Search Results (1,024)

Background/Objectives: Homologous recombination deficiency (HRD) is a predictive biomarker in high-grade serous ovarian carcinoma for platinum-based chemotherapy and PARP inhibitors. The implementation of HRD testing in routine diagnostics has generated multiple commercial assays that differ in genomic targets, bioinformatic analysis, and HRD scoring strategies. We aimed to assess the analytical performance and feasibility of a locally semi-automated workflow based on the Agilent SureSelect CD HRR17 panel with SeqOne/SomaHRD analysis, and to compare it with established commercial HRD assays currently used in routine clinical practice: Myriad MyChoice CDx and SOPHiA DDM Dx HRD Solution. Methods: Thirty high-grade serous ovarian carcinoma cases diagnosed between 2019 and 2023 were retrospectively analyzed. HRD status was assessed with the Agilent-SeqOne workflow and compared with Myriad (n = 12) and SOPHiA (n = 18). Concordance and correlation between genomic instability metrics were evaluated. Results: The Agilent/SeqOne workflow showed high concordance with both comparison workflows. Genomic instability metrics strongly correlated across assays (R² up to 0.96). A lower proportion of inconclusive classifications was observed with the Agilent/SeqOne workflow. Discordances were mainly observed in borderline cases near classification thresholds. Variant detection was highly concordant within shared genomic regions. Conclusions: The locally semi-automated HRD workflow demonstrated high analytical concordance with established commercial assays in evaluable cases. Operational advantages related to workflow flexibility and local reanalysis support its potential implementation in routine molecular diagnostics. Full article

Background: Coronary anatomical complexity is commonly used for perioperative risk assessment in patients undergoing coronary artery bypass grafting (CABG), although it may not fully reflect systemic biological vulnerability. This study aimed to evaluate the association between the Prognostic Nutritional Index (PNI), a nutritional–immune marker derived from serum albumin and lymphocyte counts, and in-hospital mortality after CABG in relation to coronary anatomical complexity and established surgical risk scores. Methods: In this single-center retrospective cohort study, 324 consecutive patients who underwent isolated CABG between April 2024 and April 2025 were analyzed. The PNI was calculated according to the standard Onodera formula using preoperative serum albumin and total lymphocyte count. Associations with in-hospital mortality were evaluated using univariable and multivariable logistic regression analyses. Discriminative performance was assessed using receiver operating characteristic curve analysis, while exploratory analyses evaluating the additional prognostic contribution of the PNI beyond surgical risk scores were performed using nested model comparison and reclassification analyses. Internal validation and calibration analyses were also performed. Results: In-hospital mortality occurred in 26 patients. Preoperative and postoperative PNI values were significantly lower in patients who experienced in-hospital mortality. In multivariable analysis, the postoperative PNI remained independently associated with in-hospital mortality, whereas the preoperative PNI lost statistical significance after adjustment for clinical, renal, and surgical risk parameters. Receiver operating characteristic analysis demonstrated modest discriminative ability for the preoperative PNI (AUC: 0.742, 95% CI: 0.661–0.823). Exploratory analyses suggested a modest improvement in model discrimination and risk classification after the addition of the PNI to STS-based models; however, the overall incremental prognostic contribution remained limited. Calibration and internal validation analyses demonstrated acceptable agreement between predicted and observed mortality risk. Conclusions: The postoperative PNI demonstrated a stronger and independent association with in-hospital mortality than the preoperative PNI, suggesting that early postoperative nutritional–immune deterioration may reflect the magnitude of perioperative physiological stress and evolving clinical deterioration after CABG. Although lower preoperative PNI values were associated with mortality in univariable analyses, this association was no longer statistically significant after adjustment for clinical, renal, and surgical risk parameters. These findings indicate that postoperative nutritional–immune status may provide complementary biological information beyond conventional risk models; however, its clinical utility requires confirmation in larger prospective multicenter studies. Full article

