Predictor Of Survival In Patients Research Articles

CTNI-65. DETERMINANTS OF LONG-TERM SURVIVAL IN PATIENTS WITH IDH-MUTANT GLIOMAS

Abstract BACKGROUND Survival times of patients with IDH-mutant gliomas is variable and can extend to decades. Many studies only provide progression-free rather than overall survival times and prognostic factors remain ill-defined. Here we explored clinical disease characteristics of long-term survivors within a large cohort of patients with extended follow-up. METHODS This single-center, retrospective analysis included 114 patients with IDH-mutant glioma that either died due to tumor-related causes or survived at least 15 years after first diagnosis. Patient characteristics, outcomes, and prognostic factors were stratified by short- (< 15 years) versus long-term (≥ 15 years) survival. Initial diagnosis was (re-)classified according to the WHO 2016 classification. Uni- and multivariate analyses were performed. RESULTS Overall, 66 patients (58%) were diagnosed with astrocytoma and 48 patients (42%) with oligodendroglioma. Median follow-up of the survivors was 16.6 years (range 15-28.9). 62 tumor-related deaths (54%) had been reported at database closure. Patients that died before reaching 15 years survival time had been diagnosed with astrocytoma in 41 (72%) and oligodendroglioma in 16 cases (28%). The long-term survivor cohort comprised 25 patients (44%) with astrocytoma and 32 patients (56%) with oligodendroglioma. Long-term survival was associated with grade 2 histology (p < 0.01), smaller tumor volumes (p < 0.01), and lack of contrast enhancement (p = 0.02). 41 long-term survivors (72%) were initially monitored by a wait-and-scan strategy as opposed to 16 patients (28%) in the cohort of short-term survival (p<0.01). In patients with oligodendroglioma, a higher KPS was associated with longer survival (HR 0.25; 95% CI 0.02-0.67; p = 0.05). Tumor resection (HR 0.43; 95% CI 0.22-0.86; p = 0.02) and wait-and-scan strategies (HR 0.34; 95% CI 0.14-0.79; p = 0.01) were associated with longer survival in patients with astrocytoma. CONCLUSION Tumor resection followed by a wait-and-scan strategy may yield excellent survival times, especially in patients with IDH-mutant grade 2 astrocytoma. Age appears not be an important predictor for survival in patients with IDH-mutant gliomas.

Integrating radiomic and 3D autoencoder-based features for Non-Small Cell Lung Cancer survival analysis

Background and objectivesThe aim of this study is to develop a radiomic and deep learning-based signature for survival analysis of patients with Non-Small Cell Lung Cancer. MethodsFour-hundred twenty-two patients from “Lung1” dataset were included in the study. A 3D convolutional autoencoder (AE) was built and features from the latent space extracted for further analysis. Radiomic features were derived from the 3D volume of the tumor region using PyRadiomics. Both radiomic and AE-based features underwent feature selection, by removing: i) highly correlated and ii) constant features. The selected variables were then used to derive both mono-domain (radiomics, AE and clinic) and multi-domain signatures fitting a Cox Proportional Hazard model with LASSO penalization and evaluated considering the concordance (C)-index as performance metric. ResultsBoth mono-domain and multi-domain signatures could significantly differentiate high risk from low risk patients. Among the mono-domain signatures, the highest hazard ratio (HR) in the test set was obtained using radiomics (HR = 1.5428) followed by the AE-based signature (HR = 1.5012) and the clinical signature (HR = 1.4770). The best overall performance was achieved by combining all three signatures, resulting in the highest HR (HR = 1.7383), while the combination of AE-based and clinical signatures yielded the highest C-index (C-index = 0.6309). ConclusionsThese preliminary results show that combining information carried by AE, radiomic and clinical domain shows potential for improving the prediction of overall survival in NSCLC patients.

Machine Learning-enhanced Signature of Metastasis-related T Cell Marker Genes for Predicting Overall Survival in Malignant Melanoma.

