Training Cohort Research Articles

Clinical outcome and prognostic factors of T4N0M0 colon cancer after R0 resection: A retrospective study

BACKGROUND Paradoxically, patients with T4N0M0 (stage II, no lymph node metastasis) colon cancer have a worse prognosis than those with T2N1-2M0 (stage III). However, no previous report has addressed this issue. AIM To screen prognostic risk factors for T4N0M0 colon cancer and construct a prognostic nomogram model for these patients. METHODS Two hundred patients with T4N0M0 colon cancer were treated at Tianjin Medical University General Hospital between January 2017 and December 2021, of which 112 patients were assigned to the training cohort, and the remaining 88 patients were assigned to the validation cohort. Differences between the training and validation groups were analyzed. The training cohort was subjected to multivariate analysis to select prognostic risk factors for T4N0M0 colon cancer, followed by the construction of a nomogram model. RESULTS The 3-year overall survival (OS) rates were 86.2% and 74.4% for the training and validation cohorts, respectively. Enterostomy (P = 0.000), T stage (P = 0.001), right hemicolon (P = 0.025), irregular review (P = 0.040), and carbohydrate antigen 199 (CA199) (P = 0.011) were independent risk factors of OS in patients with T4N0M0 colon cancer. A nomogram model with good concordance and accuracy was constructed. CONCLUSION Enterostomy, T stage, right hemicolon, irregular review, and CA199 were independent risk factors for OS in patients with T4N0M0 colon cancer. The nomogram model exhibited good agreement and accuracy.

Classifying Individuals With Rheumatic Conditions as Financially Insecure Using Electronic Health Record Data and Natural Language Processing: Algorithm Derivation and Validation.

We aimed to examine the feasibility of applying natural language processing (NLP) to unstructured electronic health record (EHR) documents to detect the presence of financial insecurity among patients with rheumatologic disease enrolled in an integrated care management program (iCMP). We incorporated supervised, rule-based NLP and statistical methods to identify financial insecurity among patients with rheumatic conditions enrolled in an iCMP (n = 20,395) in a multihospital EHR system. We constructed a lexicon for financial insecurity using data from available knowledge sources and then reviewed EHR notes from 538 randomly selected individuals (training cohort n = 366, validation cohort n = 172). We manually categorized records as having "definite," "possible," or "no" mention of financial insecurity. All available notes were processed using Narrative Information Linear Extraction, a rule-based version of NLP. Models were trained using the NLP features for financial insecurity using logistic, least absolute shrinkage operator (LASSO), and random forest performance characteristic and were compared with the reference standard. A total of 245,142 notes were processed from 538 individual patient records. Financial insecurity was present among 100 (27%) individuals in the training cohort and 63 (37%) in the validation cohort. The LASSO and random forest models performed identically and slightly better than logistic regression, with positive predictive values of 0.90, sensitivities of 0.29, and specificities of 0.98. The development of a context-driven lexicon used with rule-based NLP to extract data that identify financial insecurity is feasible for use and improved the capture for presence of financial insecurity with high accuracy. In the absence of a standard lexicon and construct definition for financial insecurity status, additional studies are needed to optimize the sensitivity of algorithms to categorize financial insecurity with construct validity.

A novel model incorporating quantitative contrast-enhanced ultrasound into PI-RADSv2-based nomogram detecting clinically significant prostate cancer

The diagnostic accuracy of clinically significant prostate cancer (csPCa) of Prostate Imaging Reporting and Data System version 2 (PI-RADSv2) is limited by subjectivity in result interpretation and the false positive results from certain similar anatomic structures. We aimed to establish a new model combining quantitative contrast-enhanced ultrasound, PI-RADSv2, clinical parameters to optimize the PI-RADSv2-based model. The analysis was conducted based on a data set of 151 patients from 2019 to 2022, multiple regression analysis showed that prostate specific antigen density, age, PI-RADSv2, quantitative parameters (rush time, wash-out area under the curve) were independent predictors. Based on these predictors, we established a new predictive model, the AUCs of the model were 0.910 and 0.879 in training and validation cohort, which were higher than those of PI-RADSv2-based model (0.865 and 0.821 in training and validation cohort). Net Reclassification Index analysis indicated that the new predictive model improved the classification of patients. Decision curve analysis showed that in most risk probabilities, the new predictive model improved the clinical utility of PI-RADSv2-based model. Generally, this new predictive model showed that quantitative parameters from contrast enhanced ultrasound could help to improve the diagnostic performance of PI-RADSv2 based model in detecting csPCa.

