Time-dependent Area Research Articles

Random survival forest algorithm for risk stratification and survival prediction in gastric neuroendocrine neoplasms.

This study aimed to construct and assess a machine-learning algorithm designed to forecast survival rates and risk stratification for patients with gastric neuroendocrine neoplasms (gNENs) after diagnosis. Data on patients with gNENs were extracted and randomly divided into training and validation sets using the Surveillance, Epidemiology, and End Results database. We developed a prediction model using 10 machine learning algorithms across 101 combinations to forecast cancer-related mortality in patients with gNENs, selecting the best model using the highest mean over a sequence of time-dependent area under the receiver operating characteristic (ROC) curve (AUC). The performance of the final model was assessed through time-dependent ROC curves for discrimination and calibration curves for calibration. The maximum selection rank method was used to determine the best prognostic risk score threshold for classifying patients into high- and low-risk groups. Afterward, Kaplan-Meier analysis and log-rank test were used to compare survival rates among these groups. Our study examined 775 patients with gNENs, dividing them into training and validation sets. A training set comprised 543 patients, with a median follow-up of 42 months and cumulative mortality rates of 40.0% at 1 year, 48.6% at 3 years, and 54.0% at 5 years. A validation set comprised 232 patients, with cumulative mortality rates of 29.1% at 1 year, 43.5% at 3 years, and 53.2% at 5 years. The optimal random survival forest (RSF) model (mtry = 4, node size = 5) achieved an AUC of 0.839 for survival prediction in the training set. Comprising 11 variables such as demographics, treatment details, tumor characteristics, T staging, N staging, and M staging, the RSF model revealed high predictive accuracy with AUCs of 0.92, 0.96, and 0.96 for 1-, 3-, and 5-year survival, respectively, which was consistently reflected in the validation set with AUCs of 0.88, 0.92, and 0.89, respectively. Moreover, patients were risk-stratified. Although our RSF model effectively stratified patients into different prognostic groups, it needs external validation to confirm its utility for noninvasive prognostic prediction and risk stratification in gNENs. Further research is required to verify its broader clinical applicability.

A preliminary study of developing an MRI-based model for postoperative recurrence prediction and treatment direction of intrahepatic cholangiocarcinoma.

To establish an MRI-based predictive model for postoperative recurrence in intrahepatic cholangiocarcinoma (iCCA) and further to evaluate the model utility in treatment direction for neoadjuvant and adjuvant therapies. Totally 114 iCCA patients with curative surgery were retrospectively included, including 38 patients in the neoadjuvant treatment, traditional surgery, and adjuvant treatment groups for each. Predictive variables associated with postoperative recurrence were identified using univariate and multivariate Cox regression analyses, and a prognostic model was formulated. Recurrence-free survival (RFS) curves were compared using log-rank test between MRI-predicted high-risk and low-risk iCCAs stratified by the optimal threshold. Tumor multiplicity (hazard ratio (HR) = 1.671 [95%CI 1.036, 2.695], P = 0.035), hemorrhage (HR = 2.391 [95%CI 1.189, 4.810], P = 0.015), peri-tumor diffusion-weighted hyperintensity (HR = 1.723 [95%CI 1.085, 2.734], P = 0.021), and positive regional lymph node (HR = 2.175 [95%CI 1.295, 3.653], P = 0.003) were independently associated with postoperative recurrence; treatment group was not significantly related to recurrence (P > 0.05). Independent variables above were incorporated for the recurrence prediction model, the 1-year and 2-year time-dependent area under the curve values were 0.723 (95%CI 0.631, 0.815) and 0.725 (95%CI 0.634, 0.816), respectively. After risk stratification, the MRI-predicted high-risk iCCA patients had higher cumulative incidences of recurrence and worse RFS than the low-risk patients (P < 0.001 for both). In the MRI-predicted high-risk patients, neoadjuvant therapy was associated with improved all-stage RFS (P = 0.034), and adjuvant therapy was associated with improved RFS after 4months (P = 0.014). The MRI-based iCCA recurrence predictive model may serve as a decision-making tool for both personalized prognostication and therapy selection.