To address the challenges of strong signal noise, pronounced cross-modal asynchrony, high subjectivity in manually defined state labels, and insufficient model stability under extreme abnormal conditions in multi-source sensor systems, a low-noise system state prediction and anomaly detection method based on multimodal sensor signals and self-supervised representation learning is proposed. Environmental sensing data, device status data, network transmission data, operational behavior data, and event log data are uniformly modeled as system state perception signals. A temporal masking-based state structure modeling method, a state-oriented contrastive learning representation constraint mechanism, and a state representation and downstream prediction task alignment strategy are designed to learn stable, transferable, and interpretable system state features. Experimental results demonstrate that the proposed method achieves the best performance in multimodal sensor state prediction and anomaly detection tasks, with mean squared error (MSE), mean absolute error (MAE), and root mean square error (RMSE) values of $0.0167$, $0.0856$, and $0.1291$, respectively, outperforming baseline models such as GARCH, MLP, LSTM, TCN, and Transformer. Meanwhile, $IC$, $RankIC$, and $AUC$ reach $0.494$, $0.460$, and $0.815$, respectively, indicating stronger state-ranking capability and improved discrimination between high-abnormality and low-abnormality states. At the classification recognition level, superior accuracy, precision, recall, and $F1$-score are also achieved by the proposed method, suggesting that potential abnormal states can be identified more accurately. Ablation experiments verify the effectiveness of multimodal fusion, temporal masking modeling, self-supervised contrastive constraints, and task alignment strategies. Robustness experiments further show that lower prediction errors and higher $AUC$ can still be maintained under high-fluctuation and extreme-shock states, demonstrating strong noise resistance, stability, and practical application potential in complex sensor system scenarios. Full article

Accurate diagnosis of potato nitrogen status is critical for optimized fertilizer management and sustaining productivity. We used data from nine field experiments (2010–2018) across major potato-producing regions in northern China to develop a regional critical nitrogen dilution curve via a Bayesian hierarchical model. The curve, Nc = 4.179 × DW^−0.417 (DW = whole-plant dry matter), provided the basis for calculating the nitrogen nutrition index (NNI), which was related to canopy spectral indices from a GreenSeeker sensor. Relationships between spectral indices and NNI were strongly growth-stage dependent. The tuber initiation–bulking period, approximately 29–70 days after emergence (DAE), represented the effective phenological window, with 29–55 DAE as the primary operational window for quantitative spectral diagnosis. Stage-specific ratio vegetation index (RVI) showed the most consistent association with NNI, whereas pooled whole-season models had low predictive power. The Bayesian framework quantified uncertainty, emphasizing that near-threshold NNI values require cautious interpretation. The resulting regional-average reference supports rapid field diagnosis of potato N status while accounting for cultivar, year, and site variability. These findings provide practical guidance for stage-specific N management and demonstrate the importance of growth-stage-aware spectral assessment in operational decision-making. Full article

