Search Results (1,292)

Background/Objectives: Endometrial polyps are a prevalent pathological condition within the uterine cavity. While the majority of lesions are classified as benign, histopathological examinations indicate a potential for malignant transformation occurring in 0.5–5% of cases. The present study aimed to evaluate the prevalence of endometrial carcinoma and pre-malignant histopathological findings (glandular hyperplasia without atypia and atypical glandular hyperplasia) associated with endometrial polyps and to identify pertinent risk factors among Polish women undergoing hysteroscopic polypectomy under local anesthesia. Methods: A retrospective cohort study was conducted at the Outpatient Hysteroscopy Center of the Heliodor Swiecicki University Hospital of Gynecology and Obstetrics, affiliated with Poznan University of Medical Sciences. The study included patients treated using the GUBBINI mini-resectoscope under local anesthesia with the Hystero-Block system (Tontarra Medizintechnik GmbH, Wurmlingen, Germany) between December 2022 and July 2026. Statistical analysis of risk factors was performed using the Kruskal–Wallis H test and Fisher’s exact test. Odds ratios for potential risk factors were also calculated. Results: The study comprised 2588 patients. Endometrial carcinoma associated with endometrial polyps was identified in 16 women (0.62%). Additional histopathological findings included glandular hyperplasia in 360 patients (13.91%) and atypical hyperplasia in 35 patients (1.35%), with the remaining patients diagnosed with benign uterine polyps (84.12%). Patients with malignant and pre-malignant lesions were significantly older than those with benign polyps (p < 0.001). Regarding polyp size, patients diagnosed with glandular hyperplasia without atypia had significantly larger polyps than those with benign lesions (p < 0.001). Estimated postmenopausal status was associated with significantly higher odds of endometrial carcinoma (Odds Ratio [OR] = 6.56, 95% Confidence Interval [CI]: 2.12–20.30, p = 0.001) and atypical glandular hyperplasia (OR = 3.13, 95% CI: 1.41–6.96, p = 0.005) compared to benign polyps. Similarly, obesity increased the odds of glandular hyperplasia without atypia (OR = 1.32, 95% CI: 1.01–1.74, p = 0.045) and atypical glandular hyperplasia (OR = 3.33, 95% CI: 1.67–6.65, p < 0.001). Furthermore, hypertension was associated with higher odds of glandular hyperplasia without atypia compared to benign polyps (OR = 1.41, 95% CI: 1.02–1.97, p = 0.040). Other evaluated risk factors, including type II diabetes mellitus, infertility, abnormal uterine bleeding, breast cancer history, hypothyroidism and polyendocrine metabolic ovarian syndrome (PMOS), did not reach statistical significance (p > 0.05). This large cohort provides important data regarding the prevalence and risk factors of carcinoma associated with endometrial polyps and pre-malignant uterine lesions in Polish women. Conclusions: Endometrial carcinoma associated with endometrial polyps is a rare entity among Polish women undergoing outpatient hysteroscopy (0.62%). Nevertheless, the primary diagnostic value of this large-scale study lies in demonstrating that despite the low prevalence of malignancy in standard “see-and-treat” outpatient settings, systematic histopathological evaluation remains an absolute diagnostic mandate, as relying solely on visual or clinical parameters is insufficient to exclude malignancy. Furthermore, identified risk factors, such as age, polyp size, estimated postmenopausal status, obesity and hypertension, serve as critical diagnostic red flags, associated with increased odds of concurrent premalignant lesions or carcinoma. Full article

AI-generated and locally manipulated images can support impersonation, forged evidence, identity-document abuse, and other forms of digital fraud, making reliable content-authenticity analysis increasingly important. This paper presents STeREx-Net, a three-class forensic framework for distinguishing real, fully synthetic, and locally tampered images using frozen diffusion-derived residual evidence, spatially aligned RGB features, multi-task prediction heads, and reviewable visual evidence. On the official balanced 60,000-image SID-Set test, the original model achieved 96.24% accuracy, 96.26% macro F1, and a multiclass Matthews correlation coefficient of 0.944. Separate matched three-seed ablations showed that diffusion-residual evidence is materially useful relative to RGB-only input; however, residual-only and simple-fusion controls outperformed the proposed fusion on the clean SID-Set, so fusion superiority is not claimed. Frozen robustness testing further revealed strong condition dependence: SID-Set macro F1 decreased from 0.9655 on clean images to 0.7987 under JPEG Q75 and 0.6330 under JPEG Q50, while generator-stratified AIS-4SD results also varied substantially. Zero-shot transfer to FantasyID failed to detect tampered samples, whereas leakage-safe restricted adaptation partially recovered tampered recall to 0.3447 and reduced the expected calibration error from 0.7815 to 0.2077, at the cost of lower real-image recall. A validation-selected localization intervention increased Dice from 0.2817 to 0.3413 but remained precision-biased and non-uniform across manipulation sizes. Quantitative explanation analysis supported in-domain decision faithfulness and benign-transformation stability, but these properties did not transfer consistently to external data. STeREx-Net should therefore be viewed as a human-supervised forensic research framework with explicitly characterized robustness, localization, and generalization boundaries rather than as a universally robust detector. Full article

