Construction and validation of nomogram to predict distant metastasis in osteosarcoma: a retrospective study

Incidence trends, overall survival, and metastasis prediction using multiple machine learning and deep learning techniques in pediatric and adolescent population with osteosarcoma and Ewing's sarcoma: nomogram and webpage

OBJECTIVE

The objective of this study was to analyze the incidence and overall survival (OS) of osteosarcoma (OSC) and Ewing's sarcoma (EWS) in a pediatric and adolescent population, employing machine learning (ML) and deep learning (DL) models to predict the likelihood of metastasis.

METHODS

Involving 2465 OSC and 1373 EWS patients aged 0-19 years, from 2004 to 2020. ML techniques-Lasso, Ridge Regression, Elastic Net, and Random Forest-were used alongside a deep learning model based on TensorFlow and Keras, to construct predictive models for metastasis. These models were optimized using grid search with cross-validation and evaluated on their performance metrics, including AUC, sensitivity, and accuracy. The variables' importance in metastasis prediction was determined using SHAP values. Statistical analysis was performed using R software, and an online nomogram was developed for clinical use.

RESULTS

The age-adjusted incidence of OSC and EWS from 2004 to 2020 showed a significant uptrend. The deep learning model, iterated 50 times, outperformed the Random Forest model in both loss and accuracy stabilization. The nomogram created demonstrated accurate survival predictions, as evidenced by its calibration curves and the distinction between high and low-risk groups.

CONCLUSION

The increasing trend in age-adjusted incidence of OSC and EWS highlights the need for continued research and improved therapeutic strategies in this domain. The study employed ML and DL models to predict distant metastasis in pediatric and adolescent patients with OSC and EWS, providing a valuable tool for prognosis. The online nomogram developed as a part of this research enhances the models' clinical utility, offering an accessible means for clinicians to predict survival outcomes effectively.

Prognostic Modeling for Bone Sarcomas Based on a Large Prospective Cohort From a Tertiary Care Cancer Center in India

PURPOSE

Outcomes of adolescents and young adults (AYA) with bone sarcomas including osteosarcoma (OGS) and Ewing sarcoma (ES) are affected by various factors including inadvertent previous treatment and poor compliance. We aimed to develop a risk-scoring system derived and validated at a tertiary care cancer center in India.

METHODS

All AYA OGS and ES cases treated at our institute with OGS-12 and Ewing's family of tumors-2001 (EFT-2001) protocols from 2011 to 2021 and 2013 to 2018, respectively, were prospectively analyzed. Weighted scores provided to each prognostic variable on the basis of approximate ratios of the beta coefficients of each factor in the multivariable model were summated to divide patients into three clinically discriminatory risk groups, validated by applying separately to derivation, validation, and whole cohorts.

RESULTS

Among 606 (81.0%) of 748 AYA with nonmetastatic OGS, significant factors included in the prognostic model were failure to complete protocol (hazard ratio [HR], 2.65), previous treatment (HR, 2.93), necrosis <90% (HR, 1.63), joint involvement (HR, 2.0), and serum alkaline phosphatase >median (204 U/L; HR, 1.63). Of 104 (39.5%) of 263 AYA with ES, significant factors were failure to complete protocol (HR, 2.84), previous treatment (HR, 6.37), necrosis <100% (HR, 8.73), and tumor size >8 cm (HR, 2.64). For 142 (38.8%) of 366 AYA with metastatic OGS, significant factors were failure to complete protocol (HR, 5.29), metastases not amenable to local treatment (HR, 1.96), necrosis <90% (HR, 1.96), and >10 metastases (HR, 2.44). For 38 (43.6%) of 82 AYA with metastatic extremity ES, significant factors were failure to complete protocol (HR, 3.88) and metastases not amenable to local treatment (HR, 10.6).

CONCLUSION

We developed simple, effective prognostic models for AYA with bone sarcomas with specific potential relevance for low- and middle-income countries.

