Predict Survival Outcomes Research Articles

Impact of IL-8 on survival after TARE in HCC: a comprehensive investigation and external validation from the SORAMIC trial.

In the treatment of hepatocellular carcinoma (HCC) with transarterial radioembolization (TARE), identifying reliable biomarkers for predicting survival outcomes remains a critical challenge. We aimed to address this gap by investigating the significance of serum cytokines associated with inflammation as potential biomarkers for the selection of patients for TARE. Our retrospective study involved 161 patients diagnosed with HCC who underwent Y90 radioembolization at our medical center between 2010 and 2020. Serum samples from a subset of 78 patients were retrospectively analyzed to determine the concentrations of pro-inflammatory cytokines. The results from the prospective SORAMIC trial were used for independent validation. With a median overall survival of 36 weeks (range 4-436), our study showed the strongest correlation between 12-week survival and IL-8 levels before treatment (p < 0.001), while other relevant interleukins, interferon-α2, INF-γ, TNF-α and MCP-1 were not associated with survival. IL-8 levels below the cut-off of 190 pg/mL were significantly associated with increased 12-week and 24-week survival, with hazard ratios of 19.01 (95% CI: 2.29-157.89) and 2.57 (95% CI: 1.05-6.31), respectively (p = 0.006 and p = 0.039, respectively). In the adjusted multivariate analysis, the 190 pg/mL cut-off for IL-8 remained independently associated with 12- (p = 0.011) and 24-week survival (p = 0.039). Similarly, the SORAMIC population showed a strong association between IL-8 levels and 36-week survival (p = 0.03). Our study emphasizes the pivotal role of IL-8 as a valuable parameter, demonstrating its potential for predicting treatment outcomes and assessing liver function in patients with HCC undergoing TARE. The robustness of these findings warrants further validation.

Comprehensive multi-omics analysis identifies chromatin regulator-related signatures and TFF1 as a therapeutic target in lung adenocarcinoma through a 429-combination machine learning approach

IntroductionLung cancer is a leading cause of cancer-related deaths, with its incidence continuing to rise. Chromatin remodeling, a crucial process in gene expression regulation, plays a significant role in the development and progression of malignant tumors. However, the role of chromatin regulators (CRs) in lung adenocarcinoma (LUAD) remains underexplored.MethodsThis study developed a chromatin regulator-related signature (CRRS) using a 429-combination machine learning approach to predict survival outcomes in LUAD patients. The CRRS model was validated across multiple independent datasets. We also investigated the impact of CRRS on the immune microenvironment, focusing on immune cell infiltration. To identify potential therapeutic targets, TFF1, a chromatin regulator, was knocked down using siRNA in LUAD cells. We assessed its impact through apoptosis analysis, proliferation assays, and in vivo tumor growth studies. Additional validation was performed using Ki67 expression and TUNEL assays.ResultsThe CRRS accurately predicted survival outcomes and was shown to modulate immune cell infiltration in the tumor microenvironment. High-risk patients demonstrated increased activity in cell cycle regulation and DNA repair pathways, along with distinct mutation profiles and immune responses compared to low-risk patients. TFF1 emerged as a key therapeutic target. Knockdown of TFF1 significantly inhibited LUAD cell proliferation, induced apoptosis, and suppressed in vivo tumor growth. Ki67 and TUNEL assays confirmed the role of TFF1 in regulating tumor growth and cell death.DiscussionThese findings highlight the potential of chromatin regulators in prognostic modeling and immune modulation in LUAD. TFF1 was identified as a promising therapeutic target, suggesting that targeting TFF1 could provide new treatment strategies. Further research is warranted to explore its full potential and therapeutic applicability.

Prediction model for survival of younger patients with breast cancer using the breast cancer public staging database

Breast cancer (BC) is a major contributor to female mortality worldwide, particularly in young women with aggressive tumors. Despite the need for accurate prognosis in this demographic, existing studies primarily focus on broader age groups, often using the SEER database, which has limitations in variable selection. This study aimed to develop an ML-based model to predict survival outcomes in young BC patients using the BC public staging database. A total of 3,401 patients with BC were included in the study. Patients were categorized as younger (n = 1574) and older (n = 1827). We applied several survival models—Random Survival Forest, Gradient Boosting Survival, Extra Survival Trees (EST), and penalized Cox models (Lasso and ElasticNet)—to compare mortality characteristics. The EST model outperformed others in predicting mortality for both age groups. Older patients exhibited a higher prevalence of comorbidities compared to younger patients. Tumor stage was the primary variable used to train the model for mortality prediction in both groups. COPD was a significant variable only in younger patients with BC. Other variables exhibited varying degrees of consistency in each group. These findings can help identify high-risk young female patients with BC who require aggressive treatment by predicting the risk of mortality.

