Prediction Of 5-year Survival Research Articles

Deep learning enabled prediction of 5-year survival in pediatric genitourinary rhabdomyosarcoma

BackgroundGenitourinary rhabdomyosarcoma (GU-RMS) is a rare, pediatric malignancy originating from embryonic mesenchyme. Current approaches to prognostication rely upon conventional statistical methods such as Cox proportional hazards (CPH) models and have suboptimal predictive ability. Given the success of deep learning approaches in other specialties, we sought to develop and compare deep learning models with CPH models for the prediction of 5-year survival in pediatric GU-RMS patients. MethodsPatients less than 20 years of age with GU-RMS were identified within the Surveillance, Epidemiology, and End Results (SEER) database (1998–2011). Deep neural networks (DNN) were trained and tested on an 80/20 split of the dataset in a 5-fold cross-validated fashion. Multivariable CPH models were developed in parallel. The primary outcomes were 5-year overall survival (OS) and disease-specific survival (DSS). Variables used for prediction were age, sex, race, primary site, histology, degree of tumor extension, tumor size, receipt of surgery, and receipt of radiation. Receiver operating characteristic curve analysis was conducted, and DNN models were tested for calibration. Results277 patients were included. The area under the curve (AUC) for the DNN models was 0.93 for OS and 0.91 for DSS. AUC for the CPH models was 0.82 for OS and 0.84 for DSS. The DNN models were well-calibrated: OS model (slope = 1.02, intercept = −0.06) and DSS model (slope = 0.79, intercept = 0.21). ConclusionsA deep learning-based model demonstrated excellent performance, superior to that of CPH models, in the prediction of pediatric GU-RMS survival. Deep learning approaches may enable improved prognostication for patients with rare cancers.

Improving five-year survival prediction via multitask learning across HPV-related cancers.

Oncology is a highly siloed field of research in which sub-disciplinary specialization has limited the amount of information shared between researchers of distinct cancer types. This can be attributed to legitimate differences in the physiology and carcinogenesis of cancers affecting distinct anatomical sites. However, underlying processes that are shared across seemingly disparate cancers probably affect prognosis. The objective of the current study is to investigate whether multitask learning improves 5-year survival cancer patient survival prediction by leveraging information across anatomically distinct HPV related cancers. Data were obtained from the Surveillance, Epidemiology, and End Results (SEER) program database. The study cohort consisted of 29,768 primary cancer cases diagnosed in the United States between 2004 and 2015. Ten different cancer diagnoses were selected, all with a known association with HPV risk. In the analysis, the cancer diagnoses were categorized into three distinct topography groups of varying specificity. The most specific topography grouping consisted of 10 original cancer diagnoses differentiated by the first two digits of the ICD-O-3 topography code. The second topography grouping consisted of cancer diagnoses categorized into six distinct organ groups. Finally, the third topography grouping consisted of just two groups, head-neck cancers and ano-genital cancers. The tasks were to predict 5-year survival for patients within the different topography groups using 14 predictive features which were selected among descriptive variables available in the SEER database. The information from the predictive features was shared between tasks in three different ways, resulting in three distinct predictive models: 1) Information was not shared between patients assigned to different tasks (single task learning); 2) Information was shared between all patients, regardless of task (pooled model); 3) Only relevant information was shared between patients grouped to different tasks (multitask learning). Prediction performance was evaluated with Brier scores. All three models were evaluated against one another on each of the three distinct topography-defined tasks. The results showed that multitask classifiers achieved relative improvement for the majority of the scenarios studied compared to single task learning and pooled baseline methods. In this study, we have demonstrated that sharing information among anatomically distinct cancer types can lead to improved predictive survival models.