In this study, we aimed to investigate disparities in the tumor immune microenvironment (TME) between primary and metastatic malignant melanoma (MM) using single-cell RNA sequencing (scRNA-seq) and to identify metastasis-related T cell marker genes (MRTMGs) for predicting patient survival using machine learning techniques. We identified 6 distinct T cell clusters in 10×scRNA-seq data utilizing the Uniform Manifold Approximation and Projection (UMAP) algorithm. Four machine learning algorithms highlighted SRGN, PMEL, GPR143, EIF4A2, and DSP as pivotal MRTMGs, forming the foundation of the MRTMGs signature. A high MRTMGs signature was found to be correlated with poorer overall survival (OS) and suppression of antitumor immunity in MM patients. We developed a nomogram that combines the MRTMGs signature with the T stage and N stage, which accurately predicts 1-year, 3-year, and 5-year OS probabilities. Furthermore, in an immunotherapy cohort, a high MRTMG signature was associated with an unfavorable response to anti-programmed death 1 (PD-1) therapy. In conclusion, primary and metastatic MM display distinct TME landscapes with different T cell subsets playing crucial roles in metastasis. The MRTMGs signature, established through machine learning, holds potential as a valuable biomarker for predicting the survival of MM patients and their response to anti-PD-1 therapy.

Bioinformatic analysis constructs an optimal prognostic index for survival-related variables (OPISV) based on whole-genome expression data in Glioblastoma

PurposeUsing clinical information and transcriptomic sequencing data from glioblastoma (GBM) patients in the TCGA database to perform gene-by-gene analysis that is aligned with individual patient characteristics and develop an optimal prognostic index of survival-related variables (OPISV) through iterative machine learning techniques to predict the prognosis of GBM patients. Study designThe TCGA dataset was utilized as the training dataset, while two GEO datasets served as independent validation cohorts. Initially, survival analysis (p < 0.001***), differential gene expression analysis (p < 0.05*), and univariate Cox regression analysis (p < 0.05*) were employed to identify genes that are highly correlated with patient prognosis and exhibit significant differences in survival status. Subsequently, incorporating the non-excludable variable of age, a multivariate Cox regression analysis was performed in a stepwise manner to construct the OPISV. Finally, logistic and LASSO regressions were used to validate the association between the identified genes and patient survival. The OPISV performance is evaluated and its potential mechanisms are explored. ResultsAge, CTSD, PTPRN, PTPRN2, NSUN5, DNAJC30 and SOX21 emerged as the optimal variables through multivariate Cox regression iterations. Further analysis characterized Age, PTPRN and DNAJC30 as independent prognostic risk factors for constructing OPISV, which is validated with external GEO datasets and GEPIA database. In OPISV_high populations, significantly upregulated GABAergic synapse function was exposed. Differential genes identified from gene clustering of the GABAergic synapse pathway and gene module highly correlated with GABAergic synapse in the WGCNA analysis are pointing unequivocally to the glioma progress. ConclusionOPISV is feasible for predicting patient survival, as it may serve as a potential mechanism underlying the involvement of GABAergic synapses in the progression of GBM.

Single- and multi-site radiomics may improve overall survival prediction for patients with metastatic lung adenocarcinoma

PurposeThe purpose of this study was to assess whether single-site and multi-site radiomics could improve the prediction of overall survival (OS) of patients with metastatic lung adenocarcinoma compared to clinicopathological model. Materials and methodsAdults with metastatic lung adenocarcinoma, pretreatment whole-body contrast-enhanced computed tomography examinations, and performance status (WHO-PS) ≤ 2 were included in this retrospective single-center study, and randomly assigned to training and testing cohorts. Radiomics features (RFs) were extracted from all measurable lesions with volume ≥ 1 cm3. Radiomics prognostic scores based on the largest tumor (RPSlargest) and the average RF values across all tumors per patient (RPSaverage) were developed in the training cohort using 5-fold cross-validated LASSO-penalized Cox regression. Intra-patient inter-tumor heterogeneity (IPITH) metrics were calculated to quantify the radiophenotypic dissimilarities among all tumors within each patient. A clinicopathological model was built in the training cohort using stepwise Cox regression and enriched with combinations of RPSaverage, RPSlargest and IPITH. Models were compared with the concordance index in the independent testing cohort. ResultsA total of 300 patients (median age: 63.7 years; 40.7% women; median OS, 16.3 months) with 1359 lesions were included (200 and 100 patients in the training and testing cohorts, respectively). The clinicopathological model included WHO-PS = 2 (hazard ratio [HR] = 3.26; P < 0.0001), EGFR, ALK, ROS1 or RET mutations (HR = 0.57; P = 0.0347), IVB stage (HR = 1.65; P = 0.0211), and liver metastases (HR = 1.47; P = 0.0670). In the testing cohort, RPSaverage, RPSlargest and IPITH were associated with OS (HR = 85.50, P = 0.0038; HR = 18.83, P = 0.0082 and HR = 8.00, P = 0.0327, respectively). The highest concordance index was achieved with the combination of clinicopathological variables and RPSaverage, significantly better than that of the clinicopathological model (concordance index = 0.7150 vs. 0.695, respectively; P = 0.0049) ConclusionSingle-site and multi-site radiomics-based scores are associated with OS in patients with metastatic lung adenocarcinoma. RPSaverage improves the clinicopathological model.

