.06 a | χ 2 = 0.388 | 0.53 a | ||||
| >65 | 13 (10.1%) | 63 (17.5%) | 11 (9.1%) | 15 (12.4%) | ||||
| ≤65 | 116 (89.9%) | 296 (82.5%) | 110 (90.9%) | 106 (87.6%) | ||||
| Marital status | χ 2 = 0.656 | 0.88 a | χ 2 = 0.340 | 0.95 a | ||||
| Other | 10 (7.8%) | 30 (8.4%) | 10 (8.3%) | 8 (6.6%) | ||||
| Married | 80 (62.0%) | 216 (60.2%) | 77 (63.6%) | 78 (64.5%) | ||||
| Single | 34 (26.4%) | 94 (26.2%) | 30 (24.8%) | 30 (24.8%) | ||||
| Widowed | 5 (3.9%) | 19 (5.3%) | 4 (3.3%) | 5 (4.1%) | ||||
| Sex | χ 2 = 1.658 | 0.19 a | χ 2 = 0.017 | 0.90 a | ||||
| Female | 66 (51.2%) | 159 (44.3%) | 61 (50.4%) | 59 (48.8%) | ||||
| Male | 63 (48.8%) | 200 (55.7%) | 60 (49.6%) | 62 (51.2%) | ||||
| Primary Site | χ 2 = 2.398 | 0.12 a | χ 2 = 0.612 | 0.43 a | ||||
| Lower limb | 69 (53.5%) | 220 (61.3%) | 67 (55.4%) | 74 (61.2%) | ||||
| Upper limb | 60 (46.5%) | 139 (38.7%) | 54 (44.6%) | 47 (38.8%) | ||||
| Grade | χ 2 = 14.359 | <0.01 a | χ 2 = 2.024 | 0.57 a | ||||
| I | 76 (58.9%) | 146 (40.7%) | 72 (59.5%) | 67 (55.4%) | ||||
| II | 43 (33.3%) | 156 (43.5%) | 42 (34.7%) | 41 (33.9%) | ||||
| III | 8 (6.2%) | 41 (11.4%) | 6 (5.0%) | 11 (9.1%) | ||||
| IV | 2 (1.6%) | 16 (4.5%) | 1 (0.8%) | 2 (1.7%) | ||||
| Laterality | χ 2 = 0.417 | 0.51 a | χ 2 = 0.265 | 0.61 a | ||||
| Left | 65 (50.4%) | 169 (47.1%) | 59 (48.8%) | 64 (52.9%) | ||||
| Right | 64 (49.6%) | 190 (52.9%) | 62 (51.2%) | 57 (47.1%) | ||||
| Tumor size | χ 2 = 26.723 | <0.01 a | χ 2 < 0.001 | 1.00 a | ||||
| >106 cm | 8 (6.2%) | 104 (29.0%) | 8 (6.6%) | 8 (6.6%) | ||||
| ≤106 cm | 121 (93.8%) | 255 (71.0%) | 113 (93.4%) | 113 (93.4%) | ||||
| Radiation | χ 2 = 0.096 | 0.75 a | χ 2 = 0.588 | 0.44 a | ||||
| None | 124 (96.1%) | 349 (97.2%) | 119 (98.3%) | 116 (95.9%) | ||||
| Yes | 5 (3.9%) | 10 (2.8%) | 2 (1.7%) | 5 (4.1%) | ||||
| Chemotherapy | χ 2 = 0.084 | 0.77 a | χ 2 = 0.254 | 0.61 a | ||||
| None | 127 (98.4%) | 350 (97.5%) | 120 (99.2%) | 118 (97.5%) | ||||
| Yes | 2 (1.6%) | 9 (2.5%) | 1 (0.8%) | 3 (2.5%) | ||||
| Survival months | t = 0.725 | 0.46 b | t = 1.945 | 0.05 b | ||||
| Mean (SD) | 94.3 (45.9) | 90.4 (47.2) | 96.9 (45.4) | 85.1 (48.7) | ||||
Abbreviation: PSM, Propensity score matching.