Identifying biomolecules and constructing a prognostic risk prediction model for recurrence in osteosarcoma

IntroductionOsteosarcoma is a high-morbidity bone cancer with an unsatisfactory prognosis. The aim of this study is to develop novel potential prognostic biomarkers and construct a prognostic risk prediction model for recurrence in osteosarcoma. MethodsBy analyzing microarray data, univariate and multivariate Cox regression analyses were performed to screen prognostic RNA signatures and to build a prognostic model. The RNA signatures were validated using Kaplan–Meier curves. Then, we developed and validated a nomogram combining age, recurrence, metastatic, and Prognostic score (PS) models to predict the individual’s overall survival at the 3- and 5-year points. Pathway enrichment of RNA was conducted based on the significant co-expressed RNAs. ResultsA total of 319 mRNAs and 14 lncRNAs were identified in the microarray data. One lncRNA (LINC00957) and six mRNAs (METL1, CA9, B3GALT4, ALDH1A1, LAMB3, and ITGB4) were identified as RNA signatures and showed good performances in survival prediction for both the training and validation cohorts. Cox regression analysis showed that the seven RNA signatures could independently predict overall survival. Furthermore, age, recurrence, metastatic, and PS models were identified as independent prognostic factors via univariate and multivariate Cox analyses (P < 0.05) and included in the prognostic nomogram. The C-index values for the 3- and 5-year overall survival predictions of the nomogram were 0.809 and 0.740, respectively. ConclusionsThe current study provides the novel potential of seven RNA candidates as prognostic biomarkers. Nomograms were constructed to provide accurate and individualized survival prediction for recurrence in osteosarcoma patients.

Integrative Analysis of DNA Methylation Identified 12 Signature Genes Specific to Metastatic ccRCC.

Background: Abnormal epigenetic alterations can contribute to the development of human malignancies. Identification of these alterations for early screening and prognosis of clear cell renal cell carcinoma (ccRCC) has been a highly sought-after goal. Bioinformatic analysis of DNA methylation data provides broad prospects for discovery of epigenetic biomarkers. However, there is short of exploration of methylation-driven genes of ccRCC.Methods: Gene expression data and DNA methylation data in metastatic ccRCC were sourced from the Gene Expression Omnibus (GEO) database. Differentially methylated genes (DMGs) at 5′-C-phosphate-G- 3′ (CpG) sites and differentially expressed genes (DEGs) were screened and the overlapping genes in DMGs and DEGs were then subject to gene set enrichment analysis. Next, the weighted gene co-expression network analysis (WGCNA) was used to search hub DMGs associated with ccRCC. Cox regression and ROC analyses were performed to screen potential biomarkers and develop a prognostic model based on the screened hub genes.Results: Three hundred and fourteen overlapping DMGs were obtained from two independent GEO datasets. The turquoise module contained 79 hub DMGs, which represent the most significant module screened by WGCNA. Furthermore, a total of 12 hub genes (CETN3, DCAF7, GPX4, HNRNPA0, NUP54, SERPINB1, STARD5, TRIM52, C4orf3, C12orf51, and C17orf65) were identified in the TCGA database by multivariate Cox regression analyses. All the 12 genes were then used to generate the model for diagnosis and prognosis of ccRCC. ROC analysis showed that these genes exhibited good diagnostic efficiency for metastatic and non-metastatic ccRCC. Furthermore, the prognostic model with the 12 methylation-driven genes demonstrated a good prediction of 5-year survival rates for ccRCC patients.Conclusion: Integrative analysis of DNA methylation data identified 12 signature genes, which could be used as epigenetic biomarkers for prognosis of metastatic ccRCC. This prognostic model has a good prediction of 5-year survival for ccRCC patients.

Comparison of the Tree-Based Machine Learning Algorithms to Cox Regression in Predicting the Survival of Oral and Pharyngeal Cancers: Analyses Based on SEER Database