Diagnostic value of combined detection of AKP, TSGF, and LDH for pediatric osteosarcoma: a case-control study

OBJECTIVE

To evaluate the diagnostic value of serum alkaline phosphatase (AKP), tumor-supplied growth factor group (TSGF), and lactate dehydrogenase (LDH) for pediatric osteosarcoma.

METHODS

A retrospective analysis of clinical data from 81 pediatric osteosarcoma patients (osteosarcoma group) and 63 patients with benign bone tumors (benign bone tumor group) admitted to Yantaishan Hospital from February 2023 to November 2023 was conducted. Basic and clinical data differences between the two groups of children were compared. A multivariate regression model was established to determine predictive factors for pediatric osteosarcoma, and the diagnostic value of identified indicators for pediatric osteosarcoma was evaluated.

RESULTS

Osteosarcoma group demonstrated significantly higher serum AKP (375.76±73.47 vs 286.12±76.50 U/L), TSGF (69.01±16.30 vs 53.57±16.37 U/mL), and LDH (269.55±66.96 vs 207.46±59.20 U/L) levels as compared to the benign bone tumor group. Correlation analysis suggested significant positive correlations between AKP (rho=0.505), TSGF (rho=406), LDH (rho=0.449) and pediatric osteosarcoma. Multivariate regression analysis showed serum AKP, TSGF, and LDH were independent predictive factor for pediatric osteosarcoma. The AUC value for AKP was 0.794, with a Youden index of 0.459; the AUC value for TSGF was 0.736, with a Youden index of 0.406; and the AUC value for LDH was 0.761, with a Youden index of 0.462. The combined use of these three biomarkers yielded an AUC of 0.886.

CONCLUSION

The combined detection of serum AKP, TSGF, and LDH can enhance the diagnostic accuracy of pediatric osteosarcoma, providing important evidence for clinical treatment.

Pituitary tumor transforming gene 1 promotes proliferation and malignant phenotype in osteosarcoma via NF-κB signaling

BACKGROUND

Pituitary tumor transforming gene (PTTG) is an oncogene reported to be actively promotes tumorigenesis in multiple tumors. Osteosarcoma (OS) is the most common primary osseous sarcoma, however, the functional significance and mechanisms underlying whether and how PTTG1 promotes OS remain largely unknown.

METHODS

Here, in our study, PTTG1 levels in clinical samples and cell lines were determined by western blotting and immunohistochemistry. The viability and migratory/invasive potential of OS cells were assessed using Cell Counting Kit-8, colony formation, wound healing, and Transwell assays. The effects of PTTG1 on NF-κB signaling pathways were evaluated both in vivo and in vitro.

RESULTS

An abnormally elevated expression of PTTG1was confirmed in human OS tissues and OS cell lines and PTTG1 levels were positively correlated with OS clinicopathological grade. We further showed that knocking down PTTG1 attenuated the viability and migratory/invasive capacity of OS cells (MG63 and HOS-8603). Additionally, the following key mechanistic principle was revealed: knockdown PTTG1-mediated OS tumorgenesis supression was associated with inactivation of the NF-κB pathway. We confirmed these results by additional nonpharmacological intervention and same conclusions were obtained in the context of opposite functional analyses. Furthermore, we also demonstrated that OS cell lines overexpressed PTTG1 showed increased tumorigenesis in athymic nude mice.

CONCLUSIONS

To sum up, the present study suggests that PTTG1 is involved in the enhancement of the malignancy and carcinogenesis of OS by regulating NF-κB signaling. Accordingly, PTTG1 likely functions as an oncogene in OS and may represent a potential therapeutic target for this cancer.

