Imaging Biomarkers Research Articles

Delta Radiomics — Potential role in Head Neck Cancer

IntroductionDelta radiomics is a tool used to assess the response of oncologic patients undergoing immunotherapy. It extracts high-dimensional quantitative features from medical images, providing information about cancer's phenotype, genotype, and tumoral microenvironment. This analysis could help avoid invasive procedures and help choose the most suitable therapeutic in multiple therapeutic options. Radiomics has gained interest as an imaging biomarker for predicting response to various immunotherapies. Delta radiomics assesses feature variations from one time point to another based on subsequent images, offering higher value for treatment-outcome prediction or patient stratification into risk categories. It has potential benefits for clinical endpoints in oncology, such as differential diagnosis, prognosis, treatment response prediction, and evaluation of side effects. Further research with prospective and multicentre studies is needed for clinical validation of delta radiomics approaches. Statement of clinical significanceIn head and neck oncology, delta radiomics can be used to enhance the precision of diagnosis, assess tumor response, forecast normal tissue toxicity, predict clinical outcome, and pinpoint characteristics for treatment modification. Patients' quality of life may be enhanced by it. It can support post-treatment surveillance. Additionally, it can support the delivery of individualized care based on a patient's reaction to medication and radiation.

Quantitative synthetic MRI for evaluation of hippocampus in patients with multiple sclerosis

ObjectiveTo identify early changes in hippocampal quantitative parameters in multiple sclerosis (MS) patients using synthetic MRI, and to correlate these changes with clinical variables. Methods45 MS patients and 26 healthy controls (HCs) underwent synthetic MRI and 3D-T1 MRI. The hippocampus volumes were assessed by using voxel-based morphometry. Synthetic MRI parameters (T1, T2, and proton density (PD)) from hippocampus and its subfield were measured and compared, and their associations with the Expanded Disability Status Scale (EDSS), Symbol Digit Modalities Test (SDMT) scores were further investigated. ResultsThere was no significant difference in hippocampal volume between MS patients and HCs. Compared with HCs, the T1, T2 and PD values of hippocampus and its subfield increased in MS patients. T2 values showed positive correlation with EDSS and negative correlation with SDMT. ConclusionsSynthetic MRI can detect subtle quantitative changes of the hippocampus in MS patients with normal hippocampal volume. Specifically, Synthetic MRI parameters may apply as potentially effective imaging biomarker for hippocampus evaluation.

Cardiovascular Magnetic Resonance Imaging of Takotsubo Syndrome: Evolving Diagnostic and Prognostic Perspectives.

Takotsubo syndrome (TS) is a temporary form of left ventricular (LV) dysfunction characterized by a distinct pattern of LV impairment, often triggered by a physical or emotional stressful event. Historically, TS was considered a benign condition due to its prompt restoration of myocardial function and generally excellent outcomes. However, recent studies have shown that complications similar to those seen after myocardial infarction can occur, necessitating careful monitoring of these patients. Among noninvasive imaging techniques, cardiovascular magnetic resonance (CMR) is becoming increasingly important in evaluating patients with TS. CMR offers a unique ability to noninvasively assess myocardial tissue characteristics, allowing for detecting the typical features of TS, such as specific wall motion abnormalities and myocardial edema. Beyond its well-established diagnostic utility in the clinical management of TS, CMR has also proven valuable in prognosis and risk stratification for these patients. Advances in CMR, including myocardial strain and parametric mapping have expanded its role in the diagnosis, prognosis, and follow-up of these patients. This review aims to provide a comprehensive overview of the potential applications of CMR in the diagnostic and prognostic evaluation of TS patients. It explores the emerging use of novel CMR imaging biomarkers that may enhance diagnosis, improve prognostic accuracy, and contribute to the overall management of these patients.

The effect of Clot Burden Score on clinical outcomes in acute ischemic stroke patients with atrial fibrillation treated with endovascular thrombectomy.