Simple SummaryFormulating accurate survival prediction models of oral and pharyngeal cancers (OPCs) is important, as they might impact the decisions of clinicians and patients. Improving the quality of these clinical prediction modelling studies can benefit the reliability of the developed models and facilitate their implementations in clinical practice. Given the growing trend on the application of machine learning methods in cancer research, we present the use of popular tree-based machine learning algorithms and compare them to the standard Cox regression as an aim to predict OPCs survival. The predictive models discussed here are based on a large cancer registry dataset incorporating various prognosis factors and different forms of bias. The comparable predictive performance between Cox and tree-based models suggested that these machine learning algorithms provide non-parametric alternatives to Cox regression and are of clinical use for estimating the survival probability of OPCs patients.This study aims to demonstrate the use of the tree-based machine learning algorithms to predict the 3- and 5-year disease-specific survival of oral and pharyngeal cancers (OPCs) and compare their performance with the traditional Cox regression. A total of 21,154 individuals diagnosed with OPCs between 2004 and 2009 were obtained from the Surveillance, Epidemiology, and End Results (SEER) database. Three tree-based machine learning algorithms (survival tree (ST), random forest (RF) and conditional inference forest (CF)), together with a reference technique (Cox proportional hazard models (Cox)), were used to develop the survival prediction models. To handle the missing values in predictors, we applied the substantive model compatible version of the fully conditional specification imputation approach to the Cox model, whereas we used RF to impute missing data for the ST, RF and CF models. For internal validation, we used 10-fold cross-validation with 50 iterations in the model development datasets. Following this, model performance was evaluated using the C-index, integrated Brier score (IBS) and calibration curves in the test datasets. For predicting the 3-year survival of OPCs with the complete cases, the C-index in the development sets were 0.77 (0.77, 0.77), 0.70 (0.70, 0.70), 0.83 (0.83, 0.84) and 0.83 (0.83, 0.86) for Cox, ST, RF and CF, respectively. Similar results were observed in the 5-year survival prediction models, with C-index for Cox, ST, RF and CF being 0.76 (0.76, 0.76), 0.69 (0.69, 0.70), 0.83 (0.83, 0.83) and 0.85 (0.84, 0.86), respectively, in development datasets. The prediction error curves based on IBS showed a similar pattern for these models. The predictive performance remained unchanged in the analyses with imputed data. Additionally, a free web-based calculator was developed for potential clinical use. In conclusion, compared to Cox regression, ST had a lower and RF and CF had a higher predictive accuracy in predicting the 3- and 5-year OPCs survival using SEER data. The RF and CF algorithms provide non-parametric alternatives to Cox regression to be of clinical use for estimating the survival probability of OPCs patients.

13-lncRNAs Signature to Improve Diagnostic and Prognostic Prediction of Hepatocellular Carcinoma.

Hepatocellular carcinoma (HCC) is a common type of cancer with a high mortality rate and is usually detected at the middle or late stage, missing the optimal treatment period. The current study aims to identify potential long non-coding RNA (lncRNAs) biomarkers that contribute to the diagnosis and prognosis of HCC. The differentially expressed lncRNAs (DElncRNAs) in HCC patients were detected from the Cancer Genome Atlas (TCGA) dataset. LncRNAs signature was screened by LASSO regression, univariate, and multivariate Cox regression. The models for predicting diagnosis and prognosis were established, respectively. The prognostic model was evaluated by Kaplan-Meier survival curve receiver operating characteristic (ROC) curve and stratified analysis. The diagnostic model was validated by ROC. The lncRNAs signature was further demonstrated by functional enrichment analysis. We found the 13-lncRNAs signature that had a good performance in predicting prognosis and could help to improve the value of diagnosis. In the training set, testing set, and entire cohort, the low-risk group had longer survival than the high-risk group (median OS: 3124 vs. 649 days, 2456 vs. 770 days and 3124 vs. 755 days). It performed well in 1-, 3-, and 5-year survival prediction. 13-lncRNAs-based risk score, age, and race were good predictors of prognosis. The AUC of diagnosis was 0.9487, 0.9265, and 0.9376, respectively. Meanwhile, the 13-lncRNAs were involved in important pathways, including the cell cycle and multiple metabolic pathways. In our study, the 13-lncRNAs signature may be a potential marker for the prognosis of HCC and improve the diagnosis.