A novel model for predicting the prognosis of postoperative intrahepatic cholangiocarcinoma patients

Intrahepatic cholangiocarcinoma (ICC) accounts for 20% of liver malignancies with a 5-year survival rate of 35% at best with limited prognostic predictors. Lung Immune Prognostic Index (LIPI) is a novel prognostic factor in pulmonary cancers. In this study, we developed a modified prognostic model from LIPI called intrahepatic immune prognostic index (IIPI) for ICC. A retrospectively study was conducted at Liver Transplant Center of West China Hospital between January 2015 and January 2023. Hematological factors and clinical features of ICC patients were collected and analyzed. The area under curve (AUC) and optimal cuff-off of each single hematological factor was calculated. In this study, derived neurtrophil to lymphocyte ratio (dNLR), arbohydrate antigen199 (CA199) and carcinoembryonic antigen (CEA) have higher AUC values. LIPI was composed of dNLR and was further modified by combing CA199 and CEA, forming the IIPI. The IIPI consists of four grades which are None, Light, Moderate and Severe. Compared to other prognostic factors, IIPI exhibited better ability to predict overall survival. The multivariate analysis indicated that cirrhosis, differentiation, hilar invasion and IIPI were independent prognostic factors for ICC patients. An IIPI-based nomogram was also established and could predict the overall survival. In addition, the subgroup analyses based on clinical prognostic factors showed that the IIPI exhibited excellent prognostic influence. IIPI model is suitable for predicting the prognosis of postoperative ICC patients. Further research is needed to explore the relationship between postoperative recurrence and metastasis of ICC patients and IIPI.

Lung and bone metastases patterns in limb osteosarcoma: Surgical treatment of primary site improves overall survival

Osteosarcoma (OS) is one of the most prevalent malignant bone tumors. The proportion of patients with limb OS was relatively high. Lung metastasis (LM) and bone metastasis are the first and second most common metastatic types of OS, respectively. A total of 270 new cases of LM, 55 new cases of bone metastases (BM), and 36 new cases of lung and BM were diagnosed in the surveillance, epidemiology and end results database from 2010 to 2019. Univariate and multivariate logistic regression analyses were used to identify the risk factors for lung and/or BM, and Cox regression analyses were performed to identify the prognostic factors for lung and/or BM. Kaplan-Meier curves and log-rank tests were used to analyze the overall survival of limb OS patients with lung and/or BM. Female sex, telangiectatic OS type, central OS type, T3 stage, N1 stage, BM, surgical treatments, radiotherapy and chemotherapy were significantly correlated with LM. T3 stage, LM, liver metastases, and radiotherapy significantly correlated with BM. The small cell OS type, T2 stage, T3 stage, N1 stage, liver metastases, and radiotherapy were significantly correlated with lung and BM. Among limb OS patients with LM, the mean survival months of older age, black race, N1 stage, BM, brain metastases, no surgery, and no chemotherapy were lower than those of the control group. In limb OS patients with LM and BM, the mean survival months in the no surgery group was lower than in the surgery group. T stage and radiotherapy significantly influence the occurrence of limb OS with lung and/or BM. Surgery at the primary site has been shown to be effective in improving the survival rate of patients with lung and/or BM.

Predicting metastasis at initial diagnosis and radiotherapy effectiveness in patients with metastatic osteosarcoma

Osteosarcoma is a primary malignant bone tumor affecting mostly children and adolescents. The overall 10 year survivals of patients with metastatic osteosarcoma are typically less than 20% in the literature and remain concerning. We aimed to develop a nomogram for predicting the risk of metastasis at initial diagnosis in patients with osteosarcoma and evaluate the effectiveness of radiotherapy in patients with metastatic osteosarcoma. Clinical and demographic data of patients with osteosarcoma were collected from the surveillance, epidemiology, and end results database. We randomly split our analytical sample into the training and validation cohorts, then established and validated a nomogram for predicting the risk of osteosarcoma metastasis at initial diagnosis. The effectiveness of radiotherapy was evaluated by performing propensity score matching in patients underwent surgery + chemotherapy and those underwent surgery + chemotherapy + radiotherapy, among patients with metastatic osteosarcoma. 1439 patients met the inclusion criteria and were included in this study. 343 of 1439 had osteosarcoma metastasis by the time of initial presentation. A nomogram for predicting the likelihood of osteosarcoma metastasis by the time of initial presentation was developed. In both unmatched and matched samples, the radiotherapy group demonstrated a superior survival profile comparing with the non-radiotherapy group. Our study established a novel nomogram to evaluate the risk of osteosarcoma with metastasis, and demonstrated that radiotherapy combined with chemotherapy and surgical resection could improve 10-year survival in patients with metastasis. These findings may guide the clinical decision-making for orthopedic surgeons.