χ2‐test.
t‐test.
FIG. 4.
(A–G) Preoperative x‐ray, CT and MR Images of patients with chondrosarcoma. X‐ray and CT showed swelling and lytic bone destruction at the end of the humerus. Annular calcification was observed in the lesion, and scallop‐like changes were observed locally in the bone cortex; MRI showed low signals on T1WI and mixed signals on T2WI and T2 fat suppression. (H) Postoperative x‐ray images. (I) Gross pathological picture of chondrosarcoma after extensive resection. (J) Pathological report: (left proximal humerus) the proximal humerus tissue was examined. The humerus head was dilated and chondrocytomatoid hyperplasia was observed under the microscope. In combination with clinical imaging, atypical chondrosarcoma was considered as grade I, and no tumor tissue was found in the soft tissues around the humerus and the broken end of the bone.
FIG. 5.
(A–H) Preoperative CT and MR Images of patients with chondrosarcoma. CT: Right middle femur bone occupying lesion. Curved calcification and small ring calcification were seen in the lesion. Scallop‐like changes were observed in the affected areas of the cortex. MRI: Low signal in T1WI, high signal in T2 fat inhibition sequence, low signal in the lesion in flakiness and line, high signal in DWI, 3D volume rendering showed high density mass occupying the bone marrow cavity and pressing the adjacent bone cortex. (I) Postoperative x‐rays. (J) Gross pathological picture of chondrosarcoma after tumor curette. (K) Chondrogenic tumor tissue was found in the bone marrow tissues submitted for examination, with small cell atypia accompanied by cartilage matrix in some areas and lobular structure formation accompanied by fibroplasia in some areas. No clear neoplastic bone formation was found in any of the bone marrow tissues submitted for examination. Chondrosarcoma (small part grade II, most of grade I) was found in combination with imaging data.
FIG. 6.
(A–H) Preoperative CT and MR Images of patients with chondrosarcoma. CT: Osteolytic lesion and soft tissue occupation of the right femoral neck. The cortical end of the bone is destroyed, and the marginal dot and small ring calcification are seen. MRI showed high signal of T2 fat inhibition sequence, and the enhanced scan showed multiple annular and lobed peripheral enhancement. (I) Postoperative x‐ray examination. (J–L) Gross pathology of radical resection of chondrosarcoma. (M) Pathological findings: tumor‐like hyperplasia of cartilage with mucous matrix, lobulated structure, massive bleeding and necrosis. The hyperplasia of some cells was active, the peripheral cells of the focal lobule were dense, the cells had atypia, and pathological nuclear division was visible, which was consistent with chondrosarcoma (grade II, local III).
Discussion
The results of this study showed that age, tumor size, tumor grade, chemoradiotherapy and surgical methods were prognostic factors affecting the survival time of patients with chondrosarcoma. As suggested in this study, the choice of surgery is also strongly influenced by other prognostic factors, and therefore directly determines the OS and CSS of a patient with chondrosarcoma. We constructed a nomogram model to predict the overall survival and death risk of patients with chondrosarcoma after surgery. From this perspective, this study is also helpful for clinicians to select different treatment strategies for different patients according to the constructed clinical model to bring maximum survival benefits to patients.
Risk Factors in Patients with Chondrosarcoma
We screened 620 patients with non‐metastatic conventional chondrosarcoma in the extremities, of which 124 were dead and the remaining 496 were alive at the end of follow‐up. Most of the 124 patients with chondrosarcoma developed distant metastases such as lung metastases during follow‐up, resulting in malignant outcomes and even death. In this study, we first demonstrated that age >65 years and tumor size >106 mm predicted a poor prognosis.
