Objective:To establish 30-day of age transcutaneous bilirubin (TcB) reference curves for healthy neonates, and to investigate regional variations in bilirubin dynamics.Methods:A multicenter retrospective cohort study was conducted. A total of 220 950 healthy neonates born at a gestational age of 35-<42 weeks, with a birth weight ≥2 000 g, who did not receive phototherapy within 60 h after birth were recruited. All of them underwent remote TcB monitoring using the Bilibaby remote jaundice monitoring system between August 1 st, 2020 and December 31 st, 2024 in 426 hospitals. TcB data were collected within the period from birth to 30-day of age. The P40, P75, and P95 of TcB values were calculated, and dynamic TcB curves for 30-day of age were constructed. Patterns of bilirubin change, rates of change, and transition outcomes were described. Regional comparisons between South and North were conducted using linear mixed-effects models for TcB trajectories and Pearson′s chi-square test for outcome differences. Results:A total of 220 950 neonates were included, of whom 101 711 (46.03%) were female. Gestational age at birth was (38.75±1.12) weeks, and birth weight was (3 272±417) g. TcB levels increased rapidly within 3-day of age, peaked at 4-6-day of age, with peak values at P40, P75, and P95 of 200.6, 239.7 and 275.4 μmol/L (11.8, 14.1 and 16.2 mg/dl), respectively. TcB levels gradually declined thereafter and stabilized after 13-day of age, with values at P40, P75, and P95 fluctuating between 147.9-159.8, 190.4-200.6, and 231.2-239.7 μmol/L (8.7-9.4, 11.2-11.8, 13.6-14.1 mg/dl), respectively. Notably, among neonates categorized as low-or low-intermediate-risk within 3-day of age, 6 700 (12.76%) progressed to intermediate-high or high risk between 4 and 30 days of age. Before 13-day of age, TcB levels in the southern regions were consistently higher than those in the northern regions ( P=0.039); from 14 to 30 days of age, the overall TcB levels had no statistically difference, but the temporal changes in TcB still showed regional differences (degrees of freedom=3, all interaction P<0.05). Among neonates classified as low-or low-intermediate risk within 3-day of age, 25 326 were from southern regions, of whom 4 254 (16.80%) progressed to intermediate-high or high risk between 4 and 30 days of age. In northern regions, 27 193 neonates were classified as low-or low-intermediate risk within 3-day of age, among whom 2 446 (8.99%) progressed to intermediate-high or high risk. The risk progression between the 2 regions had statistically difference ( χ2=716.49, P<0.001). Conclusions:A TcB percentile curve for neonates within 30-day of age was established, revealing that both the overall TcB level and its temporal trend were higher in southern than in northern newborns. These findings provide baseline data to support continuous management of neonatal jaundice.

To report a Chinese pedigree with benign hereditary chorea (BHC) related to NKX2-1 gene mutation. The proband, a 16-year-old boy, was admitted with a 15-year history of recurrent involuntary choreiform movements, exacerbated over 2 weeks. The results of whole-exome sequencing showed that the proband and his mother both had a heterozygous mutation in exon 2 of the NKX2-1 gene (c.231_232dup:p.Pro78Hisfs *24), leading to a clinical diagnosis of BHC. No patient carrying this variant was previously reported in the literature. The clinical phenotype of NKX2-1 gene defects is complex, with BHC being a characteristic manifestation of the nervous system. This group of disorders is also referred to as NKX2-1-related disorders. This article discusses the clinical features, diagnosis and treatment points of patients with NKX2-1-related disorder with reference to the current literature, aiming to enhance clinicians′ understanding and management of this disease.