Development and interpretation of a pathomics-driven ensemble model for predicting the response to immunotherapy in gastric cancer

BackgroundOnly a subset of patients with gastric cancer experience long-term benefits from immune checkpoint inhibitors (ICIs). Currently, there is a deficiency in precise predictive biomarkers for ICI efficacy. The aim...

A combined model based on radiomics features of Sonazoid contrast-enhanced ultrasound in the Kupffer phase for the diagnosis of well-differentiated hepatocellular carcinoma and atypical focal liver lesions: a prospective, multicenter study.

The objective of this study was to develop a combined model based on radiomics features of Sonazoid contrast-enhanced ultrasound (CEUS) during the Kupffer phase and to evaluate its value in differentiating well-differentiated hepatocellular carcinoma (w-HCC) from atypical benign focal liver lesions (FLLs). A total of 116 patients with preoperatively Sonazoid-CEUS confirmed w-HCC or benign FLL were selected from a prospective multiple study on the clinical application of Sonazoid in FLLs conducted from August 2020 to March 2021. According to the randomization principle, the patients were divided into a training cohort and a test cohort in a 7:3 ratio. Seventy-nine patients were used for establishing and training the radiomics model and combined model. In comparison, 37 patients were used for validating and comparing the performance of the models. The diagnostic efficacy of the models for w-HCC and atypical benign FLLs was evaluated using ROCs curves and decision curves. A combined model nomogram was created to assess its value in reducing unnecessary biopsies. Among the patients, there were 55 cases of w-HCC and 61 cases of atypical benign FLLs, including 28 cases of early liver abscess, 16 cases of atypical hepatic hemangioma, 8 cases of hepatocellular dysplastic nodules (DN), and 9 cases of focal nodular hyperplasia (FNH). The radiomics model and combined model we established had AUCs of 0.905 and 0.951, respectively, in the training cohort, and the AUCs of the two models in the test cohort were 0.826 and 0.912, respectively. The combined model outperformed the radiomics feature model significantly. Decision curve analysis demonstrated that the combined model achieved a higher net benefit within a specific threshold probability range (0.25 to 1.00). A nomogram of the combined model was developed. The combined model based on the radiomics features of Sonazoid-CEUS in the Kupffer phase showed satisfactory performance in diagnosing w-HCC and atypical benign FLLs. It can assist clinicians in timely detecting malignant FLLs and reducing unnecessary biopsies for benign diseases.

A ferroptosis-related signature predicts the clinical diagnosis and prognosis, and associates with the immune microenvironment of lung cancer

Targeting ferroptosis-related pathway is a potential strategy for treatment of lung cancer (LC). Consequently, exploration of ferroptosis-related markers is important for treating LC. We collected LC clinical data and mRNA expression profiles from TCGA and GEO database. Ferroptosis-related genes (FRGs) were obtained through FerrDB database. Expression analysis was performed to obtain differentially expressed FRGs. Diagnostic and prognostic models were constructed based on FRGs by LASSO regression, univariate, and multivariate Cox regression analysis, respectively. External verification cohorts GSE72094 and GSE157011 were used for validation. The interrelationship between prognostic risk scores based on FRGs and the tumor immune microenvironment was analyzed. Immunocytochemistry, Western blotting, and RT-qPCR detected the FRGs level. Eighteen FRGs were used for diagnostic models, 8 FRGs were used for prognostic models. The diagnostic model distinguished well between LC and normal samples in training and validation cohorts of TCGA. The prognostic models for TCGA, GSE72094, and GSE157011 cohorts significantly confirmed lower overall survival (OS) in high-risk group, which demonstrated excellent predictive properties of the survival model. Multivariate Cox regression analysis further confirmed risk score was an independent risk factor related with OS. Immunoassays revealed that in high-risk group, a significantly higher proportion of Macrophages_M0, Neutrophils, resting Natural killer cells and activated Mast cells and the level of B7H3, CD112, CD155, B7H5, and ICOSL were increased. In conclusion, diagnostic and prognostic models provided superior diagnostic and predictive power for LC and revealed a potential link between ferroptosis and TIME.

A Predictive Model for Graft Failure in Femtosecond Laser-Assisted Penetrating Keratoplasty Among Chinese Patients: A 2-Year Study.