Giuffrida et al. 14 analyzed chondrosarcoma data in the SEER database. The prognosis of patients older than 50 years was relatively poor, the 30 year survival rate was only 20%. While the age less than 50 years old, the 30 year survival rate was greater than 60%. In our study, 65 years of age was found to be the largest divide in survival by a binary classification. The reason why the cut‐off age in this study was different from others is that we pre‐screened patients with confirmed metastases and poor prognosis sites such as craniofacial spine. We believe there is another reason for the poor prognosis. For example, the study found that the elderly tend to choose no surgery or amputation surgery, which may be because of the difficulty of tolerating surgery and the difficulty of postoperative reconstruction.
Large tumor size suggests a poor prognosis. On the one hand, the tumor is too large to remove, and more people choose amputation. On the other hand, the possibility of recurrence and metastasis is increased. Previous studies used 100 mm as a metric, which is very close to the 108 mm cut‐off value set in this study. 15
Some studies have claimed that the primary prognostic factor in patients with chondrosarcoma is tumor grade. About 85% of these tumors are of low grade and overall survival is favorable. 16 According to Huang et al.'s study, Grade I (37%) and Grade II (42.8%) are close to the conclusions of this study, and the differences may be due to the higher grade of tumors in the spine, pelvic cavity and craniofacial region. 15 Advanced chondrosarcomas are more aggressive and associated with higher rates of local recurrence and mortality. Previous studies have shown that relative high grade was associated with poor survival (OS: grade III: 2.764; 1.742 to 4.386; p < 0.001; grade IV, 3.854; 2.259 to 6.574; p < 0.001), compared with the outcome of this study (OS: grade III: 1.84; 0.97 to 3.46; p = 0.11; grade IV, 3.61; 1.55 to 8.38; p = 0.05) is close. 15 In addition, this study found that Amputation was less likely to be an option for low‐grade chondrosarcoma.
It is thought that malignant cartilage cells have limited vascular connections, making delivery of chemotherapeutic agents ineffective, then chondrosarcoma is generally resistant to chemotherapy. radiation therapy is reserved for cases when adequate surgical margins cannot be achieved. 16 Surprisingly, both radiotherapy and chemotherapy were risk factors in this study. In Sina Coşkun et al.'s study, radiation was a risk factor for CSS, a result that could be attributed to patients receiving radiation having larger tumor diameter, axial tumor location, and high‐grade tumors. 17 Amer et al.'s study showed that radiotherapy was associated with reduced survival, but there was no statistically significant difference between subtypes. 18 Krochak et al. found that radiotherapy had limited efficacy. 19 But others turned out to be controversial, Like Gao et al. showed that RT was an independent protective factor and that adjuvant RT combined with surgery could improve both overall survival and cancer‐specific survival of patients with high‐grade myxoid and undifferentiated chondrosarcomas. 20 Sina Coşkun et al.'s study found that among patients diagnosed with the undifferentiated subtype, the overall survival rate was 77% in the group that received radiation, compared with only 16.6% in the group that did not receive radiation. 17 Although not statistically significant, this may suggest that combined adjuvant radiotherapy with surgery may be a good option in undifferentiated chondrosarcomas. 17 In data on patients who received radiation or chemotherapy, it was found that these patients tended not to have surgery and had higher tumor stages.
Choice of Surgical Method
Although, for all grades and subtypes of non‐metastatic chondrosarcoma, complete surgical treatment offers the only chance for cure, the most optimal type of surgical management is still debated. Wide, en‐bloc excision is the preferred surgical treatment of intermediate‐ and high‐grade chondrosarcoma cases. 21 However, wide excision can lead to considerable morbidity and a demanding reconstruction, depending on the location. On the other hand, in low‐grade chondrosarcoma, extensive intralesional curettage followed by local adjuvant treatment, for example, phenolization or cryosurgery (liquid nitrogen), and filling the cavity with bone graft has promising long term clinical results and satisfactory local control. 7 , 22 However, in some cases of low‐grade chondrosarcoma, intralesional excision may not be adequate, for example, because of large size or an intra‐articular or pelvic localization. In these cases, wide resection remains the preferable choice for local therapy. 23 , 24
We screened six surgical treatment options for chondrosarcoma of the extremities. In this study, preoperative multivariate analysis of PSM confirmed that the OS of radical resection plus extensive resection was better than that of no surgery (radical resection: HR: 0.45 [0.23–0.87] p = 0.02; extensive resection: HR: 0.30 [0.09 ~ 0.96] p = 0.04).
