Prediction Model For Overall Survival Research Articles

Development and validation of a nomogram to predict overall survival in patients with glioma: a population-based study.

The objective is to investigate the prognostic factors associated with gliomas and to develop and assess a predictive nomogram model connected to survival that may serve as an additional resource for the clinical management of glioma patients. From 2010 to 2015, participants included in the study were chosen from the Surveillance Epidemiology and End Results (SEER) database. Gliomas were definitively diagnosed in each of them. They were divided into the training group and the validation cohort at random (7/3 ratio) using a random number table. To identify the independent predictive markers for overall survival (OS), Cox regression analysis was utilized. Subsequently, the training cohort's survival-related nomogram predictive model for OS was created by incorporating the fundamental patient attributes. Following that, the training cohort's model underwent internal validation. The nomogram model's authenticity and reliability were assessed through the computation of receiver operating characteristic (ROC) curves and concordance index (C-index). To evaluate the degree of agreement between the observed and predicted values in the training and validation cohorts, calibration plots were created. Age, primary site, histological type, surgery, chemotherapy, marital status, and grade were the independent predictive factors for OS in the training cohort, according to Cox regression analysis. Moreover, the nomogram model for predicting 1-year, 3-year, and 5-year OS was built using these variables. The C-indexes of OS for glioma patients in the training cohort and internal validation cohort were found to be 0.779 (95% CI=0.769-0.789) and 0.776 (95% CI=0.760-0.792), respectively, according to the results. The ROC curves also demonstrated good discrimination. Additionally, calibration plots demonstrated a fair amount of agreement. In summary, the nomogram prediction model of OS demonstrated a moderate level of reliability in its predictive performance, offering valuable reference data to enable doctors to quickly and easily determine the survival likelihood of patients with gliomas.

Radiomic survival prediction with liver metastases in colorectal cancer.

e15584 Background: The survival outcome of colorectal cancer depends upon staging and the presence of metastases. Nonetheless, there isn’t a reliable prediction model for overall survival (OS) specifically tailored to patients with colorectal cancer who have liver metastases. Consequently, this study aims to find radiomic features that can be utilized to predict OS of colorectal cancer with liver metastases. Methods: We utilized the Colorectal Liver Metastases data set in The Cancer Imaging Archive (TCIA). Overall, 197 patients with colorectal cancer metastatic to the liver were identified. Semi-automatic segmentation of liver metastases was used to extract radiomic features including Gray-Level Co-occurrence (glcm), Gray-Level Run Length Matrix (glrlm), Gray-Level Size Zone Matrix (glszm), Gray-Level Dependence Matrix (gldm), and Neighborhood Gray-Tone Difference Matrix (ngtdm). 3D Slicer was used to extract radiomic features. We tested a Cox proportional hazards model in each radiomics feature to predict OS. Results: Total 107 radiomic features were extracted from pre-segmented liver metastases on patients’ Computed Tomography images. Among them, original shape sphericity, original glcm correlation, original gldm DependenceEntropy, and original glrlm RunEntropy were statistically significant for the prediction of OS, with a hazard ratio of -2.16 (CI 0.01-0.92; p value 0.04), 1.54 (CI 1.5-14.0; p value 0.006), 0.51 (CI 1.0-2.5; p value 0.02), and 0.52 (CI 1.0-2.6; p value 0.03), respectively. Conclusions: Our data analysis indicates crucial radiomic features that hold promise for predicting the clinical outcome of patients with colorectal cancer who have liver metastases. Future research with a larger sample size is warranted for the external validation of our results.

Integrating IASLC grading and radiomics for predicting postoperative outcomes in stage IA invasive lung adenocarcinoma.

