Lasso-Cox Model Research Articles

Deep learning-based morphological feature analysis and the prognostic association study in colon adenocarcinoma histopathological images.

Colorectal cancer (CRC) is now the third most common malignancy to cause mortality worldwide, and its prognosis is of great importance. Recent CRC prognostic prediction studies mainly focused on biomarkers, radiometric images, and end-to-end deep learning methods, while only a few works paid attention to exploring the relationship between the quantitative morphological features of patients' tissue slides and their prognosis. However, existing few works in this area suffered from the drawback of choosing the cells randomly from the whole slides, which contain the non-tumor region that lakes information about prognosis. In addition, the existing works, which tried to demonstrate their biological interpretability using patients' transcriptome data, failed to show the biological meaning closely related to cancer. In this study, we proposed and evaluated a prognostic model using morphological features of cells in the tumor region. The features were first extracted by the software CellProfiler from the tumor region selected by Eff-Unet deep learning model. Features from different regions were then averaged for each patient as their representative, and the Lasso-Cox model was used to select the prognosis-related features. The prognostic prediction model was at last constructed using the selected prognosis-related features and was evaluated through KM estimate and cross-validation. In terms of biological meaning, Gene Ontology (GO) enrichment analysis of the expressed genes that correlated with the prognostically significant features was performed to show the biological interpretability of our model.With the help of tumor segmentation, our model achieved better statistical significance and better biological interpretability compared to the results without tumor segmentation. Statistically, the Kaplan Meier (KM) estimate of our model showed that the model using features in the tumor region has a higher C-index, a lower p-value, and a better performance on cross-validation than the model without tumor segmentation. In addition, revealing the pathway of the immune escape and the spread of the tumor, the model with tumor segmentation demonstrated a biological meaning much more related to cancer immunobiology than the model without tumor segmentation. Our prognostic prediction model using quantitive morphological features from tumor regions was almost as good as the TNM tumor staging system as they had a close C-index, and our model can be combined with the TNM tumor stage system to make a better prognostic prediction. And to the best of our knowledge, the biological mechanisms in our study were the most relevant to the immune mechanism of cancer compared to the previous studies.

Using deep learning to predict survival outcome in non-surgical cervical cancer patients based on pathological images

PurposeWe analyzed clinical features and the representative HE-stained pathologic images to predict 5-year overall survival via the deep-learning approach in cervical cancer patients in order to assist oncologists in designing the optimal treatment strategies.MethodsThe research retrospectively collected 238 non-surgical cervical cancer patients treated with radiochemotherapy from 2014 to 2017. These patients were randomly divided into the training set (n = 165) and test set (n = 73). Then, we extract deep features after segmenting the HE-stained image into patches of size 224 × 224. A Lasso–Cox model was constructed with clinical data to predict 5-year OS. C-index evaluated this model performance with 95% CI, calibration curve, and ROC.ResultsBased on multivariate analysis, 2 of 11 clinical characteristics (C-index 0.68) and 2 of 2048 pathomic features (C-index 0.74) and clinical–pathomic model (C-index 0.83) of nomograms predict 5-year survival in the training set, respectively. In test set, compared with the pathomic and clinical characteristics used alone, the clinical–pathomic model had an AUC of 0.750 (95% CI 0.540–0.959), the clinical predictor model had an AUC of 0.729 (95% CI 0.551–0.909), and the pathomic model AUC was 0.703 (95% CI 0.487–0.919). Based on appropriate nomogram scores, we divided patients into high-risk and low-risk groups, and Kaplan–Meier survival probability curves for both groups showed statistical differences.ConclusionWe built a clinical–pathomic model to predict 5-year OS in non-surgical cervical cancer patients, which may be a promising method to improve the precision of personalized therapy.

Radiomics to predict the mortality of patients with rheumatoid arthritis-associated interstitial lung disease: A proof-of-concept study.

Patients with rheumatoid arthritis (RA) and interstitial lung disease (ILD) have increased mortality compared to the general population and factors capable of predicting RA-ILD long-term clinical outcomes are lacking. In oncology, radiomics allows the quantification of tumour phenotype by analysing the characteristics of medical images. Using specific software, it is possible to segment organs on high-resolution computed tomography (HRCT) images and extract many features that may uncover disease characteristics that are not detected by the naked eye. We aimed to investigate whether features from whole lung radiomic analysis of HRCT may alone predict mortality in RA-ILD patients. High-resolution computed tomographies of RA patients from January 2012 to March 2022 were analyzed. The time between the first available HRCT and the last follow-up visit or ILD-related death was recorded. We performed a volumetric analysis in 3D Slicer, automatically segmenting the whole lungs and trachea via the Lung CT Analyzer. A LASSO-Cox model was carried out by considering ILD-related death as the outcome variable and extracting radiomic features as exposure variables. We retrieved the HRCTs of 30 RA-ILD patients. The median survival time (interquartile range) was 48 months (36-120 months). Thirteen out of 30 (43.33%) patients died during the observation period. Whole line segmentation was fast and reliable. The model included either the median grey level intensity within the whole lung segmentation [high-resolution (HR) 9.35, 95% CI 1.56-55.86] as a positive predictor of death and the 10th percentile of the number of included voxels (HR 0.20, 95% CI 0.05-0.84), the voxel-based pre-processing information (HR 0.23, 95% CI 0.06-0.82) and the flatness (HR 0.42, 95% CI 0.18-0.98), negatively correlating to mortality. The correlation of grey level values to their respective voxels (HR 1.52 95% CI 0.82-2.83) was also retained as a confounder. Radiomic analysis may predict RA-ILD patients' mortality and may promote HRCT as a digital biomarker regardless of the clinical characteristics of the disease.

