Diagnostics

Objectives To assess the effectiveness of our proposed urinary segmentation method based upon image processing, and to contrast its performance against various approaches. Subjects and Methods Non-contrast-enhanced computed tomography (NCCT) scans of instances with stone disease were collected. An experienced urologist and a senior radiology resident generated ground truths by manually drawing the regions of interest (ROIs) of stone samples. The proposed method took loosely defined ROIs as input and included three steps: (1) finding mean_HU_rim, the mean Hounsfield unit (HU) of the stone rim, on each NCCT slice by an edge detection operator; (2) considering pixels with HU higher than the minimum value of all the mean_HU_rim; and (3) removing noise. The entire data was divided into a developing set of 406 samples to evaluate the performance of the proposed method as well as to fine-tune a nnUNetv2 model and a comparison test set of 125 samples to compare the performance and inference time of various segmentation methods. In addition, the reproducibility of our method was tested on the 30 samples randomly selected from the entire data. Results A total of 531 stone samples from 287 instances were included. Our method showed a high level of agreement with the ground truth in the developing set, with the median (interquartile range) Dice coefficient of 0.86 (0.81 - 0.89). The reproducibility of our method was remarkably high with the median (interquartile range) Dice coefficient of 1.0 (0.991 - 1.0). On the comparison test set, our semi-automatic segmentation outperformed the fixed thresholding and deep-learning-based methods. Lastly, the inference time of the proposed method was less than a second. Conclusion The proposed semi-automatic approach provided a consistently reliable, simple, and robust method to segment urinary stones.

Background: A 2019 report from our institution described a multilayer artificial neural network (ANN) for predicting prostate cancer at biopsy in 334 patients, trained with TensorFlow 1.x and evaluated at three fixed step counts without separating hyperparameter selection from test evaluation. We re-analyzed an expanded cohort from the same institution using contemporary machine-learning practice. Methods: We pooled all available biopsy episodes from the same institutional database (n = 526; 524 after excluding one non-binary outcome code and one record with missing digital rectal examination [DRE] data), retaining the same seven predictors used in the original report (age, prior biopsy history, PSA, prostate volume, DRE, and MRI diffusion-weighted imaging findings in the peripheral and transition zones). Because 27 patients contributed more than one biopsy episode, we used patient-ID-grouped, stratified k-fold cross-validation (StratifiedGroupKFold; scikit-learn 1.8.0) with 3 and 5 folds, repeated over 10 random partitions, to avoid leakage between folds. Four classifiers were compared: L2-regularized logistic regression, gradient boosting, random forest, and a shallow (single hidden layer) multilayer perceptron. Two outcomes were modeled: detection of any prostate cancer, and detection of clinically significant prostate cancer (Gleason score [≥] 7). Results: Any-cancer prevalence was 55.7% (292/524) and Gleason score [≥] 7 prevalence was 39.7% (208/524). With repeated 5-fold cross-validation, gradient boosting gave the highest discrimination for any prostate cancer (mean AUC 0.826, 95% CI 0.823-0.830) and for Gleason score [≥] 7 (mean AUC 0.855, 95% CI 0.852-0.859), closely followed by random forest and logistic regression (AUC 0.81-0.85). The shallow multilayer perceptron performed worse and less consistently than the other three models (any-cancer AUC 0.671; Gleason score [≥] 7 AUC 0.742) and than the deeper five-hidden-layer ANN reported in 2019. Results with 3-fold cross-validation were essentially unchanged. Conclusions: In an expanded cohort, regularized logistic regression, gradient boosting, and random forest all discriminated prostate cancer at biopsy at least as well as the previously reported multilayer ANN, using far simpler models and a methodology that separates hyperparameter tuning from performance estimation. A shallow neural network offered no advantage over these simpler alternatives in this sample size. This is a preprint; the study has not undergone external peer review.