Coronary artery calcium scans powered by artificial intelligence (AI-CAC) predicts atrial fibrillation and stroke comparably to cardiac magnetic resonance imaging: MESA

Abstract Background AI-CAC provides more actionable information than the Agatston coronary artery calcium (CAC) score. We have recently shown in the Multi-Ethnic Study of Atherosclerosis (MESA) that AI-CAC automated left atrial (LA) volumetry enabled prediction of atrial fibrillation (AF) as early as one year. In this study we evaluated the performance of AI-CAC LA volumetry versus LA measured by human experts using cardiac magnetic resonance imaging (CMRI) for predicting AF and stroke, and compared them with CHARGE-AF risk score, Agatston score, and NT-proBNP. Methods We used 15-year outcomes data from 3552 asymptomatic individuals (52.2% women, age 61.7±10.2 years) who underwent both CAC scans and CMRI in the MESA baseline examination. CMRI LA volume was previously measured by human experts. Data on BNP, CHARGE-AF risk score and the Agatston score were obtained from MESA. Discrimination was assessed using the time-dependent area under the curve (AUC). Results Over 15 years follow-up, 562 cases of AF and 140 cases of stroke accrued. The AUC for AI-CAC versus CMRI for AF and stroke were not significantly different (0.802 vs. 0.798 and 0.762 vs. 0.751 respectively, p=0.60). AI-CAC significantly improved the continuous Net Reclassification Index (NRI) for prediction of AF and stroke when added to CHARGE-AF risk score (0.28, 0.21), NT-proBNP (0.43, 0.37), and Agatston score (0.69, 0.41) respectively (p for all&amp;lt;0.0001). Conclusion AI-CAC automated LA volumetry and CMRI LA volume measured by human experts similarly predicted incident AF and stroke over 15 years. Further studies to investigate the clinical utility of AI-CAC for AF and stroke prediction are warranted.

Long-term outcomes after transcatheter aortic valve implantation: derivation and validation of the long-TAVI prediction model

Abstract Background Identification of high-risk individuals after transcatheter aortic valve implantation (TAVI) is essential to personalized care. Established risk models require a plethora of variables, fail to account for biomarker dynamics and imaging parameters, mostly focus on the short term, and offer limited predictive performance. Purpose We aimed to develop an easily applicable prediction model for long-term mortality or heart failure (HF) admission following TAVI. Methods Patients undergoing TAVI at a high-volume university centre in the United Kingdom (n = 990) were recruited into a prospective cohort study. Enrolled patients were divided into a development cohort (07/2007 – 12/2016) and a validation cohort (01/2017 – 11/2018) and followed for up to 15.1 (median 5.3, interquartile range [IQR] 2.8 – 7.4) years. Patient characteristics along with pre- and post-procedural imaging results and brain natriuretic peptide (BNP) levels before the procedure and at 3 months were recorded. A clinical risk score for long-term mortality and HF admission (i.e., Long-TAVI risk score) was derived using Cox regression and evaluated using time-dependent area under the receiver operating characteristics curve (AUC). Results Age at presentation was 79.8 (standard deviation [SD] 8.9) years, 961 (99.0%) patients had signs of heart failure, left ventricular ejection fraction was 52 (SD 11.2), and pre-TAVI BNP levels were 298.0 (interquartile range [IQR] 159.0 – 571.5) pg/mL. High BNP levels at presentation (higher vs. lower than 400 pg/mL: HR 1.50, 95% CI 1.17 – 1.92, P &amp;lt; 0.001) and dynamic changes in BNP with the first 3 months (low/high vs. low/low: 2.61, 95% CI 1.17 – 5.86, P = 0.020; high/high vs. low/low: 2.41, 95% CI 1.26 – 4.64, P = 0.008) were independent predictors mortality or HF admission. The simple 8-item Long-TAVI score accounts for periprocedural BNP dynamics and shows good discrimination (AUC 0.76, 95% CI 0.72 – 0.79) and calibration in the development cohort performing similarly well in the external validation cohort (AUC 0.70, 95% CI 0.63 – 0.77). Long-TAVI significantly outperformed the logistic EURO score (AUC 0.56, 95% CI 0.52 – 0.59, P &amp;lt; 0.001). Conclusion Both pre-TAVI BNP levels and periprocedural BNP trajectories are independently associated with long-term outcomes after TAVI. Accounting for clinical characteristics, imaging parameters, and circulating BNP levels, the newly developed Long-TAVI risk score achieves high performance and provides the first prediction model for long-term mortality or HF admission. Long-TAVI represents a novel tool for personalized risk assessment in patients undergoing TAVI.Correlation at baseline A) and FUP B)Score performance A) and risk groups B)

AI-enabled cardiac chambers volumetry in coronary calcium scans (AI-CAC) predicts heart failure and outperforms NT-proBNP: The Multi-Ethnic Study of Atherosclerosis