Machine learning models predict overall survival and progression free survival of non-surgical esophageal cancer patients with chemoradiotherapy based on CT image radiomics signatures

PurposeTo construct machine learning models for predicting progression free survival (PFS) and overall survival (OS) with esophageal squamous cell carcinoma (ESCC) patients.Methods204 ESCC patients were randomly divided into training cohort (n = 143) and test cohort (n = 61) according to the ratio of 7:3. Two radiomics models were constructed by radiomics features, which were selected by LASSO Cox model to predict PFS and OS, respectively. Clinical features were selected by univariate and multivariate Cox proportional hazards model (p < 0.05). Combined radiomics and clinical model was developed by selected clinical and radiomics features. The receiver operating characteristic curve, Kaplan Meier curve and nomogram were used to display the capability of constructed models.ResultsThere were 944 radiomics features extracted based on volume of interest in CT images. There were six radiomics features and seven clinical features for PFS prediction and three radiomics features and three clinical features for OS prediction; The radiomics models showed general performance in training cohort and test cohort for prediction for prediction PFS (AUC, 0.664, 0.676. C-index, 0.65, 0.64) and OS (AUC, 0.634, 0.646.C-index, 0.64, 0.65). The combined models displayed high performance in training cohort and test cohort for prediction PFS (AUC, 0.856, 0.833. C-index, 0.81, 0.79) and OS (AUC, 0.742, 0.768. C-index, 0.72, 0.71).ConclusionWe developed combined radiomics and clinical machine learning models with better performance than radiomics or clinical alone, which were used to accurate predict 3 years PFS and OS of non-surgical ESCC patients. The prediction results could provide a reference for clinical decision.

Nomogram of baseline CT-radiomics from small-cell lung cancer patients: evaluation of added prognostic value for overall survival and probability of distant metastasis

Objectives: To evaluate the added prognostic value of baseline CT-radiomics using nomogram for overall survival and probability of distant metastasis in small-cell lung cancer patients. Materials and methods: This retrospective study consisted of 122 patients with stage IIA-IIIB small-cell lung cancer,which 97 patients for training dataset and 25 for validation dataset. The function defined as rad_score was constructed by the linear combination of selected radiomics features from baseline CT images weighted by their respective logistic coefficients and intercept in the LASSO-Cox model. The nomogram was developed based on the above function for overall survival (OS) and calibrated by the Harrell’s concordance index (C-index). The performance of the classifiers for DM was evaluated by receiver operating characteristics (ROC) curves with the indictor of area under curves (AUC). Furthermore, survival curve depicted by Kaplan-Meier method was compared with Log-rank test between low- and high-risk group. Results: The nomogram performance of radiomics features and risk clinical factors (c-index of 0.64) don’t take advantage over the one of risk clinical factors-based alone (c-index of 0.596). The probability prediction of combination of the radiomics and clinical risk factor, radiomics alone, and clinical factors alone was shown, namely AUC of 0.673, 0.640 and 0.650, respectively. No significant different was found between ROCs (p-value &gt; 0.4, Delong test). Moreover, we compared the Kaplan-Meier curves between low- and high-risk group, and showed p&lt;0.001 with Log-rank test. Conclusion: In the study, we can not confirm the hypothesis that baseline CT-radiomics contribute to predict the OS and probability of DM significantly. Moreover, the nomogram model based on combination of radiomics and clinical parameters has disadvantage over clinical parameters alone, probably affected by heterogeneity of datasets or SCLC need more valuable information for prediction outcomes.

