Opportunistic screening for long-term muscle wasting in critically ill patients: insights from an acute pancreatitis cohort.

OBJECTIVES: To assess the feasibility of long-term muscle monitoring, we implemented an AI-guided segmentation approach on clinically indicated Computed Tomography (CT) examinations conducted throughout the hospitalization period of patients admitted to the intensive care unit (ICU) with acute pancreatitis (AP). In addition, we aimed to investigate the potential of muscle monitoring for early detection of patients at nutritional risk and those experiencing adverse outcomes. This cohort served as a model for potential integration into clinical practice. MATERIALS: Retrospective cohort study including 100 patients suffering from AP that underwent a minimum of three CT scans during hospitalization, totaling 749 assessments. Sequential segmentation of psoas muscle area (PMA) was performed and was relative muscle loss per day for the entire monitoring period, as well as for the interval between each consecutive scan was calculated. Subgroup and outcome analyses were performed including ANOVA. Discriminatory power of muscle decay rates was evaluated using ROC analysis. RESULTS: Monitoring PMA decay revealed significant long-term losses of 48.20% throughout the hospitalization period, with an average daily decline of 0.98%. Loss rates diverged significantly between survival groups, with 1.34% PMA decay per day among non-survivors vs. 0.74% in survivors. Overweight patients exhibited significantly higher total PMA losses (52.53 vs. 42.91%; p = 0.02) and average PMA loss per day (of 1.13 vs. 0.80%; p = 0.039). The first and the maximum decay rate, in average available after 6.16 and 17.03 days after ICU admission, showed convincing discriminatory power for survival in ROC analysis (AUC 0.607 and 0.718). Both thresholds for maximum loss (at 3.23% decay per day) and for the initial loss rate (at 1.98% per day) proved to be significant predictors of mortality. CONCLUSIONS: The innovative AI-based PMA segmentation method proved robust and effortless, enabling the first comprehensive assessment of muscle wasting in a large cohort of intensive care pancreatitis patients. Findings revealed significant muscle wasting (48.20% on average), particularly notable in overweight individuals. Higher rates of initial and maximum muscle loss, detectable early, correlated strongly with survival. Integrating this tool into routine clinical practice will enable continuous muscle status tracking and early identification of those at risk for unfavorable outcomes.

Reduction of false positives using zone-specific prostate-specific antigen density for prostate MRI-based biopsy decision strategies.

OBJECTIVES: To develop and test zone-specific prostate-specific antigen density (sPSAD) combined with PI-RADS to guide prostate biopsy decision strategies (BDS). METHODS: This retrospective study included consecutive patients, who underwent prostate MRI and biopsy (01/2012-10/2018). The whole gland and transition zone (TZ) were segmented at MRI using a retrained deep learning system (DLS; nnU-Net) to calculate PSAD and sPSAD, respectively. Additionally, sPSAD and PI-RADS were combined in a BDS, and diagnostic performances to detect Grade Group ≥ 2 (GG ≥ 2) prostate cancer were compared. Patient-based cancer detection using sPSAD was assessed by bootstrapping with 1000 repetitions and reported as area under the curve (AUC). Clinical utility of the BDS was tested in the hold-out test set using decision curve analysis. Statistics included nonparametric DeLong test for AUCs and Fisher-Yates test for remaining performance metrics. RESULTS: A total of 1604 patients aged 67 (interquartile range, 61-73) with 48% GG ≥ 2 prevalence (774/1604) were evaluated. By employing DLS-based prostate and TZ volumes (DICE coefficients of 0.89 (95% confidence interval, 0.80-0.97) and 0.84 (0.70-0.99)), GG ≥ 2 detection using PSAD was inferior to sPSAD (AUC, 0.71 (0.68-0.74)/0.73 (0.70-0.76); p < 0.001). Combining PI-RADS with sPSAD, GG ≥ 2 detection specificity doubled from 18% (10-20%) to 43% (30-44%; p < 0.001) with similar sensitivity (93% (89-96%)/97% (94-99%); p = 0.052), when biopsies were taken in PI-RADS 4-5 and 3 only if sPSAD was ≥ 0.42 ng/mL/cc as compared to all PI-RADS 3-5 cases. Additionally, using the sPSAD-based BDS, false positives were reduced by 25% (123 (104-142)/165 (146-185); p < 0.001). CONCLUSION: Using sPSAD to guide biopsy decisions in PI-RADS 3 lesions can reduce false positives at MRI while maintaining high sensitivity for GG ≥ 2 cancers. CLINICAL RELEVANCE STATEMENT: Transition zone-specific prostate-specific antigen density can improve the accuracy of prostate cancer detection compared to MRI assessments alone, by lowering false-positive cases without significantly missing men with ISUP GG ≥ 2 cancers. KEY POINTS: • Prostate biopsy decision strategies using PI-RADS at MRI are limited by a substantial proportion of false positives, not yielding grade group ≥ 2 prostate cancer. • PI-RADS combined with transition zone (TZ)-specific prostate-specific antigen density (PSAD) decreased the number of unproductive biopsies by 25% compared to PI-RADS only. • TZ-specific PSAD also improved the specificity of MRI-directed biopsies by 9% compared to the whole gland PSAD, while showing identical sensitivity.