Background: Response to neoadjuvant chemoradiotherapy (CRT) in locally advanced rectal cancer (LARC) varies considerably, and oxidative stress may modulate radiosensitivity. This study evaluated ischemia-modified albumin (IMA) and thiol–disulfide homeostasis as potential biochemical predictors of pathological tumor regression. Methods: A prospective observational cohort study was conducted to assess pre- and post-treatment oxidative stress biomarkers in patients with LARC receiving capecitabine-based long-course CRT. Serum IMA, native thiol, total thiol, and disulfide levels were quantified spectrophotometrically. Pathologic regression was graded according to the Modified Ryan system as good (TRG 0–1) or poor (TRG 2–3). Receiver operating characteristic (ROC) analyses, Firth-penalized logistic regression, and internal validation using cross-validation, calibration, and decision-curve analyses were performed. Results: Of 38 screened patients, 31 met eligibility criteria and completed CRT, alongside 31 matched healthy controls. Compared with controls, patients had higher baseline disulfide (15.7 ± 5.2 vs. 11.9 ± 3.1 µmol/L; p = 0.012) and IMA levels (0.886 ± 0.062 vs. 0.798 ± 0.048 ABSU; p = 0.006). Poor responders exhibited higher pre-treatment IMA (0.927 ± 0.045 vs. 0.842 ± 0.050 ABSU; p = 0.020) and disulfide levels (18.4 ± 5.2 vs. 13.0 ± 3.8 µmol/L; p = 0.012). Pre-treatment IMA demonstrated the highest predictive accuracy for poor tumor regression (AUC = 0.872; 95% CI 0.751–0.993). In multivariable Firth-penalized logistic regression, elevated baseline IMA was independently associated with poor pathological response (OR = 3.63; 95% CI 1.22–16.20; p = 0.043), whereas negative circumferential resection margin (CRM) status was independently associated with favorable regression (OR = 0.21; 95% CI 0.02–0.71; p = 0.003). The internally validated model demonstrated excellent discrimination (AUC = 0.948; 95% CI 0.866–0.966) and good calibration. Conclusions: Baseline IMA and CRM status were independently associated with pathological response after CRT in LARC. These findings suggest that oxidative-stress biomarkers may have potential value for response stratification; however, the results should be considered exploratory and require external validation in larger independent cohorts before clinical application. Full article

Artificial intelligence (AI) is reshaping ultrasound diagnosis by converting operator-dependent grayscale, Doppler, elastography, contrast-enhanced, automated-volume, and video data into reproducible decision support. In breast ultrasound, the most mature evidence involves benign–malignant lesion classification, BI-RADS risk stratification, reduction in unnecessary biopsy in selected low-risk lesions, assistance for less experienced readers, automated breast volume scanning, video-based assessment, axillary staging, and prediction of biologic markers such as molecular subtype, HER2 status, Ki-67 expression, lymphovascular invasion, and nodal metastasis. AI does not replace sonographers, radiologists, pathologists, or clinical judgment; rather, it can standardize feature extraction, prompt second-reader review, quantify uncertainty, and integrate imaging with clinical context. This review summarizes current clinical applications of AI in ultrasound diagnosis, which has a strong recent multicenter evidence base. It also discusses implementation requirements, including standardized acquisition, external validation, calibration, imaging–pathology concordance, workflow integration, data security, and equity across scanners and patient populations. Full article

Background: Biliary drainage is a key component of palliative management in patients with malignant biliary obstruction. In cases where endoscopic approaches are unsuccessful or cannot be performed, percutaneous transhepatic biliary drainage (PTBD) represents an established alternative for achieving biliary decompression. The C-reactive protein–albumin–lymphocyte (CALLY) index combines inflammatory, nutritional, and immune-related parameters into a single marker, while the modified Glasgow Prognostic Score (mGPS), based on C-reactive protein and albumin concentrations, reflects the systemic inflammatory status of the patient. This study aimed to evaluate the prognostic value of the preprocedural CALLY index and mGPS in predicting 30-day mortality among patients with advanced malignant biliary obstruction undergoing palliative PTBD. Methods: This single-center retrospective study was conducted in a total of 179 patients who underwent palliative PTBD for malignant biliary obstruction at Ankara Etlik City Hospital between December 2022 and June 2025. Results: The 30-day mortality rate was 25.1%. The cut-off value for CALLY was determined as 67 based on receiver operating characteristic (ROC) curve analysis, and mGPS was categorized as 0–1 versus 2. In univariable Cox regression analyses, pancreaticobiliary tumor type, mGPS = 2, and CALLY < 67 were associated with early mortality. In multivariable Cox analysis, CALLY ≥ 67 was independently associated with a reduced risk of 30-day mortality, whereas pancreaticobiliary tumor type was independently associated with an increased risk. In the CALLY–mGPS risk stratification, 30-day mortality rates were 8.0%, 13.5%, and 44.1% in the low-, intermediate-, and high-risk groups, respectively. Conclusions: In this retrospective cohort, preprocedural inflammation- and nutrition-based markers were found to be associated with early mortality in patients with malignant biliary obstruction undergoing PTBD. Accordingly, risk stratification using readily available parameters such as CALLY and mGPS appears feasible in the preprocedural setting. The CALLY–mGPS-based approach may provide a practical framework for clinical risk assessment; however, prospective multicenter validation, including tumor-specific subgroup analyses, is warranted. Full article