LASSO-Cox Modeling of Survival Using High-Resolution CT-Based Radiomic Features in a Cohort of COVID-19 Patients and Its Generalizability to Standard Image Reconstruction

Background: Few studies have focused on predicting the overall survival (OS) of patients affected by SARS-CoV-2 (i.e., COVID-19) using radiomic features (RFs) extracted from computer tomography (CT) images. Reconstruction of CT scans might potentially affect the values of RFs. Methods: Out of 435 patients, 239 had the scans reconstructed with a single modality, and hence, were used for training/testing, and 196 were reconstructed with two modalities were used as validation to evaluate RFs robustness to reconstruction. During training, the dataset was split into train/test using a 70/30 proportion, randomizing the procedure 100 times to obtain 100 different models. In all cases, RFs were normalized using the z-score and then given as input into a Cox proportional-hazards model regularized with the Least Absolute Shrinkage and Selection Operator (LASSO-Cox), used for feature selection and developing a robust model. The RFs retained multiple times in the models were also included in a final LASSO-Cox for developing the predictive model. Thus, we conducted sensitivity analysis increasing the number of retained RFs with an occurrence cut-off from 11% to 60%. The Bayesian information criterion (BIC) was used to identify the cut-off to build the optimal model. Results: The best BIC value indicated 45% as the optimal occurrence cut-off, resulting in five RFs used for generating the final LASSO-Cox. All the Kaplan-Meier curves of training and validation datasets were statistically significant in identifying patients with good and poor prognoses, irrespective of CT reconstruction. Conclusions: The final LASSO-Cox model maintained its predictive ability for predicting the OS in COVID-19 patients irrespective of CT reconstruction algorithms.

A novel risk score based on immune-related genes for hepatocellular carcinoma as a reliable prognostic biomarker and correlated with immune infiltration.

Immunological-related genes (IRGs) play a critical role in the immune microenvironment of tumors. Our study aimed to develop an IRG-based survival prediction model for hepatocellular carcinoma (HCC) patients and to investigate the impact of IRGs on the immune microenvironment. Differentially expressed IRGs were obtained from The Genomic Data Commons Data Portal (TCGA) and the immunology database and analysis portal (ImmPort). The univariate Cox regression was used to identify the IRGs linked to overall survival (OS), and a Lasso-regularized Cox proportional hazard model was constructed. The International Cancer Genome Consortium (ICGC) database was used to verify the prediction model. ESTIMATE and CIBERSORT were used to estimate immune cell infiltration in the tumor immune microenvironment (TIME). RNA sequencing was performed on HCC tissue specimens to confirm mRNA expression. A total of 401 differentially expressed IRGs were identified, and 63 IRGs were found related to OS on the 237 up-regulated IRGs by univariate Cox regression analyses. Finally, five IRGs were selected by the LASSO Cox model, including SPP1, BIRC5, STC2, GLP1R, and RAET1E. This prognostic model demonstrated satisfactory predictive value in the ICGC dataset. The risk score was an independent predictive predictor for OS in HCC patients. Immune-related analysis showed that the immune infiltration level in the high-risk group was higher, suggesting that the 5-IRG signature may play an important role in mediating immune escape and immune resistance in the TIME of HCC. Finally, we confirmed the 5-IRG signature is highly expressed in 65 HCC patients with good predictive power. We established and verified a new prognosis model for HCC patients based on survival-related IRGs, and the signature could provide new insights into the prognosis of HCC.

EOCSA: Predicting prognosis of Epithelial ovarian cancer with whole slide histopathological images

Ovarian cancer is one of the most serious cancers that threaten women around the world. Epithelial ovarian cancer (EOC), as the most commonly seen subtype of ovarian cancer, has rather high mortality rate and poor prognosis among various gynecological cancers. Survival analysis outcome is able to provide treatment advices to doctors. In recent years, with the development of medical imaging technology, survival prediction approaches based on pathological images have been proposed. In this study, we designed a deep framework named EOCSA which analyzes the prognosis of EOC patients based on pathological whole slide images (WSIs). Specifically, we first randomly extracted patches from WSIs and grouped them into multiple clusters. Next, we developed a survival prediction model, named DeepConvAttentionSurv (DCAS), which was able to extract patch-level features, removed less discriminative clusters and predicted the EOC survival precisely. Particularly, channel attention, spatial attention, and neuron attention mechanisms were used to improve the performance of feature extraction. Then patient-level features were generated from our weight calculation method and the survival time was finally estimated using LASSO-Cox model. The proposed EOCSA is efficient and effective in predicting prognosis of EOC and the DCAS ensures more informative and discriminative features can be extracted. As far as we know, our work is the first to analyze the survival of EOC based on WSIs and deep neural network technologies. The experimental results demonstrate that our proposed framework has achieved state-of-the-art performance of 0.980 C-index. The implementation of the approach can be found at https://github.com/RanSuLab/EOCprognosis.