Background: Vulvar syringoma is a rare benign adnexal neoplasm originating from eccrine sweat ducts. Although syringomas most commonly occur in the periorbital region, vulvar involvement is uncommon and frequently underrecognized. Due to its nonspecific clinical appearance, vulvar syringoma may mimic more common vulvar conditions, resulting in delayed diagnosis and repeated ineffective treatments. Case Presentation: A 41-year-old multiparous woman presented with a several-year history of recurrent pruritic papular lesions involving the right vulva. The lesions were repeatedly diagnosed as condyloma acuminata and treated with topical imiquimod, cryotherapy, carbon dioxide laser ablation, and local excision, without sustained improvement. Persistent symptoms and recurrence prompted repeat surgical excision. Histopathological examination revealed a well-circumscribed dermal lesion composed of multiple small eccrine ducts, cysts, and epithelial cords embedded within a dense sclerotic fibrotic stroma. The ducts were lined by two layers of bland cuboidal epithelial cells and demonstrated characteristic comma-shaped and tadpole-like configurations, confirming the diagnosis of vulvar syringoma. No cytologic atypia, increased mitotic activity, dysplasia, or malignancy was identified. The postoperative course was uncomplicated, with satisfactory wound healing and symptomatic improvement during follow-up. Conclusions: Vulvar syringoma should be considered in the differential diagnosis of persistent or treatment-resistant vulvar papules, particularly when lesions recur despite conventional therapy for presumed condyloma acuminata or other benign vulvar disorders. Early biopsy and histopathological evaluation remain essential for establishing an accurate diagnosis, preventing unnecessary interventions, and guiding appropriate management. Full article

Background: Ultra-high-frequency ultrasound (UHFUS), performed in the present study using 48 and 70 MHz transducers, enables high-resolution assessment of superficial tissues and may provide useful information on the morphology, internal architecture, vascularity, and anatomical relationships of oral cavity lesions. Methods: This retrospective single-center study included consecutive patients examined between January 2025 and June 2026. Of 137 eligible cases, five were excluded because of inadequate image or cine-loop quality, leaving 132 lesions for analysis. All examinations were performed by the same experienced radiologist using 48 and 70 MHz linear probes. Each lesion was assessed in B-mode in at least two orthogonal planes and with Color Doppler. Archived anonymized examinations were reviewed using a predefined structured form. Morphology, margins, echogenicity, echotexture, composition, posterior acoustic features, vascularity, dimensions, presumed plane of origin, and involvement of adjacent anatomic layers were evaluated. The qualitative synthesis was performed jointly by a radiologist and an oral pathologist, with disagreements resolved by consensus. Results: UHFUS allowed detailed visualization of mucosal, submucosal, muscular, and bone/periodontal interfaces and enabled qualitative characterization of lesion boundaries, internal structure, vascular patterns, and local extension. The combined use of 48 and 70 MHz probes provided complementary information according to lesion depth and tissue composition. Five representative histologically confirmed cases from the study cohort illustrated reactive/inflammatory, benign neoplastic, and malignant conditions. One additional clinically confirmed gingival fistula, not included in the analytical cohort, was presented solely as an illustrative example of a superficial fistulous tract detectable with UHFUS. Conclusions: UHFUS provides detailed, layer-based assessment of oral cavity lesions and may complement clinical examination and histopathology in lesion characterization and preoperative evaluation. Sonographic findings should be interpreted within a multimodal diagnostic framework rather than as stand-alone criteria. Full article