Formulation and validation of a baseline prognostic score for osteosarcoma treated uniformly with a non-high dose methotrexate-based protocol from a low middle income healthcare setting: a single centre analysis of 594 patients

INTRODUCTION

The outcomes of osteosarcoma in low middle income countries (LMICs) are different due to patients presenting in advanced stages, resource constraints and the use of non-high-dose-methotrexate (HDMTX)-based regimens. This study derived and validated a prognostic score for osteosarcoma that integrates biologic and social factors and is tailored for patients from an LMIC setting using a non-HDMTX-based protocol.

MATERIALS AND METHODS

A retrospective study including osteosarcoma patients enrolled for treatment at a single tertiary care centre in India between 2003-19 was conducted. Baseline biologic and social characteristics were extracted from medical records and survival outcomes were noted. The cohort was randomised into a derivation and validation cohort. Multivariable Cox regression was used to identify baseline characteristics that were independently prognostic for survival outcomes in the derivation cohort. A score was derived from the prognostic factors identified in the derivation cohort and further validated in the validation cohort with estimation of its predictive ability.

RESULTS

594 patients with osteosarcoma were eligible for inclusion in the study. Around one-third of the cohort had metastatic disease with 59% of the patients residing in rural areas. The presence of metastases at baseline (HR 3.39; p<0.001; score=3), elevated serum alkaline phosphatase (SAP) >450 IU/L (HR 1.57; p=0.001; score=1) and baseline tumour size > 10 cm (HR 1.68; p<0.001; score=1) were identified to be independent factors predicting inferior event free survival (EFS) and were included in development of the prognostic score. Patients were categorized as low risk (score 0), intermediate risk (score 1-3) and high risk (4-5). Harrell's c-indices for the score were 0.682, 0.608 and 0.657 respectively for EFS in the derivation, validation and whole cohort respectively. The timed AUC of ROC was 0.67 for predicting 18-month EFS in the derivation, validation and whole cohorts while that for 36-month EFS were 0.68, 0.66 and 0.68 respectively.

CONCLUSIONS

The study describes the outcomes among osteosarcoma patients from an LMIC treated uniformly with a non-HDMTX-based protocol. Tumor size, baseline metastases and SAP were prognostic factors used to derive a score with good predictive value for survival outcomes. Social factors did not emerge as determinants of survival.

A clinical prediction model for lung metastasis risk in osteosarcoma: A multicenter retrospective study

Background

Lung metastases (LM) have a poor prognosis of osteosarcoma. This study aimed to predict the risk of LM using the nomogram in patients with osteosarcoma.

Methods

A total of 1100 patients who were diagnosed as osteosarcoma between 2010 and 2019 in the Surveillance, Epidemiology and End Results (SEER) database were selected as the training cohort. Univariate and multivariate logistic regression analyses were used to identify independent prognostic factors of osteosarcoma lung metastases. 108 osteosarcoma patients from a multicentre dataset was as valiation data. The predictive power of the nomogram model was assessed by receiver operating characteristic curves (ROC) and calibration plots, and decision curve analysis (DCA) was utilized to interpret the accurate validity in clinical practice.

Results

A total of 1208 patients with osteosarcoma from both the SEER database(n=1100) and the multicentre database (n=108) were analyzed. Univariate and multivariate logistic regression analyses showed that Survival time, Sex, T-stage, N-stage, Surgery, Radiation, and Bone metastases were independent risk factors for lung metastasis. We combined these factors to construct a nomogram for estimating the risk of lung metastasis. Internal and external validation showed significant predictive differences (AUC 0.779, 0.792 respectively). Calibration plots showed good performance of the nomogram model.

