Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “prediction scoring”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 19 recordsLinked to original sources
PubMedDate not supplied

A lumbar disc surgery predictive score card. A retrospective evaluation.

This report assessed the relationship between patient selection and the outcome of lumbar disc surgery. Patient selecting is assessed by means of a predictive scoring technique previously reported. Because of the dynamic and ever-changing status of the lumbar spine with the passage of time, patient selectivity seems to be less critical after 5 years. For the first 5 years, however, the outcome of lumbar disc surgery seems to be directly related to patient selectivity.

Back Pain↗
PubMed2024

Blood-based DNA methylation and exposure risk scores predict PTSD with high accuracy in military and civilian cohorts.

BACKGROUND: Incorporating genomic data into risk prediction has become an increasingly useful approach for rapid identification of individuals most at risk for complex disorders such as PTSD. Our goal was to develop and validate Methylation Risk Scores (MRS) using machine learning to distinguish individuals who have PTSD from those who do not. METHODS: Elastic Net was used to develop three risk score models using a discovery dataset (n = 1226; 314 cases, 912 controls) comprised of 5 diverse cohorts with available blood-derived DNA methylation (DNAm) measured on the Illumina Epic BeadChip. The first risk score, exposure and methylation risk score (eMRS) used cumulative and childhood trauma exposure and DNAm variables; the second, methylation-only risk score (MoRS) was based solely on DNAm data; the third, methylation-only risk scores with adjusted exposure variables (MoRSAE) utilized DNAm data adjusted for the two exposure variables. The potential of these risk scores to predict future PTSD based on pre-deployment data was also assessed. External validation of risk scores was conducted in four independent cohorts. RESULTS: The eMRS model showed the highest accuracy (92%), precision (91%), recall (87%), and f1-score (89%) in classifying PTSD using 3730 features. While still highly accurate, the MoRS (accuracy = 89%) using 3728 features and MoRSAE (accuracy = 84%) using 4150 features showed a decline in classification power. eMRS significantly predicted PTSD in one of the four independent cohorts, the BEAR cohort (beta = 0.6839, p-0.003), but not in the remaining three cohorts. Pre-deployment risk scores from all models (eMRS, beta = 1.92; MoRS, beta = 1.99 and MoRSAE, beta = 1.77) displayed a significant (p < 0.001) predictive power for post-deployment PTSD. CONCLUSION: Results, especially those from the eMRS, reinforce earlier findings that methylation and trauma are interconnected and can be leveraged to increase the correct classification of those with vs. without PTSD. Moreover, our models can potentially be a valuable tool in predicting the future risk of developing PTSD. As more data become available, including additional molecular, environmental, and psychosocial factors in these scores may enhance their accuracy in predicting the condition and, relatedly, improve their performance in independent cohorts.

DNA methylation↗
PubMed2024

Blood-based DNA methylation and exposure risk scores predict PTSD with high accuracy in military and civilian cohorts.