Background and Objective: There is substantial variation in reference ranges for 24-hour urine collections, leading to diagnostic uncertainty. We aimed to curate robust evidence for expected ranges of 24-hour urine collections in the non-kidney stone forming population. Methods: We performed a systematic review (PROSPERO ID: CRD42024590784) of healthy adults without kidney stone disease undergoing 24-hour urine collections. Each component was meta-analysed in R, stratified by age, sex and ancestry, meta-regression and sensitivity analyses for study design, collection completeness and bias. An explicit eight-branch decision framework selected the recommended analysis per component. Expected ranges were derived from the pooled SD (95% interval); evidence was quantified by GRADE. Key Findings and Limitations: Seventy-two articles were included (10,284 participants). Expected ranges (mmol / 24 h unless stated): volume 1.0-2.5L, calcium 1.9-6.2, oxalate 0.2-0.6 (men) / 0.2-0.4 (women), urate 2.0-5.1(men)/2.3-3.5(women), pH 5.6-6.3, citrate 1.2-4.5, creatinine 12.2-16.3(men)/7.2-12.6(women), phosphate 19-26, sodium 98-253, potassium 28-77, magnesium 3.1-4.8, ammonium 28-46(men)/20-40(women), chloride 84-218, sulphate 17-26(men)/7-30(women), urea 234-445. Principal limitations were: high between-study heterogeneity (I-squared>95% for most components) and assay/laboratory variability (most marked for citrate and potassium). Conclusions and Clinical Implications: This is the first study to attempt a robust characterization of 24-hour urine expected ranges using meta-analysis. Six of the derived expected ranges (pH, oxalate, magnesium, sulphate, urate, creatinine) were broadly concordant with existing reference ranges, providing an evidence base for those specific components.

Obstetric anal sphincter injury (OASI) requires immediate postpartum recognition for primary repair. Diagnosis relies primarily on clinical examination, as endoanal ultrasound (EAUS) is rarely a bedside option. Therefore, objective diagnostic assessment is most needed during the first postpartum hours. However, it is challenging to obtain the anatomically verified data needed to develop diagnostic technologies for this clinical window. We evaluated the translation of a machine-learning-assisted impedance spectroscopy model into its intended immediate postpartum setting. The model lifecycle comprised three sequential phases: V1 development using anatomically verified data from a prospective multicenter study (N=152); V2 real-world adaptation, comprising post-market data collection, retraining (N=144) to adapt to acute physiological tissue changes, and subsequent separate evaluation (N=125); and prospective external validation of V2 in an unseen cohort (N=70). During this separate evaluation, V2 achieved 89.6% accuracy. In external validation, all examinations occurred within 4 hours of birth (94.3% within the first hour). Against subsequent EAUS, V2 achieved 86.2% sensitivity, 92.7% specificity, and 90.0% overall accuracy. The V2 model was prospectively externally validated in its intended immediate-postpartum setting, supporting ML-assisted impedance spectroscopy as an objective adjunct, and not a replacement, to clinical OASI assessment. Further validation is required to confirm generalizability.

Background: Indwelling urinary catheters can cause pain, urethral irritation and trauma and provide surfaces for microbial colonisation and biofilm formation. The LubriShield(R) Foley catheter incorporates a covalently bonded, superhydrophilic antifouling hydrogel designed to reduce surface friction and limit biofilm accumulation without releasing an antimicrobial agent. This study evaluated its safety and clinical performance compared with an uncoated silicone Foley catheter, including usability and patient comfort, and explored catheter-associated microbial colonisation and biofilm formation. Methods: This prospective, randomised, single-blind, single-centre controlled clinical investigation enrolled 33 adults requiring indwelling urinary catheterisation for 72 hours to 14 days. Participants received either the LubriShield(R) Foley catheter (n = 16) or an uncoated Sterimed all-silicone Foley catheter (n = 17). The primary endpoint was safety based on adverse-event reporting; the prespecified safety criterion was the absence of a serious adverse device effect in the LubriShield(R) group. Secondary endpoints included investigator-rated ease and functionality of catheter insertion and removal and participant-reported pain and irritation/discomfort assessed using 0-10 numerical rating scales. Urine microbiology, catheter colonisation and catheter-level biofilm measurements were exploratory endpoints. Analyses were performed in the as-treated population. Numerical rating scale outcomes were compared using two-sided Mann-Whitney U tests; exploratory biofilm analyses used permutation tests with bootstrap confidence intervals. Results: No serious adverse events or serious adverse device effects occurred. Nine non-serious adverse events were reported and resolved completely. At least one adverse event occurred in 4/16 (25.0%) participants in the LubriShield(R) group and 4/17 (23.5%) participants in the control group. The prespecified safety criterion, defined as the absence of a SADE in the LubriShield(R) group, was met. LubriShield(R) was rated more favourably for ease of insertion (p = 0.0096), functionality during insertion (p = 0.0051), ease of removal (p = 0.0037) and functionality during removal (p = 0.0085). Participants reported significantly less pain and irritation/discomfort during insertion, while the catheter was indwelling and during removal (all p < 0.05). Outcomes 7 {+/-} 1 days after catheter removal were similar between groups. Absolute biofilm biomass and bacterial and fungal loads did not differ significantly. In a post hoc exploratory normalised analysis, the biofilm-to-bacterial-load ratio was lower with LubriShield(R) (median 0.052 vs 0.080; permutation p = 0.043), although the bootstrap 95% confidence interval for the between-group difference included zero. Conclusions: The LubriShield(R) Foley catheter demonstrated a favourable short-term safety profile and was associated with improved investigator-rated usability and participant-reported comfort compared with an uncoated silicone Foley catheter. A post hoc exploratory analysis was compatible with reduced biofilm accumulation relative to bacterial load, but this finding is hypothesis-generating only and requires confirmation in larger, adequately powered studies with prespecified biofilm endpoints.