Intermittent body composition analysis as monitoring tool for muscle wasting in critically ill COVID-19 patients.

OBJECTIVES: SARS-CoV-2 virus infection can lead to acute respiratory distress syndrome (ARDS), which can be complicated by severe muscle wasting. Until now, data on muscle loss of critically ill COVID-19 patients are limited, while computed tomography (CT) scans for clinical follow-up are available. We sought to investigate the parameters of muscle wasting in these patients by being the first to test the clinical application of body composition analysis (BCA) as an intermittent monitoring tool. MATERIALS: BCA was conducted on 54 patients, with a minimum of three measurements taken during hospitalization, totaling 239 assessments. Changes in psoas- (PMA) and total abdominal muscle area (TAMA) were assessed by linear mixed model analysis. PMA was calculated as relative muscle loss per day for the entire monitoring period, as well as for the interval between each consecutive scan. Cox regression was applied to analyze associations with survival. Receiver operating characteristic (ROC) analysis and Youden index were used to define a decay cut-off. RESULTS: Intermittent BCA revealed significantly higher long-term PMA loss rates of 2.62% (vs. 1.16%, p < 0.001) and maximum muscle decay of 5.48% (vs. 3.66%, p = 0.039) per day in non-survivors. The first available decay rate did not significantly differ between survival groups but showed significant associations with survival in Cox regression (p = 0.011). In ROC analysis, PMA loss averaged over the stay had the greatest discriminatory power (AUC = 0.777) for survival. A long-term PMA decline per day of 1.84% was defined as a threshold; muscle loss beyond this cut-off proved to be a significant BCA-derived predictor of mortality. CONCLUSION: Muscle wasting in critically ill COVID-19 patients is severe and correlates with survival. Intermittent BCA derived from clinically indicated CT scans proved to be a valuable monitoring tool, which allows identification of individuals at risk for adverse outcomes and has great potential to support critical care decision-making.

Metadata-independent classification of MRI sequences using convolutional neural networks: Successful application to prostate MRI.

PURPOSE: The ever-increasing volume of medical imaging data and interest in Big Data research brings challenges to data organization, categorization, and retrieval. Although the radiological value chain is almost entirely digital, data structuring has been widely performed pragmatically, but with insufficient naming and metadata standards for the stringent needs of image analysis. To enable automated data management independent of naming and metadata, this study focused on developing a convolutional neural network (CNN) that classifies medical images based solely on voxel data. METHOD: A 3D CNN (3D-ResNet18) was trained using a dataset of 31,602 prostate MRI volumes with 10 different sequence types of 1243 patients. A five-fold cross-validation approach with patient-based splits was chosen for training and testing. Training was repeated with a gradual reduction in training data assessing classification accuracies to determine the minimum training data required for sufficient performance. The trained model and developed method were tested on three external datasets. RESULTS: The model achieved an overall accuracy of 99.88 % ± 0.13 % in classifying typical prostate MRI sequence types. When being trained with approximately 10 % of the original cohort (112 patients), the CNN still achieved an accuracy of 97.43 % ± 2.10 %. In external testing the model achieved sensitivities of > 90 % for 10/15 tested sequence types. CONCLUSIONS: The herein developed CNN enabled automatic and reliable sequence identification in prostate MRI. Ultimately, such CNN models for voxel-based sequence identification could substantially enhance the management of medical imaging data, improve workflow efficiency and data quality, and allow for robust clinical AI workflows.

Interactive Explainable Deep Learning Model Informs Prostate Cancer Diagnosis at MRI.

Background Clinically significant prostate cancer (PCa) diagnosis at MRI requires accurate and efficient radiologic interpretation. Although artificial intelligence may assist in this task, lack of transparency has limited clinical translation. Purpose To develop an explainable artificial intelligence (XAI) model for clinically significant PCa diagnosis at biparametric MRI using Prostate Imaging Reporting and Data System (PI-RADS) features for classification justification. Materials and Methods This retrospective study included consecutive patients with histopathologic analysis-proven prostatic lesions who underwent biparametric MRI and biopsy between January 2012 and December 2017. After image annotation by two radiologists, a deep learning model was trained to detect the index lesion; classify PCa, clinically significant PCa (Gleason score ≥ 7), and benign lesions (eg, prostatitis); and justify classifications using PI-RADS features. Lesion- and patient-based performance were assessed using fivefold cross validation and areas under the receiver operating characteristic curve. Clinical feasibility was tested in a multireader study and by using the external PROSTATEx data set. Statistical evaluation of the multireader study included Mann-Whitney U and exact Fisher-Yates test. Results Overall, 1224 men (median age, 67 years; IQR, 62-73 years) had 3260 prostatic lesions (372 lesions with Gleason score of 6; 743 lesions with Gleason score of ≥ 7; 2145 benign lesions). XAI reliably detected clinically significant PCa in internal (area under the receiver operating characteristic curve, 0.89) and external test sets (area under the receiver operating characteristic curve, 0.87) with a sensitivity of 93% (95% CI: 87, 98) and an average of one false-positive finding per patient. Accuracy of the visual and textual explanations of XAI classifications was 80% (1080 of 1352), confirmed by experts. XAI-assisted readings improved the confidence (4.1 vs 3.4 on a five-point Likert scale; P = .007) of nonexperts in assessing PI-RADS 3 lesions, reducing reading time by 58 seconds (P = .009). Conclusion The explainable AI model reliably detected and classified clinically significant prostate cancer and improved the confidence and reading time of nonexperts while providing visual and textual explanations using well-established imaging features. © RSNA, 2023 Supplemental material is available for this article. See also the editorial by Chapiro in this issue.