BACKGROUND: Incorporating genomic data into risk prediction has become an increasingly popular approach for rapid identification of individuals most at risk for complex disorders such as PTSD. Our goal was to develop and validate Methylation Risk Scores (MRS) using machine learning to distinguish individuals who have PTSD from those who do not. METHODS: Elastic Net was used to develop three risk score models using a discovery dataset (n&#x2009;=&#x2009;1226; 314 cases, 912 controls) comprised of 5 diverse cohorts with available blood-derived DNA methylation (DNAm) measured on the Illumina Epic BeadChip. The first risk score, exposure and methylation risk score (eMRS) used cumulative and childhood trauma exposure and DNAm variables; the second, methylation-only risk score (MoRS) was based solely on DNAm data; the third, methylation-only risk scores with adjusted exposure variables (MoRSAE) utilized DNAm data adjusted for the two exposure variables. The potential of these risk scores to predict future PTSD based on pre-deployment data was also assessed. External validation of risk scores was conducted in four independent cohorts. RESULTS: The eMRS model showed the highest accuracy (92%), precision (91%), recall (87%), and f1-score (89%) in classifying PTSD using 3730 features. While still highly accurate, the MoRS (accuracy&#x2009;=&#x2009;89%) using 3728 features and MoRSAE (accuracy&#x2009;=&#x2009;84%) using 4150 features showed a decline in classification power. eMRS significantly predicted PTSD in one of the four independent cohorts, the BEAR cohort (beta&#x2009;=&#x2009;0.6839, p=0.006), but not in the remaining three cohorts. Pre-deployment risk scores from all models (eMRS, beta&#x2009;=&#x2009;1.92; MoRS, beta&#x2009;=&#x2009;1.99 and MoRSAE, beta&#x2009;=&#x2009;1.77) displayed a significant (p&#x2009;<&#x2009;0.001) predictive power for post-deployment PTSD. CONCLUSION: The inclusion of exposure variables adds to the predictive power of MRS. Classification-based MRS may be useful in predicting risk of future PTSD in populations with anticipated trauma exposure. As more data become available, including additional molecular, environmental, and psychosocial factors in these scores may enhance their accuracy in predicting PTSD and, relatedly, improve their performance in independent cohorts.

Humans↗
PubMed2026

Novel genomic score predicts survival with immune checkpoint inhibition in neuroendocrine neoplasms.

BACKGROUND: Immune checkpoint inhibition (ICI) has activity in Grade 3 (G3) neuroendocrine neoplasms (NENs), but predictive biomarkers of long-term overall survival (OS) are currently lacking. METHODS: We derived a genomic scoring system using a retrospective cohort of 54 NEN patients treated with nivolumab and/or ipilimumab, alongside a chemotherapy-only control (n&#x2009;=&#x2009;15) and pan-cancer validation cohort (n&#x2009;=&#x2009;1661). An overall genomic score (GS-O) combining positive (GS-P) and negative response genes (GS-N) was calculated. RESULTS: In the ICI cohort, patients with GS-O&#x2009;&#x2265;&#x2009;2 had significantly better outcomes: median OS was not reached, compared to 3.9&#x2009;months for low scores (HR = 0.007, P&#x2009;<&#x2009;.001). After multivariate adjustment, GS-P was independently associated with improved OS while GS-N showed a trend towards worsened OS. Importantly, GS-O&#x2009;&#x2265;&#x2009;2 did not predict survival in the chemotherapy-alone cohort but was validated as independently predictive in the pan-cancer validation dataset (HR = 0.92, P&#x2009;<&#x2009;.001). CONCLUSION: This study establishes a novel genomic score to predict OS in NENs treated with ICI with important pan-cancer implications.

Humans↗
PubMed2025

Polygenic Risk Scores Predicting Estimated GFR Validated With Iohexol Clearance.

INTRODUCTION: Genome-wide association studies (GWAS) have identified hundreds of single nucleotide variants (SNVs) associated with estimated glomerular filtration rate (eGFR). eGFR has been used as a proxy phenotype because of the complexity and cost of measured GFR (mGFR) in large studies. Because eGFR is influenced by non-GFR factors, these GWAS results may be biased compared with a hypothetical study using mGFR. We aimed to investigate this by comparing aggregate measures of genetic effects on mGFR and eGFR. METHODS: We studied 1492 persons from the Renal Iohexol Clearance Survey (RENIS) cohort, a representative sample of the general population in Northern Norway without preexisting cardiovascular disease, kidney disease, or diabetes. We measured iohexol-clearance, and genotyping was performed with a microarray chip enriched for GFR-related SNVs. We compared the performance of 3 published polygenic risk scores (PGS) developed for creatinine-based eGFR (eGFRcr), narrow-sense heritability (h2) and the mean effect of SNVs on mGFR, eGFRcr, cystatin C-based eGFR (eGFRcys) and eGFRcr-cys. RESULTS: The performance of the PGS differed for mGFR and the 3 eGFRs, with best performance for prediction of eGFRcr (P < 0.05). However, when the beta coefficients of the SNVs in the 3 PGS were estimated in the RENIS-cohort, their magnitude was 11% to 46% greater for mGFR than for the 3 eGFR methods in 8 of 9 comparisons (P < 0.05). mGFR had higher h2 (0.47) than eGFRcr (0.21), eGFRcys (0.37), and eGFRcr-cys (0.42). CONCLUSIONS: SNVs with non-GFR effects on creatinine and cystatin-C influence GWAS results. The results of GWAS using eGFR should be validated using experimental and other more precise methods.