Conclusions

In this study, a nomogram model for predicting the risk of lung metastases in osteosarcoma patients was constructed and turned out to be accurate and reliable through internal and external validation. Moreover we built a webpage calculator (https://drliwenle.shinyapps.io/OSLM/) taken into account nomogram model to help clinicians make more accurate and personalized predictions.

Lung and bone metastases patterns in osteosarcoma: Chemotherapy improves overall survival

Osteosarcoma (OS) is a malignant tumor originating from the mesenchymal tissue. Simultaneous reports of lung and bone metastases (BM) in OS are rare in the literature. A total of 353 new cases of lung metastases (LM), 93 new cases of BM, and 59 new cases of LM and BM were diagnosed in the Surveillance, Epidemiology and End Results (SEER) database from 2010 to 2019. Univariate and multivariate logistic regression analyses were used to identify risk factors for LM and/or BM, and Cox regression analyses were performed to identify the prognostic factors for LM and/or BM. Kaplan-Meier (K-M) curves and log-rank tests were used to analyze the overall survival of patients with LM and/or BM. LM was diagnosed in 353 patients. Female sex, tumor size >100 mm, telangiectatic OS type, central OS type, N1 stage, other locations, BM, surgical treatments, radiotherapy and chemotherapy were significantly correlated with LM. 93 patients were diagnosed with BM. 25 to 59 years old, T1 stage, presence of LM, liver metastases, radiotherapy, and surgical treatments were significantly correlated with the BM. 59 patients were diagnosed with LM and BM. The chondroblastic OS type, small cell OS type, T1 stage, N1 stage, other locations, liver metastases, radiotherapy, and surgical treatments were significantly correlated with LM and BM. Metastases, radiotherapy, and surgery at the primary site were significantly associated with LM and/or BM. Chemotherapy at the primary site has been shown to be effective in improving the survival rate of LM and/or BM. Of the OS patients with LM, 61.47% died, and older age, BM, no surgery, and no chemotherapy were harmful to survival. 72.04% of OS patients with BM died, and N1 stage, no surgery, and no chemotherapy were harmful for survival. 69.49% of OS patients with LM and BM died, and older age and no chemotherapy were harmful for survival.

Interpretable clinical visualization model for prediction of prognosis in osteosarcoma: a large cohort data study

Background

Currently, the clinical prediction model for patients with osteosarcoma was almost developed from single-center data, lacking external validation. Due to their low reliability and low predictive power, there were few clinical applications. Our study aimed to set up a clinical prediction model with stronger predictive ability, credibility, and clinical application value for osteosarcoma.

Methods

Clinical information related to osteosarcoma patients from 2010 to 2016 was collected in the SEER database and four different Chinese medical centers. Factors were screened using three models (full subset regression, univariate Cox, and LASSO) via minimum AIC and maximum AUC values in the SEER database. The model was selected by the strongest predictive power and visualized by three statistical methods: nomogram, web calculator, and decision tree. The model was further externally validated and evaluated for its clinical utility in data from four medical centers.

Results

Eight predicting factors, namely, age, grade, laterality, stage M, surgery, bone metastases, lung metastases, and tumor size, were selected from the model based on the minimum AIC and maximum AUC value. The internal and external validation results showed that the model possessed good consistency. ROC curves revealed good predictive ability (AUC &gt; 0.8 in both internal and external validation). The DCA results demonstrated that the model had an excellent clinical predicted utility in 3 years and 5 years for North American and Chinese patients.

Conclusions

The clinical prediction model was built and visualized in this study, including a nomogram and a web calculator (https://dr-lee.shinyapps.io/osteosarcoma/), which indicated very good consistency, predictive power, and clinical application value.

Osteosarcoma immune prognostic index can indicate the nature of indeterminate pulmonary nodules and predict the metachronous metastasis in osteosarcoma patients

Purpose

The relationship between indeterminate pulmonary nodules (IPNs) and metastasis is difficult to determine. We expect to explore a predictive model that can assist in indicating the nature of IPNs, as well as predicting the probability of metachronous metastasis in osteosarcoma patients.