Abstract Introduction Coronary artery calcium (CAC) scans contain useful information beyond the Agatston CAC score that is not currently reported. We recently reported that artificial intelligence (AI)-enabled cardiac chambers volumetry in CAC scans (AI-CAC) predicted incident atrial fibrillation in the Multi-Ethnic Study of Atherosclerosis (MESA). In this study, we investigated the performance of AI-CAC for prediction of incident heart failure (HF) and compared it with 10 known clinical risk factors, NT-proBNP, and the Agatston CAC score. Methods We applied AI-CAC to 5750 CAC scans of asymptomatic individuals (52% women, White 40%, Black 26%, Hispanic 22% Chinese 12%) free of known cardiovascular disease at the MESA baseline examination (2000-2002). We used the 15-year outcomes data and compared the time-dependent area under the curve (AUC) of AI-CAC versus NT-proBNP, Agatston score, and 10 known clinical risk factors (age, gender, diabetes, current smoking, hypertension medication, systolic and diastolic blood pressure, LDL, HDL, total cholesterol, and hs-CRP) for predicting incident HF over 15 years. Results Over 15 years of follow-up, 256 HF events accrued. The AUC for predicting HF with AI-CAC (0.826) was significantly higher than NT-proBNP (0.742) and Agatston score (0.712) (p&amp;lt;.0001), and comparable to clinical risk factors (0.818, p=0.4141). AI-CAC category-free Net Reclassification Index (NRI) significantly improved on clinical risk factors (0.32), NT-proBNP (0.46), and Agatston score (0.71) for HF prediction at 15 years (p&amp;lt;0.0001). Conclusion AI-CAC significantly outperformed NT-proBNP and the Agatston CAC score, and significantly improved the AUC and category-free NRI of clinical risk factors for incident HF prediction.

Transcranial doppler to detect high-risk PFO and predict decompression illness in divers

Abstract Background It has not yet been clearly demonstrated whether it is feasible to diagnose patent foramen ovale (PFO) using transcranial doppler (TCD) and to use it in strategies to prevent decompression illness (DCI) in self-contained breathing apparatus (SCUBA) divers. Purpose DIVER-PFO study confirmed that high-risk PFO is associated with an increased risk of DCI in SCUBA divers. In this prespecified post-hoc analysis, we aimed to prospectively evaluate the accuracy of TCD to diagnose high-risk PFO and to predict the occurrence of DCI. Methods Eighty-two consecutive divers (mean age; 41 years, male 80.5%) underwent transesophageal echocardiogram (TEE) and TCD with a saline bubble test to determine the presence of PFO and were subsequently divided into high-risk and low-risk groups by TEE findings. Divers in the high-risk PFO group included those with the following conditions: PFO with atrial septal aneurysm, hypermobility, PFO size ≥2 mm, or right to left shunt at resting state. The enrolled divers were then followed-up using a self-reported questionnaire through e-mail or telephone interview for 3 years, while they were blinded to their PFO status. The primary endpoint of the study was PFO-related DCI, comprising cases of cutaneous or neurologic DCI, DCI symptoms requiring hyperbaric chamber therapy, DCI symptoms within 30 minutes after surfacing, or unexplained death occurring within 2 weeks after diving. Results PFO was detected in 47 (57.3%) and 59 (72.0%) divers by TCD and TEE, respectively. Diagnostic accuracy, sensitivity, specificity, positive predictive value, and negative predictive value of TCD were 68.3%, 67.8%, 69.6% 85.1%, and 45.7% for the diagnosis of PFO, and 65.9%, 82.8%, 56.6%, 51.1% and 85.7% for the diagnosis of high-risk PFO, respectively. PFO-related DCI occurred in 9 divers in the PFO group (PFO vs non-PFO: 6.4 vs 0.7 incidences per 10,000 person dives; p=0.03) during a mean follow-up of 25.6 months. Integrated time-dependent Area Under the Curves for TCD and TEE to predict DCI occurrence using Cox’s proportional hazards regression model were 0.714 and 0.675, respectively. Conclusions TCD was moderately accurate to diagnose high-risk PFO in divers. However, it predicted DCI as much as TEE.Study flowIncidence of DCI in the PFO vs No PFO

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

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

Predicting the survival of patients with painful tumours treated with palliative radiotherapy: a secondary analysis using the 3-variable number-of-risk-factors model

BackgroundThe 3-variable number-of-risk-factors (NRF) model is a prognostic tool for patients undergoing palliative radiotherapy (PRT). However, there is little research on the NRF model for patients with painful non-bone-metastasis tumours treated with PRT, and the efficacy of the NRF model in predicting survival is unclear to date. Therefore, we aimed to assess the prognostic accuracy of a 3-variable NRF model in patients undergoing PRT for bone and non- bone-metastasis tumours.MethodsThis was a secondary analysis of studies on PRT for bone-metastasis (BM) and PRT for miscellaneous painful tumours (MPTs), including non-BM tumours. Patients were grouped in the NRF model and survival was compared between groups. Discrimination was evaluated using a time-independent C-index and a time-dependent area under the receiver operating characteristic curve (AUROC). A calibration curve was used to assess the agreement between predicted and observed survival.ResultsWe analysed 485 patients in the BM group and 302 patients in the MPT group. The median survival times in the BM group for groups I, II, and III were 35.1, 10.1, and 3.3 months, respectively (P < 0.001), while in the MPT group, they were 22.1, 9.5, and 4.6 months, respectively (P < 0.001). The C-index was 0.689 in the BM group and 0.625 in the MPT group. In the BM group, time-dependent AUROCs over 2 to 24 months ranged from 0.738 to 0.765, while in the MPT group, they ranged from 0.650 to 0.689, with both groups showing consistent accuracy over time. The calibration curve showed a reasonable agreement between the predicted and observed survival.ConclusionsThe NRF model predicted survival moderately well in both the BM and MPT groups.