Graft failure is a major challenge in femtosecond laser-assisted penetrating keratoplasty (Fs-PKP). This study focuses on the development and validation of a clinical predictive model aimed at identifying the risk of graft failure in individuals undergoing Fs-PKP in China, offering a tailored approach to improve surgical outcomes. This retrospective cohort study at Nanjing First Hospital involved 238 patients and followed the TRIPOD statement. The cohort was divided into a training set (n = 166) and a validation set (n = 72) in a 7:3 ratio. It analyzed 23 predictor variables related to recipient, donor, and surgical factors, defining graft failure as "visually significant and irreversible corneal stromal edema, haze, or scarring." A comprehensive nomogram was created using univariate and multivariate Cox regression analyses and assessed by concordance index (C-index), time-dependent receiver operating characteristics (ROC) curve, calibration plots, and decision curve analysis (DCA). Five critical risk factors were identified: recipients' history of systemic autoimmune disorders, ocular trauma, prior penetrating keratoplasty (PKP) history, donors' diabetes history, and the endothelial cell density of the donor cornea. The nomogram showed a C-index of 0.72 (95% CI 0.65-0.79) in the training group and 0.66 (95% CI 0.55-0.76) in the validation group, indicating robust predictive accuracy. Time-dependent ROC curves, calibration plots, and DCA consistently validated the model's reliability, predictive power, and clinical utility across both training and validation cohorts. Our study developed and validated a model incorporating five key factors, enhancing preoperative prediction and management for Chinese patients with Fs-PKP graft failure.

A novel colorectal cancer screening framework with feature interpretability to identify high-risk populations for colonoscopy.

Risk assessment is of paramount importance for the detection and treatment of colorectal cancer. We developed and validated a feature interpretability screening framework to identify high-risk populations and recommend colonoscopy for them. We utilized a training cohort consisting of 1252605 participants who underwent colonoscopies in Shanghai from 2013 to 2015 to develop the screening framework. We incorporated Shapley additive explanation values into feature selection to provide interpretability for the framework. Two sampling methods were separately employed to mitigate potential model bias caused by class imbalance. Furthermore, we employed various machine learning algorithms to construct risk assessment models and compared their performance. We tested the screening models on an external validation cohort of 359462 samples and conducted comprehensive evaluation and statistical analysis of the validation results. The external validation results demonstrated that the models in the proposed framework achieved sensitivity over 0.734, specificity over 0.790, and area under the receiver operating characteristic curve ranging from 0.808 to 0.859. In the predictions of the best-performing model, the prevalence rates of colorectal cancer were 0.059% and 1.056% in the low- and high-risk groups, respectively. If colonoscopies were performed only on the high-risk group predicted by the model, only 14.36% of total colonoscopies would be needed to detect 74.86% of colorectal cancer cases. We developed and validated a novel framework to identify populations at high risk for colorectal cancer. Those classified as high risk should undergo colonoscopy for further diagnosis.

Development and Validation of a Fusion Model Based on Carotid Plaques and White Matter Lesion Burden Imaging Characteristics to Evaluate Ischemic Stroke Severity in Symptomatic Patients.