Ovarian cancer is recognized as the deadliest gynecological malignancy. Diagnosis at advanced stages and the lack of effective screening program lead to poor survival rates, dropping to 17-39 % in stage III-IV diseases. Ultrasound (US) is the primary imaging modality for ovarian structures evaluation, but it is strongly affected by the operator expertise due to the complexity of adnexal masses and the physiological variability of ovarian morphology throughout a womans lifecycle. The International Ovarian Tumor Analysis (IOTA) group introduced definitions and predictive tools to standardize gynecological US interpretation. However, these tools still rely on subjective interpretation, thus highlighting the need for more objective solutions. Recent studies have explored artificial intelligence (AI) algorithms for gynecological US, mainly focusing on adnexal masses classification. Conversely, a robust solution supporting the identification and description of healthy and tumoral ovarian structures is still lacking. This paper proposes OvAi Focus, a framework including (i) a segmentation module for the identification of healthy ovaries, plus solid and cystic components of adnexal masses; (ii) a morphology module for the extraction of IOTA-based keywords to describe adnexal masses morphology. Segmentation module results were compared to ground truth masks, showing DICE scores from 0.62 for functional ovary to 0.87 for the whole adnexal mass. Morphology module was tested through interobserver agreement analysis, obtaining Fleiss Kappa from 0.16 to 0.57 and Percent Agreement from 47 to 90 %, in line with existing literature. OvAi Focus represents an innovative solution which could help overcoming subjectivity in gynecological US imaging interpretation.

Importance There is growing interest to avoid radical cystectomy (RC) in patients with muscle-invasive bladder cancer (MIBC) who receive neoadjuvant chemotherapy and achieve pathological complete response (ypCR). To achieve this goal, molecular biomarkers will likely need to be used to enhance clinical staging given the limitations of evaluation by cystoscopy, cytology, and cross-sectional imaging. There are no studies evaluating whether safe RC avoidance (SaRCA) would be cost effective and what impacts it would have on quality of life (QoL) and survival. Objective This study models the potential economic, QoL, and survival costs/benefits of a ypCR biomarker as it relates to SaRCA using a decision analysis and Markov Model (MM). Methods/Materials A decision tree and MM was created to compare the expected costs of initial treatment, QoL, and survival under one strategy where all patients undergo RC after neoadjuvant treatment versus an alternative strategy where all patients would be subjected to the biomarker test with biomarker-positive patients (those with presumed residual disease) undergoing RC, while biomarker-negative patients (presumed complete responders) would undergo surveillance for up to 20 years. ypCR rates to neoadjuvant therapy, survival with and without RC, quality adjusted life years (QALY), and costs were abstracted from the literature. Test cost, sensitivity, and specificity were also abstracted from the literature for multiple clinical or liquid biopsy approaches. Results Broadly, SaRCA approaches are cost effective with the exception of systematic endoscopic evaluation (SEE). All testing approaches result in higher QALY and overall life expectancy compared to no testing. The cost of the test is offset by decreased usage of RC to realize a cost savings. These domains are further improved when cisplatin-based chemotherapy is replaced with emerging neoadjuvant therapies. Conclusions and relevance Modeling supports the development of accurate biomarker tests which can distinguish residual disease states to enable SaRCA. Such a biomarker could be used to avoid an expensive and risky operation, and unexpectedly would provide a survival benefit by reducing the number of perioperative mortalities in patients achieving ypCR. Development of an accurate biomarker-based test is likely to reduce cost and increase QoL and survival. An accurate biomarker test would have utility for patients, payers, hospitals, and physicians.