chronic kidney disease↗
PubMed2026

Improving the reliability of polygenic risk score-based prediction for cardiovascular and renal complications across ancestries in type 2 diabetes using Mondrian Cross-Conformal Prediction.

Polygenic risk scores (PRS) developed in European populations often show reduced predictive performance in non-European populations, limiting their clinical utility. This lack of transferability across ancestries remains a major challenge in genomic medicine and raises concerns about health equity. We aimed to evaluate whether uncertainty-aware prediction, implemented through Mondrian Cross-Conformal Prediction, improves the performance and reliability of polygenic risk score-based predictions across ancestries for nephropathy, stroke, and myocardial infarction in individuals with type 2 diabetes in a multi-ethnic cohort. We leveraged Mondrian Cross-Conformal Prediction (MCCP), an uncertainty quantification framework, combined with logistic regression applied to a multi-polygenic risk score (multiPRS) to predict the risk of nephropathy, stroke, and myocardial infarction in individuals with type 2 diabetes. Two training frameworks were evaluated: one using 4,098 individuals with type 2 diabetes of European ancestry from the ADVANCE trial for training and 17,574 White British, 1,145 South Asian, and 749 African UK Biobank participants for testing; and another using the 17,574 White British UK Biobank participants for training and the South Asian and African participants for testing. Logistic regression provided robust baseline performance across populations. On top of this baseline, MCCP did not improve performance but added capabilities absent from probability-based stratification: for each individual, it issued a prediction together with an explicit confidence and credibility level; it allowed a tolerated error level to be set in advance and delivered prediction sets respecting it in the majority of settings; and it flagged individuals for whom no reliable prediction could be made. Applying MCCP to PRS-based prediction thus enables uncertainty-aware risk stratification and improves the reliability of risk prediction across ancestries, providing a more equitable framework for clinical use.

Female↗
PubMed2026

Comparison of Performance of Publicly Available Polygenic Risk Scores to Predict Clinically Actionable Coronary Artery Calcium Scores: The BioHEART-CT Cohort.

AIM: Coronary artery disease (CAD) remains the leading cause of morbidity and mortality globally. Polygenic Risk Scores (PRS) have been trained against major adverse cardiovascular outcomes (MACE) in large cohorts. Few studies have examined the effectiveness of these CAD MACE PRS tools in detecting individuals with subclinical coronary calcification. An association would provide an opportunity for clinical translation and targeting of CT imaging to new patients at risk for subclinical disease. METHODS: An analysis of 53 publicly available CAD PRS tools was completed in participants of the BioHEART-CT Discovery 1000 cohort presenting for clinically referred CT coronary angiography (CCTA). Associations between PRS and two binary CACS outcomes reflecting clinically significant coronary calcification were assessed: a) Absolute CACS (CACS &#x2265;100 Agatston units [AU]; and b) Percentile CACS (CACS &#x2265;75th age-/sex-adjusted percentile). Models were adjusted for genetic principal components, modifiable cardiovascular risk factors, and age/sex (in Absolute CACS). A subgroup analysis was performed using Framingham Risk Score (FRS) at baseline. RESULTS: Among 803 BioHEART-CT Discovery 1000 participants, 487 (60.6%) had any detectable coronary calcium. Most PRS tools demonstrated significant association with CACS outcomes, particularly evident when PRS was modelled as a continuous predictor. For Percentile CACS, 94.3% of PRS tools were significantly associated after full adjustment (median OR per PRS SD 1.41 (IQR 1.23-1.60). Quintile-based analysis revealed that individuals in the Top Quintile PRS had up to 7.99-fold increased odds of Percentile CACS &#x2265;75th compared to those in the Bottom Quintile. Analysis by FRS group revealed positive performance, especially in individuals of Low FRS wherein incorporating a PRS increased pre-test probability from 14% to 26%. CONCLUSION: Whilst most CAD PRS tools have been developed against clinical events, we show their ability to predict clinically relevant coronary calcification. Utility appears strongest in individuals traditionally considered lower risk, presenting an opportunity for clinical translation for improved diagnosis in the primary prevention setting, with the potential to triage individuals into a CACS screening pathway.