Patients and methods

We conducted a retrospective study including 184 osteosarcoma patients at West China Hospital from January 2016 to January 2021. Hematological markers and clinical features of osteosarcoma patients were collected and analyzed.

Results

In this study, we constructed an osteosarcoma immune prognostic index (OIPI) based on the lung immune prognostic index (LIPI). Compared to other hematological markers and clinical features, OIPI had a better ability to predict metastasis. OIPI divided 184 patients into four groups, with the no-OIPI group (34 patients), the light-OIPI group (35 patients), the moderate-OIPI group (75 patients), and the severe-OIPI group (40 patients) (P < 0.0001). Subgroup analysis showed that the OIPI could have a stable predictive effect in both the no-nodule group and the IPN group. Spearman’s rank correlation test and Kruskal–Wallis test demonstrated that the OIPI was related to metastatic site and metastatic time, respectively. In addition, patients with IPNs in high-OIPI (moderate and severe) groups were more likely to develop metastasis than those in low-OIPI (none and light) groups. Furthermore, the combination of OIPI with IPNs can more accurately identify patients with metastasis, in which the high-OIPI group had a higher metastasis rate, and the severe-OIPI group tended to develop metastasis earlier than the no-OIPI group. Finally, we constructed an OIPI-based nomogram to predict 3- and 5-year metastasis rates. This nomogram could bring net benefits for more patients according to the decision curve analysis and clinical impact curve.

Conclusion

This study is the first to assist chest CT in diagnosing the nature of IPNs in osteosarcoma based on hematological markers. Our findings suggested that the OIPI was superior to other hematological markers and that OIPI can act as an auxiliary tool to determine the malignant transformation tendency of IPNs. The combination of OIPI with IPNs can further improve the metastatic predictive ability in osteosarcoma patients.

Lung Immune Prognostic Index Could Predict Metastasis in Patients With Osteosarcoma

Background

The lung immune prognostic index (LIPI), composed of serum lactate dehydrogenase (LDH) and the derived neutrophil to lymphocyte ratio (dNLR), is a novel prognostic factor of lung cancer. The prognostic effect of the LIPI has never been verified in osteosarcoma.

Methods

We retrospectively reviewed the osteosarcoma patients with metachronous metastasis from January 2016 to January 2021 in West China Hospital. We collected and analyzed the clinical data and constructed the LIPI for osteosarcoma. The correlation between the LIPI and metastasis was analyzed according to the Kaplan–Meier method and Cox regression analysis with hazard ratios (HRs) and 95% confidence intervals (CIs). Univariate analysis and multivariate analysis were conducted to clarify the independent risk factors of metastasis. The nomogram model was established by R software, version 4.1.0.

Results

The area under the curve (AUC) and best cutoff value were 0.535 and 91, 0.519, and 5.02, 0.594 and 2.77, 0.569 and 227.14, 0.59 and 158, and 0.607 and 2.05 for ALP, LMR, NLR, PLR, LDH, and dNLR, respectively. The LIPI was composed of LDH and dNLR and showed a larger AUC than other hematological factors in the time-dependent operator curve (t-ROC). In total, 184 patients, 42 (22.8%), 96 (52.2%), and 46 (25.0%) patients had LIPIs of good, moderate, and poor, respectively (P &lt; 0.0001). Univariate analysis revealed that pathological fracture, the initial CT report of suspicious nodule, and the NLR, PLR, ALP, and the LIPI were significantly associated with metastasis, and multivariate analysis showed that the initial CT report of suspicious nodule and the PLR, ALP, and LIPI were dependent risk factors for metastasis. Metastatic predictive factors were selected and incorporated into the nomogram construction, including the LIPI, ALP, PLR, initial CT report, and pathological fracture. The C-index of our model was 0.71. According to the calibration plot, this predictive nomogram could accurately predict 3- and 5-year metachronous metastasis. Based on the result of decision curve and clinical impact curve, this predictive nomogram could also help patients obtain significant net benefits.