The diagnostic value of carotid plaque characteristics based on higher-resolution vessel wall MRI (HRVW-MRI) combined with white matter lesion (WML) burden for the risk of ischemic stroke is unclear. To combine carotid plaque features and WML burden to construct a hybrid model for evaluating ischemic stroke severity and prognosis in patients with symptomatic carotid artery stenosis. Retrospective. One hundred and ninty-three patients with least one confirmed carotid atherosclerotic stenosis ≥30% and cerebrovascular symptoms within the last 2 weeks (136 in the training cohort and 57 in the test cohort). 3.0T, T2-weighted fluid attenuated inversion recovery (T2-FLAIR) and diffusion-weighted imaging (DWI); HRVW-MRI: 3D T1-weighted variable flip angle fast spin-echo sequences (VISTA), T2-weighted VISTA, simultaneous noncontrast angiography and intraplaque hemorrhage (SNAP), and contrast-enhanced T1-VISTA. The following features of the plaques or vessel wall were assessed by three MRI readers independently: calcification (CA), intraplaque hemorrhage (IPH), lipid-rich necrotic core (LRNC), ulceration, plaque enhancement (PE), maximum vessel diameter (Max VD), maximum wall thickness (Max WT), total vessel area (TVA), lumen area (LA), plaque volume, and lumen stenosis. WMLs were graded visually and categorized as absent-to-mild WMLs (Fazekas score 0-2) or moderate-severe WMLs (Fazekas score 3-6). WML volumes were quantified using a semiautomated volumetric analysis program. Modified Rankin scores (mRS) were assessed at 90 days, following an outpatient interview, or by telephone. LASSO-logistic regression analysis was performed to construct a model. The performance of the model was evaluated using receiver operating characteristic (ROC) curve analyses, calibration curves, decision curve analyses, and clinical imaging curves. Conditional logistic regression analysis was used to explore the associations between the hybrid model-derived score and the modified Rankin Scale (mRS) score at 90 days. The model was constructed using five selected features, including IPH, plaque enhancement, ulceration, NWI, and total Fazekas score in deep WMLs (DWMLs). The hybrid model yielded an area under the curve of 0.92 (95% confidence interval [CI] 0.87-0.97) in the training cohort and 0.88 (0.80-0.96) in the test cohort. Furthermore, the hybrid model-derived score (odds ratio = 1.28; 95% CI 1.06-1.53) was independently associated with the mRS score 90 days after stroke. The hybrid model constructed using MRI plaque characteristics and WML burden has potential to be an effective noninvasive method of assessing ischemic stroke severity. The model-derived score has promising utility in judging neurological function recovery. Stage 2.

A novel nomogram and prognostic factor for metastatic soft tissue sarcoma survival

BackgroundThis study represented the inaugural effort to develop predictive survival nomograms for metastatic soft tissue sarcoma (mSTS) patients in the era of immune checkpoint inhibitors.MethodFrom the Surveillance, Epidemiology, and End Results (SEER) program database, we extracted 3078 eligible patients with mSTS between 2016 and 2022. Kaplan-Meier survival analysis, univariate and multivariable Cox analyses, and univariate and multivariable logistic analyses were conducted. Subsequently, predictive nomograms were constructed. Clinical effectiveness was validated using the area under the curve (AUC), calibration curve, and decision curve analysis (DCA) methods.ResultsWe used the SEER database to include 3078 eligible patients with mSTS between 2016 and 2022. All the eligible patients were randomly allocated in a ratio of 6:4 and stratified into a training group (n = 1846) and a validation group (n = 1232). In the multivariate Cox analysis, age, race, marital status, pathological grade, histologic subtype, surgery, and chemotherapy were identified as independent prognostic factors. These factors were used to construct the nomogram to predict the 1-, 3-, and 5-year OS of mSTS patients. The C-index for the training cohort and the validation cohort was 0.722(95% confidence interval [CI]: 0.708–0.736), and 0.716(95% CI: 0.698–0.734), respectively. The calibration curves for 1-, 3-, and 5-year OS probability demonstrated excellent calibration between the predicted and the actual survival. The AUC values of the nomogram at 1-, 3-, and 5-year were 0.785, 0.767, and 0.757 in the training cohort, 0.773, 0.754, and 0.751 in the validation cohort, respectively. Furthermore, DCA indicated the favorable clinical utility of the nomogram in both cohorts. The risk stratification system was constructed using the established nomogram, which enhanced prediction accuracy, aided clinicians in identifying high-risk patients and informing treatment decisions.ConclusionThis study marked the inaugural effort in constructing predictive survival nomograms mSTS patients in the era of immune checkpoint inhibitors. The robustly constructed nomograms, alongside actual outcomes, offered valuable insights to inform follow-up management strategies.

MRI-based intratumoral and peritumoral radiomics for preoperative prediction of glioma grade: a multicenter study