coronary artery disease↗
PubMed1979

A clinical score for predicting the level of respiratory care in infants with respiratory distress syndrome.

A scoring system was developed to predict the need for transferring infants with respiratory distress syndrome (RDS) from community hospitals to specialized respiratory care centers. Five clinical and laboratory determinations (birthweight, clinical RDS score, FI02, PCO2 and pH) recorded from 100 infants with RDS during one year were utilized in a score with values ranging from 0 to 10. Application of the score to 159 infants with RDS during the following year showed that: (1) 73 per cent of infants scoring less than or equal to 3 received only oxygen by hood; (2) 75 per cent of infants scoring 4--5 required continuous positive airway pressure (CPAP); and (3) 87 per cent of infants scoring greater than or equal to 6 needed mechanical ventilation (7V). Mean scores were significantly different (p less than 0.02) for each type of respiratory therapy employed: oxygen by hood (2.30 +/- 0.19 S.E.M.); CPAP (4.27 +/- 0.16 S.E.M.); MV (6.72 +/- 0.25 S.E.M.). The accuracy and simplicity of the score make it valuable for the physician in the community hospital to assist in deciding when to transfer a neonate with RDS for more intensive respiratory therapy.

Birth Weight↗
PubMed2026

Quantitative assessment of the fingerprint evidential value using machine learning.

Fingerprints as physical evidence have long supported criminal investigation and adjudication. In practice, however, fingerprint identification relies mainly on examiners' experience. Furthermore, expert opinions tend to be categorical, even though the opinions with the same conclusion could differ substantially in evidential strength. To quantitatively assess fingerprint evidential value, this study proposes a machine learning-based framework as an interpretable decision-support tool. A lightweight residual one-dimensional convolutional neural network was constructed, incorporating channel recalibration and a similarity-driven attention mechanism to learn adaptive contribution weights for different matched minutiae (minutiae for short). Controlled experiments revealed that the predicted evidential value increased with the number of minutiae and was significantly influenced by the quality of minutiae. With 10 minutiae, the mean predicted scores were 4.49, 7.00, and 9.09 for blurred, moderately blurred, and clear minutiae, respectively. Multiple regression analysis indicated that replacing a pair of blurred minutiae with a pair of clear minutiae increased the score by 0.492, whereas replacing it with a pair of moderately blurred minutiae increased the score by only 0.216. By mapping predicted scores to graded levels of evidential strength, the framework contributes to a paradigm shift from categorical expert opinions to graded ones, helping courts evaluate fingerprint evidence more scientifically.

Humans↗
PubMed2026

Role of Polygenic Risk Scores in Predicting Cognitive Functioning after Mild Traumatic Brain Injury: A TRACK-TBI Study.