Based on independent prognostic factors discussed above, we constructed nomograms to predict patients' survival. In order to evaluate its practicability, we applied the internal validation and achieved satisfactory fitting degree (internal validation C‐index: OS, 0.787; CSS, 0.821). In addition, our nomograms displayed better accuracy than the traditional AJCC (OS, 0.640; CSS, 0.673) for predicting survival probability, so this model has higher predictive value and is widely used by clinicians.
PSM was performed to balance the effects of baseline clinicopathological differences. Among them, only local resection versus limb salvage had a significant difference, and local resection had a worse prognosis (OS HR: 1.74 [1.08–2.8] p = 0.049). No other significant differences were found. In view of retrospective studies, statistical methods, and selection bias, our findings should be interpreted with caution.
Strengths and Limitations
This study revealed the risk factors and potential causes of high mortality in patients with chondrosarcoma of the extremities, and innovatively established a nomogram model to predict the overall survival and death risk after surgery for chondrosarcoma of the extremities. Compared with previous studies, the biggest advantages of this prediction model are: it integrates multiple treatment options, including multiple surgical methods and no surgery, so it has high reference value and predictive performance. The model has high sensitivity and specificity, and can directly calculate the probability of various survival times of patients with chondrosarcoma, and provide corresponding treatment recommendations for patients.
Despite the high survival predictive power of this risk model, our study has several limitations. First, all of our samples came from our hospital's database, but there was still some inaccurate variable information that made it impossible to audit the data quality. Second, we did not collect variables such as pathological fracture, surgical margin status, complications, and recurrence, which are considered potential prognostic factors, among which surgical margin width affects local recurrence rate and disease‐specific survival, and is a key and independent prognostic factor for bone tumor survival. However, since the incisal margin was not considered in this paper, there may be a small deviation in prognosis. Third, although this is the first study to establish a survival prediction model for patients with chondrosarcoma with satisfactory accuracy of the surgical approach, additional external validation is needed. Finally, and most importantly, there is too little data available to better explore the differences in prognosis and recurrence among different surgical modalities. In the future, more data will need to be collected and integrated to improve the nomogram.
Conclusion
This study included surgical methods for the first time, as well as different samples, differences in statistical methods and the development of diagnostic and therapeutic techniques, leading to different conclusions from previous studies. Our results highlight the differences in survival between different surgical modalities for non‐traditional chondrosarcoma and suggest that the best surgical modalities should be selected for patients wherever possible. In addition, the present study did indicate that our nomograms based on basal clinicopathologic features could well predict the 1‐, 3‐, and 5‐year survival probability of patients with non‐metastatic chondrosarcoma. In the future, more samples and more detailed data are needed to study the treatment of chondrosarcoma through big data analysis and analyze the important factors affecting the prognosis.
Funding Information
This work was supported by National Natural Science Foundation of China (No. 82172011).
Conflict of Interest Statement
The authors declare that they have no conflicts of interest related to this paper.
Ethics Statement
This study was approved by the Institutional Research Ethics Board and the informed consent requirement was waived. This study was conducted according to the Declaration of Helsinki.
Author Contributions
Study design: Wenhui Wang and Junping Zhen. Methodological development: Wenhui Wang. Data acquisition and statistical analysis: Wenhui Wang. Study Research guidance and supervision: Junping Zhen. All authors contributed to the article and approved the submitted version, and wrote the manuscript.