Large language model (LLM) agents that retrieve external content and use tools are vulnerable to indirect prompt injection, in which untrusted content contains instructions intended to influence agent behavior. We evaluated four defenses and an undefended control across GPT-5.4, GPT-5.4-mini, and Claude Sonnet 4.6 on the AgentDojo banking benchmark (Tool Filter was evaluated only for the OpenAI models), reporting attack success rate (ASR), benign utility, utility under attack, operational measures, and two independent benchmark replications. Raw undefended ASR was 0/288 for GPT-5.4, 11/288 for GPT-5.4-mini, and 1/288 for Claude Sonnet 4.6; these cross-model differences require cautious interpretation because benchmark goals were not equally reachable across models. For GPT-5.4-mini, the Prompt Injection Detector and Tool Filter were associated with lower observed ASRs but also lower benign utility, and Tool Filter restricted available actions. None of the four paired GPT-5.4-mini comparisons reached significance after Holm correction; only Tool Filter had an unadjusted p-value below 0.05. Benign utility was more stable across runs than individual low-frequency attack outcomes. The findings show that defense evaluation should report attack outcomes, goal feasibility, legitimate-task utility, action availability, operational measures, and run-to-run variation. Results are limited to the evaluated benchmark, models, defenses, and conditions. Full article

Background/Objectives: Müllerian duct anomalies (MDAs) are congenital structural abnormalities of the female reproductive tract with heterogeneous and incompletely defined genetic contributions. PAX8 is a developmental transcription factor implicated in thyroid and urogenital development. This study assessed the functional consequences of a rare PAX8 variant identified by whole-exome sequencing (WES) in an MDA cohort, without presuming a causal relationship. Methods: WES was performed in 150 patients with MDAs. The PAX8 c.791T>C (p.Leu264Pro) variant was evaluated using computational predictions, cellular assays, and RNA sequencing in a 293FT transient-overexpression model. Results: The heterozygous p.Leu264Pro variant was identified in one patient with a complex septate uterine, cervical, and vaginal anomaly and was absent from 120 controls. Parental samples were unavailable. No statistically significant difference in construct-derived PAX8 mRNA abundance was detected between PAX8-WT and PAX8-L264P, and both proteins showed predominantly nuclear localization. In the direct PAX8-WT-versus-L264P RNA-seq comparison, 76 transcripts had nominal p values below 0.05, but none remained significant after Benjamini–Hochberg correction, and no gene met the prespecified differential-expression criteria. AK5, RCBTB2, and IFT88 were identified as candidate PAX8-responsive genes in this experimental system. Conclusions: No major functional difference was detected between PAX8-WT and L264P under the tested conditions; however, these findings do not establish functional equivalence or exclude smaller, tissue-specific, or developmental-stage-specific effects. The p.Leu264Pro variant remains a variant of uncertain significance with respect to MDAs, and the functional results were not used as benign evidence under ACMG/AMP criterion BS3. Full article

Objective: Patients with prostate-specific antigen (PSA) levels in the 4–10 ng/mL gray zone present a diagnostic challenge, as PSA alone cannot reliably distinguish clinically significant prostate cancer (csPCa) from benign conditions or indolent disease, potentially leading to unnecessary biopsies or missed clinically significant disease. This study aimed to develop a risk prediction model for csPCa in this population using clinical indicators and magnetic resonance imaging (MRI) quantitative data, and to evaluate its potential value for risk stratification and as a supplementary tool for biopsy decision-making. Methods: We retrospectively included 210 patients with PSA levels in the 4–10 ng/mL range and confirmed pathological diagnoses, who were admitted to our hospital between January 2018 and June 2025. csPCa was defined as a Gleason score ≥ 7. Patients were stratified by pathological diagnosis (csPCa vs. non-csPCa) and randomly divided into training and internal validation sets in a 7:3 ratio. The following variables were collected for analysis: age, PSA, apparent diffusion coefficient (ADC), and relative T2 value. Univariate and multivariate logistic regression were used to identify independent predictors of csPCa. We constructed a prediction model and nomogram. The discriminating ability, calibration, and stability of the model were assessed using receiver operating characteristic (ROC) curves, calibration curves, and Bootstrap internal validation. Decision curve analysis (DCA) was performed to evaluate the model’s net benefit across clinically relevant threshold probabilities. Risk stratification was performed based on predicted probabilities. Results: Of the 210 patients, 50 had csPCa and 160 had non-csPCa lesions. Univariate regression showed associations between age, ADC value, and relative T2 value. Multivariate logistic regression identified age and relative T2 value as independent predictors in the final model. Internal validation using 1000 Bootstrap resampling with optimism correction yielded a corrected AUC of 0.698 (95% CI: 0.669–0.716), with a mean optimism of only 0.016, indicating minimal overfitting. The area under the curve (AUC) of the model was 0.720 in the training set and 0.711 in the validation set. The calibration curve for the training set demonstrated good agreement between predicted and observed probabilities. DCA showed net benefits across certain threshold probabilities. Risk stratification based on predicted probability showed csPCa detection rates of 14.7%, 28.3%, and 47.4% in low-, intermediate- and high-risk groups, respectively. Conclusions: The logistic binary classification model based on age and relative T2 value may assist in risk assessment for csPCa in PSA gray zone patients. However, its moderate discrimination suggests it should be used as a supplementary tool rather than a standalone diagnostic test. Risk stratification based on predicted probability may help identify different csPCa risk populations, but given the exploratory nature of this single-center retrospective study, multi-center, large-sample, and prospective studies are required for external validation and optimization before clinical implementation. Full article