EDNRA regulates the tumour immune environment and predicts the efficacy and prognosis of cancer immunotherapy.

The potential role of endothelin receptor A (EDNRA) in cancer immunotherapy has been demonstrated; however, the mechanism of its therapeutic value remains to be investigated. This study aimed to reveal the potential link between cancer immunotherapy and EDNRA in human tumours. Clinical characteristics and gene expression information were acquired from the Cancer Genome Atlas database. The correlation between EDNRA expression and immune infiltration was analysed by tumour immune estimation resource (TIMER) and tumour-immune system interaction database (TISIDB). EDNRA expression in different cancer types were performed via qPCR. Immunohistochemistry was used to detect the relationships between EDNRA protein and immune checkpoints. The results have founded that EDNRA was differentially expressed in various tumours, and highly associated with patient's age and tumour stage. It is also of high potential prognostic value in predicting patient survival. It has been verified that the EDNRA, JAK-STAT, and TGF-β signalling pathways are involved in cancers. In general, EDNRA positively correlated with immunomodulatory agents, immune cell infiltration, and immunotherapy markers. Immunohistochemical analysis of breast cancer tissues showed that EDNRA was positively correlated with NRP1 expression. Furthermore, patients with low EDNRA levels showed a superior response to immunotherapy. The functional study found that EDNRA expression is upregulated in MDA-MB-231 and HepG2 cells, and knockdown of EDNRA inhibits proliferation and migration of cells. In conclusion, the immunotherapeutic function of EDNRA was elucidated in this study. EDNRA may be an important target in tumour immunotherapy and provide new insights for tumour immunotherapy.

Construction of a Prognostic Model for Mitochondria and Macrophage Polarization Correlation in Glioma Based on Single-Cell and Transcriptome Sequencing.

Numerous diseases are associated with the interplay of mitochondrial and macrophage polarization. However, the correlation of mitochondria-related genes (MRGs) and macrophage polarization-related genes (MPRGs) with the prognosis of glioma remains unclear. This study aimed to examine this relationship based on bioinformatic analysis. Glioma-related datasets (TCGA-GBMLGG, mRNA-seq-325, mRNA-seq-693, GSE16011, GSE4290, and GSE138794) were included in this study. The intersection genes were obtained by overlapping differentially expressed genes (DEGs) from differential expression analysis in GSE16011, key module genes from WGCNA, and MRGs. Subsequently, the intersection genes were further screened to obtain prognostic genes. Following this, a risk model was developed and verified. After that, independent prognostic factors were identified, followed by the construction of a nomogram and subsequent evaluation of its predictive ability. Furthermore, immune microenvironment analysis and expression validation were implemented. The GSE138794 dataset was utilized to evaluate the expression of prognostic genes at a cellular level, followed by conducting an analysis on cell-to-cell communication. Finally, the results were validated in different datasets and tissue samples from patients. ECI2, MCCC2, OXCT1, SUCLG2, and CPT2 were identified as prognostic genes for glioma. The risk model constructed based on these genes in TCGA-GBMLGG demonstrated certain accuracy in predicting the occurrence of glioma. Additionally, the nomogram constructed based on risk score and grade exhibited strong performance in predicting patient survival. Significant differences were observed in the proportion of 27 immune cell types (e.g., activated B cells and macrophages) and the expression of 32 immune checkpoints (e.g., CD70, CD200, and CD48) between the two risk groups. Single-cell RNA sequencing showed that CPT2, ECI2, and SUCLG2 were highly expressed in oligodendrocytes, neural progenitor cells, and BMDMs, respectively. The results of cell-cell communication analysis revealed that both oligodendrocytes and BMDMs exhibited a substantial number of interactions with high strength. This study revealed five genes associated with the prognosis of glioma (ECI2, MCCC2, OXCT1, SUCLG2, and CPT2), providing novel insights into individualized treatment and prognosis.