Monitoring of Cardiac Remodeling in a Mouse Model of Pressure-Overload Left Ventricular Hypertrophy with [18F]FDG MicroPET.

PURPOSE: This study aims to analyze the left ventricular function parameters, scar load, and hypertrophy in a mouse model of pressure-overload left ventricular (LV) hypertrophy over the course of 8 weeks using 2-deoxy-2-[18F]fluoro-D-glucose ([18F]FDG) micro-positron emission tomography (microPET) imaging. PROCEDURES: LV hypertrophy was induced in C57BL/6 mice by transverse aortic constriction (TAC). Myocardial hypertrophy developed after 2-4 weeks. ECG-gated microPET scans with [18F]FDG were performed 4 and 8 weeks after surgery. The extent of fibrosis was measured by histopathologic analysis. LV function parameters and scar load were calculated using QGS®/QPS®. LV metabolic volume (LVMV) and percentage injected dose per gram were estimated by threshold-based analysis. RESULTS: The fibrotic tissue volume increased significantly from 4 to 8 weeks after TAC (​1.67 vs. 3.91  mm3; P = 0.044). There was a significant increase of the EDV (4 weeks: 54 ± 15 µl, 8 weeks: 79 ± 32 µl, P < 0.01) and LVMV (4 weeks: 222 ± 24 µl, 8 weeks: 276 ± 52 µl, P < 0.01) as well as a significant decrease of the LVEF (4 weeks: 56 ± 17 %, 8 weeks: 44 ± 20 %, P < 0.01). The increase of LVMV had a high predictive value regarding the amount of ex vivo measured fibrotic tissue (R = 0.905, P < 0.001). The myocardial metabolic defects increased within 4 weeks (P = 0.055) but only moderately correlated with the fibrosis volume (R = 0.502, P = 0.021). The increase in end-diastolic volume showed a positive correlation with the fibrosis at 8 weeks (R = 0.763, P = 0.017). CONCLUSIONS: [18F]FDG-PET is applicable for serial in vivo monitoring of the TAC mouse model. Myocardial hypertrophy, the dilation of the left ventricle, and the decrease in LVEF could be reliably quantified over time, as well as the developing localized scar. The increase in volume over time is predictive of a high fibrosis load.

Attenuation of cardiac hypertrophy by G-CSF is associated with enhanced migration of bone marrow-derived cells.

Granulocyte-colony stimulating factor (G-CSF) has been shown to promote mobilization of bone marrow-derived stem cells (BMCs) into the bloodstream associated with improved survival and cardiac function after myocardial infarction. Therefore, the aim of the present study was to investigate whether G-CSF is able to attenuate cardiac remodelling in a mouse model of pressure-induced LV hypertrophy focusing on mobilization and migration of BMCs. LV hypertrophy was induced by transverse aortic constriction (TAC) in C57BL/6J mice. Four weeks after TAC procedure. Mice were treated with G-CSF (100 µg/kg/day; Amgen Biologicals) for 2 weeks. The number of migrated BMCs in the heart was analysed by flow cytometry. mRNA expression and protein level of different growth factors in the myocardium were investigated by RT-PCR and ELISA. Functional analyses assessed by echocardiography and immunohistochemical analysis were performed 8 weeks after TAC procedure. G-CSF-treated animals revealed enhanced homing of VLA-4(+) and c-kit(+) BMCs associated with increased mRNA expression and protein level of the corresponding homing factors Vascular cell adhesion protein 1 and Stem cell factor in the hypertrophic myocardium. Functionally, G-CSF significantly preserved LV function after TAC procedure, which was associated with a significantly reduced area of fibrosis compared to control animals. Furthermore, G-CSF-treated animals revealed a significant improvement of survival after TAC procedure. In summary, G-CSF treatment preserves cardiac function and is able to diminish cardiac fibrosis after induction of LV hypertrophy associated with increased homing of VLA-4(+) and c-kit(+) BMCs and enhanced expression of their respective homing factors VCAM-1 and SCF.