The Clot Burden Score (CBS) is used to assess thrombus length. The influence of CBS on functional outcome was barely analyzed in patients with acute ischemic stroke (AIS) with atrial fibrillation (AF). To assess the association between CBS and clinical outcomes in large vessel occlusion (LVO)-related patients with AF who have undergone endovascular thrombectomy (EVT). A total of 160 patients with AF were enrolled between January 2021 and April 2023. The CTA-CBS score was used to quantify the thrombus burden. The primary outcome was the modified Rankin scale (mRS) score at 90 days. A multivariate logistic regression model was used to identify prognostic predictors and determine the correlation between CTA-CBS and clinical outcomes. In the multivariable logistic regression, younger age, smaller clots, and better collateral status were the favorable prognosis factors. The odds ratios (OR) were 0.956 (95% confidence interval [CI] = 0.924-0.988, P = 0.008), 1.29 (95% CI = 1.110-1.499, P < 0.001), and 1.706 (95% CI = 1.065-2.731, P = 0.026), respectively. A smaller clot correlated with better outcomes OR of 1.29 (95% CI = 1.110-1.499, P < 0.001) for the entire cohort, 1.395 (95% CI = 1.142-1.702, P < 0.001) for bridging the EVT subgroup, and 1.171 (95% CI = 0.866-1.582, P = 0.305) for direct EVT subgroup. In LVO-related AIS patients with AF treated with EVT, lower CBS is associated with poorer functional outcomes. Notably, CBS acts as a prognostic imaging biomarker in the direct EVT subgroup and does not in bridging the EVT subgroup.

Utility of dual-energy CT and advanced multiparametric MRI based imaging biomarkers of pancreatic fibrosis in grading the severity of chronic pancreatitis.

To non-invasively quantify pancreatic fibrosis and grade severity of chronic pancreatitis (CP) on dual-energy CT (DECT) and multiparametric MRI (mpMRI). We included 72 patients (mean age:30years; 59 men) with suspected or confirmed CP from December 2019 to December 2021 graded as equivocal(n = 20), mild(n = 18), and moderate-marked(n = 34) using composite imaging and endoscopic ultrasound criteria. Study patients underwent multiphasic DECT and mpMRI of the abdomen. Normalized iodine concentration(NIC) and fat fraction(FF) on 6-minute delayed DECT, and T1 relaxation time(T1Rt), extracellular volume fraction(ECVf), intravoxel incoherent motion-based perfusion fraction(PF), and magnetization transfer ratio(MTR) on mpMRI of pancreas were compared. 20 renal donors(for DECT) and 20 patients with renal mass(for mpMRI) served as controls. NIC of pancreas in controls and progressive grades of CP were 0.24 ± 0.05, 0.80 ± 0.18, 1.06 ± 0.23, 1.40 ± 0.36, FF were 9.28 ± 5.89, 14.19 ± 5.29, 17.31 ± 5.99, 29.32 ± 12.22, T1Rt were 590.11 ± 61.13, 801.93 ± 211.01, 1006.79 ± 352.18, 1388.01 ± 312.23ms, ECVf were 0.07 ± 0.03, 0.30 ± 0.12, 0.41 ± 0.12, 0.53 ± 0.13, PF were 0.38 ± 0.04, 0.28 ± 0.07, 0.25 ± 0.09, 0.21 ± 0.05 and MTR were 0.12 ± 0.03, 0.15 ± 0.06, 0.21 ± 0.07, 0.26 ± 0.06, respectively. There were significant differences for all quantitative parameters between controls and mild CP; for NIC, PF, and ECVf between controls and progressive CP grades (p < 0.05). Area under curve for NIC, FF, T1Rt, ECVf, PF, and MTR in differentiating controls and mild CP were 1.00, 0.86, 0.95, 1.00, 0.90 and 0.84 respectively and for NIC, FF, ECVf and PF in differentiating controls and equivocal CP were 1.00, 0.76, 0.95 and 0.92 respectively. DECT and mpMRI were useful in quantifying pancreatic fibrosis and grading the severity of CP. NIC was the most accurate marker.

A Feasibility Study of [18F]F-AraG Positron Emission Tomography (PET) for Cardiac Imaging-Myocardial Viability in Ischemia-Reperfusion Injury Model.