Multi-omics Data Analyses Construct TME and Identify the Immune-Related Prognosis Signatures in Human LUAD

Lung cancer has been the focus of attention for many researchers in recent years for the leading contribution to cancer-related death worldwide, in which lung adenocarcinoma (LUAD) is the most common histological type. However, the potential mechanism behind LUAD initiation and progression remains unclear. Aiming to dissect the tumor microenvironment of LUAD and to discover more informative prognosis signatures, we investigated the immune-related differences in three types of genetic or epigenetic characteristics (expression status, somatic mutation, and DNA methylation) and considered the potential roles that these alterations have in the immune response and both the immune-related metabolic and neural systems by analyzing the multi-omics data from The Cancer Genome Atlas (TCGA) portal. Additionally, a four-step strategy based on lasso regression and Cox regression was used to construct the prognostic prediction model. For the prognostic predictions on the independent test set, the performance of the trained models (average concordance index [C-index] = 0.839) is satisfied, with average 1-year, 3-year, and 5-year areas under the curve (AUCs) equal to 0.796, 0.786, and 0.777. Finally, the overall model was constructed based on all samples, which comprised 27 variables and achieved a high degree of accuracy on the 1-year (AUC = 0.861), 3-year (AUC = 0.850), and 5-year (AUC = 0.916) survival predictions.

A Novel Ten-Gene Signature Predicting Prognosis in Hepatocellular Carcinoma.

Hepatocellular carcinoma (HCC) has a dismal long-term outcome. We aimed to construct a multi-gene model for prognosis prediction to inform HCC management. The cancer-specific differentially expressed genes (DEGs) were identified using RNA-seq data of paired tumor and normal tissue. A prognostic signature was built by LASSO regression analysis. Gene set enrichment analysis (GSEA) was performed to further understand the underlying molecular mechanisms. A 10-gene signature was constructed to stratify the TCGA and ICGC cohorts into high- and low-risk groups where prognosis was significantly worse in the high-risk group across cohorts (P < 0.001 for all). The 10-gene signature outperformed all previously reported models for both C-index and the AUCs for 1-, 3-, 5-year survival prediction (C-index, 0.84 vs 0.67 to 0.73; AUCs for 1-, 3- and 5-year OS, 0.84 vs 0.68 to 0.79, 0.81 to 0.68 to 0.80, and 0.85 vs 0.67 to 0.78, respectively). Multivariate Cox regression analysis revealed risk group and tumor stage to be independent predictors of survival in HCC. A nomogram incorporating tumor stage and signature-based risk group showed better performance for 1- and 3-year survival than for 5-year survival. GSEA revealed enrichment of pathways related to cell cycle regulation among high-risk samples and metabolic processes in the low-risk group. Our 10-gene model is robust for prognosis prediction and may help inform clinical management of HCC.

The major effects of health-related quality of life on 5-year survival prediction among lung cancer survivors: applications of machine learning

The primary goal of this study was to evaluate the major roles of health-related quality of life (HRQOL) in a 5-year lung cancer survival prediction model using machine learning techniques (MLTs). The predictive performances of the models were compared with data from 809 survivors who underwent lung cancer surgery. Each of the modeling technique was applied to two feature sets: feature set 1 included clinical and sociodemographic variables, and feature set 2 added HRQOL factors to the variables from feature set 1. One of each developed prediction model was trained with the decision tree (DT), logistic regression (LR), bagging, random forest (RF), and adaptive boosting (AdaBoost) methods, and then, the best algorithm for modeling was determined. The models’ performances were compared using fivefold cross-validation. For feature set 1, there were no significant differences in model accuracies (ranging from 0.647 to 0.713). Among the models in feature set 2, the AdaBoost and RF models outperformed the other prognostic models [area under the curve (AUC) = 0.850, 0.898, 0.981, 0.966, and 0.949 for the DT, LR, bagging, RF and AdaBoost models, respectively] in the test set. Overall, 5-year disease-free lung cancer survival prediction models with MLTs that included HRQOL as well as clinical variables improved predictive performance.

P-15 Combined gastrectomies: Survival outcomes in patients with locally advanced gastric cancer

The survival of patients with locally advanced gastric cancer (GC) takes several months. Radical operations are extremely complex and remain the prerogative of several of the best abdominal surgeons of the world. The search for an optimal treatment plan for GC patients (GCP) with stage T3-4N0-2M0 was realized. I examined factors in terms of precise prediction of 5-year survival (5YS) of locally advanced GCP after complete (R0) combined gastrectomies (G).