Background: Biochemical recurrence (BCR) occurs in up to 40% of men following radical prostatectomy (RP). Current risk models rely primarily on clinicopathologic variables and may not fully capture the biological heterogeneity associated with recurrence. The Decipher Genomic Classifier (DGC), prostate-specific membrane antigen positron emission tomography (PSMA-PET), and multiparametric magnetic resonance imaging (mpMRI) provide complementary prognostic information that may improve prediction. Objective: To develop and evaluate machine learning (ML) models integrating DGC, PSMA-PET, and mpMRI for preoperative prediction of BCR following RP. Methods: This retrospective study included patients with available preoperative DGC, PSMA-PET, mpMRI, and clinicopathologic data. Logistic regression (LR), random forest (RF), and XGBoost models were developed using single- and multimodality feature combinations. Early- and intermediate-fusion strategies were evaluated. Performance was assessed using an area under the receiver operating characteristic curve (AUC) and accuracy. Clinical utility was evaluated using decision curve analysis. Results: XGBoost consistently outperformed LR and RF. DGC achieved the highest single-modality performance (AUC 0.94, accuracy 86.7%). Among multimodal models, DGC combined with PSMA-PET using intermediate fusion achieved the best overall performance (AUC 0.93, accuracy 87.0%). Addition of mpMRI reduced performance (AUC 0.85, accuracy 83.0%). Decision curve analysis demonstrated positive net benefit across clinically relevant thresholds. Conclusion: XGBoost-based multimodal fusion improved preoperative BCR prediction following RP. DGC was the strongest individual predictor, while integration with PSMA-PET provided the best overall performance, supporting the potential of radiogenomic ML models for personalized risk stratification.

Background: Comprehensive evaluations of lung nodule detection methods remain limited, particularly regarding data from clinical routine, novel photon-counting CT (PCCT) technology, and the comparison of academic and commercial models. Materials and Methods: This retrospective study collected 1860 routine PCCT scans acquired at Hannover Medical School from 2021 to 2024. We compared four deep-learning-based lung nodule detection models, which were developed for conventional CT, namely the publicly available TotalSegmentator, nnDetection, and RadYOLO methods, as well as a commercial lung computer-aided detection (CAD) system. We evaluated the inter-model agreement in an unannotated cohort (n=1699), sensitivity in a smaller cohort of 25 manually annotated scans, and false positives as detections in 136 report-negative scans. Results: In the unannotated cohort, the commercial CAD system detected 7071, TotalSegmentator 4264, nnDetection 5254, and RadYOLO 3122 nodules. Overall, the inter-model agreement was low, and the models produced substantially different candidate nodules. Approximately 1300 nodules were detected by all models, representing 18.7% - 42.8% of each model's detections. 15.4% - 37.4% of a model's detections did not match any other model. The models showed moderate sensitivity (between 0.71, 95%-CI [0.64, 0.77], and 0.83, 95%-CI [0.77, 0.88]), and TotalSegmentator and RadYOLO produced few false positives per scan (both 0.85, 95%-CI [0.70, 1.02]). Conclusion: The lung nodules detected by four deep learning models showed high variation and a lack of consensus on clinical PCCT data. These findings highlight the importance of model benchmarking and local validation when deploying models in real-world clinical workflows.

Background and Aims: Vibration-controlled transient elastography (VCTE) and two-dimensional shear-wave elastography (2D-SWE) are used to assess liver stiffness in metabolic dysfunction-associated steatotic liver disease (MASLD); patients may be assessed by different methods over time. Because 2D-SWE fibrosis cut-offs are not uniformly validated and no biopsy or magnetic resonance elastography reference is available, we asked whether switching methods moves patients across decision thresholds. We evaluated agreement and decision-threshold interchangeability between VCTE and 2D-SWE in MASLD with obesity. Methods: In a retrospective cross-sectional agreement study at a tertiary-care center, 317 consecutive adults with MASLD underwent same-day VCTE (FibroScan) and GE LOGIQ E9/E10 2D-SWE. Continuous agreement was assessed by Bland-Altman analysis; the categorical analysis compared binary clinical decision thresholds (VCTE [&ge;]8.0/[&ge;]10.0 kPa; 2D-SWE [&ge;]7.204/[&ge;]8.060 kPa) using Cohen {kappa} and McNemar tests. Prespecified sensitivity analyses and a post-hoc recalibration were performed; VCTE served as operational reference. Results: VCTE yielded higher values (geometric mean VCTE/2D-SWE ratio, 1.23; ratio limits of agreement, 0.64 2.40; intraclass correlation coefficient, 0.41). At the lower threshold, 71 of 117 VCTE-positive patients (61%) were below the 2D-SWE threshold versus 5 of 200 (2.5%) reclassified upward (agreement 76.0%; {kappa} 0.417; P<.001); the upper threshold was similar (38 of 63, 60%, vs 8 of 254, 3.1%). Discordance persisted across cut-offs; agreement worsened descriptively across BMI strata. Recalibration removed the directional asymmetry but not the discordance (agreement unchanged, 76.0%). Conclusions: In MASLD patients, switching from VCTE to 2D-SWE reclassified decision-threshold status in ~60% of VCTE-positive patients; recalibration removed the directional bias but not the disagreement. Follow-up should use the same elastography modality and, when possible, the same platform.