PD01-10 WHOLE TRANSCRIPTOME SEQUENCING OF CLEAR CELL RENAL CELL CARCINOMA REVEALS A MULTIGENE EXPRESSION PROGNOSTIC SIGNATURE

You have accessJournal of UrologyCME1 May 2022PD01-10 WHOLE TRANSCRIPTOME SEQUENCING OF CLEAR CELL RENAL CELL CARCINOMA REVEALS A MULTIGENE EXPRESSION PROGNOSTIC SIGNATURE Srinivas Nallandhighal, Rohit Mehra, Marcin Cieslik, Amy Kasputis, Brittney Cotta, Todd Morgan, and Simpa Salami Srinivas NallandhighalSrinivas Nallandhighal More articles by this author , Rohit MehraRohit Mehra More articles by this author , Marcin CieslikMarcin Cieslik More articles by this author , Amy KasputisAmy Kasputis More articles by this author , Brittney CottaBrittney Cotta More articles by this author , Todd MorganTodd Morgan More articles by this author , and Simpa SalamiSimpa Salami More articles by this author View All Author Informationhttps://doi.org/10.1097/JU.0000000000002516.10AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: There is a critical need to develop prognostic biomarkers to improve the management of patients with clear cell renal cell carcinoma (ccRCC). This study aims to develop and validate gene expression-based biomarkers associated with recurrent disease to facilitate risk-stratification of ccRCC. METHODS: We retrospectively identified 110 patients who underwent radical nephrectomy for localized ccRCC. Patients who recurred were matched based on grade/stage to patients without recurrence. RNA next-generation sequencing (NGS) was performed on formalin-fixed paraffin-embedded (FFPE) tissue using whole-transcriptome sequencing on the illumina platform. We developed a gene signature to predict recurrence/progression-free survival (PFS) using a 15-fold lasso and elastic-net regularized linear Cox model. We derived Myriad Prolaris™ commercially available 31-gene cell cycle proliferation (mxCCP) score using RNA-seq data for each patient. Validation datasets were assembled: dataset #1 - The Cancer Genome Atlas ccRCC (TCGA, n= 382) and dataset #2 [Seishi Ogawa Japanese (n=87), International Cancer Genome Consortium (ICGC; n=81), GSE22541 (n=20) and Clinical Proteomics Tumor Analysis Consortium (CPTAC; n=91)]. Kaplan-Meier (KM) curves and multivariate Cox proportional hazard testing were then used to validate the independent prognostic impact of the gene signature on PFS and disease specific survival (DSS). RESULTS: After quality control, the training cohort comprised 50 patients with recurrence and 41 patients without, with a median follow-up 26 and 36 months, respectively. There were no significant differences between age, sex, grade, and stage between groups (all p > 0.05). We developed a 15-gene signature which was the only variable independently associated with worse PFS and DSS (PFS: HR=11.08, CI=4.9-25.1; DSS: HR=9.67, CI=3.4-27.7), adjusting for clinical-pathologic variables and mxCCP score. The 15-gene signature was also independently associated with worse PFS and DSS in both validation datasets [Validation #1 (n=382), PFS: HR=2.6, CI=1.6-4.3; DSS: HR=3; CI=1.4-6.1 and Validation #2 (n=279), PFS: HR=1.6, CI=0.7-3.6; DSS: HR=3.1; CI=1.5-6.4] adjusting for clinical-pathologic variables and mxCCP score. CONCLUSIONS: We developed and validated a novel 15-gene prognostic signature to improve risk stratification of patients with ccRCC. This signature has the potential to facilitate optimal treatment allocation and may lead to the development of novel therapeutic targets. Source of Funding: National Comprehensive Cancer Network © 2022 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 207Issue Supplement 5May 2022Page: e33 Advertisement Copyright & Permissions© 2022 by American Urological Association Education and Research, Inc.MetricsAuthor Information Srinivas Nallandhighal More articles by this author Rohit Mehra More articles by this author Marcin Cieslik More articles by this author Amy Kasputis More articles by this author Brittney Cotta More articles by this author Todd Morgan More articles by this author Simpa Salami More articles by this author Expand All Advertisement PDF DownloadLoading ...

Construction of a Comprehensive Diagnostic Scoring Model for Prostate Cancer Based on a Novel Six-Gene Panel.

Accumulating evidence indicates that the N6-methyladenosine (m6A) modification plays a critical role in human cancers. Given the current understanding of m6A modification, this process is believed to be dynamically regulated by m6A regulators. Although the discovery of m6A regulators has greatly enhanced our understanding of the mechanism underlying m6A modification in cancers, the function and role of m6A in the context of prostate cancer (PCa) remain unclear. Here, we aimed to establish a comprehensive diagnostic scoring model that can act as a complement to prostate-specific antigen (PSA) screening. To achieve this, we first drew the landscape of m6A regulators and constructed a LASSO-Cox model using three risk genes (METTL14, HNRNP2AB1, and YTHDF2). Particularly, METTL14 expression was found to be significantly related to overall survival, tumor T stage, relapse rate, and tumor microenvironment of PCa patients, showing that it has important prognostic value. Furthermore, for the sake of improving the predictive ability, we presented a comprehensive diagnostic scoring model based on a novel 6-gene panel by combining with genes found in our previous study, and its application potential was further validated by the whole TCGA and ICGC cohorts. Our study provides additional clues and insights regarding the treatment and diagnosis of PCa patients.