Patients with traumatic brain injury (TBI) and Glasgow Coma Scale scores of 13-15 (historically called mild TBI [mTBI]) commonly experience changes in cognitive functioning, including processing speed, memory, and executive functioning. In a prospective sample (N = 523) of individuals of European descent who had been treated in a U.S. level 1 trauma center for mTBI, we examined the prognostic value of four polygenic risk scores (PRS) for cognitive outcomes at 6-months postinjury. To estimate the impact of mTBI on cognition, primary cognitive outcomes were scaled as z-scores reflecting changes in performance relative to predicted preinjury performance. The PRS examined were previously developed and validated to predict cognition-related outcomes of educational attainment (Education-PRS), intelligence (Intelligence-PRS), and Alzheimer's disease (AD-mild traumatic brain injury (APOE)-PRS and AD + APOE-PRS). Both the Education-PRS and Intelligence-PRS displayed bivariate associations with all four cognitive outcomes (&#x3b2; = 0.19-0.32), whereas neither Alzheimer's disease PRS was significantly associated with any outcome. After controlling for other factors known to predict cognitive outcomes of TBI (e.g., sex, education, mTBI severity defined by a combination of Glasgow Coma Scale scores and the presence/absence of acute intracranial findings on clinical neuroimaging), the Education-PRS and Intelligence-PRS remained independently predictive of verbal episodic memory (&#x3b2; = 0.10-0.16), whereas their associations with processing speed and executive functioning were mostly nonsignificant and were mediated through educational attainment. Looking across primary z-score and secondary raw score outcomes, cognitive outcomes 6 months post-mTBI were good on average, and PRS made small independent contributions to outcome prediction. The mediation model findings may support theories of cognitive reserve, which propose that individuals with stronger preinjury cognitive processing abilities (often estimated by educational history) can better compensate for TBI. Moreover, findings indicate that PRS may contribute modestly to multivariable models predicting cognitive function after TBI.

Humans↗
PubMed2026

Addition of CAD polygenic risk score to coronary artery calcium score enhances prediction of MACE.

BACKGROUND: Coronary heart disease (CHD) is prevalent in the United States, highlighting the need for accurate risk prediction to inform primary prevention strategies. While multivariate risk models like the Framingham Risk Score and ACC/AHA Pooled Cohort Equations are commonly utilized, novel risk markers, such as the coronary artery calcium score (CACS) and polygenic risk score (PRS), are increasingly gaining recognition. OBJECTIVES: This study aimed to compare the diagnostic utility of CACS and CAD PRS, both individually and in combination, for predicting major adverse cardiovascular events (MACE). METHODS: We conducted a retrospective analysis of a cohort comprising 1,380 predominantly Caucasian participants from the Sanford Health System. CAD PRS was constructed using genetic variants, while CACS was assessed via cardiac computed tomography (CT). Statistical analyses evaluated the relationship between each modality and MACE. RESULTS: Both CAD PRS and CACS were significantly associated with future MACE. Following the adjustment for covariates, the area under the curve (AUC) for both the CACS and PRS models was comparable, indicating similar predictive capabilities for MACE. However, the combination of CAD PRS with CACS significantly enhanced predictive accuracy, outperforming either modality alone. CONCLUSIONS: This study underscores the value of integrating CACS and CAD PRS in predicting MACE. The synergistic effect of CAD PRS combined with CACS markedly improves predictive power. Further research and prospective studies are necessary to validate these findings and assess their clinical implications. Investigating the interactions between PRS and CACS will be crucial for refining cardiovascular risk prediction and optimizing prevention strategies.

cardiac genetics↗
PubMed2025

A viral clonality evenness score to predict progression to adult T-cell leukaemia in asymptomatic carriers of human T-lymphotropic virus type 1 in Japan: a retrospective longitudinal cohort study.