Background/Objectives: Falls among older adults are a major public health concern. Although instrumented posturography provides objective balance and fall-risk assessment, its cost and limited portability restrict widespread use. This study aimed to examine the construct and concurrent validity of center of pressure (COP) path length measured using a Wii Balance Board (WBB) in relation to a clinically established posturographic fall-risk construct in community-dwelling older adults and to explore its discriminatory performance across multiple sensory postural conditions. Methods: Sixty adults aged ≥ 65 years participated in this cross-sectional study. COP path length was measured using a WBB under eight postural conditions and compared with the Fall Index derived from a conventional posturography system (Tetrax^®). Functional performance was assessed using the Four Square Step Test and the Five Times Sit-to-Stand test. Pearson correlation, receiver operating characteristic (ROC), and exploratory regression analyses were performed. Results: COP path length showed significant positive correlations with the Tetrax^® Fall Index across all conditions (r = 0.349–0.561, p < 0.01) and with functional performance tests under most postural conditions (p < 0.05), except for the Normal stability, Open eyes (NO) condition. ROC analysis demonstrated acceptable-to-good discriminatory performance for classifying Tetrax^® Fall Index-based risk status (AUC = 0.783–0.865), with the NO condition showing the highest discriminatory capability (AUC = 0.865). Exploratory regression models based on selected postural conditions explained 12.1–40.7% of the variance in the reference Fall Index. Conclusions: COP path length measured using a WBB demonstrated construct validity and acceptable discriminatory capacity in relation to a conventional posturographic fall-risk construct in community-dwelling older adults. These findings support the exploratory feasibility of simplified WBB-based balance assessment approaches for community and clinical screening contexts. Further longitudinal studies incorporating prospective fall outcomes are required to establish predictive validity and broader clinical applicability. Full article

Well control during drilling requires continuous assessment of bottom-hole pressure (BHP) relative to the pressure window bounded by formation and fracture pressures. This study presents a reduced-order, physics-guided digital-twin framework for well-control decision support, kick and loss risk assessment, and hybrid BHP prediction. The framework is intended as a computational decision-support prototype rather than a fully deployed, real-time, field-validated digital twin. It combines pressure-window calculations, dimensionless risk indices, bounded machine-learning correction, scenario-based event simulation, an interactive engineering dashboard, and 3D safety-envelope visualization. The machine-learning layer was trained on a predominantly augmented drilling dataset containing 909 cases, including nine field-related baseline records and 900 synthetically generated cases, and was used as a constrained correction mechanism rather than a replacement for the physics-based model. On the held-out test set, the BHP regression model achieved R² = 0.987, MAE = 108.6 psi, and RMSE = 215.7 psi, while the well-control status classifier achieved an accuracy of 98.35%. Scenario simulations reproduced representative kick-prone and loss-prone conditions and tracked the evolution of BHP, the Pressure Safety Index, the Kick Risk Index, and the Loss Risk Index. The results show that the proposed workflow can identify underbalanced states, quantify pressure margins, evaluate mud-weight sensitivity, and support visual interpretation of well-control risk. Further field validation, real-time data integration, uncertainty quantification, and robustness testing are required before operational deployment. Full article