Supporting information
Fig. S1. Missing value recognition
Figs. S2. and S3. KM curve analysis showed that the optimal age threshold of chondrosarcoma patients.
Figs. S4. and S5 KM curve analysis showed that the optimal tumor size threshold of chondrosarcoma patients.
Fig. S6. After PSM, KM curves of Amputation and Radical resection groups.
Fig. S7. After PSM, KM curves of No surgery and Radical resection groups.
Fig. S8. After PSM, KM curves of Extensive excision and Radical resection groups.
Table S1. Comparison of baseline variables between Amputation and Radical resection in original and matched datasets of patients with chondrosarcoma.
Table S2. Comparison of baseline variables between No surgery and Radical resection in original and matched datasets of patients with chondrosarcoma.
References
- 1. Anderson WJ, Doyle LA. Updates from the 2020 World Health Organization classification of soft tissue and bone Tumours. Histopathology. 2021;78:644–657. [DOI] [PubMed] [Google Scholar]
- 2. Murphey MD, Walker EA, Wilson AJ, Kransdorf MJ, Temple HT, Gannon FH. From the archives of the AFIP: imaging of primary chondrosarcoma: radiologic‐pathologic correlation. Radiographics. 2003;23:1245–1278. [DOI] [PubMed] [Google Scholar]
- 3. Laitinen MK, Parry MC, Le Nail LR, Wigley CH, Stevenson JD, Jeys LM. Locally recurrent chondrosarcoma of the pelvis and limbs can only be controlled by wide local excision. Bone Joint J. 2019;101‐b:266–271. [DOI] [PubMed] [Google Scholar]
- 4. Dierselhuis EF, Goulding KA, Stevens M, Jutte PC, Cochrane Gynaecological, Neuro‐oncology and Orphan Cancer Group . Intralesional treatment versus wide resection for central low‐grade chondrosarcoma of the long bones. Cochrane Database Syst Rev. 2019;3:CD010778. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Song K, Shi X, Liang X, Wang H, Zou F, Lu F, et al. Risk factors for metastasis at presentation with conventional chondrosarcoma: a population‐based study. Int Orthop. 2018;42:2941–2948. [DOI] [PubMed] [Google Scholar]
- 6. Beauchamp CP. Chondrosarcoma in the United States (1973 to 2003): an analysis of 2890 cases from the SEER database. Yearbook Orthopedics. 2010;2010:212–213. [Google Scholar]
- 7. Leerapun T, Hugate RR, Inwards CY, Scully SP, Sim FH. Surgical management of conventional grade I chondrosarcoma of long bones. Clin Orthop Relat Res. 2007;463:166–172. [DOI] [PubMed] [Google Scholar]
- 8. Ozaki T, Lindner N, Hillmann A, Rödl R, Blasius S, Winkelmann W. Influence of intralesional surgery on treatment outcome of chondrosarcoma. Cancer. 1996;77:1292–1297. [DOI] [PubMed] [Google Scholar]
- 9. Compton ML, Cates JMM. Evidence‐based tumor staging of skeletal chondrosarcoma. Am J Surg Pathol. 2020;44:111–119. [DOI] [PubMed] [Google Scholar]
- 10. Wang Z, Chen G, Chen X, Huang X, Liu M, Pan W, et al. Predictors of the survival of patients with chondrosarcoma of bone and metastatic disease at diagnosis. J Cancer. 2019;10:2457–2463. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Jiang L, Gong Y, Jiang J, Zhao D. The novel dynamic nomogram and risk classification system constructed for predicting post‐surgical overall survival and mortality risk in primary chondrosarcoma: a population study based on SEER database. J Cancer Res Clin Oncol. 2023;7:15–29. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Benedetto U, Head SJ, Angelini GD, Blackstone EH. Statistical primer: propensity score matching and its alternatives. Eur J Cardiothorac Surg. 2018;53:1112–1117. [DOI] [PubMed] [Google Scholar]