BACKGROUND: Adult T-cell leukaemia/lymphoma (ATL) is a highly aggressive T-cell malignancy that occurs in approximately 2-7% of individuals with human T-lymphotropic virus type 1 (HTLV-1), after decades of asymptomatic infection. To address the urgent need for predictive biomarkers to identify asymptomatic carriers of HTLV-1 at high risk of progression to ATL, we aimed to evaluate viral clonality sequencing as a potential tool for risk stratification. METHODS: This retrospective longitudinal cohort study involved HTLV-1 carriers enrolled in the Joint Study on Predisposing Factors of ATL Development, a nationwide cohort study initiated in Japan in 2002. Participants were selected from this cohort on the basis of their baseline proviral load at the time of enrolment as an asymptomatic carrier, length of follow-up, and clinical outcome. The cohort was subdivided into three subgroups: the first comprising HTLV-1 carriers who developed ATL, the second comprising carriers with high proviral load (&#x2265;4%) who did not progress to ATL, and the third comprising carriers with low proviral load (<4%) who did not progress to ATL. DNA extracted from peripheral blood mononuclear cells collected at enrolment and at least one follow-up visit was analysed by HTLV-1 clonality sequencing and the proviral load was quantified. We calculated a viral clonality evenness (VCE) score, based on the Shannon Evenness Index, to quantify the uniformity of the clonal distribution of samples, for which 0 represents a perfectly monoclonal architecture and 1 indicates a completely polyclonal landscape. We then estimated the performance of proviral load thresholds and VCE scoring to classify the risk of progression to ATL using the area under the receiver operating characteristic curve (AUC), the accuracy, and Matthews correlation coefficient. VCEs were compared between participant subgroups with the Wilcoxon rank sum test. FINDINGS: 56 participants followed up by JSPFAD between Feb 6, 2003, and July 19, 2022, were included in this study: 17 who progressed to ATL (mean follow-up 8&#xb7;3 years [SD 4&#xb7;0]), 18 who had a high proviral load and did not progress to ATL (9&#xb7;7 years [3&#xb7;4]), and 21 who had a low proviral load and did not progress to ATL (7&#xb7;5 years [3&#xb7;0]). Clonality sequencing of samples from 39 participants who did not progress to ATL revealed hundreds to thousands of HTLV-1 integration sites at both timepoints, corresponding to multiple clones of low and uniform abundance, and these participants had high VCE scores (&#x2265;0&#xb7;694) at baseline. By contrast, most participants (14 of 17) who progressed to ATL had a single predominant clone or two to four predominant clones at both timepoints, and lower VCE scores (<0&#xb7;694) at baseline than those who did not progress (p<0&#xb7;0001). AUCs were very similar for proviral load thresholds (91 [95% CI 80-98]) and VCE scoring (91 [78-100]), although when using methods that give equal weight to every individual, VCE scoring outperformed proviral load thresholds in predicting progression to ATL (accuracy: proviral load 0&#xb7;76 [95% CI 0&#xb7;76-0&#xb7;77], VCE scoring 1&#xb7;00 [0&#xb7;99-1&#xb7;00]; Matthews correlation coefficient: proviral load 0&#xb7;23 [95% CI 0&#xb7;19-0&#xb7;24], VCE scoring 0&#xb7;91 [0&#xb7;80-1&#xb7;00]). Prediction based on VCE scoring indicated no false positives, compared with 20% when using proviral load, although VCE scoring yields a greater number of false negatives (0&#xb7;3% vs 0&#xb7;1%). INTERPRETATION: The implementation of VCE scoring in clinical practice could inform early pre-emptive therapeutic interventions, exclusively targeting individuals with HTLV-1 at high risk and aiming to prevent progression to aggressive, treatment-refractory disease. Further validation, including independent confirmation of the performance of VCE scoring in multiple populations and the characterisation of its temporal dynamics, will be crucial to determine its clinical utility and potential integration into care pathways. FUNDING: Association Jules Bordet, FNRS-T&#xe9;l&#xe9;vie, FCC, WALInnov, FLF, JSPS-KAKENHI, and CoBiA.

Humans↗
PubMed1979

Reported parental characteristics in relation to trait depression and anxiety levels in a non-clinical group.