Security operations can detect that an intrusion is under way, yet they cannot say how long an ongoing attack still needs to reach a critical objective such as data exfiltration. Prior work on multi-stage attacks identifies the active stage or predicts the next step, but does not estimate the residual time to compromise from real traffic using survival models. This paper addresses that gap through a temporal framework built on empirically measured stage durations, with three contributions. First, the MITRE ATT&CK taxonomy is given a temporal layer, in which each stage carries a duration distribution estimated empirically from the observed episodes of that stage. Second, a probability-weighted multi-path formulation combines these durations with stage-transition probabilities to estimate the time remaining before the objective. Third, the framework is validated on a real multi-stage campaign rather than on synthetic traffic, and three survival models are compared under a matched protocol as a benchmark of how learnable the durations are. Random Survival Forest, DeepSurv, and DeepHit are compared on DAPT 2020, a public advanced-persistent-threat dataset of 82,577 real network flows collected across five days. Random Survival Forest reaches a stable concordance index of 0.92, with a standard deviation of 0.006 across twenty repeated stratified splits on leakage-free features, and it retains a concordance of 0.79 when benign traffic is excluded entirely. When the three models are placed on a single concordance scale and trained on an identical subsample of 40,000 flows, DeepSurv reaches 0.955 and DeepHit 0.879, so the neural models are competitive at that scale. DeepSurv nevertheless fails to converge on the full flow set, returning no survival estimates in any of five seeds, whereas the forest fits successfully at every training size examined. A stage-transition graph recovered from the data, built from 25 observed transitions across ten multi-stage sessions, shows branching progression, and the residual time, reported at the entry to each stage, falls along the campaign, from about 139 min at reconnaissance to about 31 min at lateral movement, conditional on reaching the objective. All stage-level estimates rest on 74 episodes from a single campaign, of which 43 carry a positive duration, so cross-environment generalisation remains to be confirmed. The framework gives a security operations centre a data-driven estimate of the active attack effort that remains before compromise, supporting informed containment decisions. Full article

Background: Hysterectomy is a prevalent gynecological procedure and ensuring appropriate use of minimally invasive approaches is essential to enhance the value and quality of care. The aim of this study was to investigate the gap between algorithm-guided recommendations and routine surgical practice in a Romanian real-world cohort by identifying non-clinical factors associated with deviation from the algorithm, and assessing its impact on postoperative hospitalization. Methods: This is a retrospective analytical observational study conducted over a 4-year period (2021–2025). The database comprised 961 cases of hysterectomy performed for benign gynecological conditions using laparoscopic, abdominal and vaginal approaches in two tertiary centers. Results: The previously developed decision algorithm was implemented in a selected subgroup of approximately 332 (34.54%) patients, while the remaining 629 (65.45%) hysterectomies were performed according to surgeon preference or institutional resource availability. The mean duration of hospitalization of the entire lot was 6.88 days (SD = 2.70). In laparoscopic surgery, deviation was not associated with prolonged hospitalization, with results indicating a null effect, OR 1.08 (95% CI 0.72–1.62). In contrast, in the abdominal cohort, deviation was strongly associated with prolonged LOS, with OR 2.76 (95% CI 1.75–4.40). Conclusions: Patient-related factors shape the theoretical suitability of a surgical approach, but operator- and institution-related factors determine whether that option is realistically available. Strengthening training in minimally invasive surgery across all clinics could reduce selection bias. Full article