Background: The inflammatory prognostic index (IPI) is a novel hematological parameter that reflects both inflammatory burden and immune status. Despite several studies on this new index, the prognostic value of IPI in CABG remains unclear so far. This study aimed to evaluate whether the IPI could serve as a predictor of postoperative outcomes in patients undergoing coronary artery bypass grafting (CABG). Methods: A total of 640 patients who underwent isolated CABG between 2022 and 2025 were retrospectively analyzed. The optimal preoperative IPI cut-off value for predicting mortality was determined using receiver operating characteristic (ROC) curve analysis. The optimal IPI cut-off value was identified as 0.22 (AUC = 0.607; 95% CI: 0.468–0.747; p = 0.14). Based on this threshold, patients were categorized into two groups: high IPI (Group A, n = 293) and low IPI (Group B, n = 347). Results: Baseline demographic features, comorbid conditions, echocardiographic findings, operative variables, cardiopulmonary bypass time, cross-clamp time, and laboratory results were comparable between the groups, except for gender distribution and platelet counts. The incidence of postoperative cerebrovascular events was significantly higher in Group A. Although mortality was more frequent in Group A than in Group B (3.8% vs. 1.4%), this difference did not reach statistical significance (p = 0.06). Additionally, patients in Group A had significantly longer intensive care unit (ICU) stays (mean: 3 vs. 2.6 days, p = 0.01) and overall hospital stays (9.5 vs. 8.4 days, p = 0.002). Multivariate regression analysis demonstrated that gender, diabetes mellitus, hypertension, and IPI were not independently associated with mortality. Conclusions: The IPI was associated with longer ICU and hospital stays in isolated CABG patients. These results support the use of IPI as a potential prognostic index for CABG patients. Full article

Background/Objectives: This study evaluated whether sequential changes in cervical length (CL) can predict spontaneous preterm birth (sPTB) before 34 weeks of gestation in twin pregnancies. Methods: This retrospective study from a single tertiary-care center analyzed 349 twin pregnancies with deliveries between January 2019 and December 2023. Cervical length assessments began at 18–21 weeks, followed by biweekly serial measurements. The primary outcome was sPTB before 34 weeks. CL changes were assessed descriptively using data from patients with and without sPTB before 34 weeks. We defined the high-risk status for sPTB based on our assessment. Logistic regression models were used to compute the odds ratios (ORs) and 95% confidence intervals (CIs) to quantify the relationship between these predictors and sPTB. Diagnostic accuracy was assessed from the area under the curve using receiver operating characteristic curve analysis. Results: The sPTB rate before 34 weeks of gestation was 8.5% (18/212). In the group without -sPTB before 34 weeks of gestation, the 5th percentile CL was approximately 20 mm at 25 weeks and 15 mm at 26–27 weeks of gestation. Sequential CL measurements revealed that a rapid shortening of ≥10 mm within 2 weeks significantly predicted sPTB before 34 weeks. A decrease in CL of ≥10 mm in a 2-week interval was associated with increased odds for sPTB before 34 weeks of gestation [adjusted OR (95% CI): 6.66 (2.32–19.14)]. Conclusions: In twin pregnancies, measuring CL every 2 weeks after approximately 20 weeks of gestation may facilitate sPTB detection before 34 weeks. Full article