Prognostic value of visceral protein ratios in patients with colon cancer

This study aimed to assess different combinations of visceral proteins and to elucidate their value in predicting progression-free survival (PFS) and overall survival (OS) in patients with colon cancer. The visceral protein ratios included the albumin-globulin ratio (AGR), prealbumin-globulin ratio (PGR), and albumin-prealbumin-globulin ratio (APGR). Compared with AGR and PGR, APGR had the best time-dependent area under the receiver operating characteristic curves for predicting the outcomes. High AGR/PGR/APGR levels were associated with an increased risk of mortality. High AGR (HR=0.816, 95%CI: 0.719-0.925, p=0.001), PGR (HR=0.831, 95%CI: 0.724-0.953, p=0.008), and APGR (HR=0.789, 95%CI: 0.688-0.904, p<0.001) were independent risk factors for PFS. For every SD increase in AGR, PGR, and APGR, the risk of poor OS in patients with colon cancer was reduced by 16.9% (HR=0.831, 95%CI, 0.733–0.943; p=0.001), 15.1% (HR=0.849, 95%CI, 0.739–0.976; p=0.021), and 19.1% (HR = 0.809, 95%CI, 0.705–0.928; p=0.002), respectively. Logistic regression models showed that AGR, PGR, and APGR were independent factors that affected recurrence. Visceral protein ratios are independent predictors of PFS and OS. Compared to the existing visceral protein ratios (AGR and PGR), APGR is a more accurate and sensitive indicator for predicting the outcomes of patients with colon cancer.

Prognostic Accuracy of Transient Elastography-Based Predictors in Diabetes and Obesity: A Multicenter International Cohort Study.

Liver stiffness measurement (LSM) by vibration-controlled transient elastography (VCTE) is recommended for risk stratification of patients with nonalcoholic fatty liver disease (NAFLD). More recently, AGILE3 + and AGILE4 have combined LSM with clinical parameters to identify patients with advanced fibrosis and cirrhosis, respectively. However, there are limited data on prognostic performance of these scores in key at-risk subgroups such as those with diabetes and obesity compared to LSM alone. This is a retrospective cohort study including 1903 adult patients with NAFLD from tertiary care centers in the United States and Singapore undergoing VCTE between 2015 and 2022. Primary predictors were FAST, LSM, AGILE3 + , and AGILE4 scores and the primary outcome was liver-related events (LRE). Patients were further stratified by diabetes and obesity status. Prognostic performance was measured using the time-dependent area under the receiver operating characteristic curve (tAUC) at 5years. In total, 25 LRE occurred and the overall incidence rate of LRE was 4.4 per 1000 person-years. tAUC for predicting LRE in the overall group was significantly higher with AGILE3 + (0.94 [95% CI: 0.90-0.98]) and AGILE4 (0.94 [95% CI: 0.90-0.98]) compared to LSM (0.87 [95% CI: 0.80-0.94]) (p = 0.001 and 0.009, respectively) and FAST (0.73 [95% CI: 0.59-0.86]) (p < 0.001 for both). Similarly, tAUC was significantly higher in those with T2D for AGILE3 + compared to LSM (0.92 vs 0.86, respectively) (p = 0.015) and FAST (0.92 vs 0.73, respectively) (p = 0.008). Among people with obesity, tAUC was significantly higher for AGILE3 + compared to LSM (0.95 vs 0.89, respectively) (p = 0.005) and FAST (0.95 vs 0.76, respectively) (p = 0.0035). Though AGILE4 had a higher tAUC in these subgroups compared to LSM, it did not reach statistical significance. AGILE3 + significantly outperforms LSM and FAST for predicting LRE in patients with NAFLD including in those with diabetes or obesity.

Identifying Proteomic Prognostic Markers for Alzheimer's Disease with Survival Machine Learning: the Framingham Heart Study.