Myocardial bridging (MB) is a congenital coronary anomaly in which a segment of an epicardial coronary artery courses intramurally beneath the myocardium for part of its length, resulting in dynamic compression during systole. Once considered a largely benign incidental finding, accumulating evidence has demonstrated MB to be a clinically significant substrate for myocardial ischemia, arrhythmia, and sudden cardiac death (SCD), particularly under conditions of elevated adrenergic drive such as vigorous physical exercise. The prevalence of MB is highly dependent on the diagnostic modality used, ranging from approximately 0.5–2.5% on coronary angiography to 15–85% on post-mortem and advanced intravascular imaging studies, reflecting a substantial detection gap- in particular on routine cardiac screening tests. This review critically examines the current evidence base regarding MB in the context of competitive and recreational sport. We address: (1) the pathophysiology of exercise-induced hemodynamic compromise in MB; (2) the limitations of existing screening modalities—including electrocardiography, echocardiography, exercise stress testing, and advanced coronary imaging in identifying athletes at risk; (3) the clinical management of symptomatic MB, encompassing pharmacological therapy, coronary revascularization strategies, and catheter-based interventions; and (4) physical activity guidance for athletes with known or suspected MB. Full article

Colorectal cancer (CRC) is the third most commonly diagnosed cancer and the second leading cause of cancer-related mortality worldwide. Although screening has reduced CRC in older adults, cases in younger individuals are rising, highlighting the need for early biomarkers. Emerging research highlights the role of estrogen metabolism in CRC progression, with enzymes such as hydroxysteroid (17-beta) dehydrogenase (HSD17B) being increasingly implicated. In this study, we performed a bioinformatics analysis using publicly available datasets, including The Cancer Genome Atlas Colon Adenocarcinoma (TCGA-COAD) cohort and two independent Gene Expression Omnibus (GEO) cohorts (GSE40967 and GSE41258), to investigate the role of HSD17B enzymes in CRC. Our results suggest that HSD17B2 is frequently downregulated in precancerous lesions and early-stage CRC, which may contribute to elevated estradiol levels and a tumor-promoting microenvironment. In advanced stages, higher HSD17B2 expression levels are associated with poorer survival outcomes in retrospective cohorts. Other HSD17B enzymes also exhibit significant expression changes, further complicating the hormonal landscape of CRC. In addition, estrone, traditionally considered a weaker estrogen, emerges as a potential driver of CRC progression. Our in-silico analyses indicate that HSD17B2 and HSD17B11 warrant further investigation as candidate biomarkers for distinguishing CRC from benign and precancerous conditions, with the combination showing strong discriminatory power in Receiver Operating Characteristic (ROC) analyses. Overall, these findings highlight the potential role of estrogen metabolism in CRC and suggest that HSD17B enzymes may hold value as candidate prognostic and diagnostic indicators, though their clinical utility remains hypothetical at this stage. Experimental and clinical validation is strictly required to confirm these in silico observations and to clarify their mechanisms in CRC. Full article

Background: Unexplained cytopenia represents a common diagnostic challenge in clinical haematology, reflecting either benign conditions or early manifestations of myeloid neoplasms. Methods: In this retrospective cross-sectional study, we analysed 557 patients with cytopenia who underwent initial diagnostic evaluation. The diagnostic performance of the initial morphological assessment of the bone marrow and ancillary studies excluding genetic analyses was evaluated against a genetically supported reference diagnosis established by a multidisciplinary haematology tumour board. We examined the extent to which the initial morphological assessment could reliably detect or rule out myeloid neoplasia. The analysis was performed for the entire cohort and stratified by specific myeloid neoplasm entities. Results: In the overall cohort, the initial morphological assessment of the bone marrow showed high diagnostic performance (sensitivity, 0.864; specificity, 0.934). The misclassification analysis revealed more false negatives than false positives (31 vs. 21), with no significant asymmetry. In the entity-specific analysis, a sensitivity of 1.0 was observed for acute myeloid leukaemia, with a reduced specificity (0.400). Myelodysplastic neoplasms exhibited a higher rate of false-negative findings, whereas myelodysplastic/myeloproliferative overlap neoplasms were characterised by very high sensitivity (0.982) and limited specificity (0.632). A comparable diagnostic pattern was observed for myelofibrotic neoplasms. Conclusions: Initial morphological assessment demonstrates high overall diagnostic performance in cytopenia but shows important variation across individual entities. These findings highlight the need to consider entity-specific limitations and the role of genetic confirmation in the diagnostic work-up of specific entities. Full article