BackgroundAccurate preoperative prediction of glioma is crucial for developing individualized treatment decisions and assessing prognosis. In this study, we aimed to establish and evaluate the value of integrated models by incorporating the intratumoral and peritumoral features from conventional MRI and clinical characteristics in the prediction of glioma grade.MethodsA total of 213 glioma patients from two centers were included in the retrospective analysis, among which, 132 patients were classified as the training cohort and internal validation set, and the remaining 81 patients were zoned as the independent external testing cohort. A total of 7728 features were extracted from MRI sequences and various volumes of interest (VOIs). After feature selection, 30 radiomic models depended on five sets of machine learning classifiers, different MRI sequences, and four different combinations of predictive feature sources, including features from the intratumoral region only, features from the peritumoral edema region only, features from the fusion area including intratumoral and peritumoral edema region (VOI-fusion), and features from the intratumoral region with the addition of features from peritumoral edema region (feature-fusion), were established to select the optimal model. A nomogram based on the clinical parameter and optimal radiomic model was constructed for predicting glioma grade in clinical practice.ResultsThe intratumoral radiomic models based on contrast-enhanced T1-weighted and T2-flair sequences outperformed those based on a single MRI sequence. Moreover, the internal validation and independent external test underscored that the XGBoost machine learning classifier, incorporating features extracted from VOI-fusion, showed superior predictive efficiency in differentiating between low-grade gliomas (LGG) and high-grade gliomas (HGG), with an AUC of 0.805 in the external test. The radiomic models of VOI-fusion yielded higher prediction efficiency than those of feature-fusion. Additionally, the developed nomogram presented an optimal predictive efficacy with an AUC of 0.825 in the testing cohort.ConclusionThis study systematically investigated the effect of intratumoral and peritumoral radiomics to predict glioma grading with conventional MRI. The optimal model was the XGBoost classifier coupled radiomic model based on VOI-fusion. The radiomic models that depended on VOI-fusion outperformed those that depended on feature-fusion, suggesting that peritumoral features should be rationally utilized in radiomic studies.

Baseline 25(OH)D level is a prognostic indicator for bariatric surgery readmission: a matched retrospective cohort study

IntroductionManaging postsurgical complications is crucial in optimizing the outcomes of bariatric surgery, for which preoperative nutritional assessment is essential. In this study, we aimed to evaluate and validate the efficacy of vitamin D levels as an immunonutritional biomarker for bariatric surgery prognosis.MethodsThis matched retrospective cohort study included adult patients who underwent bariatric surgery at a tertiary medical center in China between July 2021 and June 2022. Patients with insufficient and sufficient 25(OH)D (< 30 ng/mL) were matched in a 1:1 ratio. Follow-up records of readmission at 3 months, 6 months, and 1 year were obtained to identify prognostic indicators.ResultsA matched cohort of 452 patients with a mean age of 37.14 ± 9.25 years and involving 69.47% females was enrolled. Among them, 94.25 and 5.75% underwent sleeve gastrectomy and gastric bypass, respectively. Overall, 25 patients (5.54%) were readmitted during the 1-year follow-up. The prognostic nutritional index and controlling nutritional status scores calculated from inflammatory factors did not efficiently detect malnourishment. A low 25(OH)D level (3.58 [95% CI, 1.16–11.03]) and surgery season in summer or autumn (2.68 [95% CI, 1.05–6.83]) increased the risk of 1-year readmission in both the training and validation cohorts. The area under the receiver operating characteristic curve was 0.747 (95% CI, 0.640–0.855), with a positive clinical benefit in the decision curve analyses. The relationship between 25(OH)D and 6-month readmission was U-shaped.ConclusionSerum 25(OH)D levels have prognostic significance in bariatric surgery readmission. Hence, preferable 25(OH)D levels are recommended for patients undergoing bariatric surgery.

A dosiomics model for prediction of radiation-induced acute skin toxicity in breast cancer patients: machine learning-based study for a closed bore linac

BackgroundRadiation induced acute skin toxicity (AST) is considered as a common side effect of breast radiation therapy. The goal of this study was to design dosiomics-based machine learning (ML) models for prediction of AST, to enable creating optimized treatment plans for high-risk individuals.MethodsDosiomics features extracted using Pyradiomics tool (v3.0.1), along with treatment plan-derived dose volume histograms (DVHs), and patient-specific treatment-related (PTR) data of breast cancer patients were used for modeling. Clinical scoring was done using the Common Terminology Criteria for Adverse Events (CTCAE) V4.0 criteria for skin-specific symptoms. The 52 breast cancer patients were grouped into AST 2 + (CTCAE ≥ 2) and AST 2 − (CTCAE < 2) toxicity grades to facilitate AST modeling. They were randomly divided into training (70%) and testing (30%) cohorts. Multiple prediction models were assessed through multivariate analysis, incorporating different combinations of feature groups (dosiomics, DVH, and PTR) individually and collectively. In total, seven unique combinations, along with seven classification algorithms, were considered after feature selection. The performance of each model was evaluated on the test group using the area under the receiver operating characteristic curve (AUC) and f1-score. Accuracy, precision, and recall of each model were also studied. Statistical analysis involved features differences between AST 2 − and AST 2 + groups and cutoff value calculations.ResultsResults showed that 44% of the patients developed AST 2 + after Tomotherapy. The dosiomics (DOS) model, developed using dosiomics features, exhibited a noteworthy improvement in AUC (up to 0.78), when spatial information is preserved in the dose distribution, compared to DVH features (up to 0.71). Furthermore, a baseline ML model created using only PTR features for comparison with DOS models showed the significance of dosiomics in early AST prediction. By employing the Extra Tree (ET) classifiers, the DOS + DVH + PTR model achieved a statistically significant improved performance in terms of AUC (0.83; 95% CI 0.71–0.90), accuracy (0.70), precision (0.74) and sensitivity (0.72) compared to other models.ConclusionsThis study confirmed the benefit of dosiomics-based ML in the prediction of AST. However, the combination of dosiomics, DVH, and PTR yields significant improvement in AST prediction. The results of this study provide the opportunity for timely interventions to prevent the occurrence of radiation induced AST.

