Prediction Model For Overall Survival Research Articles

Prognostic factors and survival prediction of resected non-small cell lung cancer with ipsilateral pulmonary metastases: a study based on the Surveillance, Epidemiology, and End Results (SEER) database

BackgroundPrognostic factors and survival outcomes of non-small cell lung cancer (NSCLC) with Ipsilateral pulmonary metastasis (IPM) are not well-defined. Thus, this study intended to identify the prognostic factors for these patients and construct a predictive nomogram model.MethodsOne thousand, seven hundred thirty-two patients with IPM identified between 2000 to 2019 were from the Surveillance, Epidemiology, and End Results (SEER) database. Independent prognostic factors were identified using multivariate Cox regression analyses. Nomograms were constructed to predict the overall survival (OS), C-index, the area under the curve (AUC), and the calibration curve to determine the predictive accuracy and discrimination; the decision curve analysis was used to confirm the clinical utility.ResultsPatients were randomly divided into training (n = 1213) and validation (n = 519) cohorts. In the training cohort, the multivariable analysis demonstrated that age, sex, primary tumor size, N status, number of regional lymph nodes removed, tumor grade, and chemotherapy were independent prognostic factors for IPM. We constructed a 1-year, 3-year, and 5-year OS prediction nomogram model using independent prognostic factors. The C-index of this model for OS prediction was 0.714 (95% confidence interval [CI], 0.692 to 0.773) in the training cohort and 0.695 (95% CI, 0.660 to 0.730) in the validation cohort. Based on the AUC of the receiver operating characteristic analysis, calibration plots, and decision curve analysis, we concluded that the prognosis model of IPM exhibited excellent performance. Patients with total nomogram points greater than 96 were considered high-risk.ConclusionWe constructed and internally validated a nomogram to predict 1-year, 3-year, and 5-year OS for NSCLC patients with IPM according to independent prognostic factors. This nomogram demonstrated good calibration, discrimination, clinical utility, and practical decision-making effects for the prognosis of NSCLC patients with IPM.

Predictive function of tumor burden-incorporated machine-learning algorithms for overall survival and their value in guiding management decisions in patients with locally advanced nasopharyngeal carcinoma

ObjectiveAccurate prognostic predictions and personalized decision-making on induction chemotherapy (IC) for individuals with locally advanced nasopharyngeal carcinoma (LA-NPC) remain challenging. This research examined the predictive function of tumor burden-incorporated machine-learning algorithms for overall survival (OS) and their value in guiding treatment in patients with LA-NPC. MethodsIndividuals with LA-NPC were reviewed retrospectively. Tumor burden signature-based OS prediction models were established using a nomogram and two machine-learning methods, the interpretable eXtreme Gradient Boosting (XGBoost) risk prediction model, and DeepHit time-to-event neural network. The models' prediction performances were compared using the concordance index (C-index) and the area under the curve (AUC). The patients were divided into two cohorts based on the risk predictions of the most successful model. The efficacy of IC combined with concurrent chemoradiotherapy was compared to that of chemoradiotherapy alone. ResultsThe 1 221 eligible individuals, assigned to the training (n = 813) or validation (n = 408) set, showed significant respective differences in the C-indices of the XGBoost, DeepHit, and nomogram models (0.849 and 0.768, 0.811 and 0.767, 0.730 and 0.705). The training and validation sets had larger AUCs in the XGBoost and DeepHit models than the nomogram model in predicting OS (0.881 and 0.760, 0.845 and 0.776, and 0.764 and 0.729, P < 0.001). IC presented survival benefits in the XGBoost-derived high-risk but not low-risk group. ConclusionThis research used machine-learning algorithms to create and verify a comprehensive model integrating tumor burden with clinical variables to predict OS and determine which patients will most likely gain from IC. This model could be valuable for delivering patient counseling and conducting clinical evaluations.

Overall Survival Prediction in Stereotactic Radiosurgery Patients with Glioblastoma Via a Deep-Learning Approach

