Background: Few predictive models have included circulating tumor DNA (ctDNA) indicators to predict prognosis of esophageal squamous cell carcinoma (ESCC) patients. Here, we aimed to explore whether ctDNA can be used as a predictive biomarker in nomogram models to predict the prognosis of patients with ESCC.Methods: We included 57 patients who underwent surgery and completed a 5-year follow-up. With next-generation sequencing, a 61-gene panel was used to evaluate plasma cell-free DNA and white blood cell genomic DNA from patients with ESCC. We analyzed the relationship between the mutation features of ctDNA and the prognosis of patients with ESCC, identified candidate risk predictors by Cox analysis, and developed nomogram models to predict the 2- and 5-year disease-free survival (DFS) and overall survival (OS). The area under the curve of the receiver operating characteristic (ROC) curve, concordance index (C-index), calibration plot, and integrated discrimination improvement (IDI) were used to evaluate the performance of the nomogram model. The model was compared with the traditional tumor-nodes-metastasis (TNM) staging system.Results: The ROC curve showed that the average mutant allele frequency (MAF) of ctDNA variants and the number of ctDNA variants were potential biomarkers for predicting the prognosis of patients with ESCC. The predictors included in the models were common candidate predictors of ESCC, such as lymph node stage, angiolymphatic invasion, drinking history, and ctDNA characteristics. The calibration curve demonstrated consistency between the observed and predicted results. Moreover, our nomogram models showed clear prognostic superiority over the traditional TNM staging system (based on C-index, 2-year DFS: 0.82 vs. 0.64; 5-year DFS: 0.78 vs. 0.65; 2-year OS: 0.80 vs. 0.66; 5-year OS: 0.77 vs. 0.66; based on IDI, 2-year DFS: 0.33, p <0.001; 5-year DFS: 0.18, p = 0.04; 2-year OS: 0.28, p <0.001; 5-year OS: 0.15, p = 0.04). The comprehensive scores of the nomogram models could be used to stratify patients with ESCC.Conclusions: The novel nomogram incorporating ctDNA features may help predict the prognosis of patients with resectable ESCC. This model can potentially be used to guide the postoperative management of ESCC patients in the future, such as adjuvant therapy and follow-up.

Background: Oesophageal squamous cell carcinoma (ESCC) remains a highly aggressive malignancy with limited biomarkers for monitoring tumour burden and prognosis. Circulating tumour DNA (ctDNA) has emerged as a promising tool for real‐time disease assessment, but its clinical utility in ESCC remains underexplored. Methods: In this prospective cohort study, we analysed preoperative and postoperative ctDNA from 54 treatment‐naïve ESCC patients undergoing radical surgery using a 61‐gene panel. Associations between ctDNA mutations, clinicopathological characteristics and survival outcomes were evaluated. Results: Preoperative ctDNA mutations were detected in 96.3% of patients (52/54), with Tumour Protein 53 (TP53) (59.3%), Phosphatidylinositol‐4,5‐bisphosphate 3‐kinase catalytic subunit alpha (PIK3CA) (31.5%) and Phosphatase and Tensin Homologue (PTEN) (13.0%) being most prevalent. Surgical resection significantly reduced ctDNA positivity ( p < .0001). Advanced‐stage tumours exhibited higher frequencies of PIK3CA (47.8% vs. 19.4%, p = .026) and PTEN mutations (26.1% vs. 3.2%, p = .034). Survival analysis revealed that postoperative TP53 ctDNA positivity predicted worse disease‐free survival (DFS; Hazard Ratio (HR) = 3.64, p  = .005) and overall survival (OS; HR = 3.29, p = .009), while PIK3CA positivity was associated with improved OS ( p = .032). Strikingly, preoperative PTEN ctDNA‐positive patients showed dramatically worse outcomes, with median DFS of 4.01 versus 33.27 months (HR = 7.53, p < .001) and OS of 11.80 versus 45.17 months (HR = 5.35, p < .001). In multivariate analysis, preoperative PTEN positivity remained the strongest independent prognostic factor for both DFS (HR = 7.28, p = .002) and OS (HR = 3.76, p = .028), surpassing Tumour, Node, Metastasis (TNM) stage. Conclusions: Our findings highlight the dynamic role of ctDNA in reflecting ESCC tumour burden and prognosis. While tumour‐agnostic ctDNA analysis showed limited clinical utility, gene‐specific mutations (TP53, PIK3CA and PTEN) demonstrated significant prognostic value. Preoperative PTEN ctDNA positivity emerged as a robust predictor of aggressive disease, suggesting its potential for risk stratification and personalised therapeutic strategies in ESCC.

Introduction: Digitally recorded respiratory sounds provide a practical source of physiological information for computer-assisted auscultation. Their acoustic presentation varies across subjects, recording locations, devices, filtering conditions, and environmental noise. Most classification models map acoustic representations directly to disease labels, which provides limited access to the auscultation findings associated with each prediction. This study investigated whether respiratory acoustic representations could be organized into structured evidence for classification across different label settings and evaluation levels. Methods: We developed Evi-CARD to organize respiratory sound representations into structured evidence for classification. A frozen Open Respiratory Acoustic Foundation Model (OPERA) encodes overlapping 8 s segments, from which a concept-aligned pathway estimates wheeze, crackle, inspiration, and expiration together with learned evidence weights. An attention-constrained residual pathway preserves complementary acoustic information, although its attention ceiling does not strictly bound the residual pathway’s contribution to the final logits. Separate disease-query layers adapt the shared evidence representation to each classification setting. We assessed BRACETS A2 and A3 by subject-independent five-fold validation at the recording level and KAUH-primary by subject-independent five-fold validation at the patient level for Healthy vs. Obstructive classification. Each assessment was conducted within its source cohort and was not an external-validation experiment. Results: On BRACETS, Evi-CARD achieved balanced accuracies of 77.8% for the A2 three-class task and 70.2% for the more granular A3 five-class task. On KAUH-primary, pooled patient-level out-of-fold evaluation yielded an area under the receiver operating characteristic curve of 95.1%, a balanced accuracy of 88.0%, a Macro F1 of 88.1%, and an expected calibration error of 5.1%. Across HF_Lung_V1, SPRSound, and ICBHI, the complete concept layer also provided more accurate and better-calibrated concept predictions than the evaluated acoustic-probe and concept-readout baselines. Discussion: These findings show that protocol-specific respiratory sound classification can be based on concept-associated latent evidence while retaining a constrained route for information outside the predefined concept set. However, Evi-CARD is not a strict scalar concept bottleneck, and no clinical meaning is assigned to its residual-memory slots. The evidence is limited to within-cohort cross-validation; independent-cohort testing, device- and site-stratified evaluation, and clinician review remain necessary before considering clinical deployment.