Myocardial infarction (MI) with subsequent inflammation is one of the most common heart conditions leading to progressive tissue damage. A reliable imaging marker to assess tissue viability after MI would help determine the risks and benefits of any intervention. In this study, we investigate whether a new mitochondria-targeted imaging agent, 18F-labeled 2'-deoxy-2'-18F-fluoro-9-β-d-arabinofuranosylguanine ([18F]F-AraG), a positron emission tomography (PET) agent developed for imaging activated T cells, is suitable for cardiac imaging and to test the myocardial viability after MI. To test whether the myocardial [18F]-F-AraG signal is coming from cardiomyocytes or immune infiltrates, we compared cardiac signal in wild-type (WT) mice with that of T cell deficient Rag1 knockout (Rag1 KO) mice. We assessed the effect of dietary nucleotides on myocardial [18F]F-AraG uptake in normal heart by comparing [18F]F-AraG signals between mice fed with purified diet and those fed with purified diet supplemented with nucleotides. The myocardial viability was investigated in rodent model by imaging rat with [18F]F-AraG and 2-deoxy-2[18F]fluoro-D-glucose ([18F]FDG) before and after MI. All PET signals were quantified in terms of the percent injected dose per cc (%ID/cc). We also explored [18F]FDG signal variability and potential T cell infiltration into fibrotic area in the affected myocardium with H&E analysis. The difference in %ID/cc for Rag1 KO and WT mice was not significant (p = ns) indicating that the [18F]F-AraG signal in the myocardium was primarily coming from cardiomyocytes. No difference in myocardial uptake was observed between [18F]F-AraG signals in mice fed with purified diet and with purified diet supplemented with nucleotides (p = ns). The [18F]FDG signals showed wider variability at different time points. Noticeable [18F]F-AraG signals were observed in the affected MI regions. There were T cells in the fibrotic area in the H&E analysis, but they did not constitute the predominant infiltrates. Our preliminary preclinical data show that [18F]F-AraG accumulates in cardiomyocytes indicating that it may be suitable for cardiac imaging and to evaluate the myocardial viability after MI.

Castration Levels of Testosterone Results in Atrophy of Androgen-Sensitive Perineal Muscles: A Potential Biomarker for Male Hypogonadism

ObjectivesTo evaluate MRI-based measurements of androgen-sensitive perineal/pelvic muscles in men with prostate cancer before and after androgen deprivation therapy (ADT) as a novel imaging marker for end-organ effects of hypogonadism. Diagnosing hypogonadism or testosterone deficiency (TD) requires both low serum testosterone and clinical symptoms, such as erectile dysfunction and reduced libido. However, the non-specific nature of many TD symptoms makes it challenging to initiate therapy. Objective markers of TD help to better identify patients who may benefit from testosterone supplementation; however, current markers, such as low bone mineral density, lack sensitivity. Previous studies suggest that decreased bulbocavernosus-muscle (BCM) thickness may be associated with TD, although it remains unclear if this is a correlative relationship. MethodsData was prospectively collected for patients with intermediate/high-risk localized prostate cancer enrolled in a phase II trial (NCT02430480). Patients received ADT before prostatectomy and underwent prostate MRI pre-/post-ADT. BCM, ischiocavernosus-muscle (ICM), and levator-ani-muscle (LAM) measurements were made using T2W-MRI. Paired t-tests evaluated changes in BCM/ICM/LAM width, and linear regression analyses evaluated relationships between changes in testosterone and muscle width. ResultsThirty-eight consecutive patients with pre-/post-ADT MRIs were analyzed. Baseline testosterone was 286.5ng/dl, and 36/38 patients had post-ADT testosterone <50ng/dL. Pre-ADT and post-ADT measurements of the bilateral BCM/ICM/LAM width were 7.16mm/7.95mm/5.53mm and 5.68mm/6.71mm/4.89mm, respectively (p<0.001). Decreases in testosterone predicted reduction in combined perineal muscle (BCM+ICM) width (p=0.032). ConclusionsAndrogen deprivation led to significant and relatively rapid decreases in BCM/ICM/LAM thickness. This objective biomarker of low testosterone states may help identify patients who will potentially benefit from testosterone replacement.