Predicting postoperative non-small cell lung cancer prognosis via long short-term relational regularization.

Lung cancer is the leading cause of cancer death worldwide. Prognosis of lung cancer plays a crucial role in the clinical decision-making process to optimize the treatment for patients. Most of the existing data-driven prognostic prediction models explore the relations between patient's characteristics and outcomes at a specific time interval. Although valuable, they neglect the relations between long-term and short-term prognoses and thus may limit the prediction performance. In this study, we present a novel prognostic prediction approach for postoperative NSCLC patients. Specifically, we formulate the learning objective function by exploiting the relations between long-term and short-term prognoses via a long short-term relational regularization. The regularization term is composed of two parts, i.e., the similarities between prognoses measured by patients' outcomes and the L2 -norms between the corresponding prognoses' weight vectors. Based on this regularization, the proposed method can extract critical risk factors that comprehensively consider the long-term and short-term prognoses to facilitate the estimation of clinical risks. We evaluate the proposed model on a clinical dataset containing 693 consecutive postoperative NSCLC patients with more than 5-year follow-up from 2006 to 2015. Our best models achieve 0.743, 0.709, and 0.746 AUCs for 1-year, 3-year, and 5-year survival prediction, 0.696, 0.724, and 0.736 AUCs for 1-year, 3-year, and 5-year recurrence prediction, respectively. The experimental results show the efficiency of our proposed model in improving the performances on 1-year prognostic prediction in comparison with benchmark models. By comparing with the model without the long short-term relational regularization, the proposed model extracts more consistent critical risk factors for both long-term and short-term prognoses and contains fewer unreasonable risk factors under the clinician's review. We conclude that the proposed model can effectively exploit the relations between long-term and short-term prognoses. And the risk factors recognized by the proposed model have the potentials for further prognostic prediction of postoperative non-small cell lung cancer patients.

Development of robust artificial neural networks for prediction of 5-year survival in bladder cancer

PurposeWhen exploring survival outcomes for patients with bladder cancer, most studies rely on conventional statistical methods such as proportional hazards models. Given the successful application of machine learning to handle big data in many disciplines outside of medicine, we sought to determine if machine learning could be used to improve our ability to predict survival in bladder cancer patients. We compare the performance of artificial neural networks (ANN), a type of machine learning algorithm, with that of multivariable Cox proportional hazards (CPH) models in the prediction of 5-year disease-specific survival (DSS) and overall survival (OS) in patients with bladder cancer. Subjects and methodsThe National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) 18 program database was queried to identify adult patients with bladder cancer diagnosed between 1995 and 2010, yielding 161,227 patients who met our inclusion criteria. ANNs were trained and tested on an 80/20 split of the dataset. Multivariable CPH models were developed in parallel. Variables used for prediction included age, sex, race, grade, SEER stage, tumor size, lymph node involvement, degree of extension, and surgery received. The primary outcomes were 5-year DSS and 5-year OS. Receiver operating characteristic curve analysis was conducted, and ANN models were tested for calibration. ResultsThe area under the curve for the ANN models was 0.81 for the OS model and 0.80 for the DSS model. Area under the curve for the CPH models was 0.70 for OS and 0.81 for DSS. The ANN OS model achieved a calibration slope of 1.03 and a calibration intercept of −0.04, while the ANN DSS model achieved a calibration slope of 0.99 and a calibration intercept of −0.04. ConclusionsMachine learning algorithms can improve our ability to predict bladder cancer prognosis. Compared to CPH models, ANNs predicted OS more accurately and DSS with similar accuracy. Given the inherent limitations of administrative datasets, machine learning may allow for optimal interpretation of the complex data they contain.

Albumin-Bilirubin (ALBI) Grade-Based Nomogram for Patients with Hepatocellular Carcinoma Undergoing Transarterial Chemoembolization.