Accurate and automated early survival prediction is critical for glioblastoma (GBM) patients as their poor prognosis requires timely treatment decision-making. We have developed a deep learning (DL)-based GBM overall survival (OS) prediction model based on a multi-institutional public dataset using only pre-operative basic structural multi-parametric magnetic resonance images (MRIs). The purpose of this study is to evaluate this DL-based OS prediction model with an institutional stereotactic radiosurgery (SRS) clinical trial dataset. The task of this study is to classify GBM patients into 3 OS classes: long-survivors (>15 months), mid-survivors (between 10 and 15 months) and short-survivors (< 10 months). The proposed OS prediction model is an ensemble of a ResNet-based classifier and a K-NN classifier. The ResNet-based classifier is trained in a Siamese fashion to explore inter-class differences. During testing, training sample features are implemented with a K-NN classifier to ensemble with the ResNet-based classifier. A public dataset from Medical Image Computing and Computer Assisted Intervention (MICCAI) Brain Tumor Segmentation (BraTS) challenge 2020 (235 patients) were used for model establishing and initial validation. Then the validated model was evaluated on 19 GBM patients from an institutional SRS clinical trial. Each data entry consists of pre-operative basic structural multi-parametric MRIs and survival days, as well as patient ages for BraTS data and basic clinical characteristics for institutional data. GBM sub-regions, including contrast-enhancing tumor, peri-tumoral edema, and necrotic/non-enhancing tumor core, were segmented in the multi-parametric MRIs by an in-house DL model for both datasets. The OS prediction model was trained on 90% of the segmented BraTS data and validated on the rest 10%, then further evaluated on the institutional data. The model performance was assessed by prediction accuracy (ACC) and the area under the curve (AUC). For this 3-class OS classification task, our DL-based prediction model achieved an ACC of 65.22% and an AUC of 0.81 on the BraTS dataset compared with the top-ranked result from the BraTS challenge 2020 (Rank 1st: ACC 61.7%), and an ACC of 52.63% and an AUC of 0.69 on the institutional dataset. Further analysis of the institutional dataset found that the predicted OS class had a statistically significant correlation with treatment volume (p = 0.012) and age (p = 0.006), which matches the analysis that the patients' ground truth OS class is statistical significantly correlated with treatment volume (p = 0.045). Our DL-based OS prediction model for GBM using basic structural multi-parametric pre-operative MRIs has demonstrated promising performance in both public and institutional dataset with minimal manual processing requirements. This OS prediction model can be potentially applied to assist timely clinical decision-making.

White Matter Tract Density Index Prediction Model of Overall Survival in Glioblastoma

The prognosis of overall survival (OS) in patients with glioblastoma (GBM) may depend on the underlying structural connectivity of the brain. To examine the association between white matter tracts affected by GBM and patients' OS by means of a new tract density index (TDI). This prognostic study in patients with a histopathologic diagnosis of GBM examined a discovery cohort of 112 patients who underwent surgery between February 1, 2015, and November 30, 2020 (follow-up to May 31, 2023), in Italy and 70 patients in a replicative cohort (n = 70) who underwent surgery between September 1, 2012, and November 30, 2015 (follow-up to May 31, 2023), in Germany. Statistical analyses were performed from June 1, 2021, to May 31, 2023. Thirteen and 12 patients were excluded from the discovery and the replicative sets, respectively, because of magnetic resonance imaging artifacts. The density of white matter tracts encompassing GBM. Correlation, linear regression, Cox proportional hazards regression, Kaplan-Meier, and prediction analysis were used to assess the association between the TDI and OS. Results were compared with common prognostic factors of GBM, including age, performance status, O6-methylguanine-DNA methyltransferase methylation, and extent of surgery. In the discovery cohort (n = 99; mean [SD] age, 62.2 [11.5] years; 29 female [29.3%]; 70 male [70.7%]), the TDI was significantly correlated with OS (r = -0.34; P < .001). This association was more stable compared with other prognostic factors. The TDI showed a significant regression pattern (Cox: hazard ratio, 0.28 [95% CI, 0.02-0.55; P = .04]; linear: t = -2.366; P = .02). and a significant Kaplan-Meier stratification of patients as having lower or higher OS based on the TDI (log-rank test = 4.52; P = .03). Results were confirmed in the replicative cohort (n = 58; mean [SD] age, 58.5 [11.1] years, 14 female [24.1%]; 44 male [75.9%]). High (24-month cutoff) and low (18-month cutoff) OS was predicted based on the TDI computed in the discovery cohort (accuracy = 87%). In this study, GBMs encompassing regions with low white matter tract density were associated with longer OS. These findings indicate that the TDI is a reliable presurgical outcome predictor that may be considered in clinical trials and clinical practice. These findings support a framework in which the outcome of GBM depends on the patient's brain organization.

Impact of sarcopenia on the survival of patients undergoing liver transplantation for decompensated liver cirrhosis.