BACKGROUND:In the follow-up after unicompartmental knee arthroplasty,some patients have knee pain,among which postmenopausal obese women are the most common. As an important index to measure the degree of body obesity,whether body mass index is related to the curative effect after unicompartmental knee arthroplasty and whether obesity will affect the function of knee joint after operation are worthy of further study.OBJECTIVE:To evaluate the clinical efficacy of postmenopausal obese women undergoing medial unicompartmental knee arthroplasty,and to determine the influence of body mass index on the quality of life after unicompartmental knee arthroplasty.METHODS:Female postmenopausal patients who underwent medial unicompartmental knee arthroplasty for the first time due to medial knee pain from January 2017 to January 2019 in the Fourth Clinical Medical College of Xinjiang Medical University were enrolled. A total of 270 cases were included according to inclusion and exclusion criteria,and were divided into 4 groups according to preoperative body mass index:There were 42 cases in normal group (body mass index 18.5-22.9 kg/m2),58 cases in overweight group (body mass index 23.0-24.9 kg/m2),122 cases in obese group (body mass index 25.0-29.9 kg/m2),and 48 cases in severely obese group (body mass index ≥30 kg/m2). Hospital for Special Surgery score,Western Ontario and McMaster Universities Osteoarthritis Index score,knee range of motion,visual analog scale score,and hip-knee-ankle angle were compared before,after and at the last time in each group. Patients were followed up to record the time of use of prostheses after surgery and reasons for failure or revision. The effective utilization rate of prostheses was calculated and compared in each group. Survival curve was used for statistical analysis of the effective utilization rate of prostheses.RESULTS AND CONCLUSION:(1) There was no significant difference in postoperative follow-up time,knee joint range of motion,visual analog scale score,and hip-knee-ankle angle between the groups (P>0.05). (2) The Hospital for Special Surgery score and Western Ontario and McMaster Universities Osteoarthritis Index score of each group in final follow-up were significantly improved compared with those before surgery (P<0.05),and the postoperative effect was obvious in each group (P<0.05). Regarding Hospital for Special Surgery score,the improvement effect was worse in the severely obese group. (3) The comparison of hip-knee-ankle angle between all groups immediately after surgery and the last follow-up showed that there were significant differences between the other groups at two time points (P<0.05) except the normal group (P>0.05). (4) The effective utilization rate of prosthesis in normal,overweight,obesity,and severely obese groups was 100％,95％,94％,and 94％,respectively,and there was no significant difference between the groups (x2=2.532,P=0.469). (5) It is indicated that body mass index of postmenopausal obese women had no significant effect on the effective utilization rate of medial unicompartmental prosthesis. Obesity is an important factor affecting the postoperative knee function score and the effective utilization rate of prosthesis.Weight should be properly controlled before and after unicompartmental knee arthroplasty. At the same time,female body mass index ≥ 30 kg/m2 is not the best indication for unicompartmental knee arthroplasty. It is suggested that female patients undergoing unicompartmental knee arthroplasty should controlbody mass index below 30 kg/m2.

Objective:To investigate the value of habitat radiomic features based on gadolinium ethoxybenzyl diethylenetriamine pentaacetic acid (Gd-EOB-DTPA)-enhanced MRI in establishing a predictive model for cytokeratin 19 (CK19) expression in hepatocellular carcinoma (HCC) and to evaluate its role in prognostic risk stratification.Methods:This multicenter case-control study retrospectively enrolled 489 patients with pathologically confirmed HCC who underwent Gd-EOB-DTPA-enhanced MRI between June 2016 and June 2024. Among them, 346 patients from the First Affiliated Hospital of Soochow University were divided into a training cohort ( n=245) and an internal test cohort ( n=101) via stratified sampling at a 7∶3 ratio. And 143 patients from Nantong Third Hospital Affiliated to Nantong University served as an external validation cohort. The training cohort included 53 CK19-positive and 192 CK19-negative patients. The internal test cohort included 21 CK19-positive and 80 CK19-negative patients. The external validation cohort included 30 CK19-positive and 113 CK19-negative patients. Univariate logistic regression analysis was performed to identify potential factors associated with CK19 expression, and a