Background & AimsWe previously developed multiple machine-learning and regression-based risk prediction models using retrospective electronic health records (EHR) to estimate near-term risk of pancreatic cancer (PC). This pilot study aimed to assess the feasibility of an early detection strategy combining initial EHR-based risk prediction with downstream imaging and blood-based testing. MethodsThis prospective study included adults aged 50-84 years with an estimated [&ge;]1% 18-month risk of PC, identified between January 2021 and March 2023 using four previously validated models. Patients with a history or active suspicion of PC were excluded. Participants completed cross-sectional imaging and CA19-9 labs at baseline and again at 18 months. Outcomes included enrollment rates, test completion rate, and burden of incidental findings. ResultsOf 1,251 eligible patients, 102 (8.2%) enrolled (median age 69 years [IQR 62-75]; 58% women; 48% Hispanic, 34% White, 11% Black, and 7% Asian). At baseline, 89 (87%) completed imaging and 98 (96%) completed CA19-9. At 18 months, 79% completed both imaging and CA19-9. Ten (10%) participants underwent additional clinical evaluations triggered by imaging findings. Evaluations included an endoscopic ultrasound, renal imaging, and FibroScan. CA19-9 results were normal in most participants, though 18% had elevated levels at baseline. ConclusionsReal-time EHR-based ML algorithms can be combined with blood- and imaging-based surveillance for early detection of PC with acceptable burden of incidental findings. However, improved patient engagement is a key aspect that will need to be addressed in order to successfully scale an AI-guided strategy for early detection.

Background: Robot-assisted radical prostatectomy is the standard minimally invasive treatment for localized prostate cancer. Conventional multiport robotic surgery is widely applied yet associated with more abdominal trauma and postoperative discomfort. Single-port robotic surgery can further minimize surgical injury. The domestic KangDuo robotic system has been gradually used in urological oncology, but high-level prospective evidence comparing single-port and multiport extraperitoneal radical prostatectomy remains scarce. This trial aims to compare the efficacy and safety of these two surgical approaches. Methods: This is a single-center, prospective, 1:1 randomized, non-inferiority trial. Eligible patients diagnosed with cT1c-cT2cN0M0 localized prostate cancer will be randomized to receive either single-port (experimental group) or multiport (control group) extraperitoneal robotic radical prostatectomy using the KangDuo system. All participants will receive standardized enhanced recovery after surgery (ERAS) protocols. The primary endpoint is postoperative positive surgical margin rate. Perioperative parameters, postoperative recovery, medium-term functional recovery and oncological outcomes will be assessed comprehensively. Expected results: We hypothesize that the single-port approach is non-inferior to the multiport approach in terms of oncological radical effect. The single-port technique is expected to provide better minimally invasive benefits, with comparable perioperative safety, urinary/sexual function recovery and mid-term oncological outcomes between groups. Conclusions: This randomized trial will provide high-quality clinical evidence for the application of domestic single-port robotic surgery. The findings will facilitate individualized surgical selection for patients with localized prostate cancer. Trial Registration: Chinese Clinical Trial Registry, ChiCTR2600125669 (http://www.chictr.org.cn/showproj.aspx?proj=2600125669).