Identification and validation of a prognostic model based on three TLS-Related genes in oral squamous cell carcinoma

BackgroundThe tertiary lymphoid structures (TLSs) have an immunomodulatory function and have a positive impact on the survival outcomes of patients with oral squamous cell carcinoma (OSCC). However, there is a lack of standard approaches for quantifying TLSs and prognostic models using TLS-related genes (TLSRGs). These limitations limit the widespread use of TLSs in clinical practice.MethodsA convolutional neural network was used to automatically detect and quantify TLSs in HE-stained whole slide images. By employing bioinformatics and diverse statistical methods, this research created a prognostic model using TCGA cohorts and explored the connection between this model and immune infiltration. The expression levels of three TLSRGs in clinical specimens were detected by immunohistochemistry. To facilitate the assessment of individual prognostic outcomes, we further constructed a nomogram based on the risk score and other clinical factors.ResultsTLSs were found to be an independent predictor of both overall survival (OS) and disease-free survival in OSCC patients. A larger proportion of the TLS area represented a better prognosis. After analysis, we identified 69 differentially expressed TLSRGs and selected three pivotal TLSRGs to construct the risk score model. This model emerged as a standalone predictor for OS and exhibited close associations with CD4 + T cells, CD8 + T cells, and macrophages. Immunohistochemistry revealed high expression levels of CCR7 and CXCR5 in TLS + OSCC samples, while CD86 was highly expressed in TLS- OSCC samples. The nomogram demonstrates excellent predictive ability for overall survival in OSCC patients.ConclusionsThis is the first prognostic nomogram based on TLSRGs, that can effectively predict survival outcomes and contribute to individual treatment strategies for OSCC patients.

Enhancing Machine Learning Models Through PCA, SMOTE-ENN, and Stochastic Weighted Averaging

Predicting survival outcomes in critical accidents has been a focal point in machine learning research. This study addresses several limitations of existing methods, including insufficient management of data imbalance, lack of emphasis on hyperparameter tuning, and proneness to overfitting. Many existing models struggle to generalize effectively on imbalanced datasets or depend on default hyperparameter settings, resulting in biased predictions. By integrating Principal Component Analysis (PCA), hyperparameter optimization, and resampling methods, as well as combining Edited Nearest Neighbors (ENN) with the Synthetic Minority Oversampling Technique (SMOTE), the model significantly improves predictive accuracy and model generalization. An ensemble model combining seven machine learning algorithms—Logistic Regression, Support Vector Machine, KNN, Random Forest, XGBoost, LightGBM, and CatBoost—was applied to predict survival outcomes. Stochastic Weighted Averaging (SWA) was applied to mitigate overfitting and enhance generalization. The accuracy increased from 91.97% to 94.89% after SWA was applied in this specific scenario. The combination of PCA-based dimensionality reduction, hyperparameter tuning, and resampling techniques (ENN + SMOTE) ensured the model handled data imbalance and optimized predictive accuracy. The final model demonstrated excellent performance, with Area Under the Curve (AUC) and Average Precision (AP) values both reaching 0.98, indicating high accuracy and precision. These improvements were validated using the Titanic dataset in a binary classification problem of predicting passenger survival. The results emphasize that ensemble learning, enhanced by SWA, offers a powerful framework for handling imbalanced and complex datasets, providing significant advancements in predictive modeling accuracy. This study provides insights into how machine learning techniques can be effectively combined to solve classification challenges in real-world scenarios.

Primary diffuse large B-cell lymphoma of bone in adults: A SEER population-based study.