The International Association for the Study of Lung Cancer (IASLC) Pathology Committee introduced a histologic grading system for invasive lung adenocarcinoma (LUAD) in 2020. The IASLC grading system, hinging on the evaluation of predominant and high-grade histologic patterns, has proven to be practical and prognostic for invasive LUAD. However, there are still limitations in evaluating the prognosis of stage IA LUAD. Radiomics may serve as a valuable complement. To establish a model that integrates IASLC grading and radiomics, aimed at predicting the prognosis of stage IA LUAD. We conducted a retrospective analysis of 628 patients diagnosed with stage IA LUAD who underwent surgical resection between January 2015 and December 2018 at our institution. The patients were randomly divided into the training set (n=439) and testing set (n=189) at a ratio of 7:3. Overall survival (OS) and disease-free survival (DFS) were taken as the end points. Radiomics features were obtained by PyRadiomics. Feature selection was performed using the least absolute shrinkage and selection operator (LASSO). The prediction models for OS and DFS were developed using multivariate Cox regression analysis, and the models were visualized through nomogram plots. The model's performance was evaluated using area under the curves (AUC), concordance index (C-index), calibration curves, and survival decision curve analysis (DCA). In total, nine radiomics features were selected for the OS prediction model, and 15 radiomics features were selected for the DFS prediction model. Patients with high radiomics scores were associated with a worse prognosis (p<0.001). We built separate prediction models using radiomics or IASLC alone, as well as a combined prediction model. In the prediction of OS, we observed that the combined model (C-index: 0.812±0.024, 3 years AUC: 0.692, 5 years AUC: 0.792) achieved superior predictive performance than the radiomics (C-index: 0.743±0.038, 3 years AUC: 0.633, 5 years AUC: 0.768) and IASLC grading (C-index: 0.765±0.042, 3 years AUC: 0.658, 5 years AUC: 0.743) models alone. Similar results were obtained in the models for DFS. The combination of radiomics and IASLC pathological grading proves to be an effective approach for predicting the prognosis of stage IA LUAD. This has substantial clinical relevance in guiding treatment decisions for early-stage LUAD.

Prognostic factors and novel prediction models for overall survival of patients with submandibular gland cancer: A population-based retrospective cohort study

BackgroundAccurately predicting the survival rate of submandibular gland cancer (SGC) is of significant importance for guiding treatment decision-making and improving patient outcomes. This study was aimed to identify the independent prognostic factors of overall survival (OS) in SGC patients, and develop novel prediction models to aid clinicians in predicting the survival probability. Materials and methodsPatients diagnosed with primary SGC after the year 2010 were extracted from SEER database and then randomly allocated into training and test samples in a 7:3 ratio. Uni- and multi-variable COX analyses were employed using the training sample to ascertain independent prognostic factors for OS. Subsequently, graphic and online dynamic nomograms were established basing on the independent prognostic factors. We utilized C-index, calibration curve, receiver operating characteristic (ROC) curve, and area under ROC curve (AUC) value to evaluate the discrimination capacity and the consistency between predicted and actual survival. ResultsA total of 527 SGC patients were included (369 assigned to training group and 158 assigned to test group). The multivariable COX analysis showed that age, sex, marital status, tumor histology, summary stage, metastases to bone, and tumor size were independently associated with OS. Novel graphical and online dynamic (URL: https://yangxg1209.shinyapps.io/overall_survival_submandibular_gland_tumor/) nomograms were established. The C-indices (training: 0.77, 95%CI 0.71–0.84; test: 0.77, 95%CI 0.68–0.85) indicate favorable discrimination ability of the model, and the calibration curves demonstrated favorable consistency between the predicted and actual survival rates. ConclusionsOur study identified the independent prognostic factors influencing OS in patients with SGC, and successfully established and validated novel nomograms, which provide accurate prediction of survival rates and allows for personalized risk assessment.