Care and overprotection appear to reflect the principal dimensions underlying parental behaviours and attitudes. In previous studies of neurotically depressed patients and of a non-clinical group, subjects who scored their parents as lacking in care and/or overprotective had the greater depressive experience. The present study of another non-clinical group (289 psychology students) replicated those findings in regard to trait depression levels. In addition, associations between those parental dimensions and trait anxiety scores were demonstrated. Multiple regression analyses established that 9-10% of the variance in mood scores was accounted for by scores on those parental dimensions. Low maternal care scores predicted higher levels of both anxiety and depression, while high maternal overprotection scores predicted higher levels of anxiety but not levels of depression. Maternal influences were clearly of greater relevance than paternal influences.

Anxiety↗
PubMed2026

Topologically distinct intratumoral heterogeneity scores for predicting high-risk pathological grades in invasive lung adenocarcinoma: A multicenter study across four institutions.

High-risk subtypes of invasive lung adenocarcinoma (IAC), particularly micropapillary- or solid-predominant patterns, are closely associated with poor prognosis. This multicenter retrospective study developed and validated a predictive model for the preoperative identification of these high-risk subtypes using topologically distinct intratumoral heterogeneity (ITH) scores derived from CT images. The study included 1,051 patients with IAC. Two complementary ITH scores were developed: a two-dimensional ITH score, which integrated local radiomics features with global pixel distribution patterns on the largest cross-sectional CT slice, and a three-dimensional ITH score, which extended this quantification across the entire tumor volume. Clinicoradiological features and ITH scores were incorporated as model inputs to construct six base machine learning classifiers and a final stacking ensemble classifier. Model interpretability and robustness were evaluated using SHapley Additive exPlanations (SHAP)-based ablation analyses. An independent dataset from The Cancer Imaging Archive (TCIA) was used for external validation to investigate associations between ITH scores and pathological characteristics, genomic features, recurrence-free survival, and overall survival. The stacking ensemble classifier achieved the best predictive performance, with an area under the receiver operating characteristic curve of 0.875, outperforming models based solely on radiomics features (0.834) or clinicoradiological features (0.792). SHAP analysis identified the 3D ITH score as the most influential contributor to model output, and TCIA validation showed that higher 3D ITH scores were associated with more aggressive tumor biology and poorer survival outcomes. The topologically distinct 3D ITH score may provide a clinically meaningful imaging biomarker for preoperative risk stratification in IAC.

Journal Article↗
PubMed2026

A multi-scale fusion model based on multi-phase contrast-enhanced CT for predicting pancreatic cancer resectability.

Purpose.Develop a multi-scale fusion model (MSFM) based on multi-phase contrast-enhanced computed tomography (CECT) to predict pancreatic cancer (PC) resectability, thereby assisting expert decision-making.Methods.This retrospective study enrolled 280 patients with PC from four institutions, which were randomly divided into a training cohort (202 patients) and an independent test cohort (78 patients). Three-phase CECT images (arterial, venous, and delayed phases) were used for modeling. The MSFM comprises two sub-networks: (1) a multi-phase fusion network for extracting cross-phase shared fusion features, (2) a phase-specific branch network for capturing phase-specific features; and a post-fusion strategy to generate the final predictive score by integrating the shared fusion features and three groups of phase-specific features. Additionally, a human-machine fusion deep learning model (HMfDL) was constructed by fusing the predictive score of the MSFM with expert assessments.Results.In the independent test, the MSFM achieved an AUC (area under the receiver operating characteristic curve) of 0.8385 (95% CI: 0.7521-0.9249), accuracy of 84.62%, sensitivity of 72.00%, and specificity of 90.57%. This performance outperformed single-phase models (AUC range: 0.7638-0.7781), two-phase models (AUC range: 0.7826-0.7864), and ten states-of-the-art classifiers (AUC range: 0.7404-0.7796). The HMfDL further improved the performance, reaching an AUC of 0.8626 (95% CI: 0.7853-0.9400), accuracy of 91.03%, sensitivity of 80.00%, and specificity of 96.23%. Notably, the HMfDL corrected 58.82% of misdiagnosis made by experts.Conclusions. The MSFM effectively fuses multi-phase CECT to enable highly accurate predictions of PC resectability, and provides valuable support for expert decision-making through HMfDL.

Humans↗