A nomogram based on nutritional and inflammatory parameters to predict DMFS and identify beneficiaries of adjuvant chemotherapy in IVA-stage nasopharyngeal carcinoma

ObjectiveThis study aims to develop a nomogram integrating inflammation (NLR), Prognostic Nutritional Index (PNI), and EBV DNA (tumor burden) to achieve personalized treatment and prediction for stage IVA NPC. Furthermore, it endeavors to pinpoint specific subgroups that may derive significant benefits from S-1 adjuvant chemotherapy.MethodsA total of 834 patients diagnosed with stage IVA NPC were enrolled in this study and randomly allocated into training and validation cohorts. Multivariate Cox analyses were conducted to identify independent prognostic factors for constructing the nomogram. The predictive and clinical utility of the nomogram was assessed through measures including the AUC, calibration curve, DCA, and C-indexes. IPTW was employed to balance baseline characteristics across the population. Kaplan-Meier analysis and log-rank tests were utilized to evaluate the prognostic value.ResultsIn our study, we examined the clinical features of 557 individuals from the training cohort and 277 from the validation cohort. The median follow-up period was 50.1 and 49.7 months, respectively. For the overall cohort, the median follow-up duration was 53.8 months. The training and validation sets showed 3-year OS rates of 87.7% and 82.5%, respectively. Meanwhile, the 3-year DMFS rates were 95.9% and 84.3%, respectively. We created a nomogram that combined PNI, NRI, and EBV DNA, resulting in high prediction accuracy. Risk stratification demonstrated substantial variations in DMFS and OS between the high and low risk groups. Patients in the high-risk group benefited significantly from the IC + CCRT + S-1 treatment. In contrast, IC + CCRT demonstrated non-inferior 3-year DMFS and OS compared to IC + CCRT + S-1 in the low-risk population, indicating the possibility of reducing treatment intensity.ConclusionsIn conclusion, our nomogram integrating NLR, PNI, and EBV DNA offers precise prognostication for stage IVA NPC. S-1 adjuvant chemotherapy provides notable benefits for high-risk patients, while treatment intensity reduction may be feasible for low-risk individuals.

Cross-site validation of lung cancer diagnosis by electronic nose with deep learning: a multicenter prospective study

BackgroundAlthough electronic nose (eNose) has been intensively investigated for diagnosing lung cancer, cross-site validation remains a major obstacle to be overcome and no studies have yet been performed.MethodsPatients with lung cancer, as well as healthy control and diseased control groups, were prospectively recruited from two referral centers between 2019 and 2022. Deep learning models for detecting lung cancer with eNose breathprint were developed using training cohort from one site and then tested on cohort from the other site. Semi-Supervised Domain-Generalized (Semi-DG) Augmentation (SDA) and Noise-Shift Augmentation (NSA) methods with or without fine-tuning was applied to improve performance.ResultsIn this study, 231 participants were enrolled, comprising a training/validation cohort of 168 individuals (90 with lung cancer, 16 healthy controls, and 62 diseased controls) and a test cohort of 63 individuals (28 with lung cancer, 10 healthy controls, and 25 diseased controls). The model has satisfactory results in the validation cohort from the same hospital while directly applying the trained model to the test cohort yielded suboptimal results (AUC, 0.61, 95% CI: 0.47─0.76). The performance improved after applying data augmentation methods in the training cohort (SDA, AUC: 0.89 [0.81─0.97]; NSA, AUC:0.90 [0.89─1.00]). Additionally, after applying fine-tuning methods, the performance further improved (SDA plus fine-tuning, AUC:0.95 [0.89─1.00]; NSA plus fine-tuning, AUC:0.95 [0.90─1.00]).ConclusionOur study revealed that deep learning models developed for eNose breathprint can achieve cross-site validation with data augmentation and fine-tuning. Accordingly, eNose breathprints emerge as a convenient, non-invasive, and potentially generalizable solution for lung cancer detection.Clinical trial registrationThis study is not a clinical trial and was therefore not registered.