- 13. Liang J, Hu Z, Zhan C, Wang Q. Using propensity score matching to balance the baseline characteristics. J Thorac Oncol. 2021;16:e45–e46. [DOI] [PubMed] [Google Scholar]
- 14. Giuffrida AY, Burgueno JE, Koniaris LG, Gutierrez JC, Duncan R, Scully SP. Chondrosarcoma in the United States (1973 to 2003): an analysis of 2890 cases from the SEER database. J Bone Joint Surg Am. 2009;91:1063–1072. [DOI] [PubMed] [Google Scholar]
- 15. Huang R, Sun Z, Zheng H, Yan P, Hu P, Yin H, et al. Identifying the prognosis factors and predicting the survival probability in patients with non‐metastatic chondrosarcoma from the SEER database. Orthop Surg. 2019;11:801–810. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Ferguson JL, Turner SP. Bone cancer: diagnosis and treatment principles. Am Fam Physician. 2018;98:205–213. [PubMed] [Google Scholar]
- 17. Sina Coşkun H, Erdoğan F, Büyükceran İ, Dabak N. Evaluation of prognostic factors affecting survival in chondrosarcoma treatment and comparison with literature. Joint Dis Relat Surg. 2022;33:440–448. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Amer KM, Munn M, Congiusta D, Abraham JA, Basu Mallick A. Survival and prognosis of chondrosarcoma subtypes: SEER database analysis. J Orthop Res. 2020;38:311–319. [DOI] [PubMed] [Google Scholar]
- 19. Krochak R, Harwood AR, Cummings BJ, Quirt IC. Results of radical radiation for chondrosarcoma of bone. Radiother Oncol. 1983;1:109–115. [DOI] [PubMed] [Google Scholar]
- 20. Gao Z, Lu T, Song H, Gao Z, Ren F, Ouyang P, et al. Prognostic factors and treatment options for patients with high‐grade chondrosarcoma. Med Sci Monit. 2019;25:8952–8967. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Fiorenza F, Abudu A, Grimer RJ, Carter SR, Tillman RM, Ayoub K, et al. Risk factors for survival and local control in chondrosarcoma of bone. J Bone Joint Surg Br. 2002;84:93–99. [DOI] [PubMed] [Google Scholar]
- 22. Veth R, Schreuder B, Vanbeem H, Pruszczynski M, Derooy J. Cryosurgery in aggressive, benign, and low‐grade malignant bone tumours. Lancet Oncol. 2005;6:25–34. [DOI] [PubMed] [Google Scholar]
- 23. Shemesh SS, Acevedo‐Nieves JD, Pretell‐Mazzini J. Treatment strategies for central low‐grade chondrosarcoma of long bones: a systematic review of the literature and meta‐analysis. Musculoskelet Surg. 2018;102:95–109. [DOI] [PubMed] [Google Scholar]
- 24. Verdegaal SH, Brouwers HF, van Zwet EW, Hogendoorn PCW, Taminiau AHM. Low‐grade chondrosarcoma of long bones treated with intralesional curettage followed by application of phenol, ethanol, and bone‐grafting. J Bone Joint Surg Am. 2012;94:1201–1207. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Fig. S1. Missing value recognition
Figs. S2. and S3. KM curve analysis showed that the optimal age threshold of chondrosarcoma patients.
Figs. S4. and S5 KM curve analysis showed that the optimal tumor size threshold of chondrosarcoma patients.
Fig. S6. After PSM, KM curves of Amputation and Radical resection groups.
Fig. S7. After PSM, KM curves of No surgery and Radical resection groups.
Fig. S8. After PSM, KM curves of Extensive excision and Radical resection groups.
Table S1. Comparison of baseline variables between Amputation and Radical resection in original and matched datasets of patients with chondrosarcoma.
Table S2. Comparison of baseline variables between No surgery and Radical resection in original and matched datasets of patients with chondrosarcoma.