Surgical treatment and prognosis analysis of hilar cholangiocarcinoma

Objective: To investigate the surgical treatment effect and prognostic factors of hilar cholangiocarcinoma. Methods: This is an ambispective cohort study. From August 2005 to December 2022,data of 510 patients who diagnosed with hilar cholangiocarcinoma and underwent surgical resection at the Hepatobiliary Center of the First Affiliated Hospital of Nanjing Medical University were retrospectively collected. In the cohort,there were 324 males and 186 females,with an age of (M (IQR)) 63(13)years (range:25 to 85 years). The liver function at admission was Child-Pugh A (343 cases,67.3%) and Child-Pugh B (167 cases,32.7%). Three hundred and seventy-two(72.9%) patients had jaundice symptoms and the median total bilirubin was 126.3(197.6) μmol/L(range: 5.4 to 722.8 μmol/L) at admission. Two hundred and fourty-seven cases (48.4%) were treated with percutaneous transhepatic cholangial drainage or endoscopic nasobiliary drainage before operation. The median bilirubin level in the drainage group decreased from 186.4 μmol/L to 85.5 μmol/L before operation. Multivariate Logistic regression was used to identify the influencing factors for R0 resection,and Cox regression was used to construct multivariate prediction models for overall survival(OS) and disease-free survival(DFS). Results: Among 510 patients who underwent surgical resection,Bismuth-Corlett type Ⅲ-Ⅳ patients accounted for 71.8%,among which 86.1% (315/366) underwent hemi-hepatectomy,while 81.9% (118/144) underwent extrahepatic biliary duct resection alone in Bismuch-Corlett type Ⅰ-Ⅱ patients. The median OS time was 22.8 months, and the OS rates at 1-,3-,5-and 10-year were 72.2%,35.6%,24.8% and 11.0%,respectively. The median DFS time was 15.2 months,and the DFS rates was 66.0%,32.4%,20.9% and 11.0%,respectively. The R0 resection rate was 64.5% (329/510), and the OS rates of patients with R0 resection at 1-,3-,5-and 10-year were 82.5%, 48.6%, 34.4%, 15.2%,respectively. The morbidity of Clavien-Dindo grade Ⅲ-Ⅴ complications was 26.1%(133/510) and the 30-day mortality was 4.3% (22/510). Multivariate Logistic regression indicated that Bismuth-Corlett type Ⅰ-Ⅲ (P=0.009), hemi-hepatectomy and extended resection (P=0.001),T1 and T2 patients without vascular invasion (T2 vs. T1:OR=1.43 (0.61-3.35),P=0.413;T3 vs. T1:OR=2.57 (1.03-6.41), P=0.010;T4 vs. T1, OR=3.77 (1.37-10.38), P<0.01) were more likely to obtain R0 resection. Preoperative bilirubin,Child-Pugh grade,tumor size,surgical margin,T stage,N stage,nerve infiltration and Edmondson grade were independent prognostic factors for OS and DFS of hilar cholangiocarcinoma patients without distant metastasis. Conclusions: Radical surgical resection is necessary to prolong the long-term survival of hilar cholangiocarcinoma patients. Hemi-hepatectomy and extended resection,regional lymph node dissection and combined vascular resection if necessary,can improve R0 resection rate.

A refined prediction model for survival in hepatocellular carcinoma patients treated with transarterial chemoembolization.

Transarterial chemoembolization (TACE) is widely performed as a major treatment for hepatocellular carcinoma (HCC) patients, and there is a need to stratify patients for whom the most benefit from the treatment. This study aimed to develop a refined prediction model for overall survival (OS) in patients undergoing TACE as a first-line treatment in a large cohort and validate its performance. A total of 2,632 patients with HCC of Barcelona Clinic Liver Cancer stage A or B who underwent TACE between 2008 and 2017 were enrolled. The patients were randomly assigned to a training cohort (n = 1,304) or a validation cohort (n = 1,328). Independent predictors of OS were used to develop a prediction model. The median age of patients in the entire cohort was 63 years, with the majority having hepatitis B virus (56.6%) and being classified as Child-Pugh class A (82.4%). We developed a new prognostic model, called the TACE-prognostic (TP) score, based on tumor burden (sum of the largest tumor diameter and tumor number), alpha-fetoprotein, and Albumin-Bilirubin grade. Patients were classified into five risk groups according to TP scores, with median survival significantly differentiated in both training and validation cohorts (P < 0.001). The new model consistently outperformed other currently available models in both the training and validation cohorts. This newly developed TP scoring system has the potential to be a useful tool in identifying ideal candidates of TACE and predicting OS with favorable performance and discrimination. However, further external validation is needed to confirm its effectiveness.