Patients with decompensated cirrhosis (DC) are prone to skeletal muscle loss, namely, sarcopenia, before liver transplantation (LT). While sarcopenia is reportedly associated with adverse outcomes after LT, these findings are limited owing to mixed diseases and retrospective data. We investigated the association between sarcopenia and 1-year overall survival (OS) in patients with DC after LT and established and validated a prediction model for postoperative OS based on sarcopenia. Overall, 222 consecutive patients who underwent LT at our centre were registered between September 2020 and June 2022. Third lumbar spine skeletal muscle mass index was measured using computed tomography. Patients were divided into sarcopenia and non-sarcopenia groups according to the skeletal muscle mass index, and baseline data and postoperative outcomes were collected, compared and analysed. The primary outcome was the 1-year OS after LT. We established a dynamic nomogram of the OS predictive model. We included 177DC patients [mean (standard deviation) age, 50.2±9.3years; 52 women (29.4%)]; 73 (41.2%) had sarcopenia. The mean (standard deviation) body mass index was 22.6±4.5kg/m2 , 28 (15.8%) patients had weight loss ≥5% within 6months before admission, and the mean (standard deviation) model for end-stage liver disease (MELD) score was 18.4±7.9. Patients with sarcopenia had a longer duration of intensive care unit stay (4.1±2.2 vs. 3.1±1.1days, P=0.008), higher rate of major complications (45.2% vs. 22.1%, P=0.001) and higher postoperative mortality (15.1% vs. 2.9%, P=0.003) than those without sarcopenia. The median 1-year OS after surgery was shorter in patients with sarcopenia than in those without (P<0.001). Sarcopenia [hazard ratio (HR), 2.54; 95% confidence interval (CI), 1.54-5.63; P=0.022], weight loss ≥5% (HR, 2.46; 95% CI, 1.39-5.09; P=0.015) and MELD score (HR, 1.05; 95% CI, 1.01-1.09; P=0.009) were independent risk factors associated with 1-year OS. The area under the curve of the established dynamic nomogram was 0.774, the calibration curve showed good consistency, and analysis of the decision curve showed more clinical benefits than the MELD score alone. High-risk patients (>102.9 points calculated using the nomogram) had a significantly reduced survival rate. Sarcopenia is associated with adverse outcomes after LT in patients with DC. High-risk patients should be classified by dynamic nomogram upon admission.

Association between Contrast-Enhanced Computed Tomography Radiomic Features, Genomic Alterations and Prognosis in Advanced Lung Adenocarcinoma Patients.

Non-invasive methods to assess mutational status, as well as novel prognostic biomarkers, are warranted to foster therapy personalization of patients with advanced non-small cell lung cancer (NSCLC). This study investigated the association of contrast-enhanced Computed Tomography (CT) radiomic features of lung adenocarcinoma lesions, alone or integrated with clinical parameters, with tumor mutational status (EGFR, KRAS, ALK alterations) and Overall Survival (OS). In total, 261 retrospective and 48 prospective patients were enrolled. A Radiomic Score (RS) was created with LASSO-Logistic regression models to predict mutational status. Radiomic, clinical and clinical-radiomic models were trained on retrospective data and tested (Area Under the Curve, AUC) on prospective data. OS prediction models were trained and tested on retrospective data with internal cross-validation (C-index). RS significantly predicted each alteration at training (radiomic and clinical-radiomic AUC 0.95-0.98); validation performance was good for EGFR (AUC 0.86), moderate for KRAS and ALK (AUC 0.61-0.65). RS was also associated with OS at univariate and multivariable analysis, in the latter with stage and type of treatment. The validation C-index was 0.63, 0.79, and 0.80 for clinical, radiomic, and clinical-radiomic models. The study supports the potential role of CT radiomics for non-invasive identification of gene alterations and prognosis prediction in patients with advanced lung adenocarcinoma, to be confirmed with independent studies.

Development and validation of prediction model for overall survival in patients with lymphoma: a prospective cohort study in China

ObjectiveThe survival of patients with lymphoma varies greatly among individuals and were affected by various factors. The aim of this study was to develop and validate a prognostic model for predicting overall survival (OS) in patients with lymphoma.MethodsWe conducted a prospective longitudinal cohort study in China between January 2014 and December 2018 (n = 1,594). After obtaining the follow-up data, we randomly split the cohort into the training cohort (n = 1,116) and the validation cohort (n = 478). The least absolute shrinkage and selection operator (LASSO) regression analysis was used to select the predictors of the model. Cox stepwise regression analysis was used to identify independent prognostic factors, which were finally displayed as static nomogram and web-based dynamic nomogram. We calculated the concordance index(C-index) to describe how the predicted survival of objectively confirmed prognosis. The calibration plot is used to evaluate the prediction accuracy and discrimination ability of the model. Net reclassification index (NRI) and decision curve analysis (DCA) curves were also used to evaluate the prediction ability and net benefit of the model.ResultsNine variables in the training cohort were considered to be independent risk factors for patients with lymphoma in the final model: age, Ann Arbor Stage, pathologic type, B symptoms, chemotherapy, targeted therapy, lactate dehydrogenase (LDH), β2-microglobulin and C-reactive protein (CRP). The C-indices of OS were 0.749 (95% CI, 0.729–0.769) in the training cohort and 0.731 (95% CI, 0.762–0.700) in the validation cohort. A good agreement between prediction by nomogram and actual observation was shown in the calibration curve for the probability of survival in both the training cohort and validation cohorts. The areas under curve (AUC) of the area under the receiver operating characteristic (ROC) curves for 1-year, 3-year, and 5-year OS were 0.813, 0.800, and 0.762, respectively, in the training cohort, and 0.802, 0.768, and 0.721, respectively, in the validation cohort. Compared with the Ann Arbor Stage system, NRI and DCA showed that the model had a higher predictive capacity and net benefit.ConclusionThe prediction models reliably estimate the outcome of patients with lymphoma. The model had high discrimination and calibration, which provided a simple and reliable tool for the survival prediction of the patients, and it might help patients benefit from personalized intervention.