Unveiling the molecular complexity of proliferative diabetic retinopathy through scRNA-seq, AlphaFold 2, and machine learning

BackgroundProliferative diabetic retinopathy (PDR), a major cause of blindness, is characterized by complex pathogenesis. This study integrates single-cell RNA sequencing (scRNA-seq), Non-negative Matrix Factorization (NMF), machine learning, and AlphaFold 2 methods to explore the molecular level of PDR.MethodsWe analyzed scRNA-seq data from PDR patients and healthy controls to identify distinct cellular subtypes and gene expression patterns. NMF was used to define specific transcriptional programs in PDR. The oxidative stress-related genes (ORGs) identified within Meta-Program 1 were utilized to construct a predictive model using twelve machine learning algorithms. Furthermore, we employed AlphaFold 2 for the prediction of protein structures, complementing this with molecular docking to validate the structural foundation of potential therapeutic targets. We also analyzed protein−protein interaction (PPI) networks and the interplay among key ORGs.ResultsOur scRNA-seq analysis revealed five major cell types and 14 subcell types in PDR patients, with significant differences in gene expression compared to those in controls. We identified three key meta-programs underscoring the role of microglia in the pathogenesis of PDR. Three critical ORGs (ALKBH1, PSIP1, and ATP13A2) were identified, with the best-performing predictive model demonstrating high accuracy (AUC of 0.989 in the training cohort and 0.833 in the validation cohort). Moreover, AlphaFold 2 predictions combined with molecular docking revealed that resveratrol has a strong affinity for ALKBH1, indicating its potential as a targeted therapeutic agent. PPI network analysis, revealed a complex network of interactions among the hub ORGs and other genes, suggesting a collective role in PDR pathogenesis.ConclusionThis study provides insights into the cellular and molecular aspects of PDR, identifying potential biomarkers and therapeutic targets using advanced technological approaches.

Nomogram built based on machine learning to predict recurrence in early-stage hepatocellular carcinoma patients treated with ablation

IntroductionTo analyze the risk factors affecting recurrence in early-stage hepatocellular carcinoma (HCC) patients treated with ablation and then establish a nomogram to provide a clear and accessible representation of the patients’ recurrence risk.MethodsCollect demographic and clinical data of 898 early-stage HCC patients who underwent ablation treatment at Beijing You’an Hospital, affiliated with Capital Medical University from January 2014 to December 2022. Patients admitted from 2014 to 2018 were included in the training cohort, while 2019 to 2022 were in the validation cohort. Lasso and Cox regression was used to screen independent risk factors for HCC patients recurrence, and a nomogram was then constructed based on the screened factors.ResultsAge, gender, Barcelona Clinic Liver Cancer (BCLC) stage, tumor size, globulin (Glob) and γ-glutamyl transpeptidase (γ-GT) were finally incorporated in the nomogram for predicting the recurrence-free survival (RFS) of patients. We further confirmed that the nomogram has optimal discrimination, consistency and clinical utility by the C-index, Receiver Operating Characteristic Curve (ROC), calibration curve and Decision Curve Analysis (DCA). Moreover, we divided the patients into different risk groups and found that the nomogram can effectively identify the high recurrence risk patients by the Kaplan-Meier curves.ConclusionThis study developed a nomogram using Lasso-Cox regression to predict RFS in early-stage HCC patients following ablation, aiding clinicians in identifying high-risk groups for personalized follow-up treatments.

Identification of vitamin D-related signature for predicting the clinical outcome and immunotherapy response in hepatocellular carcinoma.