Heart and Lung Dose as Predictors of Overall Survival in Patients With Locally Advanced Lung Cancer. A National Multicenter Study

IntroductionIt is an ongoing debate how much lung and heart irradiation impact overall survival (OS) after definitive radiotherapy for lung cancer. This study uses a large national cohort of patients with locally advanced NSCLC to investigate the association between OS and irradiation of lung and heart. MethodsTreatment plans were acquired from six Danish radiotherapy centers, and patient characteristics were obtained from national registries. A hybrid segmentation tool automatically delineated the heart and substructures. Dose-volume histograms for all structures were extracted and analyzed using principal component analyses (PCAs). Parameter selection for a multivariable Cox model for OS prediction was performed using cross-validation based on bootstrapping. ResultsThe population consisted of 644 patients with a median survival of 26 months (95% confidence interval [CI]: 24–29). The cross-validation selected two PCA variables to be included in the multivariable model. PCA1 represented irradiation of the heart and affected OS negatively (hazard ratio, 1.14; 95% CI: 1.04–1.26). PCA2 characterized the left-right balance (right atrium and left ventricle) irradiation, showing better survival for tumors near the right side (hazard ratio, 0.92; 95% CI: 0.84–1.00). Besides the two PCA variables, the multivariable model included age, sex, body-mass index, performance status, tumor dose, and tumor volume. ConclusionsBesides the classic noncardiac risk factors, lung and heart doses had a negative impact on survival, while it is suggested that the left side of the heart is a more radiation dose–sensitive region. The data indicate that overall heart irradiation should be reduced to improve the OS if possible.

Prognostic biomarker HIF1α and its correlation with immune infiltration in gliomas.

Certain glioma subtypes, such as glioblastoma multiforme or low-grade glioma, are common malignant intracranial tumors with high rates of relapse and malignant progression even after standard therapy. The overall survival (OS) is poor in patients with gliomas; hence, effective prognostic prediction is crucial. Herein, the present study aimed to explore the potential role of hypoxia-inducible factor 1 subunit alpha (HIF1α) in gliomas and investigate the association between HIF1α and infiltrating immune cells in gliomas. Data from The Cancer Genome Atlas were evaluated via RNA sequencing, clinicopathological, immunological checkpoint, immune infiltration and functional enrichment analyses. Validation of protein abundance was performed using paraffin-embedded samples from patients with glioma. A nomogram model was created to forecast the OS rates at 1, 3 and 5 years after cancer diagnosis. The association between OS and HIF1α expression was estimated using Kaplan-Meier survival analysis and the log-rank test. Finally, HIF1α expression was validated using western blotting, reverse transcription-quantitative PCR, Cell Counting Kit-8 and Transwell assays. The results demonstrated that HIF1α expression was significantly upregulated in gliomas compared with normal human brain glial cells. Immunohistochemistry staining demonstrated differential expression of the HIF1α protein. Moreover, glioma cell viability and migration were inhibited via HIF1α downregulation. HIF1α impacted DNA replication, cell cycling, DNA repair and the immune microenvironment in glioma. HIF1α expression was also positively associated with several types of immune cells and immunological checkpoints and with neutrophils, plasmacytoid dendritic cells and CD56bright cells. The Kaplan-Meier survival analyses further demonstrated a strong association between high HIF1α expression and poor prognosis in patients with glioma. Analysis of the receiver operating characteristic curves demonstrated that HIF1α expression accurately differentiated paired normal brain cells from tumor tissues. Collectively, these findings suggested the potential for HIF1α to be used as a novel prognostic indicator for patients with glioma and that OS prediction models may help in the future to develop effective follow-up and treatment strategies for these patients.

Risk stratification in gastric cancer lung metastasis: Utilizing an overall survival nomogram and comparing it with previous staging.