Bayesian hierarchical lasso Cox model: A 9-gene prognostic signature for overall survival in gastric cancer in an Asian population.

ObjectiveGastric cancer (GC) is one of the most common tumour diseases worldwide and has poor survival, especially in the Asian population. Exploration based on biomarkers would be efficient for better diagnosis, prediction, and targeted therapy.MethodsExpression profiles were downloaded from the Gene Expression Omnibus (GEO) database. Survival-related genes were identified by gene set enrichment analysis (GSEA) and univariate Cox. Then, we applied a Bayesian hierarchical lasso Cox model for prognostic signature screening. Protein-protein interaction and Spearman analysis were performed. Kaplan–Meier and receiver operating characteristic (ROC) curve analysis were applied to evaluate the prediction performance. Multivariate Cox regression was used to identify prognostic factors, and a prognostic nomogram was constructed for clinical application.ResultsWith the Bayesian lasso Cox model, a 9-gene signature included TNFRSF11A, NMNAT1, EIF5A, NOTCH3, TOR2A, E2F8, PSMA5, TPMT, and KIF11 was established to predict overall survival in GC. Protein-protein interaction analysis indicated that E2F8 was likely related to KIF11. Kaplan-Meier analysis showed a significant difference between the high-risk and low-risk groups (P<0.001). Multivariate analysis demonstrated that the 9-gene signature was an independent predictor (HR = 2.609, 95% CI 2.017–3.370), and the C-index of the integrative model reached 0.75. Function enrichment analysis for different risk groups revealed the most significant enrichment pathway/term, including pyrimidine metabolism and respiratory electron transport chain.ConclusionOur findings suggested that a novel prognostic model based on a 9-gene signature was developed to predict GC patients in high-risk and improve prediction performance. We hope our model could provide a reference for risk classification and clinical decision-making.

DNA Repair-Related Gene Signature in Predicting Prognosis of Colorectal Cancer Patients.

Background: Increasing evidence have depicted that DNA repair–related genes (DRGs) are associated with the prognosis of colorectal cancer (CRC) patients. Thus, the aim of this study was to evaluate the impact of DNA repair–related gene signature (DRGS) in predicting the prognosis of CRC patients. Method: In this study, we retrospectively analyzed the gene expression profiles from six CRC cohorts. A total of 1,768 CRC patients with complete prognostic information were divided into the training cohort (n = 566) and two validation cohorts (n = 624 and 578, respectively). The LASSO Cox model was applied to construct a prediction model. To further validate the clinical significance of the model, we also validated the model with Genomics of Drug Sensitivity in Cancer (GDSC) and an advanced clear cell renal cell carcinoma (ccRCC) immunotherapy data set. Results: We constructed a prognostic DRGS consisting of 11 different genes to stratify patients into high- and low-risk groups. Patients in the high-risk groups had significantly worse disease-free survival (DFS) than those in the low-risk groups in all cohorts [training cohort: hazard ratio (HR) = 2.40, p < 0.001, 95% confidence interval (CI) = 1.67–3.44; validation-1: HR = 2.20, p < 0.001, 95% CI = 1.38–3.49 and validation-2 cohort: HR = 2.12, p < 0.001, 95% CI = 1.40–3.21). By validating the model with GDSC, we could see that among the chemotherapeutic drugs such as oxaliplatin, 5-fluorouracil, and irinotecan, the IC50 of the cell line in the low-risk group was lower. By validating the model with the ccRCC immunotherapy data set, we can clearly see that the overall survival (OS) of the objective response rate (ORR) with complete response (CR) and partial response (PR) in the low-risk group was the best. Conclusions: DRGS is a favorable prediction model for patients with CRC, and our model can predict the response of cell lines to chemotherapeutic agents and potentially predict the response of patients to immunotherapy.

A Risk Model Based on Immune-Related Genes Predicts Prognosis and Characterizes the Immune Landscape in Esophageal Cancer.

Aberrant immune gene expression has been shown to have close correlations with the occurrence and progression of esophageal cancer (EC). We aimed to generate a prognostic signature based on immune-related genes (IRGs) capable of predicting prognosis, immune checkpoint gene (ICG) expressions, and half-inhibitory concentration (IC50) for chemotherapy agents for EC patients. Transcriptome, clinical, and mutation data on tumorous and paratumorous tissues from EC patients were collected from The Cancer Genome Atlas (TCGA) database. Then, we performed differential analysis to identify IRGs differentially expressed in EC. Their biofunctions and related pathways were explored using Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses. These gene expression profiling data were merged with survival information and subjected to univariate Cox regression to select prognostic genes, which were then included in a Lasso-Cox model for signature generation (risk score calculation). Patients were divided into the high- and low-risk groups using the median risk score as a cutoff. The accuracy of the signature in overall survival prediction was assessed, so were its performances in predicting ICG expressions and IC50 for chemotherapy and targeted therapy agents and immune cell landscape characterization. Fifteen prognostic IRGs were identified, seven of which were optimal for risk score calculation. As expected, high-risk patients had worse overall survival than low-risk individuals. Significant differences were found in tumor staging, immune cell infiltration degree, frequency of tumor mutations, tumor mutation burden (TMB), and immune checkpoint gene expressions between high- vs. low-risk patients. Further, high-risk patients exhibited high predicted IC50 for paclitaxel, cisplatin, doxorubicin, and erlotinib compared to low-risk patients. The seven-IRG-based signature can independently and accurately predict overall survival and tumor progression, characterize the tumor immune microenvironment (TIME) and estimate ICG expressions and IC50 for antitumor therapies. It shows the potential of guiding personalized treatment for EC patients.