The prognosis of patients with hepatocellular carcinoma (HCC) undergoing transarterial chemoembolization (TACE) is highly heterogeneous because of variable characteristics of tumor burden and liver dysfunction. We aimed to propose and validate an albumin-bilirubin (ALBI) grade-based prognostic nomogram for HCC patients undergoing TACE. A total of 1051 patients with HCC undergoing TACE were randomly assigned to derivation (n = 525) and validation (n = 526) set in this retrospective study based on prospective data. The multivariate Cox proportional hazards model in derivation set was used to generate the nomogram. The predictive accuracy of the nomogram was evaluated by discrimination and calibration tests. In multivariate analysis, presence of ascites, ALBI grade 2-3, serum ɑ-fetoprotein level ≥ 400ng/mL, total tumor volume ≥ 396cm3, presence of vascular invasion, and poor performance status were independently associated with decreased survival of patients in the derivation set. Each patient had an individualized score from 0 to 41 by adding up the points from these six prognostic predictors. The nomogram generated from the derivation set had a concordance index of 0.72 (95% confidence interval [CI] 0.63-0.82). Discrimination test in the validation set provided a good concordance index 0.72 (95% CI 0.62-0.81), and the calibration plots consistently matched the ideal 45-degree reference line for 3- and 5-year survival prediction. The ALBI grade-based prognostic model can well discriminate the survival in HCC patients undergoing TACE. The proposed easy-to-use nomogram may accurately predict the survival at 3 and 5years for individual HCC patient in the precision medicine era.

Development and Validation of an Individualized Nomogram for Predicting Overall Survival in Patients With Typical Lung Carcinoid Tumors.

We aim to develop and validate an effective nomogram prognostic model for patients with typical lung carcinoid tumors using a large patient cohort from the Surveillance, Epidemiology, and End Results (SEER) database. Data from patients with typical lung carcinoid tumors between 2010 and 2015 were selected from the SEER database for retrospective analysis. Univariate and multivariate Cox analysis was performed to clarify independent prognostic factors. Next, a nomogram was formulated to predict the probability of 3- and 5-year overall survival (OS). Concordance indexes (c-index), receiver operating characteristic analysis and calibration curves were used to evaluate the model. The selected patients were randomly divided into a training and a validation cohort. A nomogram was established based on the training cohort. Cox analysis results indicated that age, sex, T stage, N stage, surgery, and bone metastasis were independent variables for OS. All these factors, except surgery, were included in the nomogram model for predicting 3- and 5-year OS. The internally and externally validated c-indexes were 0.787 and 0.817, respectively. For the 3-year survival prediction, receiver operating characteristic analysis showed that the areas under the curve in the training and validation cohorts were 0.824 and 0.795, respectively. For the 5-year survival prediction, the area under the curve in the training and validation cohorts were 0.812 and 0.787, respectively. The calibration plots for probability of survival were in good agreement. The nomogram brings us closer to personalized medicine and the maximization of predictive accuracy in the prediction of OS in patients with typical lung carcinoid tumors.

The prognostic significance of the advanced lung cancer inflammation index(ALI) in patients with advanced biliary tract cancer: A retrospective study.

e16613 Background: Intrahepatic cholangiocarcinoma is the second most common primary cancer, but the prognosis is poor and aggressive therapies provide only limited benefit when it is in advanced stage. Many studies were conducted to find significant prognostic factors to predict clinical outcomes, but there is no definitive parameter or scoring model yet in this disease. This study is a retrospective analysis to test various kinds of prognostic scoring systems including the advanced lung cancer inflammation (ALI) index, a new prognostic model, for patients treated for advanced intrahepatic cholangiocarcinoma. Methods: We retrospectively searched medical records of patients who were diagnosed with advanced intrahepatic cholangiocarcinoma between January 2012 and December 2017 and actively treated at a single tertiary regional cancer center. Patients who received only supportive care, or had active infection concomitantly were excluded. 60 patients were identified, and we reviewed and analyzed their baseline characteristics and clinical outcomes. Various types of inflammation or nutritional scoring systems, neutrophil to lymphocyte ratio (NLR), Onodera’s prognostic nutritional index (OPNI), the advanced lung cancer inflammation (ALI) index were concurrently calculated and significance for patients’ survival was evaluated. Results: In 50 patients with metastatic stage and 10 patients with locally advanced disease, the median overall survival was 13.97 months (95% confidence interval (CI) 7.99 – 19.95). ALI of 28.5 was determined as the optimal cut-off value for prediction of 1-year survival by receiver operating characteristics (ROC) curve with an AUC value of 0.671 (sensitivity and specificity were 77.8% and 60.6%, respectively). 26 patients with ALI &gt; 28.5 had significantly better overall survival compared with 34 patients with ALI≤28.5 (19.2 months vs. 8.73 months, p= .007). NLR above 2.8 only showed mildly significant difference for overall survival (OS), and the OPNI failed to predict prognosis of these patients. In multivariate analysis, there was no independent prognostic factor observed to have association with overall survival. Conclusions: The advanced lung cancer inflammation (ALI) index showed a potential to be a new prognostic model for advanced intrahepatic cholangiocarcinoma. Further cohort studies to expand population size may be worth.