Purpose Restaging transurethral resection of bladder tumor (reTURBT) is the standard-of-care for some patients with high-risk non-muscle invasive bladder cancer (NMIBC). However, it may lead to complications and biomarkers to avoid unnecessary reTURBT are needed. Materials and Methods Patients with high-risk NMIBC undergoing reTURBT were prospectively enrolled. Whole-exome sequencing along with low-pass whole-genome sequencing were performed on index TURBT samples to generate mutational profiles and copy number variation. Up to 50 personalized mutations and a fixed panel of 500 hotspot mutations were used for detecting variants from urine samples collected prior to reTURBT. Using a prespecified algorithm, tumor fraction (TF) and copy number burden (CNB) score were used to measure minimal residual disease and correlated with reTURBT pathology. Results Overall, 72/76 urine samples collected from patients with high-risk NMIBC prior to reTURBT passed quality check. From the indext TURBT specimens, a median of 39 personalized variants from the index TURBT were used for disease tracking. Overall, median TF was 0.206 vs. 0.001 in patients with and without residual tumor, respectively (p<0.001). Similarly, CNB score was elevated (7.34 vs. 4.02, p<0.001). utDNA achieved sensitivity of 97.8% and specificity of 69.2%, with area under the curve (AUC) of 0.932 in predicting disease found on reTURBT. Analyses of the index and reTURBT tissue and urine mutational profiles provided evidence for underdiagnosis of subclinical disease at the time of reTURBT. Conclusion Paired personalized and panel utDNA can be used to accurately detect residual disease in patients with high-risk NMIBC with potentially wide-ranging clinical applications.

Background: Low-dose computed tomography (LDCT) screening reduces lung cancer mortality but is limited by low uptake and associated with high rates of false-positives and indeterminate-nodules. Breath volatile organic compound (VOC) analysis is a non-invasive candidate biomarker approach that could complement LDCT, but prior work has relied on laboratory-based high-resolution mass spectrometry (HRMS), limiting point-of-care deployment. Methods: In this pilot study, breath samples were collected from 40 patients with treatment-naive, pathologically confirmed non-small cell lung cancer (NSCLC) and 25 lung-cancer-screening-eligible healthy controls. Paired samples were analyzed via a compact point-of-care GC-MS platform (CLARION) and a laboratory HRMS reference. Diagnostic classification models were built independently for each platform using elastic net logistic regression with leave-one-out cross-validation, and performance was evaluated by area under the receiver operating characteristic curve (AUC). Results: CLARION identified 103 VOCs across breath specimens, compared to over 900 identified by HRMS. Despite this difference in panel size, CLARION achieved diagnostic performance nearly identical to HRMS for distinguishing NSCLC cases from controls (AUC 0.864 vs. 0.863). Compared to controls, performance statistics were similar for early-stage NSCLC (AUC 0.854 vs. 0.841) and adenocarcinoma (AUC 0.770 vs. 0.787). VOCs of interest include p-cymene, phenol, propylbenzene, tetradecane, {beta}-ocimene, 2,3-dihydro-indole, and 1-methylthio-(Z)-1-propene. Conclusion: A compact, point-of-care breath GC-MS platform achieved diagnostic performance for NSCLC detection comparable to a laboratory HRMS reference despite a substantially smaller detected VOC panel. These findings support continued development of point-of-care breath VOC testing as a non-invasive, field-deployable complement to LDCT-based lung cancer screening.

Accurate classification of lung adenocarcinoma growth patterns from hematoxylin and eosin (H\&E)-stained whole-slide images is essential for treatment planning and prognosis. We develop and evaluate an exponentially weighted ensemble deep learning approach that combines five architectures (EfficientNet-B3, DeiT3, Swin Transformer, Vision Transformer, ConvNeXt) with weights determined by a softmax transformation of per-model Quadratic Weighted Kappa (QWK) scores. Using 25,545 image patches at 40x magnification from the Dartmouth-Hitchcock dataset, we evaluated performance with stratified 5-fold cross-validation. Individual model performance ranged from Cohen's kappa kappa = 0.5563 (ConvNeXt) to kappa = 0.9561 (EfficientNet-B3). The ensemble achieved 96.52% accuracy and kappa = 0.9648, exceeding the best individual model by delta kappa = +0.0087 with statistical significance (McNemar's test, chi^2 = 50.84, p < 10^-12; bootstrap 95% CI on delta kappa [+0.0056, +0.0117]. Pairwise disagreement rates of 16.6--43.2% between architectures indicate that the models capture different aspects of tissue morphology, and the framework automatically down-weights unreliable components, as illustrated by ConvNeXt receiving only 5.7% ensemble weight despite a single-fold training failure. The approach offers a clinically grounded method for ensemble aggregation in digital pathology.