Primary diffuse large B-cell lymphoma of the bone (PB-DLBCL) is an extremely rare type of extra-nodal lymphoma. The clinical characteristics, management, and survival outcomes of adult PB-DLBCL patients remain poorly defined. To explore the clinical manifestations, staging, therapeutic options, prognostic factors and outcomes of adult patients with PB-DLBCL and to create a model to predict survival outcomes. Data of adult PB-DLBCL patients were obtained from the Surveillance, Epidemiology, and End Results (SEER) Program 18 registries database from 2000 to 2018. The Kaplan-Meier survival analysis was conducted to calculate survival rates. Univariate Cox regression, best subset selection (BESS), and least absolute shrinkage and selection operator (LASSO), followed by backward stepwise multivariable Cox regression, were used to construct the nomogram. The nomograms were evaluated using the concordance index (C-index), calibration curves and decision curve analysis (DCA). Diffuse large B-cell lymphoma (DLBCL) (67.51%) was the most frequent type of primary bone lymphoma. The most involved sites were the spine and lower-limb long bones. For the whole cohort, the 3-, 5-, 10- and 15-year overall survival (OS) rates were 74.9%, 70.5%, 60.0%, and 49.9%, and corresponding disease-specific survival (DSS) rates were 79.7%, 77.8%, 75.1%, and 71.4%, respectively. For OS, age, Ann Arbor stage, primary site and therapy were confirmed as final factors to develop the nomogram in adult PB-DLBCL patients, whereas for DSS, Age, marital status, Ann Arbor stage, number of bone lesions, therapy and year of diagnosis were confirmed as final factors in developing the nomogram. The nomograms demonstrated good accuracy and clinical utility. Established nomograms can accurately predict the survival of patients with PB-DLBCL and help clinicians optimize treatment.

Construction and interpretation of machine learning-based prognostic models for survival prediction among intestinal-type and diffuse-type gastric cancer patients

BackgroundGastric cancer is one of the most common malignant tumors worldwide, with high incidence and mortality rates, and it has a complex etiology and complex pathological features. Depending on the tumor type, gastric cancer can be classified as intestinal-type and diffuse-type gastric cancer, each with distinct pathogenic mechanisms and clinical presentations. In recent years, machine learning techniques have been widely applied in the medical field, offering new perspectives for the diagnosis, treatment, and prognosis of gastric cancer patients.MethodsThis study recruited 2158 gastric cancer patients and constructed prognostic prediction models for both intestinal-type and diffuse-type gastric cancer. Clinical pathological data were collected from patients, and machine learning algorithms were used for feature selection and model construction. The performance of the models was validated with training and testing datasets. The Shapley additive explanations (SHAP) values were used to interpret the model predictions and identify the main factors that influence patient survival.ResultsIn the prognostic model for intestinal-type gastric cancer, the gradient boosting decision tree (GBDT) model demonstrated the best performance, with key features including pTNM, CA125, tumor size, CA199, and PALB. Similarly, in the prognostic model for diffuse-type gastric cancer, the GBDT model was utilized, with key features comprising pTNM, Borrmann type IV disease, lymphocyte (LYM), lactate dehydrogenase (LDH), potassium (K), perineural invasion (PNI), tumor size, and whole stomach location. Risk stratification analysis revealed that the prognosis of high-risk patients was significantly worse than that of low-risk patients.ConclusionMachine learning shows great potential in predicting survival outcomes of gastric cancer patients, providing strong support for the development of personalized treatment plans.

Correlation analysis of Ki67 changes with survival outcomes in breast cancer before and after neoadjuvant therapy based on residual cancer Burden grade

PurposeThis study aims to investigate the change of Ki67 value pre- and post-neoadjuvant therapy (NAT) and evaluate its potential value in predicting survival outcomes in different molecular subtypes of breast cancer. MethodsA total of 257 breast cancer patients who underwent NAT at Renmin Hospital of Wuhan University from July 2019 to Sep 2023 were included in this study. The Ki67 index of the patients was re-interpreted by two attending physicians, and the changes of Ki67 value pre- and post-NAT were compared. Chi-square test (χ2) and logistic regression were conducted to examine the correlation between various characteristics and the efficacy of NAT. Disease-free survival (DFS) was calculated using the Kaplan-Meier curve and compared using the log-rank test. ResultsPatients with higher histological grade, negative expression of estrogen receptor (ER) or progesterone receptor (PR), positive expression of human epidermal growth receptor 2 (HER2), higher pretreatment Ki67 index, absence of lymph node metastasis, and those with HER2 positive and triple-negative breast cancer were associated with improved efficacy of NAT. Our study identified that the optimal cut-off value for the changes in Ki67 index pre- and post-NAT related to the effectiveness of NAT was “-88.19 %” in whole chort, which was related to the aforementioned clinical characteristics. Besides, the optimal cut-off values for the luminal, HER2-enriched and triple-negative subtypes were “-91.83 %”, “-46.12 %” and “-81.67 %”, respectively. Survival analysis demonstrated that the changes in Ki67 value were significantly associated with DFS in the HER2-enriched and triple-negative subtype, but not in the luminal subtype. ConclusionsPreoperative clinicopathological features and changes in Ki67 value pre-and post-NAT can contribute to providing patients with a more accurate prognosis.