Prognostic Value of Adding Blood and Lymphatic Vessel Invasion to the 8th Classification of TNM in Lung Cancer in Stages I and II

ObjectivesExpanding TNM staging system for lung cancer with the addition of new prognostic factors could enhance patient stratification and survival prediction. The goal of this study is to assess if TNM prognosis capacity could be improved by incorporating other pathological characteristics of surgical specimen. MethodsWe retrospectively reviewed lung cancer resections, stages I–II, performed between January 1st 2010 and May 1st 2019. We collected clinical variables and pathological characteristics, including vascular, lymphovascular and perineural invasion, STAS, necrosis and stromal features. Mortality and recurrence-free survival were assessed with univariable and multivariable Cox analysis. We explored how these factors would modify the TNM Harrel's index. Results629 tumors were analyzed. Median overall survival was 53.9 months. Median recurrence-free survival was 47.6 months. Specific survival at 3, 5 and 10 years was 90, 83 and 74%. Recurrence-free survival at 3, 5 and 10 years was 76, 70 and 65%.The multivariable analysis showed that overall survival was significantly related to TNM classification (p<0.0002), vascular infiltration (HR 1.93, CI 1.42–2.64, p<0.0001), lymphovascular invasion (HR 1.88, CI 1.30–2.71, p<0.0015) and necrosis (HR 1.74, CI 1.24–2.45, p<0.0025). Harrell's index for TNM was 0.6139. Adding vascular, lymphovascular invasion and necrosis, it increased up to 0.6531.The multivariable analysis showed that specific survival was significantly related to TNM classification (p<0.001), vascular infiltration (HR 2.23, CI 1.44–3.46, p<0.001) and lymphovascular invasion (HR 1.85, CI 1.09–3.13, p<0.021). Harrell's index for TNM was 0.6645. Adding vascular and lymphovascular invasion, it increased up to 0.7103. Recurrence-free survival was related to TNM, vascular infiltration (HR 1.48, CI 1.05–2.09, p<0.023) and lymphovascular invasion (HR 2.40, CI 1.64–3.50, p<0.001). Harrell's index for TNM was 0.6264. Adding vascular and lymphovascular invasion, it increased up to 0.6794. ConclusionsIncluding vascular and angiolymphatic invasion in the staging system classification could better stratify patients at risk of recurrence and tumor-related death.

Development of a web-based tool for estimating individualized survival curves in glioblastoma using clinical, mRNA, and tumor microenvironment features with fusion techniques.

Glioblastoma (GBM), one of the most common brain tumors, is known for its low survival rates and poor treatment responses. This study aims to create a robust predictive model that integrates multiple feature types, including clinical data, RNA expression, and tumor microenvironment data, using fusion techniques to enhance model performance. We obtained data from the SEER database to assess the impact of nine demographic and clinical features on the survival of 58,495 GBM patients and built predictive machine learning models. Additionally, mRNA expression data from 600 GBM patients from TCGA, CGGA, and GEO were analyzed. We used Cox regression and LASSO to create a gene signature, which was compared against 13 published signatures for accuracy. Twenty-one machine learning models were applied to predict survival at multiple time points. Finally, we integrated multiple feature types using fusion techniques and developed a Shiny app to provide survival predictions for GBM patients. Using the SEER database, we constructed machine learning models based on nine clinical variables: age, gender, marital status, race, tumor site, laterality, surgery, chemotherapy, and radiation therapy. The best-performing model achieved AUC values of 0.775, 0.728, 0.692, and 0.683 for predicting survival at 6, 12, 18, and 24months in the testing cohort. In the merged TCGA, CGGA, and GEO cohorts, we identified 11 genes to develop predictive models. These 11 genes outperformed 13 other published gene signatures in predicting the prognosis of GBM. When incorporating mRNA features, tumor microenvironment features, and clinical variables into the fusion models, the AUC values for predicting survival at 6, 12, 18, and 24months were 0.641, 0.624, 0.655, and 0.637, respectively. A user-friendly tool for predicting the survival curve of individual GBM patients is available at https://zzubioinfo.shinyapps.io/mlGBM/ . Our study provides a web-based tool that includes two modules: one for predicting survival curves using only clinical variables, and another that integrates multiple feature types for more comprehensive predictions.