Automatic prediction model of overall survival in prostate cancer patients with bone metastasis using deep neural networks

Abstract Objectives Bone is the most common site of metastasis in prostate cancer (PCa) patients and is correlated with poor prognosis and increasing economic burden. Few studies have analyzed the prognostic prediction for metastatic PCa patients with the assistance of neural networks. Methods Four convolutional neural network (CNN) models are developed and evaluated to predict the overall survival (OS) of PCa patients with bone metastasis. All the CNN models are first trained with 64 samples and evaluated with 10 samples; two models use only bone scan images and two models use both bone scan images and clinical parameters (CPs). The predictions of the best models are compared with those by two urology surgeons on 20 test samples. Results Our best models can predict OS of PCa patients with bone metastasis with AUC=0.8022 by using only bone scan images and AUC=0.8132 by using both bone scan images and CPs on 20 test samples. The best Youden indexes of the two models are 0.6263 and 0.7142, respectively, which are 0.3077 and 0.3131 higher than that of the urologists’ average Youden index, which indicate that CNN models exhibit significant advantages. Conclusions CNN models are suitable to predict OS in PCa patients with bone metastasis using bone scan images and CPs. Our models show better performance in terms of accuracy and stability than urology surgeons.

A web-based prediction model for overall survival of elderly patients with epithelial ovarian cancer: A population-based study.

e17570 Background: Epithelial ovarian cancer (EOC) has been extensively studied. However, no prediction model has been carried out on the prognosis of elderly patients. Our study aims to explore survival patterns in elderly EOC patients. Methods: Data were extracted from the Surveillance, Epidemiology, and End Results (SEER) database. Patients over 65 with EOC between 2004 and 2014 were included for model training and validation, while 2015 for external validation. COX proportional-hazards regression model was used to identify risk factors and applied them to the prediction model. We then evaluated performance of our model through receiver operating characteristic (ROC) curve, calibration curve and decision curve analysis (DCA). Risk stratification system was used to reflect the survival. To facilitate daily use, we developed an online application. Results: A total of 6141 patients were included. COX regression models indicated 10 independent risk factors. Our model was proved to have good performance with AUC of ROC curve: 0.768(95CI:0.749-0.788), 0.748 (95CI:0.733-0.763), 0.766 (95CI:0.750-0.781) for 1-, 3-, 5-year OS in the training cohort and 0.776 (95CI:0.757-0.796), 0.760 (95CI:0.745-0.775), 0.784 (95CI:0.769-0.799) for CSS. In the validation cohort, 0.800 (95CI:0.773-0.827), 0.770 (95CI:0.748-0.793), 0.787 (95CI:0.765-0.810) for OS and 0.808 (95CI:0.780-0.836), 0.780 (95CI:0.758-0.803), 0.801 (95CI:0.779-0.824) for CSS. In the risk stratification system, the low-risk group had the highest survival rate. Almost all patients in the low- and intermediate-risk groups underwent surgery, about 1/3 patients in the high-risk group did not receive surgery. Their survival rates were not optimistic, and patients over 85 were more likely to be classified as high-risk. Conclusions: A predictive model was constructed to evaluate survival for elderly EOC patients and demonstrated good predictive value that we can apply to help physicians make clinical decisions and plan treatment. [Table: see text]

MCM10: An effective treatment target and a prognostic biomarker in patients with uterine corpus endometrial carcinoma.

Molecular profiling has been applied for uterine corpus endometrial carcinoma (UCEC) management for many years. The aim of this study was to explore the role of MCM10 in UCEC and construct its overall survival (OS) prediction models. Data from TCGA, GEO, cbioPotal and COSMIC databases and the methods, such as GO, KEGG, GSEA, ssGSEA and PPI, were employed to bioinformatically detect the effects of MCM10 on UCEC. RT-PCR, Western blot and immunohistochemistry were used to validate the effects of MCM10 on UCEC. Based on Cox regression analysis using the data from TCGA and our clinical data, two OS prediction models for UCEC were established. Finally, the effects of MCM10 on UCEC were detected in vitro. Our study revealed that MCM10 was variated and overexpressed in UCEC tissue and involved in DNA replication, cell cycle, DNA repair and immune microenvironment in UCEC. Moreover, silencing MCM10 significantly inhibited the proliferation of UCEC cells in vitro. Importantly, based on MCM10 expression and clinical features, the OS prediction models were constructed with good accuracy. MCM10 could be an effective treatment target and a prognostic biomarker for UCEC patients. The OS prediction models might help establish the strategies of follow-up and treatment for UCEC patients.