P16 status or response to induction chemotherapy, which predicts survival outcomes in Chinese oropharyngeal cancer treated with definitive radiotherapy?

PurposeTo identify whether p16 status or response to induction chemotherapy (IC) predicts the radiotherapy (RT) response and survival outcomes in Chinese oropharyngeal squamous cell carcinoma (OPSCC). MethodsA total of 211 patients, including 128 p16-positive and 83 p16-negative were analyzed. All patients underwent IC followed by definitive RT or concurrent chemoradiotherapy (CCRT). Propensity score matching (PSM) was used to eliminate the baseline variations. ResultsAge, sex, smoking history, alcohol history, and primary site were unbalanced between different p16 status subgroups. Before PSM, the objective response rates to IC between p16-positive and p16-negative groups were 80.5 % and 85.5 % (p = 0.344). After RT, the complete response (CR) rates were 73.4 % and 66.3 %, respectively (p = 0.264). IC-sensitive (IC-s) subgroups had a higher percentage of RT-CR rate than the IC-resistant (IC-r) subgroups in both p16-positive and p16-negative patients. IC-s showed significant improvement in cancer-specific survival (CSS) (92.9 % vs. 53.6 %, p < 0.0001), progression-free survival (PFS) (p < 0.0001), locoregional relapse-free survival (LRFS) (p < 0.0001) and distant metastasis-free survival (DMFS) (p = 0.025). After PSM, the CR rates among different p16 groups remained comparable following RT (71.2 % vs. 65.8 %, p = 0.476). Before or after PSM, CSS, PFS, LRFS, and DMFS were similar between different p16 status either in IC-s or IC-r subgroups (p > 0.05). IC-r was independently associated with shorter PFS (HR = 2.661, p = 0.002) and LRFS (HR = 2.876, p = 0.002; HR = 2.78, p = 0.018). ConclusionsResponse to IC is an important predictor of prognosis in Chinese OPSCC treated with definitive RT. Poor response to IC is associated with unsatisfactory outcomes either in p16-positive or p16-negative OPSCC.

The Neutrophil Percentage-to-Albumin Ratio as a Biomarker for All-Cause and Diabetes-Cause Mortality Among Diabetes Patients: Evidence From the NHANES1988-2018.

Neutrophil percentage-to-albumin ratio (NPAR) was significantly correlated with diabetes-related complications. There are little data about NPAR and mortality risk in individuals with diabetes. This study included 3858 diabetes patients from the National Health and Nutrition Examination Survey (NHANES) conducted from 1988 to 2018. Using a restricted cubic spline (RCS), the relationship between the NPAR and mortality risk was shown. Multivariable Cox regression models were used to evaluate the relationship between the NPAR and diabetes-cause and all-cause death. An examination of the time-dependent receiver operating characteristic curve (ROC) was used to assess how well the NPAR predicted survival outcomes. Among 3858 diabetes individuals, a total of 1198 (31.1%) died over a mean follow-up of 7.86 years; of these, 326 (8.4%) had diabetes-related deaths and 872 (22.6%) had deaths from other causes. The RCS regression analysis showed a positive linear association between the NPAR and all-cause and diabetes-cause mortality. High NPAR group had a significantly higher risk of all-cause and diabetes-cause mortality in univariate and multivariate analysis. Compared with low NPAR group, high NPAR group had a low survival rate of diabetes cases in all-cause death and diabetes-cause mortality with area under the curve of the 3-, 5-, and 10-year ROC curve being 0.725, 0.739, and 0.734 for all-cause mortality and 0.754, 0.752, and 0.745 for diabetes-cause mortality, respectively. In summary, we examined 3858 diabetes patients from NHANES database (1998-2018) and suggested NPAR as a biomarker for all-cause and diabetes-cause mortality prediction.