The apparent diffusion coefficient can serve as a predictor of survival in patients with gliomas

Background and purposeMagnetic resonance imaging is indispensable for the preoperative diagnosis of glioma. This study aimed to investigate the role of the apparent diffusion coefficient values as predictors of survival in patients with gliomas.Methods and materialsA retrospective analysis was conducted on 101 patients with gliomas who underwent surgery between 2015 and 2020. Diffusion-weighted MRI was performed before the surgery. The regions of interest were categorized into parenchymal area, non-enhancing peritumoral area, and necrotic or cystic area. All the patients were divided into three subgroups: the parenchyma group, the non-enhancing peritumoral signal abnormality group, and the necrosis or cyst group. Univariate and multivariate analyses were performed using COX regression.ResultsIn the parenchymal group, Ki67, P53, IDH, and the high or low ADC values were identified as independent prognosticators for disease-free survival, while Ki67, IDH, and the high or low ADC values for overall survival. In the non-enhancing peritumoral signal abnormality group, Ki67, P53, IDH, and the ADC parenchymal area/ADC non−enhancing peritumoral area ratio were identified as independent prognostic factors for disease-free survival, while Ki67, IDH, and the ADC parenchymal area/ADC non−enhancing peritumoral area ratio for overall survival. In the necrosis or cyst group, Ki67 was significantly associated with disease-free survival, while Ki67 and the ADC value of the necrotic or cystic area for overall survival.ConclusionsThe ADC values, including the ADC value in the parenchymal area, the ADC parenchymal area/ADC non−enhancing peritumoral area ratio, and the ADC value in the necrotic or cystic area, can serve as an efficient and potential index for predicting the survival of patients with glioma.

Electrocardiographic markers of depolarization and repolarization in lead V1 predict right ventricular dysfunction in patients with heart failure

Abstract Background Right ventricular (RV) dysfunction is an important predictor of survival in patients with heart failure (HF) with both preserved (HFpEF) and reduced ejection fraction (HFrEF). To date there is limited data on the value of the electrocardiogram in predicting RV function. Purpose Given these considerations the aim of our study was to assess the possible relationship between several electrocardiographic (ECG) parameters and right ventricular function evaluated by echocardiography. Methods We included 109 patients (age 61±14 years, 68men) of whom 89 had HF (HFpEF N=62 and HFrEF N=27) and 20 patients without HF. Twelve-lead ECG and echocardiograms were performed in all patients. We assessed echocardiographic parameters of global and regional RV function. Results Time to Intrinsicoid deflection in lead V1 (TID-V1) and the QTc interval measured in lead V1 (QTc-V1) were significantly more prolonged in patients with HF vs. patients with no HF (36±24ms vs. 23±15ms; p=0.03, and 450±40ms vs. 427±20ms; p=0.047 respectively). TID-V1 correlated significantly with the tricuspid annulus systolic excursion (TAPSE) and RV fractional area change (FAC; r=-0.31; p=0.02, and r=-0.35; p=0.029 respectively). QTc-V1 correlated with parameters of global and regional RV function including FAC, RV apex longitudinal strain, RV free wall strain, and global RV strain (r=-0.42; p=0.0004, r=0.35; p=0.0032, r=0.21; p=0.045, and r=0.41; p=0.0005). RV dysfunction as defined by RV-FAC&amp;lt;35% was predicted by TID-V1 &amp;gt; 30ms with a sensitivity of 61% and specificity of 75%; AUC=0.752 and by QTc-V1&amp;gt;440ms with a sensitivity of 80% and specificity of 57%; AUC=0.888, and by a multivariate logistic regression model that also included TID-V1 and QTcV1, age, TAPSE, and LV ejection fraction with an AUC=0.967 (Figure). Conclusions Lead V1 time to intrinsic deflection and QTc-V1 predict RV dysfunction independent of echocardiographic parameters of ventricular function in patients with HF.

Developmental-status-aware transcriptional decomposition establishes a cell state panorama of human cancers