Prediction of risk and overall survival of pancreatic cancer from blood soluble immune checkpoint-related proteins.

Immune checkpoint inhibition holds promise as a novel treatment for pancreatic ductal adenocarcinoma (PDAC). The clinical significance of soluble immune checkpoint (ICK) related proteins have not yet fully explored in PDAC. We comprehensively profiled 14 soluble ICK-related proteins in plasma in 70 PDAC patients and 70 matched healthy controls. Epidemiological data of all subjects were obtained through structured interviews, and patients' clinical data were retrieved from electronical health records. We evaluated the associations between the biomarkers with the risk of PDAC using unconditional multivariate logistic regression. Consensus clustering (k-means algorithm) with significant biomarkers was performed to identify immune subtypes in PDAC patients. Prediction models for overall survival (OS) in PDAC patients were developed using multivariate Cox proportional hazards regression. Harrell's concordance index (C-index), time-dependent receiver operating characteristic (ROC) curve and calibration curve were utilized to evaluate performance of prediction models. Gene expressions of the identified ICK-related proteins in tumors from TCGA were analyzed to provide insight into underlying mechanisms. Soluble BTLA, CD28, CD137, GITR and LAG-3 were significantly upregulated in PDAC patients (all q < 0.05), and elevation of each of them was correlated with PDAC increased risk (all p < 0.05). PDAC patients were classified into soluble immune-high and soluble immune-low subtypes, using these 5 biomarkers. Patients in soluble immune-high subtype had significantly poorer OS than those in soluble immune-low subtype (log-rank p = 9.7E-03). The model with clinical variables and soluble immune subtypes had excellent predictive power (C-index = 0.809) for the OS of PDAC patients. Furthermore, the immune subtypes identified with corresponding genes' expression in PDAC tumor samples in TCGA showed an opposite correlation with OS to that of immune subtypes based on blood soluble ICK-related proteins (log-rank p =0.02). The immune-high subtype tumors displayed higher cytolytic activity (CYT) score than immune-low subtype tumors (p < 2E-16). Five soluble ICK-related proteins were identified to be significantly associated with the risk and prognosis of PDAC. Patients who were classified as soluble immune-low subtype based on these biomarkers had better overall survival than those of the soluble immune-high subtype.

Mining glycosylation-related prognostic lncRNAs and constructing a prognostic model for overall survival prediction in glioma: A study based on bioinformatics analysis.

Dysregulation of protein glycosylation plays a crucial role in the development of glioma. Long noncoding RNA (lncRNAs), functional RNA molecules without protein-coding ability, regulate gene expression and participate in malignant glioma progression. However, it remains unclear how lncRNAs are involved in glycosylation glioma malignancy. Identification of prognostic glycosylation-related lncRNAs in gliomas is necessary. We collected RNA-seq data and clinicopathological information of glioma patients from the cancer genome atlas and Chinese glioma genome atlas. We used the "limma" package to explore glycosylation-related gene and screened related lncRNAs from abnormally glycosylated genes. Using univariate Cox analyses Regression and least absolute shrinkage and selection operator analyses, we constructed a risk signature with 7 glycosylation-related lncRNAs. Based on the median risk score (RS), patients with gliomas were divided into low- and high-risk subgroups with different overall survival rates. Univariate and multivariate Cox analyses regression analyses were performed to assess the independent prognostic ability of the RS. Twenty glycosylation-related lncRNAs were identified by univariate Cox regression analyses. Two glioma subgroups were identified using consistent protein clustering, with the prognosis of the former being better than that of the latter. Least absolute shrinkage and selection operator analysis identified 7 survival RSs for glycosylation-related lncRNAs, which were identified as independent prognostic markers and predictors of glioma clinicopathological features. Glycosylation-related lncRNAs play an important role in the malignant development of gliomas and may help guide treatment options.

Construction of diagnostic and prognostic models based on gene signatures of nasopharyngeal carcinoma by machine learning methods.