Objective: The purpose of this study was to estimate the negative predictive value (NPV) of screening for group B streptococcus (GBS) colonization in pregnant patients undergoing antepartum hospitalization for GBS colonization status at the time of preterm delivery. Study Design: This prospective, observational cohort study compared GBS colonization status upon initial hospital admission to that at the time of delivery. Pregnant patients at 22 to 35 weeks gestation admitted to the antepartum unit at a tertiary care hospital underwent standard screening for GBS colonization. When preterm labor progressed or iatrogenic preterm delivery was indicated, the GBS colonization test was repeated. Comparison of the sequential test results was performed to determine the NPV of the antepartum screening test for the intrapartum status. Results: 159 eligible patients were enrolled in the study, and 100 completed the study and were included in the analysis. The average gestational age at admission was 30 weeks 1 day (95% confidence interval [CI] 29w4d to 30w6d) and the average duration of pregnancy latency in the study group was 17.5 days (95% CI 15.1 to 19.8). GBS colonization rate at the time of admission was 18% and at the time of delivery was 20%. The NPV of GBS screening at admission was 91.5% (95% CI 83.2 to 96.5%) and the positive predictive value (PPV) was 72.2% (95% CI 46.4 to 90.3%). Conclusion: In a cohort of pregnant patients with preterm pregnancy complications, GBS screening at the time of antepartum hospital admission has an NPV of 91.5% (95% CI 83.2 to 96.5%) for GBS colonization at the time of preterm delivery. This is comparable to the published NPV of routine GBS screening for colonization status at term delivery.

Purpose: Videourodynamics (VUDS) are used to assess bladder dysfunction and risk of upper urinary tract deterioration in patients with spina bifida (SB), but interpretation is subject to interrater variability. Machine learning models have previously classified bladder dysfunction severity and predicted incident hydronephrosis using VUDS data. This study evaluated the performance of adapted versions of these models in an independent external cohort. Materials and Methods: VUDS data were collected from SB patients who underwent VUDS at a single institution between 2016 and 2025. Previously developed machine learning frameworks were adapted for implementation in the independent dataset. Three models were utilized: a deep learning convolutional neural network model using pressure-volume data, a deep learning imaging model using fluoroscopic imaging data, and an ensemble model that averaged the risk data from the pressure-volume and fluoroscopic data. The models were evaluated for prediction of incident hydronephrosis and classification of bladder dysfunction severity compared to expert pediatric urologist reviewers. Results: 70 patients were included in the hydronephrosis cohort, of whom 13 (18%) developed incident hydronephrosis. The ensemble model outperformed individual modalities, with a concordance index of 0.77 (pressure-volume only: 0.70, imaging only: 0.72) and an overall AUROC of 0.75 (pressure-volume: 0.67, imaging: 0.74). For bladder dysfunction classification in 95 VUDS studies, the ensemble model achieved 71% accuracy compared to 60% for pressure-volume data alone and 66% for imaging alone, with no substantial disagreements with expert reviewers. Conclusions: The machine learning models that were developed to predict incident hydronephrosis and classify the severity of bladder dysfunction using VUDS data were able to be adapted to an external, independent dataset and perform with similar discrimination and accuracy as previously reported.

Background: Lung cancer is leading cause of cancer related death globally. It is second most prevalent cancer among women worldwide and ranks third among females in Nepal. Contributing factors are smoking, tobacco use, air pollution, and delayed diagnosis. Image-guided fine needle aspiration cytology (FNAC) is rapid diagnostic technique for evaluating lung lesions. It is minimally invasive procedure with less complications. This study examine histocytologic makeup of lung lesions and link the results. Materials and Methods: This cross-sectional observational study included 65 patients irrespective of age and sex presenting with lung masses at Chitwan Medical College and Teaching Hospital from April 2023 to September 2024. After clinical and radiologic evaluation, all cases underwent image-guided FNAC and biopsy. Only specimens with unequivocal malignant features were classified positive. Histopathology served as diagnostic reference standard. Results: FNAC diagnosed 90.8% as malignant and 9.2% as benign. Biopsy confirmed malignancy in 92.3% of cases. FNAC demonstrated a sensitivity of 98.33%, specificity of 100%, positive predictive value(PPV) of 100%, and negative predictive value (NPV) of 83.33%. Concordance between FNAC and histopathological subtyping was 98.46%. Adenocarcinoma was most common subtype, followed by Squamous cell carcinoma(SCC) and small cell carcinoma. Smoking was most common contributing factor associated with malignancy. Conclusion and implications: Image-guided FNAC is an excellent diagnostic accuracy tool which possess higher level of concordance with biopsy in evaluating lung masses. It should be considered as frontline diagnostic tool, especially in resource limited settings. Keywords: FNAC, Lung cancer, Biopsy, SCC, Adenocarcinoma, Small cell carcinoma, Nepal