Prognostic Value of Postoperative Circulating Tumor DNA in Patients With Early- and Intermediate-Stage Hepatocellular Carcinoma.

Majority of patients with resected early- and intermediate-stage liver cancer will experience postoperative recurrence. This study aimed to investigate the application of ctDNA sequencing in the postoperative period of hepatocellular carcinoma. A total of 96 patients with liver cancer were enrolled in this study. Postoperative peripheral blood samples were collected from all patients after surgery and analyzed using hybridization capture-based next-generation sequencing. Identification of at least one somatic mutation in the peripheral blood was defined as ctDNA+. Five genetic features in tumor tissues were associated with disease-free survival (DFS) using Lasso-Cox model. The area under the receiver operating characteristic curve was 0.813 and 0.882 in training and validation cohorts, respectively. The recurrence rate in ctDNA+ and ctDNA- groups was 60.9% and 27.8%, respectively. Multivariate Cox regression analysis showed that the postoperative ctDNA was an independent prognostic predictor of DFS (HR [hazard ratio]: 6.074, 95% Cl [confidence interval]: 2.648-13.929, P<0.001) and overall survival (OS) (HR: 4.829, 95% CI: 1.508-15.466, P=0.008). Combined ctDNA with AFP improved prediction performance. The median DFS was 2.0, and 8.0 months in ctDNA+/AFP-H and ctDNA+/AFP-L groups, respectively; while ctDNA-/AFP-H and ctDNA-/AFP-L groups had not reached the median time statistically (Log-rank test, P < 0.0001). Furthermore, ctDNA- patients had better prognosis than ctDNA+ patients irrespective of tumor stage. Postoperative ctDNA sequencing has great prognostic value in patients with liver cancer. Patients with positive ctDNA should receive more intensive disease monitoring and more aggressive treatment strategies to improve the survival time.

Predicting survival in patients with glioblastoma using MRI radiomic features extracted from radiation planning volumes.

Quantitative image analysis using pre-operative magnetic resonance imaging (MRI) has been able to predict survival in patients with glioblastoma (GBM). The study explored the role of postoperative radiation (RT) planning MRI-based radiomics to predict the outcomes, with features extracted from the gross tumor volume (GTV) and clinical target volume (CTV). Patients with IDH-wildtype GBM treated with adjuvant RT having MRI as a part of RT planning process were included in the study. 546 features were extracted from each GTV and CTV. A LASSO Cox model was applied, and internal validation was performed using leave-one-out cross-validation with overall survival as endpoint. Cross-validated time-dependent area under curve (AUC) was constructed to test the efficacy of the radiomics model, and clinical features were used to generate a combined model. Analysis was done for the entire group and in individual surgical groups-gross total excision (GTR), subtotal resection (STR), and biopsy. 235 patients were included in the study with 57, 118, and 60 in the GTR, STR, and biopsy subgroup, respectively. Using the radiomics model, binary risk groups were feasible in the entire cohort (p < 0.01) and biopsy group (p = 0.04), but not in the other two surgical groups individually. The integrated AUC (iAUC) was 0.613 for radiomics-based classification in the biopsy subgroup, which improved to 0.632 with the inclusion of clinical features. Imaging features extracted from the GTV and CTV regions can lead to risk-stratification of GBM undergoing biopsy, while the utility in other individual subgroups needs to be further explored.

A novel machine learning derived RNA-binding protein gene-based score system predicts prognosis of hepatocellular carcinoma patients.