Background: Psychosocial assessment is central to lung transplant evaluation. Structured tools such as the Stanford Integrated Psychosocial Assessment for Transplantation (SIPAT) can be used either to support exclusionary decisions or to guide psychosocial prehabilitation by identifying modifiable targets for intervention. We examined how SIPAT functions in a program that explicitly prioritizes remediation of modifiable psychosocial risks. Methods: We conducted a retrospective observational cohort study of consecutive adult lung transplant candidates evaluated at a single center in Poland between December 2021 and November 2025. Psychosocial risk was assessed using SIPAT (locally translated), including total and domain scores, candidate categories, and binary indicators of clinically relevant alcohol, illicit substance, and nicotine-related risk. The primary endpoint was a pragmatic program outcome, defined as ever being listed (including transplanted) versus not being listed. Analyses focused on describing psychosocial risk profiles and their relationship to the program pathway rather than on building a predictive model of listing decisions. Results: In 491 candidates (mean age 57.2 years; 40.5% women), psychosocial burden was generally low (mean total SIPAT 12.4, SD 6.8), and most patients were rated as excellent or good candidates. SIPAT total, domain scores, and candidate categories were not meaningfully associated with ever being listed. Substance-related risk indicators were not independently associated with listing status after adjustment for age and sex. Cluster analyses based on SIPAT domain scores identified a higher-risk psychosocial profile, but cluster membership was not associated with listing status. ROC analyses showed that neither SIPAT total score nor domain scores discriminated listed from non-listed candidates, supporting the interpretation that SIPAT was not used as a binary gatekeeping tool in this program. Conclusions: In this prehabilitation-oriented program, SIPAT did not operate as a binary gatekeeping instrument for listing. Instead, it primarily served to identify modifiable psychosocial targets that trigger tailored support. These findings support using SIPAT as a structured roadmap for psychosocial prehabilitation rather than a stand-alone exclusion tool. Full article

Background: Moderate-to-severe preterm brain injury (PBI), including intraventricular hemorrhage (IVH) and periventricular leukomalacia (PVL), remains an important cause of adverse neurodevelopmental outcomes in preterm infants. Early risk stratification using routinely collected clinical data may help prioritize surveillance in vulnerable infants. Methods: We retrospectively included 318 preterm infants admitted between 2015 and 2024 as the development cohort. Thirty-three candidate predictors derived from perinatal factors, first laboratory tests within 24 h of admission, and selected early hospitalization variables were evaluated. Seven machine-learning algorithms were developed using stratified 10 × 5 nested cross-validation with prespecified preprocessing, class-balancing, and feature-selection procedures. Candidate models were compared primarily using the mean fold-level area under the receiver operating characteristic curve (AUROC). After model selection, the finalized LightGBM model was calibrated using Platt scaling, and its pooled out-of-fold (OOF) performance was summarized. Two prespecified thresholds (Youden and high-sensitivity) were used for risk stratification. A small independent temporal cohort of 35 infants was used for preliminary external validation. Results: PBI occurred in 62/318 infants (19.5%) in the development cohort and 6/35 infants (17.1%) in the temporal external cohort. During candidate-model comparison, LightGBM achieved the highest mean fold-level AUROC (0.768, 95% CI 0.708–0.825). The finalized 14-feature LightGBM model, evaluated using pooled OOF predictions after Platt calibration, yielded an AUROC of 0.747 (95% CI 0.679–0.811), a PR-AUC of 0.392, and a Brier score of 0.136. At the Youden threshold (0.18), sensitivity was approximately 0.70 and specificity approximately 0.85; at the high-sensitivity threshold (0.10), sensitivity was approximately 0.95 and specificity approximately 0.50. Key predictors included ventilation status and early physiologic and laboratory indicators. In the small temporal external cohort (n = 35), the AUROC was 0.897 (95% CI 0.672–1.000); however, this high point estimate should not be overinterpreted because of the limited sample size, wide confidence interval, and suboptimal calibration, and should therefore be considered preliminary. Conclusions: We developed an interpretable LightGBM model using routinely available early postnatal and early hospitalization data to support risk stratification for PBI in preterm infants. The model showed moderate internal discrimination and a positive net benefit across clinically relevant thresholds. Preliminary temporal external validation in a small cohort yielded highly uncertain estimates; larger multicenter studies are needed to confirm generalizability, refine calibration, and determine the most appropriate implementation strategy before routine clinical use. Full article