clinical-radiologic model was constructed. The k-means clustering algorithm was applied to segment target HCC lesions into 3 subregions. Radiomic features were extracted and selected from these habitat subregions. Habitat radiomics models were constructed for the arterial phase (AP), portal venous phase, hepatobiliary phase (HBP), and combined phases (CP). Multivariate logistic regression analysis identified independent clinical and radiologic predictors of CK19 expression, and the optimal habitat model score was integrated to build a clinical-radiologic-habitat combined model. The area under the receiver operating characteristic curve (AUC) was used to evaluate model predictive performance. Recurrence-free survival (RFS) was analyzed using the Kaplan-Meier method and the differences in survival curves were compared with the log-rank test. Results:Univariate logistic regression analysis revealed that alpha-fetoprotein (AFP) ( OR=2.629, 95% CI 1.412-4.896, P=0.002), AP enhancement ( OR=3.636, 95% CI 1.642-8.052, P=0.001), AP peritumoral enhancement ( OR=2.219, 95% CI 1.084-4.542, P=0.029), and HBP peritumoral hypointensity ( OR=2.010, 95% CI 1.004-4.021, P=0.049) were potential factors associated with CK19 expression, which were incorporated into the clinical-radiologic model. In the internal and external validation cohorts, the AUC of the clinical-radiologic model was 0.690 (95% CI 0.590-0.778) and 0.650 (95% CI 0.565-0.727), respectively. The habitat radiomics model based on CP images demonstrated the highest performance. It achieved AUC of 0.729 (95% CI 0.622-0.836) and 0.725 (95% CI 0.607-0.842) in the internal and external validation cohorts, respectively. Multivariate analysis identified AFP ( OR=2.494, 95% CI 1.163-5.348, P=0.019), AP enhancement ( OR=5.230, 95% CI 1.868-14.643, P=0.002) and habitat radiomics model score ( OR=4.105, 95% CI 2.643-6.368, P<0.001) as independent predictors of CK19 positivity. Based on these factors, a combined clinical-radiologic-habitat combined model was established. The clinical-radiologic-habitat combined model achieved AUCs of 0.767 (95% CI 0.671-0.846) and 0.730 (95% CI 0.649-0.801) in the internal and external validation cohorts, respectively. Significant differences in RFS were observed between the CK19-positive group (25.1 month) and CK19-negative group (51.0 month) as predicted by the clinical-radiologic-habitat model ( χ2=4.17, P=0.041). Conclusion:The clinical-radiologic-habitat combined model based on Gd-EOB-DTPA-enhanced MRI habitat radiomics demonstrates good predictive performance for CK19 expression in HCC and offers valuable prognostic stratification for clinical practice.

Objective:To explore the value of the deep learning model based on gadolinium ethoxybenzyl diethylenetriamine pentaacetic acid (Gd-EOB-DTPA) enhanced MRI in preoperatively predicting vessels completely encapsulating tumor clusters (VETC) in hepatocellular carcinoma (HCC).Methods:This study adopted a case-control design to retrospectively analyze 420 patients with HCC confirmed by postoperative pathology who underwent Gd-EOB-DTPA enhanced MRI between June 2016 and March 2023. A total of 420 patients were divided into a training set ( n=305) from the First Affiliated Hospital of Soochow University and an external validation set ( n=115) from Affiliated Nantong Hospital 3 of Nantong University. Based on postoperative pathological findings, patients were stratified into VETC-positive and VETC-negative groups. The training set comprised 161 VETC-positive cases and 144 VETC-negative cases, while the external validation set included 55 VETC-positive cases and 60 VETC-negative cases. Tumor regions of interest in arterial, portal venous, and hepatobiliary phases were manually delineated using ITK-SNAP software. Pre-trained Vgg19, Densenet121, and Vision Transformer (ViT) models were employed for transfer learning, extracting deep learning features from each image. Feature data were processed using FAE software, and 12 logistic regression models (arterial phase, portal venous phase, hepatobiliary phase, and combined three-phase models) were constructed to select the optimal deep learning model. Independent predictors in clinical characteristics were identified through univariate and multivariate logistic analyses to establish a clinical model for predicting VETC pattern. Subsequently, a clinical-deep learning fusion model was developed by integrating these clinical predictors with the optimal deep learning features. Model performance in predicting VETC-positive HCC was evaluated using receiver operating characteristic curves, calibration curves, and decision curve analysis (DCA). Results:In the external validation set, the area under the curve (AUC) of the Vgg19 model in the