Hepatocellular carcinoma (HCC) is one of the most common cancers globally, seriously endangering people health. Vitamin D was significantly associated with tumor progression and patients' prognosis. Integrative 10 machine learning algorithms were used to develop a Vitamin D-related signature (VRS) with one training cohort and 3 testing cohorts. The performance of VRS in predicting the immunology response was verified using several predicting approaches. The optimal VRS was constructed by stepCox + superPC algorithm. VRS acted as a risk factor for HCC patients. HCC patients with high-risk score had a poor clinical outcome and the AUCs of 1-, 3-, and 5-year ROC were 0.786, 0.755, and 0.786, respectively. A higher level of CD8 + cytotoxic T cells and B cells was obtained in HCC patients with low-risk score. There is higher PD1&CTLA4 immunophenoscore and TMB score in low-risk score in HCC patients. Lower TIDE score and tumor escape score was found in HCC cases with low-risk score. The IC50 value of camptothecin, docetaxel, crizotinib, dasatinib, and erlotinib was lower in HCC cases with high-risk score. HCC patients with high-risk score had a higher score of cancer-related hallmarks, including angiogenesis, glycolysis, and NOTCH signaling. Our study proposed a novel VRS for HCC, which served as an indicator for predicting clinical outcome and immunotherapy responses in HCC.

Nomogram for predicting post-therapy recurrence in BCLC A/B hepatocellular carcinoma with Child-Pugh B cirrhosis

IntroductionThis study conducts a retrospective analysis on patients with BCLC stage A/B hepatocellular carcinoma (HCC) accompanied by Child-Pugh B cirrhosis, who underwent transarterial chemoembolization (TACE) in combination with local ablation therapy. Our goal was to uncover risk factors contributing to post-treatment recurrence and to develop and validate an innovative 1-, 3-, and 5-year recurrence free survival (RFS) nomogram.MethodsData from 255 BCLC A/B HCC patients with Child-Pugh B cirrhosis treated at Beijing You’an Hospital (January 2014 - January 2020) were analyzed using random survival forest (RSF), LASSO regression, and multivariate Cox regression to identify independent risk factors for RFS. The prognostic nomogram was then constructed and validated, categorizing patients into low, intermediate, and high-risk groups, with RFS assessed using Kaplan-Meier curves.ResultsThe nomogram, integrating the albumin/globulin ratio, gender, tumor number, and size, showcased robust predictive performance. Harrell’s concordance index (C-index) values for the training and validation cohorts were 0.744 (95% CI: 0.703–0.785) and 0.724 (95% CI: 0.644–0.804), respectively. The area under the curve (AUC) values for 1-, 3-, and 5-year RFS in the two cohorts were also promising. Calibration curves highlighted the nomogram’s reliability and decision curve analysis (DCA) confirmed its practical clinical benefits. Through meticulous patient stratification, we also revealed the nomogram’s efficacy in distinguishing varying recurrence risks.ConclusionThis study advances recurrence prediction in BCLC A/B HCC patients with Child-Pugh B cirrhosis following TACE combined with ablation. The established nomogram accurately predicts 1-, 3-, and 5-year RFS, facilitating timely identification of high-risk populations.

Machine learning and multi-omics data reveal driver gene-based molecular subtypes in hepatocellular carcinoma for precision treatment.

The heterogeneity of Hepatocellular Carcinoma (HCC) poses a barrier to effective treatment. Stratifying highly heterogeneous HCC into molecular subtypes with similar features is crucial for personalized anti-tumor therapies. Although driver genes play pivotal roles in cancer progression, their potential in HCC subtyping has been largely overlooked. This study aims to utilize driver genes to construct HCC subtype models and unravel their molecular mechanisms. Utilizing a novel computational framework, we expanded the initially identified 96 driver genes to 1192 based on mutational aspects and an additional 233 considering driver dysregulation. These genes were subsequently employed as stratification markers for further analyses. A novel multi-omics subtype classification algorithm was developed, leveraging mutation and expression data of the identified stratification genes. This algorithm successfully categorized HCC into two distinct subtypes, CLASS A and CLASS B, demonstrating significant differences in survival outcomes. Integrating multi-omics and single-cell data unveiled substantial distinctions between these subtypes regarding transcriptomics, mutations, copy number variations, and epigenomics. Moreover, our prognostic model exhibited excellent predictive performance in training and external validation cohorts. Finally, a 10-gene classification model for these subtypes identified TTK as a promising therapeutic target with robust classification capabilities. This comprehensive study provides a novel perspective on HCC stratification, offering crucial insights for a deeper understanding of its pathogenesis and the development of promising treatment strategies.