Clinical and imaging predictors for the development of diabetes mellitus following a single episode of acute pancreatitis in youth

BackgroundAcute pancreatitis (AP) increases the risk of diabetes mellitus (DM). Our aim was to identify clinical, laboratory and imaging predictors of preDM/DM in youth post index AP. MethodsThis was a prospective cohort study of patients ≤21 years-old with an index admission for AP and follow up at 3 and/or 12 months. Clinical laboratory values, imaging findings, admission course, and plasma chemokine and cytokine measures collected at index admission were tested for association with preDM/DM development. A multivariable regression model was used to predict preDM/DM. ResultsAmong 187 enrolled participants, 137 (73 %) and 144 (77 %) underwent DM screening at 3 and 12 months respectively, and 137 (73 %) had imaging available. PreDM/DM occurred in 22/137 (16 %; preDM n = 21, DM n = 1) at 3 months and 23/144 (16 %; preDM n = 18, DM n = 5) participants at 12 months. Univariate associations with preDM/DM at 12 months included: severe AP (SAP) (52 % preDM/DM vs. 17 % no DM; p = 0.0008), median [IQR] IL-6 (910 pg/ml [618–3438] vs. 196 pg/ml [71–480], p < 0.05) and CRP (4.16 mg/L [1.67–10.7] vs. 1.55 mg/L [0.4–3.68], p = 0.1) at time of AP attack. The optimal multivariable model to predict preDM/DM included with clinical variables was severe acute pancreatitis (SAP), c reactive protein (CRP), interleukin-6 (IL-6), and age [AUC = 0.80; (0.70, 0.88)]. Including imaging markers, the ideal model included SAP, CRP, IL-6, subcutaneous fat area, age and presence of autoimmune disease with an AUC [0.82 (0.71, 0.90)]. ConclusionsDevelopment of preDM/DM following an index AP episode can be predicted by baseline AP severity, baseline CRP, IL-6 levels, and subcutaneous fat area.

Pregnancy alters fatty acid metabolism, glucose regulation, and detoxification of the liver in synchrony with biomechanical property changes

Pregnancy places a metabolic burden on the body including the liver, which is responsible for ensuring adequate nutrition for the maternal and fetal systems. To gain a better understanding of liver adaptation, this study investigates metabolic shifts occurring in livers of pregnant rats. Metabolic capacities of the livers of pregnant and non-pregnant female Wistar rats were assessed using comprehensive metabolic models. Kinetic metabolic models were generated for each animal based on protein abundance data from proteomics analysis allowing for a subject-specific assessment of hepatic metabolic functions. Data are available via ProteomeXchange with identifier PXD050758. Additionally, tissue stiffness, viscosity, and water diffusion obtained from magnetic resonance imaging and elastography were correlated with metabolic capabilities to study the relationship between metabolic function and biophysical properties. Proteome profiling revealed differences in protein expression in the livers of pregnant and non-pregnant animals. Functional analysis showed significant variations in metabolic capacities. Livers of pregnant rats had reduced capacities in carbohydrate and fatty acid metabolism, along with altered urea synthesis. Additionally, there were associations between metabolic functions and biophysical properties highlighting potential links between changes in liver structure and metabolic capacities during pregnancy. In summary, our work reveals extensive hepatic metabolic changes in pregnant rats. The liver adapts its metabolic capacities to ensure whole-body metabolic homeostasis but may struggle to counteract nutritional challenges, such as hypoglycemia. The study, employing a personalized approach combining proteomics, kinetic modeling, and advanced imaging, sheds light on the intricate interplay between hepatic adaptations and medical imaging markers, providing a foundation for further investigations into the implications for maternal and fetal health.

Psychosocial Health and the Association Between Cerebral Small Vessel Disease Markers With Dementia: The ARIC Study.

Associations between magnetic resonance imaging markers of cerebral small vessel disease (CSVD) and dementia risk in older adults have been established, but it remains unclear how lifestyle factors, including psychosocial health, may modify this association. Social support and social isolation were assessed among participants of the community-based ARIC (Atherosclerosis Risk in Communities) Study, via self-reported questionnaires (1990-1992). Following categorization of both factors, participants were classified as having strong or poor mid-life social relationships. At visit 5 (2011-2013), participants underwent 3T brain magnetic resonance imaging quantifying CSVD measures: white matter hyperintensity volume, microbleeds (subcortical), infarcts (lacunar), and white matter integrity (diffusion tensor imaging). Incident dementia cases were identified from the time of imaging through December 31, 2020 with ongoing surveillance. Associations between CSVD magnetic resonance imaging markers and incident dementia were evaluated using Cox proportional-hazard regressions adjusted for demographic and additional risk factors (from visit 2). Effect modification by mid-life social relationships was evaluated. Of the 1977 participants with magnetic resonance imaging, 1617 participants (60.7% women; 26.5% Black participants; mean age at visit 2, 55.4 years) were examined. In this sample, mid-life social relationships significantly modified the association between white matter hyperintensity volume and dementia risk (P interaction=0.001). Greater white matter hyperintensity volume was significantly associated with risk of dementia in all participants, yet, more substantially in those with poor (hazard ratio, 1.84 [95% CI, 1.49-2.27]) versus strong (hazard ratio, 1.26 [95% CI, 1.08-1.47]) mid-life social relationships. Although not statistically significant, subcortical microbleeds in participants with poor mid-life social relationships were associated with a greater risk of dementia, relative to those with strong social relationships, in whom subcortical microbleeds were no longer associated with elevated dementia risk. The elevated risk of dementia associated with CSVD may be reduced in participants with strong mid-life social relationships. Future studies evaluating psychosocial health through the life course and the mechanisms by which they modify the relationship between CSVD and dementia are needed.