arterial phase, portal venous phase, hepatobiliary phase, and combined three-phase, respectively were 0.799,0.756,0.789,0.821, which were higher than those of Densenet121 (AUC: 0.544,0.581,0.544,0.583) and ViT (AUC: 0.740,0.752,0.785,0.767) model. The three-phase combined Vgg19 model achieved the highest AUC of 0.821 (95% CI 0.746-0.897). Multivariate logistic regression identified alpha-fetoprotein level ( OR=1.826,95% CI 1.069-3.120, P=0.028) and tumor diameter ( OR=1.329,95% CI 1.206-1.466, P<0.001) as independent predictors of VETC-positive HCC, forming the clinical model with an AUC of 0.789 (95% CI 0.703-0.859). The clinical-deep learning fusion model further achieved the AUC of 0.825 (95% CI 0.749-0.900). Calibration curves confirmed high concordance between predicted and actual probabilities for the three-phase Vgg19 model, while DCA revealed greater net clinical benefit for the combined Vgg19 and fusion models compared with the clinical model alone. Conclusions:The deep learning model based on Gd-EOB-DTPA enhanced MRI can be used to predict VETC of HCC preoperatively, among which the three-phase combined Vgg19 model and the clinical-deep learning model provide high predictive value.

BACKGROUND:In the follow-up after unicompartmental knee arthroplasty,some patients have knee pain,among which postmenopausal obese women are the most common. As an important index to measure the degree of body obesity,whether body mass index is related to the curative effect after unicompartmental knee arthroplasty and whether obesity will affect the function of knee joint after operation are worthy of further study.OBJECTIVE:To evaluate the clinical efficacy of postmenopausal obese women undergoing medial unicompartmental knee arthroplasty,and to determine the influence of body mass index on the quality of life after unicompartmental knee arthroplasty.METHODS:Female postmenopausal patients who underwent medial unicompartmental knee arthroplasty for the first time due to medial knee pain from January 2017 to January 2019 in the Fourth Clinical Medical College of Xinjiang Medical University were enrolled. A total of 270 cases were included according to inclusion and exclusion criteria,and were divided into 4 groups according to preoperative body mass index:There were 42 cases in normal group (body mass index 18.5-22.9 kg/m2),58 cases in overweight group (body mass index 23.0-24.9 kg/m2),122 cases in obese group (body mass index 25.0-29.9 kg/m2),and 48 cases in severely obese group (body mass index ≥30 kg/m2). Hospital for Special Surgery score,Western Ontario and McMaster Universities Osteoarthritis Index score,knee range of motion,visual analog scale score,and hip-knee-ankle angle were compared before,after and at the last time in each group. Patients were followed up to record the time of use of prostheses after surgery and reasons for failure or revision. The effective utilization rate of prostheses was calculated and compared in each group. Survival curve was used for statistical analysis of the effective utilization rate of prostheses.RESULTS AND CONCLUSION:(1) There was no significant difference in postoperative follow-up time,knee joint range of motion,visual analog scale score,and hip-knee-ankle angle between the groups (P>0.05). (2) The Hospital for Special Surgery score and Western Ontario and McMaster Universities Osteoarthritis Index score of each group in final follow-up were significantly improved compared with those before surgery (P<0.05),and the postoperative effect was obvious in each group (P<0.05). Regarding Hospital for Special Surgery score,the improvement effect was worse in the severely obese group. (3) The comparison of hip-knee-ankle angle between all groups immediately after surgery and the last follow-up showed that there were significant differences between the other groups at two time points (P<0.05) except the normal group (P>0.05). (4) The effective utilization rate of prosthesis in normal,overweight,obesity,and severely obese groups was 100％,95％,94％,and 94％,respectively,and there was no significant difference between the groups (x2=2.532,P=0.469). (5) It is indicated that body mass index of postmenopausal obese women had no significant effect on the effective utilization rate of medial unicompartmental prosthesis. Obesity is an important factor affecting the postoperative knee function score and the effective utilization rate of prosthesis.Weight should be properly controlled before and after unicompartmental knee arthroplasty. At the same time,female body mass index ≥ 30 kg/m2 is not the best indication for unicompartmental knee arthroplasty. It is suggested that female patients undergoing unicompartmental knee arthroplasty should controlbody mass index below 30 kg/m2.