Conclusion

We first demonstrated the metastatic predictive effect of the LIPI on osteosarcoma. This LIPI-based model is useful for clinicians to predict metastasis in osteosarcoma patients and could help conduct timely intervention and facilitate personalized management of osteosarcoma patients.

An External-Validated Prediction Model to Predict Lung Metastasis among Osteosarcoma: A Multicenter Analysis Based on Machine Learning

BACKGROUND

Lung metastasis greatly affects medical therapeutic strategies in osteosarcoma. This study aimed to develop and validate a clinical prediction model to predict the risk of lung metastasis among osteosarcoma patients based on machine learning (ML) algorithms.

METHODS

We retrospectively collected osteosarcoma patients from the Surveillance Epidemiology and End Results (SEER) database and from four hospitals in China. Six ML algorithms, including logistic regression (LR), gradient boosting machine (GBM), extreme gradient boosting (XGBoost), random forest (RF), decision tree (DT), and multilayer perceptron (MLP), were applied to build predictive models for predicting lung metastasis using patient's demographics, clinical characteristics, and therapeutic variables from the SEER database. The model was internally validated using 10-fold cross-validation to calculate the mean area under the curve (AUC) and the model was externally validated using the Chinese multicenter osteosarcoma data. Relative importance ranking of predictors was plotted to understand the importance of each predictor in different ML algorithms. The correlation heat map of predictors was plotted to understand the correlation of each predictor, selecting the 10-fold cross-validation with the highest AUC value in the external validation ROC curve to build a web calculator.

RESULTS

Of all enrolled patients from the SEER database, 17.73% (194/1094) developed lung metastasis. The multiple logistic regression analysis showed that sex, N stage, T stage, surgery, and bone metastasis were all independent risk factors for lung metastasis. In predicting lung metastasis, the mean AUCs of the six ML algorithms ranged from 0.711 to 0.738 in internal validation and 0.697 to 0.729 in external validation. Among the six ML algorithms, the extreme gradient boosting (XGBoost) model had the highest AUC value with an average internal AUC of 0.738 and an external AUC of 0.729. The best performing ML algorithm model was used to build a web calculator to facilitate clinicians to calculate the risk of lung metastasis for each patient.

CONCLUSIONS

The XGBoost model may have the best prediction effect and the online calculator based on this model can help doctors to determine the lung metastasis risk of osteosarcoma patients and help to make individualized medical strategies.

A Nomogram for Predicting Cancer-Specific Survival of Osteosarcoma and Ewing's Sarcoma in Children: A SEER Database Analysis

Background

Osteosarcoma (OSC) and Ewing's sarcoma (EWS) are children's most common primary bone tumors. The purpose of the study is to develop and validate a new nomogram to predict the cancer-specific survival (CSS) of childhood OSC and EWS.

Methods

The clinicopathological information of all children with OSC and EWS from 2004 to 2018 was downloaded from the Surveillance, Epidemiology, and End Results (SEER) database. Univariate and multivariate Cox regression analyses were used to screen children's independent risk factors for CSS. These risk factors were used to construct a nomogram to predict the CSS of children with OSC and EWS. A series of validation methods, including calibration plots, consistency index (C-index), and area under the receiver operating characteristic curve (AUC), were used to validate the accuracy and reliability of the prediction model. Decision curve analysis (DCA) was used to validate the clinical application efficacy of predictive models. All patients were divided into low- and high-risk groups based on the nomogram score. Kaplan-Meier curve and log-rank test were used to compare survival differences between the two groups.

Results

A total of 2059 children with OSC and EWS were included. All patients were randomly divided into training cohort 60% (N = 1215) and validation cohort 40% (N = 844). Univariate and multivariate analysis suggested that age, surgery, stage, primary site, tumor size, and histological type were independent risk factors. Nomograms were established based on these factors to predict 3-, 5-, and 8-years CSS of children with OSC and EWS. The calibration plots showed that the predicted value was highly consistent with the actual value. In the training cohort and validation cohort, the C-index was 0.729 (0.702–0.756) and 0.735 (0.702–0.768), respectively. The AUC of the training cohort and the validation cohort also showed similar results. The DCA showed that the nomogram had good clinical value.