Enhancing medical image security through machine learning and dual watermarking-based technique

ABSTRACT As the world becomes increasingly digital, healthcare is no exception. With the ease of sharing e-healthcare records over open networks, smart healthcare systems have become a popular way to manage patient information. But as the popularity of these systems has grown, so has the concern for their security. That's where image security techniques come in. In this paper we have developed a new approach to secure e-patient records like DICOM images. By combining the redundant discrete wavelets transform (RDWT), Hessenberg Decomposition (HD), and randomized singular value decomposition (RSVD). We developed a robust and dual watermarking scheme. This scheme uses multiple watermarks, including Electronic Patient Record (EPR) as text and images, to ensure high-level authentication. In order to attain a balance between imperceptibility and robustness, a PSO-based optimization of scale factor is employed and Turbo code is utilized to encode the EPR and minimize channel noise. Additionally, the marked image undergoes encryption using a 3D chaotic-based encryption technique, and the extracted watermark is denoised through a deep neural network. The result shows that the proposed scheme is both secure and reliable. With this dual watermarking scheme, we have made great strides in securing e-patient records.

Quantitative electroencephalography to assess post-stroke functional disability: A systematic review and meta-analysis

ObjectiveQuantitative electroencephalography (QEEG) is a non-invasive, reliable and easily accessible modality to assess neuronal activity. QEEG in acute stroke may predict short and long-term functional outcomes. The role of individual indices has not been studied in a meta-analysis. We aim to assess individual QEEG-derived indices to predict post-stroke disability. MethodsWe included studies (sample size ≥ 10) with stroke patients who underwent EEG and a follow-up outcome assessment was available either in the form of modified Rankin scale (mRS) or National Institute of Stroke scale (NIHSS) or Fugl-Meyer scale (FMA). QEEG indices analysed were delta-alpha ratio (DAR), delta-theta-alpha-beta ratio (DTABR), brain symmetry index (BSI) and pairwise derived brain symmetry (pdBSI). ResultsNine studies (8 had only ischemic stroke, and one had both ischemic and haemorrhagic stroke), including 482 participants were included for meta-analysis. Higher DAR was associated with worse mRS (n=300, Pearson's r 0.26, 95 % CI 0.21-0.31). Higher DTABR was associated with worse mRS (n=337, r=0.32, 95 % CI 0.26-0.39). Higher DAR was associated with higher NIHSS (n=161, r=0.42, 95 % CI0.24-0.6). Higher DTABR was associated with higher NIHSS (n=158, r=0.49, 95 % CI 0.31-0.67). ConclusionsQEEG-derived indices DAR and DTABR have the potential to assess post-stroke disability. Adding QEEG to the clinical and imaging biomarkers in the acute phase may help in better prediction of post-stroke recovery. RegistryPROSPERO 2022 CRD42022292281

Comparative analysis of COVID-19 pneumonia in pregnant versus matched non-pregnant women: radiologic, laboratory, and clinical perspectives.