The chlorophyll-a (chl-a) concentration is a major indicator of marine ecosystem status, harmful algal blooms, and marine primary productivity. In coastal waters, however, complex hydrodynamic and ecological conditions lead to highly variable chl-a dynamics, driven by diverse and interacting mechanisms, posing substantial challenges for chl-a forecasts. To assess the applicability of machine learning approaches in predicting chl-a under complex coastal environments, we present a case study in the Taiwan Strait, where harmful algal blooms occur a few times every year. Based on satellite remote sensing data, a spatiotemporal imputation and prediction framework (STIMP), temporal models (Transformer, CrossFormer, Tsmixer), and spatiotemporal models (MTGNN and PredRNN) were applied to simulate chl-a spatiotemporal variability. A hydrodynamic–biogeochemical model was compared with these machine learning approaches to assess the model skills in coastal chl-a simulations. Results indicate that machine learning models trained with satellite data exhibit reasonable predictive skill offshore with pronounced seasonal variability and low data missing ratio, while their performance weakens in regions where seasonal signals are masked by short-term chl-a fluctuations with more missing data. In contrast, the hydrodynamic–biogeochemical model represents short-term variations in chl-a in nearshore regions with higher temporal resolution and accounts for the underlying mechanisms of phytoplankton biomass accumulation and die-off. When trained with model output, the machine learning approach shows improved performance in coastal chl-a forecasts, with much higher computational efficiency compared to the hydrodynamic–biogeochemical model. This study highlights the advantage of mechanistic and machine learning models in deciphering the spatiotemporal scales and governing mechanisms of chl-a variability in coastal regions and extracting spatiotemporal variability with computational efficiency, respectively. With input data of sufficient temporal resolution (e.g., daily to 3 days) and duration (5–10 years), a combination of the machine learning and mechanistic modeling approaches is recommended for operational coastal phytoplankton bloom forecasting. Full article

Background: Pathological complete response (pCR) to neoadjuvant chemotherapy (NACT) in breast cancer is influenced by multiple tumor- and host-related factors, and readily available pretreatment biomarkers of response are still limited. This study aimed to evaluate the association between pretreatment systemic inflammatory and nutritional parameters and pCR assessed by the Miller–Payne grading system, with a specific focus on the independent predictive value of pretreatment serum albumin compared with established inflammatory ratios. Methods: A total of 226 patients with breast carcinoma who received NACT between May 2017 and September 2023 were retrospectively evaluated. Pretreatment laboratory parameters—including neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), C-reactive protein (CRP), serum albumin, and the CRP/albumin ratio (CAR)—were recorded. Pathological response was assessed using the Miller–Payne grading system by two breast pathologists blinded to laboratory data. Univariable and multivariable logistic regression and receiver operating characteristic (ROC) curve analyses were performed, complemented by bootstrap validation of the optimal cut-off, a sensitivity analysis using the contemporary ypT0/is ypN0 definition of pCR, and a subgroup analysis by molecular subtype. Results: pCR was observed in 41 patients (18.1%). Pretreatment serum albumin levels were significantly lower in responders than in non-responders (p = 0.027), whereas NLR, PLR, CRP, and CAR were not significantly associated with response. In multivariable analysis, pretreatment serum albumin, Ki-67, and HER2 status emerged as independent predictors of pCR. ROC analysis demonstrated moderate discriminatory ability for albumin (AUC = 0.64); the optimal cut-off was 4.22 g/dL (bootstrap 95% CI 3.50–4.53 g/dL), with values below this threshold associated with a higher likelihood of pCR. The association between low pretreatment albumin and pCR was particularly pronounced in the triple-negative subgroup (3.30 vs. 4.02 g/dL, p = 0.027). The albumin signal remained significant under the stricter ypT0/is ypN0 definition of pCR in univariable analysis (OR 0.47, p = 0.045). Conclusions: Pretreatment serum albumin, independent of systemic inflammatory ratios, is associated with pCR to NACT in breast cancer and may serve as a candidate biomarker for pretreatment risk stratification, particularly when interpreted alongside established tumor-related predictors such as Ki-67 and HER2 status. The association appears especially relevant in the triple-negative subgroup, suggesting that patients with TNBC and low pretreatment serum albumin may warrant heightened multidisciplinary attention during NACT. Validation in larger, prospective, multicenter cohorts is needed before routine clinical implementation. Full article