BackgroundCancer cells evolve under unique functional adaptations that unlock transcriptional programs embedded in adult stem and progenitor-like cells for progression, metastasis, and therapeutic resistance. However, it remains challenging to quantify the stemness-aware cell state of a tumor based on its gene expression profile.MethodsWe develop a developmental-status-aware transcriptional decomposition strategy using single-cell RNA-sequencing-derived tissue-specific fetal and adult cell signatures as anchors. We apply our method to various biological contexts, including developing human organs, adult human tissues, experimentally induced differentiation cultures, and bulk human tumors, to benchmark its performance and to reveal novel biology of entangled developmental signaling in oncogenic processes.ResultsOur strategy successfully captures complex dynamics in developmental tissue bulks, reveals remarkable cellular heterogeneity in adult tissues, and resolves the ambiguity of cell identities in in vitro transformations. Applying it to large patient cohorts of bulk RNA-seq, we identify clinically relevant cell-of-origin patterns and observe that decomposed fetal cell signals significantly increase in tumors versus normal tissues and metastases versus primary tumors. Across cancer types, the inferred fetal-state strength outperforms published stemness indices in predicting patient survival and confers substantially improved predictive power for therapeutic responses.ConclusionsOur study not only provides a general approach to quantifying developmental-status-aware cell states of bulk samples but also constructs an information-rich, biologically interpretable, cell-state panorama of human cancers, enabling diverse translational applications.

Prognostic significance of Standard Uptake Value (SUVmax) and primary tumor size predicting patient survival in vulvar tumors.

Prognostic significance of Standard Uptake Value (SUVmax) and primary tumor size predicting patient survival in vulvar tumors.

ANGIOGENIC FACTORS AS PREDICTORS OF SURVIVAL IN PATIENTS WITH BONE SARCOMAS

Introduction. Primary bone sarcomas are relatively rare neoplasms. They occur predominantly in adolescents and young adults, are characterized by an aggressive clinical course and a high metastatic potential, have a poor prognosis. In this regard, the problem of stratifying the risk of metastasis and death in order to select the optimal treatment tactics always remains relevant. The study of such markers as endostatin, vascular endothelial growth factor (VEGF), insulin-like growth factors IGF-1, IGF-2 as potentially informative predictors of survival in patients with bone sarcomas is pathogenetically determined by their involvement in the processes of tumor growth and metastasis. Aim – to analyze the relationship between pre-treatment serum levels of endostatin, VEGF, IGF-1 and IGF-2 with overall survival rates in patients with bone sarcomas. Material and methods. An analysis of overall survival rates was carried out in 134 patients with malignant bone tumors aged from 1 to 73 years (Me=27.0 years; Q1–Q3: 18.0 – 43.0 years), among them 87 men (64.9%) and 47 women (35.1%). All patients comprised 4 groups: osteosarcoma (n=58), undifferentiated pleomorphic sarcoma (n=5), Ewing sarcoma (n=29), chondrosarcoma (n=42). In all of them, before the start of specific antitumor treatment, the levels of endostatin, VEGF, IGF-1, and IGF-2 in the blood serum were determined using enzyme immunoassay. The follow-up period ranged from 1 to 229 months (Me=23.0 months; Q1–Q3: 10.0–121.0 months). To analyze overall survival, life tables and Kaplan–Meier curves were constructed. The effect of the studied serum markers on survival rates was assessed using the logrank test and Cox proportional hazards regression model. Results. During observation, death occurred in 51 patients (38.1%). In the group of deceased patients, serum levels of endostatin and IGF-2 were statistically significantly increased compared to the group of surviving patients (p 0.001 and p=0.011, respectively). Kaplan-Meier survival analysis using the logrank test revealed statistically significant differences in overall survival between the four groups of patients with different serum levels of endostatin, IGF-1 and IGF-2. The Cox regression method established the prognostic significance of endostatin as a predictor of overall survival (HR=1.012; p=0.013). Conclusion. Endostatin, IGF-1, IGF-2 can serve as predictors of overall survival of patients with bone sarcomas.

18Fluorodeoxyglucose positron emission tomography (18F-FDG PET)–derived tumoral and peritumoral radiomic parameters can predict pathological subtype and survival in esophageal carcinoma