Protein abundance levels, sensitive to both physiological changes and external interventions, are useful for assessing the Alzheimer's disease (AD) risk and treatment efficacy. However, identifying proteomic prognostic markers for AD is challenging by their high dimensionality and inherent correlations. Our study analyzed 1128 plasma proteins, measured by the SOMAscan platform, from 858 participants 55 years and older (mean age 63 years, 52.9% women) of the Framingham Heart Study (FHS) Offspring cohort. We conducted regression analysis and machine learning models, including LASSO-based Cox proportional hazard regression model (LASSO) and generalized boosted regression model (GBM), to identify protein prognostic markers. These markers were used to construct a weighted proteomic composite score, the AD prediction performance of which was assessed using time-dependent area under the curve (AUC). The association between the composite score and memory domain was examined in 339 (of the 858) participants with available memory scores, and in an independent group of 430 participants younger than 55 years (mean age 46, 56.7% women). Over a mean follow-up of 20 years, 132 (15.4%) participants developed AD. After adjusting for baseline age, sex, education, and APOE ε4+ status, regression models identified 309 proteins (P ≤ 0.2). After applying machine learning methods, nine of these proteins were selected to develop a composite score. This score improved AD prediction beyond the factors of age, sex, education, and APOE ε4+ status across 15 to 25 years of follow-up, achieving its peak AUC of 0.84 in the LASSO model at the 22-year follow-up. It also showed a consistent negative association with memory scores in 339 participants (beta = -0.061, P = 0.046), 430 independent participants (beta = -0.060, P = 0.018), and the pooled 769 samples (beta = -0.058, P = 0.003). These findings highlight the utility of proteomic markers in improving AD prediction and emphasize the complex pathology of AD. The composite score may aid early AD detection and efficacy monitoring, warranting further validation in diverse populations.

Exploring clinical factors to predict the survival of patients with resectable non-small cell lung cancer with neoadjuvant immunotherapy.

The goal was to explore clinical factors and build a predictive model for the disease-free and overall survival of patients with non-small cell lung cancer (NSCLC) receiving neoadjuvant chemotherapy combined with immune checkpoint inhibitors. Inclusion criteria for patients in this multicentre study were as follows: (i) Patients who were diagnosed with stages I-III NSCLC after a bronchoscopy biopsy or puncture; (ii) patients who were examined with computed tomography/positron emission tomography-computed tomography before treatment and surgery; (iii) patients who received neoadjuvant chemotherapy combined with immune checkpoint inhibitors for 2 to 6 cycles preoperatively; (iv) patients whose peripheral blood indicators and tumour markers were assessed before treatment and preoperatively; (v) patients who underwent radical lung cancer surgery after neoadjuvant therapy. Cases were divided into high- and low-risk groups according to 78 clinical indicators based on a 10-fold Least Absolute Shrinkage and Selection Operator selection. We used Cox proportional hazards models to predict disease-free and overall survival. Then, we used time-dependent area under the curve and decision curve analyses to examine the accuracy of the results. Data were collected continuously, and 212 and 85 cases were randomly assigned to training and testing sets, respectively. The area under the curve for the prediction of disease-free survival (training: 1 year, 0.83; 2 years, 0.81; 3 years, 0.83 versus testing: 1 year, 0.65; 2 years, 0.66; 3 years, 0.70), overall survival (training: 1 year, 0.86; 2 years, 0.85; 3 years, 0.86 versus testing: 1 year, 0.66; 2 years, 0.57; 3 years, 0.70) were determined. The coefficient factors including pathological response; preoperative tumour maximum diameter; preoperative lymph shorter diameter; preoperative tumour and lymph maximum standardized uptake value; change in tumour standardized uptake value preoperatively; and blood-related risk factors were favourably associated with prognosis (P < 0.001). Our prediction model, which integrated data from preoperative positron emission tomography-CT, preoperative blood parameters and pathological response, was able to make highly accurate predictions for disease-free and overall survival in patients with NSCLC receiving neoadjuvant immunity with chemical therapy.

BioFusionNet: Deep Learning-Based Survival Risk Stratification in ER+ Breast Cancer Through Multifeature and Multimodal Data Fusion.

Breast cancer is a significant health concern affecting millions of women worldwide. Accurate survival risk stratification plays a crucial role in guiding personalised treatment decisions and improving patient outcomes. Here we present BioFusionNet, a deep learning framework that fuses image-derived features with genetic and clinical data to obtain a holistic profile and achieve survival risk stratification of ER+ breast cancer patients. We employ multiple self-supervised feature extractors (DINO and MoCoV3) pretrained on histopathological patches to capture detailed image features. These features are then fused by a variational autoencoder and fed to a self-attention network generating patient-level features. A co-dual-cross-attention mechanism combines the histopathological features with genetic data, enabling the model to capture the interplay between them. Additionally, clinical data is incorporated using a feed-forward network, further enhancing predictive performance and achieving comprehensive multimodal feature integration. Furthermore, we introduce a weighted Cox loss function, specifically designed to handle imbalanced survival data, which is a common challenge. Our model achieves a mean concordance index of 0.77 and a time-dependent area under the curve of 0.84, outperforming state-of-the-art methods. It predicts risk (high versus low) with prognostic significance for overall survival in univariate analysis (HR=2.99, 95% CI: 1.88-4.78, p 0.005), and maintains independent significance in multivariate analysis incorporating standard clinicopathological variables (HR=2.91, 95% CI: 1.80-4.68, p 0.005).