Purpose: To quantify the sampling error of conventional 2D regions of interest (ROIs) compared with whole-organ 3D volumetric CT attenuation for opportunistic hepatic steatosis screening, and to evaluate its dependency on intraparenchymal spatial heterogeneity. Materials and Methods: In this retrospective computational study (CLAIM-compliant), 971 abdominal CT scans were analyzed across a multicenter derivation cohort (TotalSegmentator v2, n = 841) and an external validation cohort (Medical Segmentation Decathlon Task03, n = 130). Automated 3D liver segmentation was performed using nnU-Net, followed by an 8-mm spherical 3D erosion to eliminate peripheral partial volume effects and major hilar vessels. Three clinical sampling paradigms (1, 3, and 5 axial 2D ROIs of 15 mm diameter) were modeled through 500 stochastic Monte Carlo iterations per scan (485,500 total simulations). Absolute sampling error (|Delta HU| = |HU_ROI - HU_3D|) and diagnostic misclassification rates at the standard <= 40 HU steatosis threshold were computed. Results: In the derivation cohort, the conventional 3-ROI approach yielded a mean absolute error of 6.61 +/- 7.83 HU (median: 2.92 HU), significantly underperforming compared to the 5-ROI strategy (5.21 +/- 6.00 HU; Wilcoxon paired test, p < 0.001) but improving on the single-ROI approach (11.32 +/- 14.16 HU; p < 0.001). Sampling error correlated strongly with parenchymal spatial heterogeneity (Pearson r = 0.850, 95% CI: [0.830-0.868], p < 0.001). In borderline attenuation zones (35-45 HU), diagnostic misclassification reached 45.2% in heterogeneous livers. Findings were corroborated on external validation, confirming the reproducible decay of sampling error (|Delta HU| = 7.38 HU for 1 ROI vs. 5.14 HU for 3 ROI, p < 0.001). Conclusion: Manual 2D ROI sampling exhibits an intrinsic metrological uncertainty exceeding clinically relevant thresholds (5 HU) in heterogeneous livers, misclassifying nearly half of borderline cases. Opportunistic screening requires fully automated 3D volumetric extraction embedded into clinical PACS workflows.

Purpose: To prospectively validate a semi-supervised learning framework with a lesion-only teacher model (RG-SSL-LOC) for scalable clinically significant prostate cancer detection on biparametric MRI (bpMRI) and assess its added value in multimodal models. Materials and Methods: A multicenter dataset of 13,706 bpMRI examinations (13,630 patients, 27 centers) was used for model development/validation. Three segmentation models (fully supervised learning [FSL], a state-of-the-art report-guided semi-supervised approach [RG-SSL], and the proposed RG-SSL-LOC) were evaluated at lesion- and case-level on external retrospective, external prospective, and internal prospective cohorts. Predictions from the best-performing model were combined with clinico-radiologic variables in a multimodal approach. All case-level results were compared with PI-RADS. Results: At lesion level, RG-SSL-LOC achieved higher median Dice than FSL and RG-SSL (0.49 vs 0.41 and 0.40; both p<.001). At case level, RG-SSL-LOC achieved area-under-the-curve (AUC) values of 0.83, 0.82, and 0.87 in the external retrospective, external prospective, and internal prospective cohorts, respectively. Compared with FSL, AUCs were 0.84 (p=.237), 0.80 (p=.020), and 0.84 (p<.001); compared with RG-SSL, AUCs were 0.83 (p=.929), 0.82 (p=.652), and 0.86 (p=.007); compared with PI-RADS, AUCs were 0.78 (p=.055), 0.83 (p=.652) and 0.86 (p=.480). Combined with clinico-radiological variables, RG-SSL-LOC significantly improved AUC versus clinico-radiological variables alone in the external retrospective (0.85 vs 0.80, p=.002), external prospective (0.87 vs 0.84, p=.008), and internal prospective (0.91 vs 0.88, p<.001) cohorts; in the latter, it reduced unnecessary biopsies by 15.19%. Conclusion: RG-SSL-LOC achieves better segmentation quality than other methods, demonstrates robust prospective multicenter performance and improves multimodal detection.