Conclusion

We constructed a new nomogram to predict the CSS of OSC and EWS in children. This predictive model has good accuracy and reliability and can help doctors and patients develop clinical strategies.

Development and Validation of Nomograms to Assess Risk Factors and Overall Survival Prediction for Lung Metastasis in Young Patients with Osteosarcoma: A SEER-Based Study

BACKGROUND

To establish two nomograms to quantify the diagnostic factors of lung metastasis (LM) and their role in assessing prognosis in young patients with LM osteosarcoma.

METHODS

A total of 618 osteosarcoma young patients from 2010 to 2015 were included from the Surveillance, Epidemiology, and End Results (SEER) database. Another 131 patients with osteosarcoma from local hospitals were also collected as an external validation set. Patients were randomized into training sets (n = 434) and validation sets (n = 184) with a ratio of 7:3. Univariate and multivariate logistic regression analyses were used to identify the risk factor for LM and were used to construct the nomogram. Risk variables for the overall survival rate of patients with LM were evaluated by Cox regression. Another nomogram was also constructed to predict survival rates. The results were validated using bootstrap resampling and retrospective research on 131 osteosarcoma young patients from 2010 to 2019 at three local hospitals.

RESULTS

There were 114 (18.45%) patients diagnosed as LM at initial diagnosis. The multivariate logistic regression analysis suggested that T stage, N stage, and bone metastasis were independent risk factors for LM in newly diagnosed young osteosarcoma patients (P < 0.001). The ROC analysis revealed that area under the curve (AUC) values were 0.751, 0.821, and 0.735 in the training set, internal validation set, and external validation set, respectively, indicating good predictive discrimination. The multivariate Cox proportional hazard regression analysis suggested that age, surgery, chemotherapy, primary site, and bone metastasis were prognostic factors for young osteosarcoma patients with LM. The time-dependent ROC curves showed that the AUCs for predicting 1-year, 2-year, and 3-year survival rates were 0.817, 0.792, and 0.815 in the training set and 0.772, 0.807, and 0.804 in the internal validation set, respectively. As for the external validation set, the AUCs for predicting 1-year, 2-year, and 3-year survival rates were 0.787, 0.818, and 0.717.

CONCLUSIONS

The nomograms can help clinicians strengthen their personal decision-making and can improve the prognosis of osteosarcoma patients.

Risk Factors, Prognostic Factors, and Nomograms for Distant Metastasis in Patients With Newly Diagnosed Osteosarcoma: A Population-Based Study

BACKGROUND

Osteosarcoma is the most common bone cancer, mainly occurring in children and adolescents, among which distant metastasis (DM) still leads to a poor prognosis. Although nomogram has recently been used in tumor areas, there are no studies focused on diagnostic and prognostic evaluation of DM in primary osteosarcoma patients.

METHODS

The data of osteosarcoma patients diagnosed between 2004 and 2015 were extracted from the Surveillance, Epidemiology, and End Results (SEER) database. Univariate and multivariate logistic regression analyses were used to identify independent risk factors for DM in osteosarcoma patients, and univariate and multivariate Cox proportional hazards regression analyses were used to determine independent prognostic factors of osteosarcoma patients with DM. We then established two novel nomograms and the results were evaluated by receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA).

RESULT

A total of 1,657 patients with osteosarcoma were included, and 267 patients (16.11%) had DM at the time of diagnosis. The independent risk factors for DM in patients with osteosarcoma include age, grade, T stage, and N stage. The independent prognostic factors for osteosarcoma patients with DM are age, chemotherapy and surgery. The results of ROC curves, calibration, DCA, and Kaplan-Meier (K-M) survival curves in the training, validation, and expanded testing sets, confirmed that two nomograms can precisely predict occurrence and prognosis of DM in osteosarcoma patients.

CONCLUSION

Two nomograms are expected to be effective tools for predicting the risk of DM for osteosarcoma patients and personalized prognosis prediction for patients with DM, which may benefit clinical decision-making.