Diagnostic models based on gene signatures of nasopharyngeal carcinoma (NPC) were constructed by random forest (RF) and artificial neural network (ANN) algorithms. Least absolute shrinkage and selection operator (Lasso)-Cox regression was used to select and build prognostic models based on gene signatures. This study contributes to the early diagnosis and treatment, prognosis, and molecular mechanisms associated with NPC. Two gene expression datasets were downloaded from the Gene Expression Omnibus (GEO) database, and differentially expressed genes (DEGs) associated with NPC were identified by gene expression differential analysis. Subsequently, significant DEGs were identified by a RF algorithm. ANN were used to construct a diagnostic model for NPC. The performance of the diagnostic model was evaluated by area under the curve (AUC) values using a validation set. Lasso-Cox regression examined gene signatures associated with prognosis. Overall survival (OS) and disease-free survival (DFS) prediction models were constructed and validated from The Cancer Genome Atlas (TCGA) database and the International Cancer Genome Consortium (ICGC) database. A total of 582 DEGs associated with NPC were identified, and 14 significant genes were identified by the RF algorithm. A diagnostic model for NPC was successfully constructed using ANN, and the validity of the model was confirmed on the training set AUC =0.947 [95% confidence interval (CI): 0.911-0.969] and the validation set AUC =0.864 (95% CI: 0.828-0.901). The 24-gene signatures associated with prognosis were identified by Lasso-Cox regression, and prediction models for OS and DFS of NPC were constructed on the training set. Finally, the ability of the model was validated on the validation set. Several potential gene signatures associated with NPC were identified, and a high-performance predictive model for early diagnosis of NPC and a prognostic prediction model with robust performance were successfully developed. The results of this study provide valuable references for early diagnosis, screening, treatment and molecular mechanism research of NPC in the future.

A visualized dynamic prediction model for overall survival in patients diagnosed with brain metastases from lung squamous cell carcinoma.

Patients presenting with brain metastases (BMs) from lung squamous cell carcinoma (LUSC) often encounter an extremely poor prognosis. A well-developed prognostic model would assist physicians in patient counseling and therapeutic decision-making. Patients with LUSC who were diagnosed with BMs between 2000 and 2018 were reviewed in the Surveillance, Epidemiology, and End Results (SEER) database. Using the multivariate Cox regression approach, significant prognostic factors were identified and integrated. Bootstrap resampling was used to internally validate the model. An evaluation of the performance of the model was conducted by analyzing the area under the curve (AUC) and calibration curve. A total of 1812 eligible patients' clinical data was retrieved from the database. Patients' overall survival (OS) was significantly prognosticated by five clinical parameters. The nomogram achieved satisfactory discrimination capacity, with 3-, 6-, and 9-month AUC values of 0.803, 0.779, and 0.760 in the training cohort and 0.796, 0.769, and 0.743 in the validation cohort. As measured by survival rate probabilities, the calibration curve agreed well with actual observations. There was also a substantial difference in survival curves between the different prognostic groups stratified by prognostic scores. For ease of access, the model was deployed on a web-based server. In this study, a nomogram and a web-based predictor were developed to assist physicians with personalized clinical decisions and treatment of patients who presented with BMs from LUSC.

Abstract 5395: Improving survival prediction using flexible late fusion machine learning framework for multi-omics data integration

Abstract Improving cancer patients Overall Survival (OS) prognosis is critical for personalization of treatment using model-identified drivers of cancer progression. Current cancer prognosis models largely rely on clinical and demographic patient characteristics. Adding ‘omics’-based modalities can help improve patient OS prediction and lead to better disease categorization and understanding. We introduce a data driven methodology for combining multi-omics and clinical data, including clinical/demographics, mutations, gene expression, long non-coding and micro-RNA expression, DNA methylation, and proteomics for improving prediction of OS in cancer patients. High dimensionality of ‘omics’ modalities present challenges to combining them into one model. We propose a late stage modalities fusion where we construct a separate data driven model for OS prediction for each modality, later combining individual predictions in a final linear OS prediction model. With a limited number of patients to develop the model, such an approach helps to better protect against overfitting, and allows to account for different degrees of informativeness of modalities by weighting them according to individual success. We introduce a robust machine learning pipeline with rigorous training, testing and evaluation capabilities, and demonstrate its effectiveness on a suit of TCGA data. When comparing early vs. late fusion of omics and clinical modalities for survival prediction using NSCLC TCGA data, we observe the C-index improvement from 0.57±.04 to 0.61±.01. Best individual modality performance was at 0.59±.02 using clinical modality. Dominant modalities in unimodal survival analysis varied between cancers, with clinical, RNA, and miRNA for LUAD, and clinical, RNA, and RPPA for LUSC. Using pan-cancer TCGA data for survival prediction, the best C-index = 0.77 was achieved using multi-omics model, followed by 0.76±.01 for clinical, 0.75±.01 for RNA seq, and 0.73±.01 for RPPA unimodal models. Citation Format: Nikos Nikolaou, Domingo Salazar, Harish RaviPrakash, Miguel Goncalves, Gustavo Alonso Arango Argoty, Nikolay Burlutsky, Natasha Markuzon, Etai Jacob. Improving survival prediction using flexible late fusion machine learning framework for multi-omics data integration. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5395.