How Does the Skeletal Oncology Research Group Algorithm's Prediction of 5-year Survival in Patients with Chondrosarcoma Perform on International Validation?

How Does the Skeletal Oncology Research Group Algorithm's Prediction of 5-year Survival in Patients with Chondrosarcoma Perform on International Validation?

Systematic Construction and Validation of a Metabolic Risk Model for Prognostic Prediction in Acute Myelogenous Leukemia.

Background: Acute myelogenous leukemia (AML) is a heterogeneous disease with recurrent gene mutations and variations in disease-associated gene expression, which may be useful for prognostic prediction.Methods: RNA matrix and clinical data of AML were downloaded from GEO, TCGA, and TARGET databases. Prognostic metabolic genes were identified by LASSO analysis to establish a metabolic model. Prognostic accuracy of the model was quantified by time-dependent receiver operating characteristic curves and the area under the curve (AUC). Survival analysis was performed by log-rank tests. Enriched pathways in different metabolic risk statuses were evaluated by gene set enrichment analyses (GSEA).Results: We identified nine genes to construct a prognostic model of shorter survival in the high-risk vs. low-risk group. The prognostic model showed good predictive efficacy, with AUCs for 5-year overall survival of 0.78 (0.73–0.83), 0.76 (0.62–0.89), and 0.66 (0.57–0.75) in the training, adult external, and pediatric external cohorts, respectively. Multivariable analysis demonstrated that the metabolic signature had independent prognostic value with hazard ratios of 2.75 (2.06–3.66), 1.89 (1.09–3.29), and 1.96 (1.00–3.84) in the training, adult external, and pediatric external cohorts, respectively. Combining metabolic signatures and classic prognostic factors improved 5-year overall survival prediction compared to the prediction by classic prognostic factors (p < 0.05). GSEA revealed that most pathways were metabolism-related, indicating potential mechanisms.Conclusion: We identified dysregulated metabolic features in AML and constructed a prognostic model to predict the survival of patients with AML.

Construction of a novel gene-based model for prognosis prediction of clear cell renal cell carcinoma

BackgroundClear cell renal cell carcinoma (ccRCC) comprises the majority of kidney cancer death worldwide, whose incidence and mortality are not promising. Identifying ideal biomarkers to construct a more accurate prognostic model than conventional clinical parameters is crucial.MethodsRaw count of RNA-sequencing data and clinicopathological data were acquired from The Cancer Genome Atlas (TCGA). Tumor samples were divided into two sets. Differentially expressed genes (DEGs) were screened in the whole set and prognosis-related genes were identified from the training set. Their common genes were used in LASSO and best subset regression which were performed to identify the best prognostic 5 genes. The gene-based risk score was developed based on the Cox coefficient of the individual gene. Time-dependent receiver operating characteristic (ROC) and Kaplan–Meier (KM) survival analysis were used to assess its prognostic power. GSE29609 dataset from GEO (Gene Expression Omnibus) database was used to validate the signature. Univariate and multivariate Cox regression were performed to screen independent prognostic parameters to construct a nomogram. The predictive power of the nomogram was revealed by time-dependent ROC curves and the calibration plot and verified in the validation set. Finally, Functional enrichment analysis of DEGs and 5 novel genes were performed to suggest the potential biological pathways.ResultsPADI1, ATP6V0D2, DPP6, C9orf135 and PLG were screened to be significantly related to the prognosis of ccRCC patients. The risk score effectively stratified the patients into high-risk group with poor overall survival (OS) based on survival analysis. AJCC-stage, age, recurrence and risk score were regarded as independent prognostic parameters by Cox regression analysis and were used to construct a nomogram. Time-dependent ROC curves showed the nomogram performed best in 1-, 3- and 5-year survival predictions compared with AJCC-stage and risk score in validation sets. The calibration plot showed good agreement of the nomogram between predicted and observed outcomes. Functional enrichment analysis suggested several enriched biological pathways related to cancer.ConclusionsIn our study, we constructed a gene-based model integrating clinical prognostic parameters to predict prognosis of ccRCC well, which might provide a reliable prognosis assessment tool for clinician and aid treatment decision-making in the clinic.