This study aimed to assess the severity and outcomes of COVID-19 in pregnant women, focusing on laboratory and radiological discrepancies between pregnant women and matched nonpregnant women. In this retrospective cross-sectional analysis, we matched 107 nonpregnant women with 66 pregnant women in terms of age, comorbidities, and the interval between symptom onset and hospital admission. Demographic, clinical, laboratory, and radiological data were collected, and chest CT scans were evaluated using a severity scale ranging from 0 to 5. Logistic regression and adjusted Cox regression models were used to assess the impact of various factors on pregnancy status and mortality rates. Differences in several laboratory parameters, including the neutrophil-to-lymphocyte ratio, liver aminotransferases, alkaline phosphatase, urea, triglycerides, cholesterol, HbA1c, ferritin, coagulation profiles, and blood gases, were detected. Radiologic exams revealed that nonpregnant women had sharper opacities, whereas pregnant women presented with hazy opacities and signs of crypt-organizing pneumonia. A notable difference was also observed in the pulmonary artery diameter. The mortality rate among pregnant women was 4.62%, which was comparable to the 5.61% reported in nonpregnant patients. Compared with nonpregnant patients, pregnancy did not significantly affect the severity or mortality of COVID-19. Our study revealed discernible differences in specific laboratory and imaging markers between pregnant and nonpregnant COVID-19 patients.

Mitral annular disjunction as the cause of malignant ventricular arrhythmia in a young adult: a case report

Mitral annular disjunction (MAD) is associated with ventricular arrhythmias and sudden cardiac death (VA/SCD). Risk prediction for VA/SCD in individuals with MAD includes clinical, ECG and imaging markers. Imaging assessment for MAD includes transthoracic echocardiography (TTE) and cardiac magnetic resonance imaging (CMRI) and there are certain features in each modality to aid both diagnosis and risk assessment. A 23-year-old female with no medical history presented to hospital following out of hospital cardiac arrest. Effective bystander CPR and successful cardioversion from ventricular fibrillation to sinus rhythm by paramedics achieved return of spontaneous circulation. Focused transthoracic echocardiography and cardiac magnetic resonance imaging identified MAD. Notably, there were no specific high risk clinical or imaging features. The patient recovered completely and was discharged with implantable defibrillator for secondary prevention. MAD is a rare cause of VA/SCD. Arrhythmias secondary to MAD are more likely with abnormalities on cardiac imaging, including late gadolinium enhancement, fibrosis of the mitral annulus or papillary muscles, mitral valve pathology and ventricular ectopic beats. In this case, none of the traditional risk factors for VA/SCD were present. Thus, MAD remains an important differential in all patients with an otherwise unexplained cardiac arrest, even if high risk features are not present.

Systemic Metabolic and Volumetric Assessment via Whole-Body [18F]FDG-PET/CT: Pancreas Size Predicts Cachexia in Head and Neck Squamous Cell Carcinoma.

Background/Objectives: Cancer-associated cachexia in head and neck squamous cell carcinoma (HNSCC) is challenging to diagnose due to its complex pathophysiology. This study aimed to identify metabolic biomarkers linked to cachexia and survival in HNSCC patients using [18F]FDG-PET/CT imaging and machine learning (ML) techniques. Methods: We retrospectively analyzed 253 HNSCC patients from Vienna General Hospital and the MD Anderson Cancer Center. Automated organ segmentation was employed to quantify metabolic and volumetric data from [18F]FDG-PET/CT scans across 29 tissues and organs. Patients were categorized into low weight loss (LoWL; grades 0-2) and high weight loss (HiWL; grades 3-4) groups, according to the weight loss grading system (WLGS). Machine learning models, combined with Cox regression, were used to identify survival predictors. Shapley additive explanation (SHAP) analysis was conducted to determine the significance of individual features. Results: The HiWL group exhibited increased glucose metabolism in skeletal muscle and adipose tissue (p = 0.01), while the LoWL group showed higher lung metabolism. The one-year survival rate was 84.1% in the LoWL group compared to 69.2% in the HiWL group (p < 0.01). Pancreatic volume emerged as a key biomarker associated with cachexia, with the ML model achieving an AUC of 0.79 (95% CI: 0.77-0.80) and an accuracy of 0.82 (95% CI: 0.81-0.83). Multivariate Cox regression confirmed pancreatic volume as an independent prognostic factor (HR: 0.66, 95% CI: 0.46-0.95; p < 0.05). Conclusions: The integration of metabolic and volumetric data provided a strong predictive model, highlighting pancreatic volume as a key imaging biomarker in the metabolic assessment of cachexia in HNSCC. This finding enhances our understanding and may improve prognostic evaluations and therapeutic strategies.

A Robust [18F]-PSMA-1007 Radiomics Ensemble Model for Prostate Cancer Risk Stratification.