This study investigates the atmospheric corrosion behavior of high-lead bronze alloys within a simulated aggressive environment characterized by high humidity, acidity, chlorides, oxygen, and CO₂. The evolution of surface morphology and corrosion products was systematically monitored using colorimetric analysis and micro-Raman spectroscopy. Results indicate that the initial patina primarily comprised cuprite (Cu₂O) and cassiterite (SnO₂), which are predominantly benign phases. During the early corrosion stage, lead oxidation and carbonation prevailed, generating abundant bright-white cerussite. Subsequently, as copper-driven corrosion became dominant, these white deposits diminished and were progressively replaced by characteristic green “bronze disease”, identified as atacamite (Cu₂(OH)₃Cl). The findings reveal that preferential lead corrosion is likely to induce localized pitting, thereby accelerating degradation of the copper substrate. Consequently, higher lead content may reduce the overall corrosion resistance of bronze artifacts under these specific conditions. These results offer experimental insights into atmospheric corrosion mechanisms and inform the development of evidence-based conservation strategies for bronze cultural heritage. Full article

Background/Objectives: In Germany, women have direct access to gynaecologists without GP referral, making the GP’s role in female-specific conditions largely secondary. It has not previously been examined whether the sex of the GP is nonetheless associated with the documented prevalence of these conditions in primary care records. Methods: This retrospective cross-sectional study used data from the IQVIA™ Disease Analyzer. All women aged ≥18 years with at least one visit to one of 1183 GPs in Germany in 2025 were included. Women attending female and male GPs were matched 1:1 by propensity score on age, statutory health insurance, practice size, consultation frequency in 2025, and the van Walraven comorbidity score (excluding malignancy), yielding 661,485 women per group. Associations between GP sex and the documented prevalence of eight female-specific conditions were examined using conditional logistic regression within the age strata clinically relevant to each condition (from 18–30 to >75 years), with the age × GP-sex interaction tested in a single model per condition and E-values calculated to assess robustness to unmeasured confounding. Results: The association between GP sex and documented prevalence was strongly age-dependent, with a significant age × GP-sex interaction for seven of eight conditions (all p < 0.001). Female GP sex was associated with higher documented prevalence in mid-life, including menopausal and perimenopausal disorders (46–60 years: OR 1.37, 95% CI 1.28–1.46), endometriosis (31–45 years: OR 1.15, 1.03–1.28) and benign neoplasm of the female breast and genital organs (46–60 years: OR 1.16, 1.04–1.29); in the oldest women, the pattern attenuated or reversed (menopausal disorders > 75 years: OR 0.81, 0.72–0.90). No association was observed for breast cancer or female genital organ cancer at any age. Conclusions: In German primary care, the association between GP sex and the documented prevalence of female-specific disorders is age-dependent—most pronounced for conditions that require active clinical recognition and patient disclosure in mid-life, and absent for specialist-managed malignancies. Full article

Background/Objectives: Circulating tumor-associated cells (CTACs) are rare among peripheral blood nucleated cells (PBNCs), creating a challenge for image-based multi-cancer detection. We evaluated a predefined CTAC-detection pipeline incorporating Attention U-Net segmentation, post-processing, cytological feature extraction, and Random Forest classification. Methods: Model suitability was explored in asymptomatic individuals and patients with advanced solid tumors. Clinical performance was assessed in a case–control cohort of therapy-naive stage I/II cancers, benign conditions, and asymptomatic individuals, followed by four prospective cohort evaluations performed within the same laboratory and imaging workflow: recurrent cancer with low radiological tumor burden, peri-operative solid tumors, suspected cancer, and asymptomatic screening. PBNCs were stained with EpCAM/Hoechst 33342 and imaged. Pathologists’ review established ground truth annotations. Results: The model had 90.68% sensitivity and 99.53% specificity in the exploratory study. In the case–control cohort, sensitivity was 88.65% in therapy-naive stage I/II cancers, while specificity was 78.95% in benign conditions and >99.9% in asymptomatic individuals. In the prospective cohorts, CTAC detection sensitivity was 91.96% in pretreated low tumor burden cases; CTACs were detected in 100% of pre-surgery specimens and 29.41% of post-surgery specimens; and in suspected cancer cases, the Positive Predictive Value (PPV) and Negative Predictive Value (NPV) were 96.34% and 32.35%, respectively. In the asymptomatic screening cohort, 44/7183 participants were CTAC-positive; 16 had confirmed Stage I/II cancer, 10 had no radiologically detectable disease at the available assessment, and 18 remained unresolved. The conservative lower-bound PPV was 36.36%, and the NPV was 99.97%; estimates remain provisional pending complete follow-up. Conclusions: The integrated Attention U-Net/feature-based classification pipeline demonstrated consistent CTAC detection across the evaluated cohorts and supports its potential clinical utility for cancer detection. Full article