A nomogram to predict cancer-specific mortality in adult patients with malignant meningioma: a competing risk analysis

BackgroundComprehensive investigations of the prognosis factors and treatment strategies with adjustment of competing causes of death for patients with malignant meningioma (MM) is still lacking.Patient and methodThe surveillance, Epidemiology, and End Results (SEER) database were used to include adult patients with this rare disease between 2004 and 2018. The probability of MM-caused mortality (MMCM) and non-MM-caused mortality (non-MMCM) were presented by cumulative incidence function curves. Then, the association between variates with non-MMCM was evaluated by the cox proportional hazard model, and the prognostic factors of MMCM were identified by Fine-Gray competing risk regression model. Furthermore, a nomogram was developed to predict the 1-year, 2-year, and 5-year MMCM and the performance was tested by a time-dependent area under the receiver operating characteristic (ROC) curve and calibration.Result577 patients were included, with a median age of 62 (18–100) years old and a median overall survival time of 36 (0–176) months. The percentage of non-MMCM was 15.4% (n = 89) in the entire population and 21.7% (n = 54) in elderly patients. The multivariable Cox proportional hazard regression model revealed that older age and other tumor(s) before or after MM had an independently significant association with higher non-MMCM. After adjustment of competing causes of death, the multivariable Fine-gray regression model identified age group ≥ 65 year, tumor size > 5.3 cm, recurrent MM, and histologic type 9530/3 (Meningioma, malignant) had an independently significant association with higher MMCM. Compared with gross total (GTR) of tumor, subtotal resection of tumor (HR 1.66, 95%CI 1.08–2.56, P = 0.02), partial resection of lobe (HR 2.26, 95%CI 1.32–3.87, P = 0.003), and gross total resection of lobe (HR 1.69, 95%CI 1.12–2.51, P = 0.01) had an independently significant association with higher MMCM.ConclusionThe competing risk nomogram including age group, tumor size, initial status, histologic type, and extent of resection is discriminative and clinically useful. This study emphasized the importance of the GTR of tumor in the treatment of MM patients, which had a significantly lower incidence of MMCM compared with biopsy, STR of tumor, partial resection of lobe, and GTR of lobe.

CT-based multimodal deep learning for non-invasive overall survival prediction in advanced hepatocellular carcinoma patients treated with immunotherapy

ObjectivesTo develop a deep learning model combining CT scans and clinical information to predict overall survival in advanced hepatocellular carcinoma (HCC).MethodsThis retrospective study included immunotherapy-treated advanced HCC patients from 52 multi-national in-house centers between 2018 and 2022. A multi-modal prognostic model using baseline and the first follow-up CT images and 7 clinical variables was proposed. A convolutional-recurrent neural network (CRNN) was developed to extract spatial-temporal information from automatically selected representative 2D CT slices to provide a radiological score, then fused with a Cox-based clinical score to provide the survival risk. The model’s effectiveness was assessed using a time-dependent area under the receiver operating curve (AUC), and risk group stratification using the log-rank test. Prognostic performances of multi-modal inputs were compared to models of missing modality, and the size-based RECIST criteria.ResultsTwo-hundred seven patients (mean age, 61 years ± 12 [SD], 180 men) were included. The multi-modal CRNN model reached the AUC of 0.777 and 0.704 of 1-year overall survival predictions in the validation and test sets. The model achieved significant risk stratification in validation (hazard ratio [HR] = 3.330, p = 0.008), and test sets (HR = 2.024, p = 0.047) based on the median risk score of the training set. Models with missing modalities (the single-modal imaging-based model and the model incorporating only baseline scans) can still achieve favorable risk stratification performance (all p < 0.05, except for one, p = 0.053). Moreover, results proved the superiority of the deep learning-based model to the RECIST criteria.ConclusionDeep learning analysis of CT scans and clinical data can offer significant prognostic insights for patients with advanced HCC.Critical relevance statementThe established model can help monitor patients’ disease statuses and identify those with poor prognosis at the time of first follow-up, helping clinicians make informed treatment decisions, as well as early and timely interventions.Key PointsAn AI-based prognostic model was developed for advanced HCC using multi-national patients.The model extracts spatial-temporal information from CT scans and integrates it with clinical variables to prognosticate.The model demonstrated superior prognostic ability compared to the conventional size-based RECIST method.Graphical

Novel nomogram and risk stratification for peritoneal recurrence after curative resection in gastric cancer