The aim of this study is to investigate the role of [18F]-PSMA-1007 PET in differentiating high- and low-risk prostate cancer (PCa) through a robust radiomics ensemble model. This retrospective study included 143 PCa patients who underwent [18F]-PSMA-1007 PET/CT imaging. PCa areas were manually contoured on PET images and 1781 image biomarker standardization initiative (IBSI)-compliant radiomics features were extracted. A 30 times iterated preliminary analysis pipeline, comprising of the least absolute shrinkage and selection operator (LASSO) for feature selection and fivefold cross-validation for model optimization, was adopted to identify the most robust features to dataset variations, select candidate models for ensemble modelling, and optimize hyperparameters. Thirteen subsets of selected features, 11 generated from the preliminary analysis plus two additional subsets, the first based on the combination of robust and fine-tuning features, and the second only on fine-tuning features were used to train the model ensemble. Accuracy, area under curve (AUC), sensitivity, specificity, precision, and f-score values were calculated to provide models' performance. Friedman test, followed by post hoc tests corrected with Dunn-Sidak correction for multiple comparisons, was used to verify if statistically significant differences were found in the different ensemble models over the 30 iterations. The model ensemble trained with the combination of robust and fine-tuning features obtained the highest average accuracy (79.52%), AUC (85.75%), specificity (84.29%), precision (82.85%), and f-score (78.26%). Statistically significant differences (p < 0.05) were found for some performance metrics. These findings support the role of [18F]-PSMA-1007 PET radiomics in improving risk stratification for PCa, by reducing dependence on biopsies.

Personalized prediction of immunotherapy response in lung cancer patients using advanced radiomics and deep learning

BackgroundLung cancer (LC) is a leading cause of cancer-related mortality, and immunotherapy (IO) has shown promise in treating advanced-stage LC. However, identifying patients likely to benefit from IO and monitoring treatment response remains challenging. This study aims to develop a predictive model for progression-free survival (PFS) in LC patients with IO based on clinical features and advanced imaging biomarkers.Materials and methodsA retrospective analysis was conducted on a cohort of 206 LC patients receiving IO treatment. Pre-treatment computed tomography images were used to extract advanced imaging biomarkers, including intratumoral and peritumoral-vasculature radiomics. Clinical features, including age, gene status, hematology, and staging, were also collected. Key radiomic and clinical features for predicting IO outcomes were identified using a two-step feature selection process, including univariate Cox regression and chi-squared test, followed by sequential forward selection. The DeepSurv model was constructed to predict PFS based on clinical and radiomic features. Model performance was evaluated using the area under the time-dependent receiver operating characteristic curve (AUC) and concordance index (C-index).ResultsCombining radiomics of intratumoral heterogeneity and peritumoral-vasculature with clinical features demonstrated a significant enhancement (p < 0.001) in predicting IO response. The proposed DeepSurv model exhibited a prediction performance with AUCs ranging from 0.76 to 0.80 and a C-index of 0.83. Furthermore, the predicted personalized PFS curves revealed a significant difference (p < 0.05) between patients with favorable and unfavorable prognoses.ConclusionsIntegrating intratumoral and peritumoral-vasculature radiomics with clinical features enabled the development of a predictive model for PFS in LC patients with IO. The proposed model’s capability to estimate individualized PFS probability and differentiate the prognosis status held promise to facilitate personalized medicine and improve patient outcomes in LC.

Preoperative Prediction of Her-2 and Ki-67 Status in Gastric Cancer Using 18F-FDG PET/CT Radiomics Features of Visceral Adipose Tissue.

Aims/Background Immunohistochemistry (IHC) is the main method to detect human epidermal growth factor receptor 2 (Her-2) and Ki-67 expression levels. However, IHC is invasive and cannot reflect their expression status in real-time. This study aimed to build radiomics models based on visceral adipose tissue (VAT)'s 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography/computed tomography (PET/CT) imaging, and to evaluate the relationship between radiomics features of VAT and positive expression of Her-2 and Ki-67 in gastric cancer (GC). Methods Ninety patients with GC were enrolled in this study. 18F-FDG PET/CT radiomics features were calculated using the PyRadiomics package. Two methods were employed to reduce radiomics features. The machine learning models, logistic regression (LR), and support vector machine (SVM), were constructed and estimated by the receiver operator characteristic (ROC) curve. The correlation of outstanding features with Ki-67 and Her-2 expression status was evaluated. Results For the Ki-67 set, the area under of the receiver operator characteristic curve (AUC) and accuracy were 0.86 and 0.79 for the LR model and 0.83 and 0.69 for the SVM model. For the Her-2 set, the AUC and accuracy were 0.84 and 0.86 for the LR model and 0.65 and 0.85 for the SVM model. The LR model for Ki-67 exhibited outstanding prediction performance. Three wavelet transform features were correlated with Her-2 expression status (p all < 0.001), and one wavelet transform feature was correlated with the expression status of Ki-67 (p = 0.042). Conclusion 18F-FDG PET/CT-based radiomics models of VAT demonstrate good performance in predicting Her-2 and Ki-67 expression status in patients with GC. Radiomics features can be used as imaging biomarkers for GC.