Gastric cancer (GC) is prevalent and aggressive, especially when patients have distant lung metastases, which often places patients into advanced stages. By identifying prognostic variables for lung metastasis in GC patients, it may be possible to construct a good prediction model for both overall survival (OS) and the cumulative incidence prediction (CIP) plot of the tumour. To investigate the predictors of GC with lung metastasis (GCLM) to produce nomograms for OS and generate CIP by using cancer-specific survival (CSS) data. Data from January 2000 to December 2020 involving 1652 patients with GCLM were obtained from the Surveillance, epidemiology, and end results program database. The major observational endpoint was OS; hence, patients were separated into training and validation groups. Correlation analysis determined various connections. Univariate and multivariate Cox analyses validated the independent predictive factors. Nomogram distinction and calibration were performed with the time-dependent area under the curve (AUC) and calibration curves. To evaluate the accuracy and clinical usefulness of the nomograms, decision curve analysis (DCA) was performed. The clinical utility of the novel prognostic model was compared to that of the 7th edition of the American Joint Committee on Cancer (AJCC) staging system by utilizing Net Reclassification Improvement (NRI) and Integrated Discrimination Improvement (IDI). Finally, the OS prognostic model and Cox-AJCC risk stratification model modified for the AJCC system were compared. For the purpose of creating the OS nomogram, a CIP plot based on CSS was generated. Cox multivariate regression analysis identified eleven significant prognostic factors (P < 0.05) related to liver metastasis, bone metastasis, primary site, surgery, regional surgery, treatment sequence, chemotherapy, radiotherapy, positive lymph node count, N staging, and time from diagnosis to treatment. It was clear from the DCA (net benefit > 0), time-dependent ROC curve (training/validation set AUC > 0.7), and calibration curve (reliability slope closer to 45 degrees) results that the OS nomogram demonstrated a high level of predictive efficiency. The OS prediction model (New Model AUC = 0.83) also performed much better than the old Cox-AJCC model (AUC difference between the new model and the old model greater than 0) in terms of risk stratification (P < 0.0001) and verification using the IDI and NRI. The OS nomogram for GCLM successfully predicts 1- and 3-year OS. Moreover, this approach can help to appropriately classify patients into high-risk and low-risk groups, thereby guiding treatment.

Nomogram for predicting the overall survival and cancer-specific survival of patients with intraductal carcinoma of the prostate

PurposeIntraductal carcinoma of the prostate (IDC-P) is a histological subtype that differs from conventional acinar adenocarcinoma in terms of its origin, appearance, and pathological features. For IDC-P, there is currently no recognized best course of action, and its prognosis is unclear. The goal of this study is to analyze independent prognostic factors in IDC-P patients and to develop and validate a nomogram to predict overall survival (OS) and cancer-specific survival (CSS).MethodsClinical data for IDC-P patients were collected from the Surveillance, Epidemiology, and End Results database. To identify the independent variables influencing prognosis, multivariate Cox regression analysis was performed. A nomogram model was created utilizing these variables after comparing the variations in OS and CSS among various subgroups using Kaplan‒Meier curves. Internal validation of the nomograms was verified using the bootstrap resampling method.ResultsThe study included 280 IDC-P patients in total. Marital status, summary stage, grade, and the presence of lung metastases were significant factors impacting OS, and CSS was significantly influenced by marital status, summary stage, AJCC stage, the presence of lung metastases, the presence of bone metastases, and PSA according to univariate and multivariate Cox regression models (P < 0.05). Nomogram models were created to estimate OS and CSS using these parameters. The OS prediction model’s C-index was 0.744, whereas the CSS prediction model’s C-index was 0.831.ConclusionWe developed and verified nomogram models for the prediction of 1-, 3-, and 5-year OS and CSS in patients with IDC-P. These nomograms serve as a resource for evaluating patient prognosis, therapy, and diagnosis, ultimately improving clinical decision-making accuracy.

Interpretable machine learning model for prediction of overall survival in laryngeal cancer

Background: The mortality rates of laryngeal squamous cell carcinoma cancer (LSCC) have not significantly decreased in the last decades. Objectives: We primarily aimed to compare the predictive performance of DeepTables with the state-of-the-art machine learning (ML) algorithms (Voting ensemble, Stack ensemble, and XGBoost) to stratify patients with LSCC into chance of overall survival (OS). In addition, we complemented the developed model by providing interpretability using both global and local model-agnostic techniques. Methods: A total of 2792 patients in the Surveillance, Epidemiology, and End Results (SEER) database diagnosed with LSCC were reviewed. The global model-agnostic interpretability was examined using SHapley Additive exPlanations (SHAP) technique. Likewise, individual interpretation of the prediction was made using Local Interpretable Model Agnostic Explanations (LIME). Results: The state-of-the-art ML ensemble algorithms outperformed DeepTables. Specifically, the examined ensemble algorithms showed comparable weighted area under receiving curve of 76.9, 76.8, and 76.1 with an accuracy of 71.2%, 70.2%, and 71.8%, respectively. The global methods of interpretability (SHAP) demonstrated that the age of the patient at diagnosis, N-stage, T-stage, tumor grade, and marital status are among the prominent parameters. Conclusions: A ML model for OS prediction may serve as an ancillary tool for treatment planning of LSCC patients.