A Clinical Prediction Model of Overall Survival for Patients with Cervical Cancer Aged 25–69 Years

Aims: This study aims to develop a prediction tool for the overall survival of cervical cancer patients. Methods: We obtained 4116 female patients diagnosed with cervical cancer aged 25-69 during 2008-2019 from the Surveillance, Epidemiology, and End Results Program. The overall survival between groups was illustrated by the Kaplan-Meier method and compared by a log-rank test adjusted by the Bonferroni-Holm method. We first performed the multivariate Cox regression analysis to evaluate the predictive values of the variables. A prediction model was created using cox regression based on the training set, and the model was presented as a nomogram. The proposed nomogram was designed to predict the 1-year, 3-year, and 5-year overall survival of patients with cervical cancer. Besides the c-index, time-dependent receiver operating curves, and calibration curves were created to evaluate the accuracy of the nomogram at the timepoint of one year, three years, and five years. Results: With a median follow-up of 54 (28, 92) months, 1045 (25.39%) patients were deceased. Compared with alive individuals, the deceased were significantly older and the primary site was more likely to be the cervix uteri site, large tumor size, higher grade, and higher combined summary stage (all p values < 0.001). In the multivariate Cox regression, age at diagnosis, race, tumor size, grade, combined summary stage, pathology, and surgery treatment were significantly associated with the all-cause mortality for patients with cervical cancer. The proposed nomogram showed good performance with a C-index of 0.82 in the training set. The 1-year, 3-year, and 5-year areas under the curves (with 95% confidence interval) of the receiver operating curves were 0.88 (0.84, 0.91), 0.84 (0.81, 0.87), and 0.83 (0.80, 0.86), respectively. Conclusions: This study develops a prediction nomogram model for the overall survival of cervical cancer patients with a good performance. Further studies are required to validate the prediction model further.

Clinical usefulness of the lymphocyte-to-monocyte ratio and aggregate index of systemic inflammation in patients with esophageal cancer: a retrospective cohort study

BackgroundMultiple perioperative inflammatory markers are considered important factors affecting the long-term survival of esophageal cancer (EC) patients. Hematological parameters, whether single or combined, have high predictive value.AimTo investigate the inflammatory status of patients with preoperative EC using blood inflammatory markers, and to establish and validate competing risk nomogram prediction models for overall survival (OS) and progression-free survival (PFS) in EC patients.MethodsA total of 508 EC patients who received radical surgery (RS) treatment in The First Affiliated Hospital of Zhengzhou University from August 5, 2013, to May 1, 2019, were enrolled and randomly divided into a training cohort (356 cases) and a validation cohort (152 cases). We performed least absolute shrinkage and selection operator (LASSO)-univariate Cox- multivariate Cox regression analyses to establish nomogram models. The index of concordance (C-index), time-dependent receiver operating characteristic (ROC) curves, time-dependent area under curve (AUC) and calibration curves were used to evaluate the discrimination and calibration of the nomograms, and decision curve analysis (DCA) was used to evaluate the net benefit of the nomograms. The relative integrated discrimination improvement (IDI) and net reclassification improvement (NRI) were calculated to evaluate the improvement in predictive accuracy of our new model compared with the AJCC staging system and another traditional model. Finally, the relationship between systemic inflammatory response markers and prognostic survival was explored according to risk plot, time-dependent AUC, Kaplan–Meier and restricted cubic spline (RCS).ResultsBased on the multivariate analysis for overall survival (OS) in the training cohort, nomograms with 10 variables, including the aggregate index of systemic inflammation (AISI) and lymphocyte-to-monocyte ratio (LMR), were established. Time-dependent ROC, time-dependent AUC, calibration curves, and DCA showed that the 1-, 3-, and 5 year OS and PFS probabilities predicted by the nomograms were consistent with the actual observations. The C-index, NRI, and IDI of the nomograms showed better performance than the AJCC staging system and another prediction model. Moreover, risk plot, time-dependent AUC, and Kaplan–Meier showed that higher AISI scores and lower LMR were associated with poorer prognosis, and there was a nonlinear relationship between them and survival risk.ConclusionAISI and LMR are easy to obtain, reproducible and minimally invasive prognostic tools that can be used as markers to guide the clinical treatment and prognosis of patients with EC.

Epidemiology, antimicrobial resistance, and mortality risk factors of carbapenem resistant gram-negative bacteria in hematopoietic stem cell transplantation recipients.