Objective:To investigate the effects of individual versus connected microdroplet culture modes in time-lapse (TL) incubators on embryo development parameters and pregnancy outcomes in patients undergoing whole embryo culture to blastocyst stage.Methods:Using a retrospective cohort study, clinical data from 3 507 fresh blastocyst transfer cycles were analyzed. These cycles involved patients who underwent assisted reproductive technology treatment with whole embryo culture to blastocyst stage at the Reproductive Medical Center of Northwest Women's and Children's Hospital between January 2019 and December 2023. Based on different culture modes, patients were divided into two groups, connected group ( n=2 446, using connected microdroplet culture) and individual group ( n=1 061, using individual microdroplet culture). Baseline characteristics, embryo development parameters, pregnancy outcomes, and neonatal outcomes were compared between the two groups. Generalized linear models (GLM) were used to adjust for confounding factors and analyze the effect of culture mode. Results:Embryo development assessment showed the day 3 (D3) high-quality embryo rate in the connected group [60.12% (12 136/20 187)] was significantly lower than that in the individual group [63.62% (4 705/7 395), P<0.001], whereas the high-quality blastocyst formation rate [34.93% (7 052/20 187)] and the available blastocyst formation rate [56.07% (11 319/20 187)] were both significantly higher than those in the individual group [33.08% (2 446/7 395), P=0.004; 51.45% (3 805/7 395), P<0.001], with statistically significant differences. The implantation rate [67.40% (1 774/2 632)], the clinical pregnancy rate [70.20% (1 717/2 446)], and the live birth rate [60.66% (1 469/2 446)] in the connected group were all significantly higher than those in the individual group [63.40% (724/1 142), P=0.017; 66.73% (708/1 061), P=0.041; 55.89% (593/1 061), P=0.021], with statistically significant differences. Neonatal outcomes showed no statistically significant difference between the two groups (all P>0.05). After adjusting for confounding factors using GLM, connected culture was an independent influencing factor for D3 high-quality embryo rate (a MD=-0.017, 95% CI: -0.034-0.000, P=0.046), high-quality blastocyst formation rate (a MD=-0.020, 95% CI: 0.002-0.037, P=0.026), available blastocyst formation rate (a MD=0.032, 95% CI: 0.015-0.048, P<0.001), live birth rate (a OR=1.182, 95% CI: 1.006-1.388, P=0.042). However, it had no effect on D3 available embryo rate, clinical pregnancy rate, or early miscarriage rate (all P>0.05). Conclusion:In TL incubator systems, individual and connected microdroplet culture modes exert different effects at various stages of embryo development. Individual microdroplet culture can significantly enhance cleavage-stage embryo quality, whereas the connected microdroplet culture was more beneficial for enhancing the blastocyst formation rate and quality, ultimately improving the live birth rate without increasing neonatal risks.