Prognostic risk model of LIHC T-cells based on scRNA-seq and RNA-seq and the regulation of the tumor immune microenvironment

BackgroundT-cell-related genes play a crucial role in LIHC development. However, a reliable prognostic profile based on risk models of these genes has yet to be identified.MethodsSingle-cell datasets from both tumor and normal tissue samples were obtained from the GEO database. We identified T-cell marker genes and developed a genetic risk model using the TCGA-LIHC dataset, which was subsequently validated with an independent GEO dataset. We also explored the relationship between risk model predictions and immune responses.ResultsWe constructed a prognostic risk model using eight gene features identified through screening 860 T-cell marker genes via scRNA-seq and RNA-seq, which was subsequently integrated with the TCGA dataset. Its validity was independently confirmed using GEO and ICGC datasets. The TCGA dataset was stratified into high-risk and low-risk groups based on the risk model. Multivariate Cox regression analysis confirmed the risk score as an independent prognostic factor. GSEA indicated ribosomal transporter metabolism enrichment in the high-risk group and significant transcriptional activation in the low-risk group. ESTIMATE analysis showed higher ESTIMATE, immune, and stromal scores in the low-risk group, which also exhibited lower tumor purity than the high-risk group. Immunophenotyping revealed distinct patterns of immune cell infiltration and an immunosuppressive environment in the high-risk group.ConclusionsThis study introduces a T-cell marker-based prognostic risk model for LIHC patients. This model effectively predicted survival outcomes and immunotherapy effectiveness in LIHC patients, aligning with diverse immune responses and the distinct immunological profiles observed in the high-risk group.

Development and validation of a prognostic model for colon cancer based on mitotic gene signatures and immune microenvironment analysis

BackgroundMitotic processes play a pivotal role in tumor progression and immune responses. However, the correlation between mitosis-related genes, clinical outcomes, and the tumor microenvironment (TME) in colon cancer remains unclear. This study aims to develop a prognostic and therapeutic significance model for colon cancer based on mitosis-related genes.MethodsRNA expression profiles and clinical data of 453 colon cancer patients were downloaded from The Cancer Genome Atlas (TCGA). Mitosis-related genes were selected from the MsigDb database. The gene model was constructed using differential analysis, univariate and multivariate Cox regression, and Lasso regression analyses. The predictive model was validated using data from the GSE17536, GSE17537, and GSE39582 datasets. Predictive accuracy was evaluated via Receiver Operating Characteristic (ROC) curves, while nomograms were developed by integrating clinical and pathological features. Gene set enrichment analysis explored biological processes and pathways linked to the model. TME was assessed using ESTIMATE, and the proportion and function of immune cells were analyzed through CIBERSORT. Drug sensitivity analysis was conducted using the CTRP database.ResultsA predictive model based on 17 mitosis-related genes (KIFC1, CCNF, EME1, CDC25C, ORC1, CCNJL, ANKRD53, MEIS2, FZD3, TPD52L1, MAPK3, CDKN2A, EDN3, NPM2, PSRC1, INHBA, BIRC5) was created. The model exhibited robust predictive performance across both training and validation cohorts. Nomograms for predicting 3-, 5-, and 7-year survival rates in colon cancer (COAD) patients were generated. The model's correlation with immune cell infiltration and function was highlighted.ConclusionThe mitosis-related gene model serves as a valuable indicator for predicting survival outcomes in colon cancer patients.

Deciphering the prognostic and therapeutic significance of BAG1 and BAG2 for predicting distinct survival outcome and effects on liposarcoma

Liposarcoma (LPS) is the second most common kind of soft tissue sarcoma, and a heterogeneous malignant tumor derived from adipose tissue. Up to now, the prognostic value of BAG1 or BAG2 in LPS has not been defined yet. Expression profiling data of LPS patients were collected from TCGA and GEO database. Survival curves were plotted to verify the outcome differences of patients based on BAG1 or BAG2 expression. Univariate and multivariate Cox regression models were used to analyze the prognostic ability of BAG1 or BAG2. Chaperone’s regulators BAG1 and BAG2 were identified as prognostic biomarkers for LPS patients, which exhibited distinct expression patterns and survival outcome prediction performances. Patients with high BAG2 expression and/or low BAG1 expression had worse prognosis. Enrichment analysis showed that BAG1 was involved in negative regulation of TGF-β signaling. Low expression of BAG1 was associated with high abundance of regulatory T cells (Tregs). The 2-gene signature model further confirmed the improved risk assessment performance of BAG1 and BAG2: high risk patients displayed poor prognosis. BAG1 and BAG2 are supposed to be potential prognostic biomarkers for LPS and have impacts on liposarcomagenesis and immune infiltration in distinctive manners, which may function as potential therapy targets (BAG1 agonists/BAG2 inhibitors) for LPS.