The aim of this study is to investigate the importance of the quantitative parameters of the tumoral and peritumoral regions in prediction of pathological subtypes and 1-year survival in patients with esophageal carcinoma. A total of 103 patients with esophageal squamous cell carcinoma (SCC) and adenocarcinoma (AC) and in whom 18fluorodeoxyglucose positron emission tomography/computerized tomography (18F-FDG PET/CT) was performed were included in the study. One-year progression-free survival (PFS) and overall survival times of all patients were noted. Primary tumor and peritumoral area were drawn with manual segmentation on the 18F-FDG PET images. Seventy-three quantitative parameters were extracted from tumoral and peritumoral volumes of interest by using software. The differences of parameters of tumoral and peritumoral regions were determined statistically between pathological subtypes and between 1-year survivors and non-survivors. Diagnostic models were created with the statistically significant parameters. The model that consists of a tumoral and a peritumoral parameter could classify the pathological subtypes in 61% of the patients correctly (area under the curve [AUC]: 0.706, 61.2% accuracy, 53.7% sensitivity, and 75% specificity). The model that was created with 2 tumoral parameters could classify the 1-year survival in 66% of the patients correctly (AUC: 0.695, 66% accuracy, 73.6% sensitivity, and 56.1% specificity). The model consisting of a tumoral and a peritumoral parameter detected 1-year PFS in 66% of the patients accurately (AUC: 0.687, 66% accuracy, 72.4% sensitivity, and 55.6% specificity). The quantitative parameters obtained from tumoral and peritumoral regions can provide information about pathological subtypes and 1-year survival in esophageal carcinoma.

Dynamic prediction of kidney allograft and patient survival using post-transplant estimated glomerular filtration rate trajectory.

Allograft loss is the most feared outcome of kidney transplant recipients. We aimed to develop a dynamic Bayesian model using estimated glomerular filtration rate (eGFR) trajectories to predict long-term allograft and patient survivals. We used data from the Australian and New Zealand Dialysis and Transplant registry and included all adult kidney transplant recipients (1980-2017) in Australia (derivation cohort) and New Zealand (NZ, validation cohort). Using a joint model, the temporal changes of eGFR trajectories were used to predict patient and allograft survivals. The cohort composed of 14 915 kidney transplant recipients [12 777 (86%) from Australia and 2138 (14%) from NZ] who were followed for a median of 8.9 years. In the derivation cohort, eGFR trajectory was inversely associated with allograft loss [every 10ml/min/1.73 m2 reduction in eGFR, adjusted hazard ratio [HR, 95% credible intervals (95%CI) 1.31 (1.23-1.39)] and death [1.12 (1.10-1.14)]. Similar estimates were observed in the validation cohort. The respective dynamic area under curve (AUC) (95%CI) estimates for predicting allograft loss at 5-years post-transplantation were 0.83 (0.75-0.91) and 0.81 (0.68-0.93) for the derivation and validation cohorts. This straightforward model, using a single metric of eGFR trajectory, shows good model performance, and effectively distinguish transplant recipients who are at risk of death and allograft loss from those who are not. This simple bedside tool may facilitate early identification of individuals at risk of allograft loss and death.

VASARI-AUTO: PERFORMANT, EQUITABLE, EffiCIENT, ECONOMIC, AND SURVIVAL PREDICTIVE FEATURISATION OF GLIOMA MRI

Abstract AIMS The VASARI (Visually AcceSAble Rembrandt Images) MRI feature set is a quantitative system designed to standardize glioma imaging descriptions. Though effective, deriving VASARI is time-consuming and therefore seldom used in clinical practice. This is however a problem that performant machine-learning software could plausibly automate. METHOD In 100 patients, two consultant neuroradiologists independently quantified VASARI features. In parallel, we developed VASARI-auto, an automated VASARI-labelling software applied to both open-source hand-segmented lesion masks and our openly available tumour segmentation model. We quantified: 1) agreement across neuroradiologists and VASARI-auto; 2) calibration of patient equity; 3) an economic workforce analysis; and 4) fidelity in predicting patient overall survival. Neuroradiologists were blinded to VASARI-auto software development and evaluations, and software developers were blinded to neuroradiologist labelling. RESULTS Segmentation and VASARI-auto were equally performant regardless of age or sex (mean segmentation Dice coefficient 0.947). A modest inter-rater variability between neuroradiologists (mean Cohen’s Kappa 0.49) was comparable between neuroradiologists and VASARI-auto (mean Cohen’s Kappa 0.41), with higher agreement across VASARI-auto methods (mean Cohen’s Kappa 0.94). The time taken for neuroradiologists to derive VASARI features was substantially higher than VASARI-auto (mean time per case 317 vs. 3 seconds, p&amp;lt;0.0001). A hospital workforce analysis of our centre forecast that three years of VASARI glioma featurisation would demand 771 consultant neuroradiologist workforce hours (£40,789) or 8.6 hours of computing time (£3.76 of power) with VASARI-auto. The best-performing survival model utilised VASARI-auto features (R2 0.25), opposed to those derived by consultant neuroradiologists (R2 0.21). CONCLUSION VASARI-auto is a highly efficient automated labelling system for VASARI featurisation with equitable performance regardless of patient age or sex, a favourable economic profile if used as a decision support tool, and with non-inferior fidelity in downstream patient survival prediction. Future work should integrate such tools to enhance patient care.