Explainable deep learning-based survival prediction for non-small cell lung cancer patients undergoing radical radiotherapy

Background and purposeSurvival is frequently assessed using Cox proportional hazards (CPH) regression; however, CPH may be too simplistic as it assumes a linear relationship between covariables and the outcome. Alternative, non-linear machine learning (ML)-based approaches, such as random survival forests (RSFs) and, more recently, deep learning (DL) have been proposed; however, these techniques are largely black-box in nature, limiting explainability. We compared CPH, RSF and DL to predict overall survival (OS) of non-small cell lung cancer (NSCLC) patients receiving radiotherapy using pre-treatment covariables. We employed explainable techniques to provide insights into the contribution of each covariable on OS prediction. Materials and methodsThe dataset contained 471 stage I-IV NSCLC patients treated with radiotherapy. We built CPH, RSF and DL OS prediction models using several baseline covariable combinations. 10-fold Monte-Carlo cross-validation was employed with a split of 70%:10%:20% for training, validation and testing, respectively. We primarily evaluated performance using the concordance index (C-index) and integrated Brier score (IBS). Local interpretable model-agnostic explanation (LIME) values, adapted for use in survival analysis, were computed for each model. ResultsThe DL method exhibited a significantly improved C-index of 0.670 compared to the CPH and a significantly improved IBS of 0.121 compared to the CPH and RSF approaches. LIME values suggested that, for the DL method, the three most important covariables in OS prediction were stage, administration of chemotherapy and oesophageal mean radiation dose. ConclusionWe show that, using pre-treatment covariables, a DL approach demonstrates superior performance over CPH and RSF for OS prediction and use explainable techniques to provide transparency and interpretability.

Establishment of Prognosis Nomogram for Primary Splenic Diffuse Large B Cell Lymphoma: A Study Based on SEER Database.

Despite being the most common primary tumor of the spleen, in the past, few studies have predicted the prognosis of primary spleen diffuse large B cell lymphoma. This study aimed to establish a nomogram prediction model of overall survival in primary DLBCL of the spleen. We screened out 347 patients with primary splenic DLBCL from surveillance, epidemiology, and end results database. According to the Cox regression results (age, Ann Arbor Stage, splenectomy and chemotherapy was the independent risk factor for primary splenic DLBCL), the nomogram was constructed. We evaluated the predictive ability of nomogram with C-Index (training cohort: 0.719 [0.669-0.769]; validation cohort: 0.711 [0.641-0.781]) and 3-year/5-year receiver operating characteristic area under curve (3-year/5-year ROCAUC, training cohort: 0.731/0.742; validation cohort: 0.721/0.742). Calibratioin plot shows that our predicted values fluctuate around the actual value, indicating good agreement with nomogram. The decision curve analysis (DCA) results showed that our nomogram could benefit more than Ann Arbor Stage for predicts the prognosis of the primary splenic DLBCL. The Kaplan-Meier and landmark analysis showed that a great discrimination between high-risk group and low-risk group (P < 0.05) and indicating that our nomogram has the good ability to identify high-risk patients. In this study, a nomogram prediction model for primary spleen DLBCL was established, which has good ability of prediction and generalization. It can help clinicians carry out individualized treatment measures.