Likelihood adaptively incorporated external aggregate information with uncertainty for survival data.

Population-based cancer registry databases are critical resources to bridge the information gap that results from a lack of sufficient statistical power from primary cohort data with small to moderate sample size. Although comprehensive data associated with tumor biomarkers often remain either unavailable or inconsistently measured in these registry databases, aggregate survival information sourced from these repositories has been well documented and publicly accessible. An appealing option is to integrate the aggregate survival information from the registry data with the primary cohort to enhance the evaluation of treatment impacts or prediction of survival outcomes across distinct tumor subtypes. Nevertheless, for rare types of cancer, even the sample sizes of cancer registries remain modest. The variability linked to the aggregated statistics could be non-negligible compared with the sample variation of the primary cohort. In response, we propose an externally informed likelihood approach, which facilitates the linkage between the primary cohort and external aggregate data, with consideration of the variation from aggregate information. We establish the asymptotic properties of the estimators and evaluate the finite sample performance via simulation studies. Through the application of our proposed method, we integrate data from the cohort of inflammatory breast cancer (IBC) patients at the University of Texas MD Anderson Cancer Center with aggregate survival data from the National Cancer Data Base, enabling us to appraise the effect of tri-modality treatment on survival across various tumor subtypes of IBC.

Prediction of preoperative lymph-vascular space invasion and survival outcomes of cervical squamous cell carcinoma by utilizing 18 F-FDG PET/CT imaging at early stage.

To establish nomograms for predicting preoperative lymph-vascular space invasion (LVSI) and survival outcomes of cervical squamous cell carcinoma (CSCC) based on PET/CT radiomics. One hundred and twenty-three patients with CSCC and LVSI status were enrolled retrospectively. Independent predictors of LVSI were identified through clinicopathological factors and PET/CT metabolic parameters. We extracted 1316 features from PET and CT volume of interest, respectively. Additionally, four models (PET-RS: radiomic signature of PET only; CT-RS: radiomic signature of CT only; PET/CT-RS + clinical data; PET/CT-RS: radiomic signature of PET and CT) were established to predict LVSI status. Calculation of radiomics scores of PET/CT was executed for assessment of the survival outcomes, followed by development of nomograms with radiomics (NR) or without radiomics (NWR). One hundred and twenty-three patients with pathologically confirmed CSCC had been categorized into two sets (training and testing sets). It was found that only maximum standardized uptake value (SUV max ) and squamous cell carcinoma antigen were independent predictors of LVSI. Meanwhile, the PET/CT-RS + clinical data outperformed the other three models in the training set [area under the curve (AUC): 0.91 vs. 0.861 vs. 0.81 vs. 0.814] and the testing set (AUC: 0.885 vs. 0.857 vs. 0.783 vs. 0.798). Additionally, SUV max and LVSI had been demonstrated to be independent prognostic indicators for progression-free survival and overall survival. Decision curve analysis and calibration curve indicated that NRs were superior to NWRs. The survival outcomes were assessed. PET/CT-based radiomic signature nomogram enables a new method for preoperative prediction of LVSI and survival prognosis for patients with CSCC.

Machine learning applications in breast cancer survival and therapeutic outcome prediction based on multi-omic analysis.

The high heterogeneity within and between breast cancer patients complicates treatment determination and prognosis assessment. Treatment decision-making is influenced by various factors, such as tumor subtype, histological grade, and genotype, necessitating personalized treatment strategies. Prognostic outcomes vary significantly depending on patient-specific conditions. As a critical branch of artificial intelligence, machine learning efficiently handles large datasets and automates decision-making processes. The introduction of machine learning offers new solutions for breast cancer treatment selection and prognosis assessment. In the field of cancer therapy, traditional methods for predicting treatment and survival outcomes often rely on single or few biomarkers, limiting their ability to capture the complexity of biological processes comprehensively. Machine learning analyzes patients' multi-omic data and the intricate patterns of variations during cancer initiation and progression to predict patients' survival and treatment outcomes. Consequently, it facilitates the selection of appropriate therapeutic interventions to implement early intervention and improve treatment efficacy for patients. Here, we first introduce common machine learning methods, and then elaborate on the application of machine learning in the field of survival prediction and prognosis from two aspects: evaluating survival and predicting treatment outcomes for breast cancer patients. The aim is to provide breast cancer patients with precise treatment strategies to improve therapeutic outcomes and quality of life.