Peritoneal recurrence (PR) in gastric cancer after curative resection has poor prognosis. Therefore, we aimed to construct a nomogram to predict PR, and establish PR score for risk stratification to guide adjuvant chemotherapy. A total of 315 patients with gastric cancer after radical surgery were included, and randomly stratified into training group (n = 221) and validation group (n = 94). Univariate and multivariate analyses were used to determine predictive factors of PR. The nomogram was constructed to predict the risk of PR. We utilized the time-dependent area under the receiver operating characteristic (ROC) curves (AUCs), calibration curves, and decision curve analysis (DCA) to evaluate the performance of the nomogram. Multivariate analysis showed that tumor site, N stage, preoperative CEA, and postoperative CA199 were independent predictors of PR. A nomogram was constructed to predict PR based on these factors. The AUC value was 0.755 in the training group and 0.715 in the validation group. The calibration curves showed good agreement between prediction and observation in the training and validation groups. The decision curve analysis displayed a good net benefit of the nomogram. The novel PR score was developed and patients were stratified into the low-, medium-, and high -risk groups. For the high-risk group, postoperative adjuvant chemotherapy significantly improved patients' overall survival (OS) and disease-free survival (DFS). The establishment of nomogram facilitates the prediction of PR after radical gastrectomy, and a novel PR score may help guide adjuvant chemotherapy for gastric cancer.

Homotypic cell-in-cell structure as a novel prognostic predictor in non-small cell lung cancer and frequently localized at the invasive front

Homotypic cell-in-cell structures (hoCICs) are associated with tumor proliferation, invasion, and metastasis and is considered a promising prognostic marker in various cancers. However, the role of hoCICs in non-small cell lung cancer (NSCLC) remains unclear. Tumor tissue sections were obtained from 411 NSCLC patients. We analyzed the relationship between clinicopathological variables and the number of hoCICs. LASSO and multivariate Cox regression analysis were employed to identify prognostic factors for NSCLC. The impact of hoCICs on overall survival (OS) and disease-free survival (DFS) was assessed using the Kaplan–Meier curves and log-rank test. Prognostic models for OS and DFS were developed and validated using the C-index, time-dependent area under the curve (AUC), net reclassification improvement (NRI), integrated discrimination improvement (IDI), calibration curves and decision curve analysis (DCA). Among the cohort, 56% of patients had hoCICs while 44% did not. Notably, hoCICs were primarily found at the tumor invasion front. Male gender, smoking, squamous cell carcinoma, low differentiation, tumor size ≥ 3 cm, advanced TNM stage, lymph node metastasis, pleural invasion, vascular invasion, necrosis, P53 mutation, and high expression of Ki-67 were identified as relative risk factors for hoCICs. Furthermore, hoCICs was found to be a significant prognostic factor for both OS and DFS, with higher frequencies of hoCICs correlating with poorer outcomes. We constructed nomograms for predicting 1-, 3-, and 5-year OS and DFS based on hoCICs, and the calibration curves showed good agreement between the predicted and actual outcomes. The results of the C-index, time-dependent AUC, NRI, IDI, and DCA analyses demonstrated that incorporating hoCICs into the prognostic model significantly enhanced its predictive power and clinical applicability. HoCICs indicated independent perdictive value for OS and DFS in patients with NSCLC. Furthermore, the frequent localization of hoCICs at the tumor invasion front suggested a strong association between hoCICs and tumor invasion as well as metastasis.

Integrating Clinical, Genetic, and Electrocardiogram-Based Artificial Intelligence to Estimate Risk of Incident Atrial Fibrillation.

AF risk estimation is feasible using clinical factors, inherited predisposition, and artificial intelligence (AI)-enabled electrocardiogram (ECG) analysis. To test whether integrating these distinct risk signals improves AF risk estimation. In the UK Biobank prospective cohort study, we estimated AF risk using three models derived from external populations: the well-validated Cohorts for Aging in Heart and Aging Research in Genomic Epidemiology AF (CHARGE-AF) clinical score, a 1,113,667-variant AF polygenic risk score (PRS), and a published AI-enabled ECG-based AF risk model (ECG-AI). We estimated discrimination of 5-year incident AF using time-dependent area under the receiver operating characteristic (AUROC) and average precision (AP). Among 49,293 individuals (mean age 65±8 years, 52% women), 825 (2.4%) developed AF within 5 years. Using single models, discrimination of 5-year incident AF was higher using ECG-AI (AUROC 0.705 [95%CI 0.686-0.724]; AP 0.085 [0.071-0.11]) and CHARGE-AF (AUROC 0.785 [0.769-0.801]; AP 0.053 [0.048-0.061]) versus the PRS (AUROC 0.618, [0.598-0.639]; AP 0.038 [0.028-0.045]). The inclusion of all components ("Predict-AF3") was the best performing model (AUROC 0.817 [0.802-0.832]; AP 0.11 [0.091-0.15], p<0.01 vs CHARGE-AF+ECG-AI), followed by the two component model of CHARGE-AF+ECG-AI (AUROC 0.802 [0.786-0.818]; AP 0.098 [0.081-0.13]). Using Predict-AF3, individuals at high AF risk (i.e., 5-year predicted AF risk >2.5%) had a 5-year cumulative incidence of AF of 5.83% (5.33-6.32). At the same threshold, the 5-year cumulative incidence of AF was progressively higher according to the number of models predicting high risk (zero: 0.67% [0.51-0.84], one: 1.48% [1.28-1.69], two: 4.48% [3.99-4.98]; three: 11.06% [9.48-12.61]), and Predict-AF3 achieved favorable net reclassification improvement compared to both CHARGE-AF+ECG-AI (0.039 [0.015-0.066]) and CHARGE-AF+PRS (0.033 [0.0082-0.059]). Integration of clinical, genetic, and AI-derived risk signals improves discrimination of 5-year AF risk over individual components. Models such as Predict-AF3 have substantial potential to improve prioritization of individuals for AF screening and preventive interventions.