MRI-based radiomics model for preoperative prediction of 5-year survival in patients with hepatocellular carcinoma

BackgroundRecurrence is the major cause of mortality in patients with resected HCC. However, without a standard approach to evaluate prognosis, it is difficult to select candidates for additional therapy.MethodsA total of 201 patients with HCC who were followed up for at least 5 years after curative hepatectomy were enrolled in this retrospective, multicentre study. A total of 3144 radiomics features were extracted from preoperative MRI. The random forest method was used for radiomics signature building, and five-fold cross-validation was applied. A radiomics model incorporating the radiomics signature and clinical risk factors was developed.ResultsPatients were divided into survivor (n = 97) and non-survivor (n = 104) groups based on the 5-year survival after surgery. The 30 most survival-related radiomics features were selected for the radiomics signature. Preoperative AFP and AST were integrated into the model as independent clinical risk factors. The model demonstrated good calibration and satisfactory discrimination, with a mean AUC of 0.9804 and 0.7578 in the training and validation sets, respectively.ConclusionsThis radiomics model is a valid method to predict 5-year survival in patients with HCC and may be used to identify patients for clinical trials of perioperative therapies and for additional surveillance.

International validation of the Amsterdam model for survival prediction after resected pancreatic cancer

Background: A recent systematic review identified only two prediction models for survival in resectable pancreatic cancer of high methodological quality. The objective of this study was to validate one of these, the Amsterdam prediction model (including differentiation grade, positive lymph node ratio, margin status and adjuvant therapy), in a large international cohort. Methods: All consecutive patients who underwent pancreatoduodenectomy for pancreatic cancer between 2000-2017 at 10 tertiary hospitals from 8 countries were included. Calibration and model performance for prediction of 3-year and 5-year survival were evaluated and compared to the 8th edition of the TNM staging system, using Uno’s C-statistic, the Receiver Operating Characteristics (ROC) curves and area under the curves (AUC). Results: A total of 2,203 patients were included (mean age 65 years, 54% male), with a median overall survival of 23.6 months. The validation cohort was demographically comparable to the design cohort, but had a higher lymph node yield (median of 19 vs. 8) and more N2 disease (22.0% vs. 40.3%), and a higher use of adjuvant therapy (76.0% vs. 35.1%). Calibration analysis demonstrated that the predictions of 3-year survival did not systematically deviate from the observed survival rates with a slope of 0.85 and an intercept of 0.04. Uno’s C-statistic was 0.65 in the validation cohort, compared to 0.67 in the derivation cohort. ROC-curves showed an AUC of 0.70 and 0.72 for 3- and 5-year survival, compared to 0.65 and 0.68 for the TNM 8, respectively. Analysis of model performance among the different countries demonstrated a robust C-statistic, ranging from 0.61 to 0.67. Conclusions: This international multicenter study validated the previous predictions of the Amsterdam model for 3- and 5-year survival after pancreatoduodenectomy for pancreatic cancer. Model performance was fairly accurate in terms of calibration and discrimination, and the model demonstrated robustness within the different countries.