Prognostic impact of body mass index on metastatic HER2-positive breast cancer survival

Background. The association between body mass index and prognosis in patients with HER2-positive metastatic breast cancer (mBC) is unclear. Purpose – the first purpose of our study was to determine whether BMI is an independent prognostic factor for progression-free survival (PFS) and overall survival (OS) in HER2-positive mBC patients. The second objective was to assess the mutual impact of baseline clinicopathological characteristics on survival outcomes. Materials and Methods. The study group included patients treated at the Sumy Regional Clinical Oncology Center. We considered the underweight patients if their BMI was &lt;18.5 kg/m2, normal weight 18.5–24.9 kg/m2, overweight 25.0–29.9 kg/m2, and obese ≥ 30 kg/m2. The following formula was used for calculations: weight/height2 (kilograms/meter2). Information about the height, weight, and other clinicopathological characteristics of the patient at the time of the start of drug therapy was taken from the primary medical documentation. Pearson’s test and Chi2 test for categorical variables were used to compare baseline clinicopathological characteristics in groups with normal, overweight, and obesity. The Kaplan-Meier method was used to establish the medians of PFS and OS. Cox regression analysis assessed the mutual impact on the survival of various clinicopathological characteristics. The statistical significance threshold was considered P ≤0.05. Results. Seventy-eight patients with HER2-positive mBC were enrolled and categorized according to their BMI in normal weight, overweight, and obese groups. BMI had a different impact on PFS and OS. Median PFS was 14.9, 11.9, and 14.2 months for normal-weight, overweight, and obese patients, respectively. There is no statistically significant difference in PFS between groups (Р = 0.110). Multivariable regression analysis confirmed no impact of BMI on PFS. Nevertheless, metastases in bones and kidneys and the nuclear grade were statistically significant determinants of PFS. Median OS was 26.4, 28.8, and 37.9 months for normal-weight, overweight, and obese patients, respectively. BMI, kidney metastases, and nuclear grade were prognostically significant determinants of OS. Conclusions. BMI is an independent prognostic factor of OS in patients with HER2-positive mBC. Obesity is associated with better OS but does not impact PFS. Metastases in bones and kidneys and the nuclear grade can predict survival outcomes.

Development and Validation of a New Prognostic Model for Predicting Survival Outcomes in Patients with Acute-on-chronic Liver Failure.

Early determination of prognosis in patients with acute-on-chronic liver failure (ACLF) is crucial for optimizing treatment options and liver allocation. This study aimed to identify risk factors associated with ACLF and to develop new prognostic models that accurately predict patient outcomes. We retrospectively selected 1,952 hospitalized patients diagnosed with ACLF between January 2010 and June 2018. This cohort was used to develop new prognostic scores, which were subsequently validated in external groups. The study included 1,386 ACLF patients and identified six independent predictors of 28-day mortality through multivariate analysis (all p < 0.05). The new score, based on a multivariate regression model, demonstrated superior predictive accuracy for both 28-day and 90-day mortalities, with Areas under the ROC curves of 0.863 and 0.853, respectively (all p < 0.05). This score can be used to stratify the risk of mortality among ACLF patients with ACLF, showing a significant difference in survival between patients categorized by the cut-off value (log-rank (Mantel-Cox) χ2 = 487.574 and 606.441, p = 0.000). Additionally, the new model exhibited good robustness in two external cohorts. This study presents a refined prognostic model, the Model for end-stage liver disease-complication score, which accurately predicts short-term mortality in ACLF patients. This model offers a new perspective and tool for improved clinical decision-making and short-term prognostic assessment in ACLF patients.