Development and Validation of a Nomogram to Predict Overall Survival in Stage I–III Colorectal Cancer Patients after Radical Resection with Normal Preoperative Serum Carcinoembryonic Antigen

We aimed to develop a clinical predictive model for predicting the overall survival (OS) in stage I–III CRC patients after radical resection with normal preoperative CEA. This study included 1082 consecutive patients. They were further divided into a training set (70%) and a validation set (30%). The selection of variables for the model was informed by the Akaike information criterion. After that, the clinical predictive model was constructed, evaluated, and validated. The net reclassification index (NRI) and integrated discrimination improvement (IDI) were employed to compare the models. Age, histologic type, pT stage, pN stage, carbohydrate antigen 242 (CA242), and carbohydrate antigen 125 (CA125) were selected to establish a clinical prediction model for OS. The concordance index (C-index) (0.748 for the training set and 0.702 for the validation set) indicated that the nomogram had good discrimination ability. The decision curve analysis highlighted that the model has superior efficiency in clinical decision-making. NRI and IDI showed that the established nomogram markedly outperformed the TNM stage. The new clinical prediction model was notably superior to the AJCC 8th TNM stage, and it can be used to accurately assess the OS of stage I–III CRC patients undergoing radical resection with normal preoperative CEA.

NIMG-13. ROBUSTNESS OF PROGNOSTIC STRATIFICATION IN DE NOVO GLIOBLASTOMA PATIENTS ACROSS 22 GEOGRAPHICALLY DISTINCT INSTITUTIONS: INSIGHTS FROM THE RESPOND CONSORTIUM

Abstract PURPOSE Glioblastoma is the most prevalent primary malignant brain tumor in adults, with a median overall survival (OS) of approximately 15 months and only limited advancements in prognostication and survival prediction. This study aims to evaluate an AI-based prognostic stratification model for OS prediction trained on the ReSPOND consortium data and to validate its performance on an independent dataset. METHODS The AI model was trained on a cohort of 2,293 glioblastoma patients from 22 institutions across three continents. For validation, an independent cohort of 78 treatment-naïve patients was used from three institutions. Preoperative structural MRI scans were utilized for feature extraction. Automated segmentation defined three tumor sub-compartments: enhancing, necrotic, and peritumoral T2-FLAIR abnormality. The AI predictor incorporated variables such as patient age, normalized tumor sub-compartment volume, spatial distribution characteristics, and morphologic descriptors. The overall survival predictor index provided a continuous value between 0 and 1 for patient stratification that higher values indicating longer predicted OS. Generalizability was assessed using Leave-One-Cohort-Out-Cross-Validation (LOCOCV) for training data, and the model was subsequently applied to the validation cohort. RESULTS Survival analysis demonstrated a concordance index of 0.64 for LOCOCV training data and 0.59 for the independent validation data, indicating effective prognostic stratification of patients. CONCLUSION Multi-parametric AI assisted image analysis extracts prognostic biomarkers, which correlate with OS in glioblastoma patients. The generalizability of this method was validated using the extensive centralized glioblastoma imaging dataset registry from the ReSPOND consortium and an independent dataset, demonstrating its generalizability across diverse patient populations and acquisition settings. This model holds promise for robust prognostic stratification and prediction in de novo glioblastoma patients.

Construction of a Nomogram to Predict Overall Survival in Patients with Early-Onset Hepatocellular Carcinoma: A Retrospective Cohort Study.

Hepatocellular carcinoma (HCC) is a widespread and impactful cancer which has pertinent implications worldwide. Although most cases of HCC are typically diagnosed in individuals aged ≥60 years, there has been a notable rise in the occurrence of HCC among younger patients. However, there is a scarcity of precise prognostic models available for predicting outcomes in these younger patients. A retrospective analysis was conducted to investigate early-onset hepatocellular carcinoma (EO-LIHC) using data from the Surveillance, Epidemiology, and End Results (SEER) database from 2004 to 2018. The analysis included 1392 patients from the SEER database and our hospital. Among them, 1287 patients from the SEER database were assigned to the training cohort (n = 899) and validation cohort 1 (n = 388), while 105 patients from our hospital were assigned to validation cohort 2. A Cox regression analysis showed that age, sex, AFP, grade, stage, tumor size, surgery, and chemotherapy were independent risk factors. The nomogram developed in this study demonstrated its discriminatory ability to predict the 1-, 3-, and 5-year overall survival (OS) rates in EO-LIHC patients based on individual characteristics. Additionally, a web-based OS prediction model specifically tailored for EO-LIHC patients was created and validated. Overall, these advancements contribute to improved decision-making and personalized care for individuals with EO-LIHC.

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.