BackgroundAlthough the expression of RNA-binding protein (RBP) genes in hepatocellular carcinoma (HCC) varies and is associated with tumor progression, there has been no overview study with multiple cohorts and large samples. The HCC-associated RBP genes need to be more accurately identified, and their clinical application value needs to be further explored.MethodsFirst, we used the robust rank aggregation (RRA) algorithm to extract HCC-associated RBP genes from nine HCC microarray datasets and verified them in The Cancer Genome Atlas Liver Hepatocellular Carcinoma (TCGA-LIHC) cohort and International Cancer Genome Consortium (ICGC) Japanese liver cancer (ICGC-LIRI-JP) cohort. In addition, the copy number variation (CNV), single-nucleotide variant (SNV), and promoter-region methylation data of HCC-associated RBP genes were analyzed. Using the random forest algorithm, we constructed an RBP gene–based prognostic score system (RBP-score). We then evaluated the ability of RBP-score to predict the prognosis of patients. The relationships between RBP-score and other clinical characteristics of patients were analyzed.ResultsThe RRA algorithm identified 30 RBP mRNAs with consistent expression patterns across the nine HCC microarray datasets. These 30 RBP genes were defined as HCC-associated RBP genes. Their mRNA expression patterns were further verified in the TCGA-LIHC and ICGC-LIRI-JP cohorts. Among these 30 RBP genes, some showed significant copy number gain or loss, while others showed differences in the methylation levels of their promoter regions. Some RBP genes were risk factors or protective factors for the prognosis of patients. We extracted 10 key HCC-associated RBP genes using the random forest algorithm and constructed an RBP-score system. RBP-score effectively predicted the overall survival (OS) and disease-free survival (DFS) of HCC patients and was associated with the tumor, node, metastasis (TNM) stage, α-fetoprotein (AFP), and metastasis risk. The clinical value of RBP-score was validated in datasets from different platforms. Cox analysis suggested that a high RBP-score was an independent risk factor for poor prognosis in HCC patients. We also successfully established a combined RBP-score+TNM LASSO-Cox model that more accurately predicted the prognosis.ConclusionThe RBP-score system constructed based on HCC-associated RBP genes is a simple and highly effective prognostic evaluation tool. It is suitable for different subgroups of HCC patients and has cross-platform characteristics. Combining RBP-score with the TNM staging system or other clinical parameters can lead to an even greater clinical benefit. In addition, the identified HCC-associated RBP genes may serve as novel targets for HCC treatment.

M6A regulator expression profile predicts the prognosis, benefit of adjuvant chemotherapy, and response to anti-PD-1 immunotherapy in patients with small-cell lung cancer

BackgroundSmall cell lung cancer (SCLC) is lethal and possesses limited therapeutic options. Platinum-based chemotherapy—with or without immune checkpoint inhibitors (anti-PDs)—is the current first-line therapy for SCLCs; however, its associated outcomes are heterogeneous. N6-methyladenosine (m6A) is a novel and decisive factor in tumour progression, chemotherapy resistance, and immunotherapy response. However, m6A modification in SCLC remains poorly understood.MethodsWe systematically explored the molecular features and clinical significance of m6A regulators in SCLC. We then constructed an m6A regulator-based prognostic signature (m6A score) based on our examination of 256 cases with limited-stage SCLC (LS-SCLC) from three different cohorts—including an independent cohort that contained 150 cases with qPCR data. We additionally evaluated the relationships between the m6A score and adjuvant chemotherapy (ACT) benefits and the patients’ responses to anti-PD-1 treatment. Immunohistochemical (IHC) staining and the HALO digital pathological platform were used to calculate CD8+ T cell density.ResultsWe observed abnormal somatic mutations and expressions of m6A regulators. Using the LASSO Cox model, a five-regulator-based (G3BP1, METTL5, ALKBH5, IGF2BP3, and RBM15B) m6A score was generated from the significant regulators to classify patients into high- and low-score groups. In the training cohort, patients with high scores had shorter overall survival (HR, 5.19; 2.75–9.77; P < 0.001). The prognostic accuracy of the m6A score was well validated in two independent cohorts (HR 4.6, P = 0.006 and HR 3.07, P < 0.001). Time-dependent ROC and C-index analyses found the m6A score to possess superior predictive power than other clinicopathological parameters. A multicentre multivariate analysis revealed the m6A score to be an independent prognostic indicator. Additionally, patients with low scores received a greater survival benefit from ACT, exhibited more CD8+ T cell infiltration, and were more responsive to cancer immunotherapy.ConclusionsOur results, for the first time, affirm the significance of m6A regulators in LS-SCLC. Our multicentre analysis found that the m6A score was a reliable prognostic tool for guiding chemotherapy and immunotherapy selections for patients with SCLC.

Abstract 14074: Machine Learning Applied on ECG Improves Cardiovascular Risk Prediction When Added to Coronary Artery Calcium Scoring in Patients With Cancer