Survival prediction in diffuse large B-cell lymphoma patients: multimodal PET/CT deep features radiomic model utilizing automated machine learning

PurposeWe sought to develop an effective combined model for predicting the survival of patients with diffuse large B-cell lymphoma (DLBCL) based on the multimodal PET-CT deep features radiomics signature (DFR-signature).Methods369 DLBCL patients from two medical centers were included in this study. Their PET and CT images were fused to construct the multimodal PET-CT images using a deep learning fusion network. Then the deep features were extracted from those fused PET-CT images, and the DFR-signature was constructed through an Automated machine learning (AutoML) model. Combined with clinical indexes from the Cox regression analysis, we constructed a combined model to predict the progression-free survival (PFS) and the overall survival (OS) of patients. In addition, the combined model was evaluated in the concordance index (C-index) and the time-dependent area under the ROC curve (tdAUC).ResultsA total of 1000 deep features were extracted to build a DFR-signature. Besides the DFR-signature, the combined model integrating metabolic and clinical factors performed best in terms of PFS and OS. For PFS, the C-indices are 0.784 and 0.739 in the training cohort and internal validation cohort, respectively. For OS, the C-indices are 0.831 and 0.782 in the training cohort and internal validation cohort.ConclusionsDFR-signature constructed from multimodal images improved the classification accuracy of prognosis for DLBCL patients. Moreover, the constructed DFR-signature combined with NCCN-IPI exhibited excellent potential for risk stratification of DLBCL patients.

New model to predict survival in advanced pancreatic ductal adenocarcinoma patients by measuring GGT and LDH levels and monocyte count.

Pancreatic ductal adenocarcinoma (PDAC) is a lethal cancer with a poor survival outcome. Predicting patient survival allows physicians to tailor treatments to specific individuals. Thus, a simple and cost-effective prognosis model is sorely needed. This retrospective study assesses the prognostic value of blood biomarkers in advanced and metastatic PDAC patients (n=96) from Spain. Cut-off points for hematological parameters were calculated and correlated with overall survival (OS) using Kaplan-Meier, log-rank test, robust Cox proportional hazards and logistic regressions. In univariate analysis, individuals with low levels of GGT, LDH, ALP, leukocyte-, neutrophil- and monocyte counts showed significantly longer survival than patients with higher levels. In multivariate analysis, lower levels of GGT (HR (95%CI), 2.734 (1.223-6.111); p=0.014), LDH (HR (95%CI), 1.876 (1.035-3.400); p=0.038) and monocyte count (HR (95%CI), 1.657 (1.095-2.506); p = 0.017) remained significantly beneficial. In consequence, we propose a prognostic model based on logistic regression (AUC=0.741) of these three biomarkers as a pioneer tool to estimate OS in PDAC. This study has demonstrated that the joint use of GGT (<92.00), LDH (<220.00) and monocyte count (<800) are independent positive prognostic factors in PDAC that can predict one-year survival in a novel prognostic logistic model.

DynProfiler: a Python package for comprehensive analysis and interpretation of signaling dynamics leveraged by deep learning techniques.

Signaling dynamics encode important features and regulatory mechanisms of biological systems, and recent studies have reported the use of simulated signaling dynamics with mechanistic modeling as biomarkers for human diseases. Given the success of deep learning techniques, it is expected that they can extract informative patterns from simulation results more effectively than traditional approaches involving manual feature selection, which can be used for subsequent analyses, such as patient stratification and survival prediction. Here, we propose DynProfiler, which utilizes the entire signaling dynamics, including intermediate variables, as input and leverages deep learning techniques to extract informative features without requiring any labels. Furthermore, DynProfiler incorporates a modern explainable AI solution to provide quantitative time-dependent importance scores for each dynamics. Using simulated dynamics of patients with breast cancer as an example, we demonstrate DynProfiler's ability to extract high-quality features that can predict mortality risk and identify important dynamics, highlighting upregulated phosphorylated GSK3β as a biomarker for poor prognosis. Overall, this tool can be useful for clinical application, as well as for elucidating biological system dynamics. The DynProfiler Python library is available in GitHub at https://github.com/okadalabipr/DynProfiler.