An update of the severe trauma scoring system using the Korean National Emergency Department Information System (NEDIS) database

BackgroundVarious scoring systems are utilized to assess severe trauma patients, with one of the most commonly used tools being the International Classification of Diseases Injury Severity Score (ICISS) criteria derived from the Survival Risk Ratio (SRR) calculated using diagnostic codes. This study aimed to redefine the severe trauma scoring system in Korea based on the SRR for diagnostic codes, and subsequently evaluate its performance in predicting survival outcomes for trauma patients. MethodsThis study included trauma patients who visited Level 1 and 2 emergency departments (EDs) between January 2016 and December 2019, utilizing the Korean National Emergency Department Information System (NEDIS) database. The primary outcome of this study was in-hospital mortality. The new SRR-2020 value was calculated for each of the 865 trauma diagnosis codes (Korean Standard Classification of Diseases [KCD-7] codes, 4-digit format), and the patient-specific ICISS-2020 value was derived by multiplying the corresponding SRR-2020 value based on patient diagnosis. We compared the predictive performance for in-hospital mortality between severe trauma patients with an ICISS <0.9 based on the newly developed ICISS-2020 version and those defined by the previously used ICISS-2015 version. ResultsA total of 3,841,122 patients were enrolled, with an in-hospital mortality rate of 0.5 %. Severe trauma patients with ICISS-2020 < 0.9 accounted for 5.3 % (204,897 cases) that was lower than ICISS-2015 < 0.9 accounting for 15.3 % (587,801 cases). Among the 20,619 in-hospital mortality cases, 81.4 % had ICISS-2020 < 0.9, and 88.6 % had ICISS-2015 < 0.9. When comparing predictive performance for in-hospital mortality between the two ICISS versions, ICISS-2020 showed higher accuracy (0.95), specificity (0.95), positive predictive value (PPV) (0.08), positive likelihood ratio (LR+) (16.53), and area under the receiver operating characteristic curve (AUROC) (0.96) than ICISS-2015 for accuracy (0.85), sensitivity (0.88), specificity (0.85), PPV (0.03), LR+ (5.94), and AUROC (0.94). However, regarding sensitivity, ICISS-2020 < 0.9 showed a lower value of 0.81 compared to ICISS-2015 < 0.9, which was 0.88. The negative predictive value (NPV) was 1.00 for both versions. ConclusionsThe newly developed ICISS-2020, utilizing a nationwide emergency patient database, demonstrated relatively good performance (accuracy, specificity, PPV, LR+, and AUROC) in predicting survival outcomes for patients with trauma.

MR radiomics unveils neoadjuvant chemo-responsiveness with insights into selective treatment de-intensification in HPV-positive oropharyngeal carcinoma

BackgroundAccurate prediction of neoadjuvant chemotherapy (NAC) response allows for NAC-guided personalized treatment de-intensification in HPV-positive oropharyngeal squamous cell carcinoma (OPSCC). In this study, we aimed to apply baseline MR radiomic features to predict NAC response to help select NAC-guided de-intensification candidates, and to explore biological underpinnings of response-oriented radiomics. MethodsPre-treatment MR images and clinical data of 131 patients with HPV-positive OPSCC were retrieved from Fudan University Shanghai Cancer Center. Patients were divided into training cohort (n = 47), validation cohort 1 (n = 49) from NAC response-adapted de-intensification trial (IChoice-01, NCT04012502) and real-world validation cohort 2 (n = 35). NAC prediction model using linear support vector machine (SVM) was built and validated. Subsequent nomograms combined radiomics and clinical characteristics were established to predict survival outcomes. RNA-seq and proteomic data were compared to interpret the molecular features underlying radiomic signatures with differential NAC response. FindingsFor NAC response prediction, the fusion model with both oropharyngeal and nodal signatures achieved encouraging performance to predict good responders in the training cohort (AUC 0·89, 95% CI, 0·79-0·95) and validation cohort 1 (AUC 0·71, 95% CI, 0·59-0·83). For prognosis prediction, radiomics-based nomograms exhibited satisfactory discriminative ability between low-risk and high-risk patients (PFS, C-index 0·85, 0·76 and 0·83; OS, C-index 0·79, 0·76 and 0·87, respectively) in three cohorts. Expression analysis unveiled NAC poor responders had predominantly enhanced keratinization while good responders were featured by upregulated immune response and oxidative stress. InterpretationThe MR-based radiomic models and prognostic models efficiently discriminate among patients with different NAC response and survival risk, which help candidate selection in HPV-positive OPSCC with regard to personalized treatment de-intensification.