Introduction: Patients with cancer are at increased cardiovascular (CV) risk. Both Coronary Artery Calcium (CAC) and 12-lead ECG features predict CV risk in the general population. However, the impact of these measures has not been investigated in patients with cancer. Hypothesis: We hypothesized that machine learning-based analysis of automatically-extracted ECG features, when added to CAC, can improve CV risk stratification in patients with cancer. Methods: We analyzed 1825 cancer patients with at least 1 CV risk factor who underwent CAC scoring and a standard 12-ECG as part of the CLARIFY (ClinicalTrials.gov Identifier: NCT04075162) registry. Events were defined as myocardial infarction, revascularization, stroke or death. A total of 445 ECG features were automatically extracted using a commercially available software (ECG Muse, GE healthcare). A LASSO-Cox method was used with 10-fold cross validation to select the most important ECG features with a novel ECG score derived using individual feature coefficients. We evaluated the ability of the combined ECG score with age, sex, and CAC to prognosticate a CV event in comparison to a model that was limited to age, sex, and CAC alone. The models were compared using an ANOVA test. Results: Over a median follow-up of 14 months, 224 patients had an event. Using the LASSO-Cox model, 25 ECG features were selected. The novel ECG score was associated with CV events (HR 8.6, 95% CI: 5.1-14.3, P&lt;0.001), which was not attenuated after adjusting for age, sex, and CACS (HR 7.3, 95% CI: 4.3-12.4, P&lt;0.001). Addition of ECG score to a model including age, sex, and CAC improved discrimination (C-Index 0.63 vs 0.72, P=0.0001). Conclusions: A novel ML-based ECG risk score using automatically extracted ECG features can improve CV risk stratification when added to CAC and demographics in patients with cancer.

Comprehensive analysis of N6 -methyladenosine-related long non-coding RNAs for prognosis prediction in liver hepatocellular carcinoma.

BackgroundLiver hepatocellular carcinoma (LIHC) is a lethal cancer. This study aimed to identify the N6‐methyladenosine (m6A)‐targeted long non‐coding RNA (lncRNA) related to LIHC prognosis and to develop an m6A‐targeted lncRNA model for prognosis prediction in LIHC.MethodsThe expression matrix of mRNA and lncRNA was obtained, and differentially expressed (DE) mRNAs and lncRNAs between tumor and normal samples were identified. Univariate Cox and pathway enrichment analyses were performed on the m6A‐targeted lncRNAs and the LIHC prognosis‐related m6A‐targeted lncRNAs. Prognostic analysis, immune infiltration, and gene DE analyses were performed on LIHC subgroups, which were obtained from unsupervised clustering analysis. Additionally, a multi‐factor Cox analysis was used to construct a prognostic risk model based on the lncRNAs from the LASSO Cox model. Univariate and multivariate Cox analyses were used to assess prognostic independence.ResultsA total of 5031 significant DEmRNAs and 292 significant DElncRNAs were screened, and 72 LIHC‐specific m6A‐targeted binding lncRNAs were screened. Moreover, a total of 29 LIHC prognosis‐related m6A‐targeted lncRNAs were obtained and enriched in cytoskeletal, spliceosome, and cell cycle pathways. An 11‐m6A‐lncRNA prognostic model was constructed and verified; the top 10 lncRNAs included LINC00152, RP6‐65G23.3, RP11‐620J15.3, RP11‐290F5.1, RP11‐147L13.13, RP11‐923I11.6, AC092171.4, KB‐1460A1.5, LINC00339, and RP11‐119D9.1. Additionally, the two LIHC subgroups, Cluster 1 and Cluster 2, showed significant differences in the immune microenvironment, m6A enzyme genes, and prognosis of LIHC.ConclusionThe m6A‐lncRNA prognostic model accurately and effectively predicted the prognostic survival of LIHC. Immune cells, immune checkpoints (ICs), and m6A enzyme genes could act as novel therapeutic targets for LIHC.

Predicting All-Cause Mortality Risk in Atrial Fibrillation Patients: A Novel LASSO-Cox Model Generated From a Prospective Dataset.

Background: Although mortality remains high in patients with atrial fibrillation (AF), there have been limited studies exploring machine learning (ML) models on mortality risk prediction in patients with AF.Objectives: This study sought to develop an ML model that captures important variables in order to predict all-cause mortality in AF patients.Methods: In this single center prospective study, an ML-based mortality prediction model was developed and validated using a dataset of 2,012 patients who experienced AF from November 2018 to February 2020 at the First Affiliated Hospital of Shantou University Medical College. The dataset was randomly divided into a training set (70%, n = 1,223) and a validation set (30%, n = 552). A total of 122 features were collected for variable selection. Least absolute shrinkage and selection operator (LASSO) and random forest (RF) algorithms were used for variable selection. Ten ML models were developed using variables selected by LASSO or RF. The best model was selected and compared with conventional risk scores. A nomogram and user-friendly online tool were developed to facilitate the mortality predictions and management recommendations.Results: Thirteen features were selected by the LASSO regression algorithm. The LASSO-Cox model achieved an area under the curve (AUC) of 0.842 in the training dataset, and 0.854 in the validation dataset. A nomogram based on eight independent features was developed for the prediction of survival at 30, 180, and 365 days following discharge. Both the time dependent receiver operating characteristic (ROC) and decision curve analysis (DCA) showed better performances of the nomogram compared to the CHA2DS2-VASc and HAS-BLED models.Conclusions: The LASSO-Cox mortality predictive model shows potential benefits in death risk evaluation for AF patients over the 365-day period following discharge. This novel ML approach may also provide physicians with personalized management recommendations.