RadShap: An Explanation Tool for Highlighting the Contributions of Multiple Regions of Interest to the Prediction of Radiomic Models.

Explaining the decisions made by a radiomic model is of significant interest, as it can provide valuable insights into the information learned by complex models and foster trust in well-performing ones, thereby facilitating their clinical adoption. Promising radiomic approaches that aggregate information from multiple regions within an image currently lack suitable explanation tools that could identify the regions that most significantly influence their decisions. Here we present a model- and modality-agnostic tool (RadShap, https://github.com/ncaptier/radshap), based on Shapley values, that explains the predictions of multiregion radiomic models by highlighting the contribution of each individual region. Methods: The explanation tool leverages Shapley values to distribute the aggregative radiomic model's output among all the regions of interest of an image, highlighting their individual contribution. RadShap was validated using a retrospective cohort of 130 patients with advanced non-small cell lung cancer undergoing first-line immunotherapy. Their baseline PET scans were used to build 1,000 synthetic tasks to evaluate the degree of alignment between the tool's explanations and our data generation process. RadShap's potential was then illustrated through 2 real case studies by aggregating information from all segmented tumors: the prediction of the progression-free survival of the non-small cell lung cancer patients and the classification of the histologic tumor subtype. Results: RadShap demonstrated strong alignment with the ground truth, with a median frequency of 94% for consistently explained predictions in the synthetic tasks. In both real-case studies, the aggregative models yielded superior performance to the single-lesion models (average [±SD] time-dependent area under the receiver operating characteristic curve was 0.66 ± 0.02 for the aggregative survival model vs. 0.55 ± 0.04 for the primary tumor survival model). The tool's explanations provided relevant insights into the behavior of the aggregative models, highlighting that for the classification of the histologic subtype, the aggregative model used information beyond the biopsy site to correctly classify patients who were initially misclassified by a model focusing only on the biopsied tumor. Conclusion: RadShap aligned with ground truth explanations and provided valuable insights into radiomic models' behaviors. It is implemented as a user-friendly Python package with documentation and tutorials, facilitating its smooth integration into radiomic pipelines.

Novel insight into neurofilament light chain and rhythm outcomes after catheter ablation of new-onset atrial fibrillation: A prospective cohort study

BackgroundAtrial fibrillation (AF) is an age-related disorder closely linked to autonomic nervous system dysfunction. Neurofilament light chain (NFL) protein is a biomarker for neurodegenerative diseases. ObjectiveThe purpose of this study was to evaluate the predictive value of NFL in forecasting AF recurrence after ablation. MethodsPatients newly diagnosed with AF who underwent catheter ablation were included. Serum NFL levels were measured using enzyme-linked immunosorbent assay. The primary outcome was AF recurrence during follow-up. ResultsA total of 215 consecutive patients were enrolled, with average follow-up period of 10.69 months. During this period, 29 patients experienced AF recurrence. Multivariate Cox regression analysis revealed that high NFL levels (≥300 pg/mL) were an independent predictor of recurrence risk (adjusted hazard ratio [HR] 3.756; 95% confidence interval [CI] 1.392–10.136). The associations between NFL levels and AF recurrence were consistent across subgroups defined by age (>65 years), gender, hypertension, and paroxysmal AF. Restricted cubic spline analysis showed a consistent linear relationship across the entire range of NFL levels. Furthermore, incorporating NFL into the CHA2DS2-VASc score model significantly improved the prediction of recurrent AF risk, as demonstrated by time-dependent area under the curve and decision curve analysis. Notable enhancements were also observed in terms of net reclassification improvement (HR 0.464; 95% CI 0.226–0.675; P <.05) and integrated discrimination improvement (HR 0.087; 95% CI 0.017–0.183; P = .08). ConclusionNFL may serve as an effective biomarker for risk stratification and therapeutic decision-making in patients with new-onset AF who have undergone catheter ablation.