Carbapenem resistant gram-negative bacteria (CRGNB) infection is more and more frequent in patients after hematopoietic stem cell transplantation (HSCT), and the prognosis is very poor. The purpose of this study was to investigate the clinical characteristics and risk factors for mortality with CRGNB infection in HSCT recipients, and to provide useful information for guiding the application of antibiotics and improving the prognosis in the future. Electronic medical records of CRGNB infected patients who underwent HSCT in Xiangya Hospital from January 1, 2015 to June 30, 2022 were collected. At the same time, 1:1 case-control matching was performed according to gender, age and disease type. The epidemiological characteristics and drug resistance of patients with CRGNB infection and non-CRGNB infection were compared. Logistic regression and Cox regression analysis were used to determine the risk factors for CRGNB acquisition and death respectively, and a prediction model of overall survival was constructed by R language. The crude infection rate of CRGNB in HSCT recipients was 7.42%, and the mortality rate was 47.1%. CRGNB was resistant to most commonly used antibiotics. Time interval from diagnosis to transplantation >180 days (HR=7.886, 95% CI 2.624-23.703, P=0.000), septic shock (HR=6.182, 95% CI 2.605-14.671, P=0.000), platelet count < 20 × 109/L (HR=2.615, 95% CI 1.152-5.934, P=0.022) and total bilirubin > 34.2 μmol/L (HR=7.348, 95% CI 2.966-18.202, P=0.000) at the initial stage of infection were 4 independent risk factors associated with mortality. CRGNB infection has become a serious threat to HSCT recipients. Clinicians should pay high attention to it and actively seek personalized treatment strategies suitable for local medical conditions.

The AUGIS Survival Predictor: Prediction of Long-Term and Conditional Survival After Esophagectomy Using Random Survival Forests.

The aim of this study was to develop a predictive model for overall survival after esophagectomy using pre/postoperative clinical data and machine learning. For patients with esophageal cancer, accurately predicting long-term survival after esophagectomy is challenging. This study investigated survival prediction after esophagectomy using a RandomSurvival Forest (RSF) model derived from routine data from a large, well-curated, national dataset. Patients diagnosed with esophageal adenocarcinoma or squamous cell carcinoma between 2012 and 2018 in England and Wales who underwent an esophagectomy were included. Prediction models for overall survival were developed using the RSF method and Cox regression from 41 patient and disease characteristics. Calibration and discrimination (time-dependent area under the curve) were validated internally using bootstrap resampling. The study analyzed 6399 patients, with 2625 deaths during follow-up. Median follow-up was 41 months. Overall survival was 47.1% at 5 years. The final RSF model included 14 variables and had excellent discrimination with a 5-year time-dependent area under the receiver operator curve of 83.9% [95% confidence interval (CI) 82.6%-84.9%], compared to 82.3% (95% CI 81.1%-83.3%) for the Cox model. The most important variables were lymph node involvement, pT stage, circumferential resection margin involvement (tumor at < 1 mm from cut edge) and age. There was a wide range of survival estimates even within TNM staging groups, with quintiles of prediction within Stage 3b ranging from 12.2% to 44.7% survival at 5 years. An RSF model for long-term survival after esophagectomy exhibited excellent discrimination and well-calibrated predictions. At a patient level, it provides more accuracy than TNM staging alone and could help in the delivery of tailored treatment and follow-up.

Deep learning based on preoperative magnetic resonance (MR) images improves the predictive power of survival models in primary spinal cord astrocytomas.

Prognostic models for spinal cord astrocytoma patients are lacking due to the low incidence of the disease. Here, we aim to develop a fully automated deep learning (DL) pipeline for stratified overall survival (OS) prediction based on preoperative MR images. A total of 587 patients diagnosed with intramedullary tumors were retrospectively enrolled in our hospital to develop an automated pipeline for tumor segmentation and OS prediction. The automated pipeline included a T2WI-based tumor segmentation model and 3 cascaded binary OS prediction models (1-year, 3-year, and 5-year models). For the tumor segmentation model, 439 cases of intramedullary tumors were used to model training and testing using a transfer learning strategy. A total of 138 patients diagnosed with astrocytomas were included to train and test the OS prediction models via 10 × 10-fold cross-validation using CNNs. The dice of the tumor segmentation model with the test set was 0.852. The results indicated that the best input of OS prediction models was a combination of T2W and T1C images and the tumor mask. The 1-year, 3-year, and 5-year automated OS prediction models achieved accuracies of 86.0%, 84.0%, and 88.0% and AUCs of 0.881 (95% CI 0.839-0.918), 0.862 (95% CI 0.827-0.901), and 0.905 (95% CI 0.867-0.942), respectively. The automated DL pipeline achieved 4-class OS prediction (<1 year, 1-3 years, 3-5 years, and >5 years) with 75.3% accuracy. We proposed an automated DL pipeline for segmenting spinal cord astrocytomas and stratifying OS based on preoperative MR images.