Glymphatic dysfunction in multiple sclerosis and its association with disease pathology and disability.

The role of the glymphatic system in multiple sclerosis (MS)-related disability remains underexplored. Diffusion-tensor image analysis along the perivascular space (DTI-ALPS) offers a non-invasive method to assess glymphatic function. To evaluate glymphatic function in MS patients with lower and higher disability. This study included 118 MS patients who underwent structural, diffusion-weighted imaging, and clinical assessment. The participants were divided into lower (MS-L, n = 57) and higher disability (MS-H, n = 61) subgroups. Brain parenchymal fraction (BPF), lesion load (LL), and DTI-ALPS index were measured. Subgroup differences and correlations between DTI-ALPS index and other measures were explored. Logistic regression was performed to evaluate BPF, LL, and DTI-ALPS index in classifying lower and higher disability patients. Significant differences in DTI-ALPS index between MS-H and MS-L (d = -0.71, false discovery rate-corrected p-value (p-FDR) = 0.001) were found. The DTI-ALPS index correlated significantly with disease duration (rp = -0.29, p-FDR = 0.002) and EDSS (rsp = -0.35, p-FDR = 0.0002). It also showed significant correlations with BPF and LL. DTI-ALPS index and LL were significant predictors of disability subgroup (DTI-ALPS: odds ratio (OR) = 1.77, p = 0.04, LL: OR = 0.94, p = 0.02). Our findings highlight DTI-ALPS index as an imaging biomarker in MS, suggesting the involvement of glymphatic impairment in MS pathology, although further research is needed to elucidate its role in contributing to MS-related disability.

Exploring prognostic biomarkers in pathological images of colorectal cancer patients via deep learning.

Hematoxylin and eosin (H&E) whole slide images provide valuable information for predicting prognostic outcomes in colorectal cancer (CRC) patients. However, extracting prognostic indicators from pathological images is challenging due to the subtle complexities of phenotypic information. We trained a weakly supervised deep learning model on data from 640 CRC patients in the prostate, lung, colorectal, and ovarian (PLCO) cancer screening trial dataset and validated it using data from 522 CRC patients in the cancer genome atlas (TCGA) dataset. We created the colorectal cancer risk score (CRCRS) to assess patient prognosis, visualized the pathological phenotype of the risk score using Grad-CAM, and employed multiomics data from the TCGA CRC cohort to investigate the potential biological mechanisms underlying the risk score. The overall survival analysis revealed that the CRCRS served as an independent prognostic indicator for both the PLCO cohort (p < 0.001) and the TCGA cohort (p < 0.001), with its predictive efficacy remaining unaffected by the clinical staging system. Additionally, satisfactory chemotherapeutic benefits were observed in stage II/III CRC patients with high CRCRS but not in those with low CRCRS. A pathomics nomogram constructed by integrating the CRCRS with the tumor-node-metastasis (TNM) staging system enhanced prognostic prediction accuracy compared with using the TNM staging system alone. Noteworthy features of the risk score were identified, such as immature tumor mesenchyme, disorganized gland structures, small clusters of cancer cells associated with unfavorable prognosis, and infiltrating inflammatory cells associated with favorable prognosis. The TCGA multiomics data revealed potential correlations between the CRCRS and the activation of energy production and metabolic pathways, the tumor immune microenvironment, and genetic mutations in APC, SMAD2, EEF1AKMT4, EPG5, and TANC1. In summary, our deep learning algorithm identified the CRCRS as a prognostic indicator in CRC, providing a significant approach for prognostic risk stratification and tailoring precise treatment strategies for individual patients.