Objective:To investigate the effects of individual versus connected microdroplet culture modes in time-lapse (TL) incubators on embryo development parameters and pregnancy outcomes in patients undergoing whole embryo culture to blastocyst stage.Methods:Using a retrospective cohort study, clinical data from 3 507 fresh blastocyst transfer cycles were analyzed. These cycles involved patients who underwent assisted reproductive technology treatment with whole embryo culture to blastocyst stage at the Reproductive Medical Center of Northwest Women's and Children's Hospital between January 2019 and December 2023. Based on different culture modes, patients were divided into two groups, connected group ( n=2 446, using connected microdroplet culture) and individual group ( n=1 061, using individual microdroplet culture). Baseline characteristics, embryo development parameters, pregnancy outcomes, and neonatal outcomes were compared between the two groups. Generalized linear models (GLM) were used to adjust for confounding factors and analyze the effect of culture mode. Results:Embryo development assessment showed the day 3 (D3) high-quality embryo rate in the connected group [60.12% (12 136/20 187)] was significantly lower than that in the individual group [63.62% (4 705/7 395), P<0.001], whereas the high-quality blastocyst formation rate [34.93% (7 052/20 187)] and the available blastocyst formation rate [56.07% (11 319/20 187)] were both significantly higher than those in the individual group [33.08% (2 446/7 395), P=0.004; 51.45% (3 805/7 395), P<0.001], with statistically significant differences. The implantation rate [67.40% (1 774/2 632)], the clinical pregnancy rate [70.20% (1 717/2 446)], and the live birth rate [60.66% (1 469/2 446)] in the connected group were all significantly higher than those in the individual group [63.40% (724/1 142), P=0.017; 66.73% (708/1 061), P=0.041; 55.89% (593/1 061), P=0.021], with statistically significant differences. Neonatal outcomes showed no statistically significant difference between the two groups (all P>0.05). After adjusting for confounding factors using GLM, connected culture was an independent influencing factor for D3 high-quality embryo rate (a MD=-0.017, 95% CI: -0.034-0.000, P=0.046), high-quality blastocyst formation rate (a MD=-0.020, 95% CI: 0.002-0.037, P=0.026), available blastocyst formation rate (a MD=0.032, 95% CI: 0.015-0.048, P<0.001), live birth rate (a OR=1.182, 95% CI: 1.006-1.388, P=0.042). However, it had no effect on D3 available embryo rate, clinical pregnancy rate, or early miscarriage rate (all P>0.05). Conclusion:In TL incubator systems, individual and connected microdroplet culture modes exert different effects at various stages of embryo development. Individual microdroplet culture can significantly enhance cleavage-stage embryo quality, whereas the connected microdroplet culture was more beneficial for enhancing the blastocyst formation rate and quality, ultimately improving the live birth rate without increasing neonatal risks.

Objective:To investigate the value of habitat radiomic features based on gadolinium ethoxybenzyl diethylenetriamine pentaacetic acid (Gd-EOB-DTPA)-enhanced MRI in establishing a predictive model for cytokeratin 19 (CK19) expression in hepatocellular carcinoma (HCC) and to evaluate its role in prognostic risk stratification.Methods:This multicenter case-control study retrospectively enrolled 489 patients with pathologically confirmed HCC who underwent Gd-EOB-DTPA-enhanced MRI between June 2016 and June 2024. Among them, 346 patients from the First Affiliated Hospital of Soochow University were divided into a training cohort ( n=245) and an internal test cohort ( n=101) via stratified sampling at a 7∶3 ratio. And 143 patients from Nantong Third Hospital Affiliated to Nantong University served as an external validation cohort. The training cohort included 53 CK19-positive and 192 CK19-negative patients. The internal test cohort included 21 CK19-positive and 80 CK19-negative patients. The external validation cohort included 30 CK19-positive and 113 CK19-negative patients. Univariate logistic regression analysis was performed to identify potential factors associated with CK19 expression, and a clinical-radiologic model was constructed. The k-means clustering algorithm was applied to segment target HCC lesions into 3 subregions. Radiomic features were extracted and selected from these habitat subregions. Habitat radiomics models were constructed for the arterial phase (AP), portal venous phase, hepatobiliary phase (HBP), and combined phases (CP). Multivariate logistic regression analysis identified independent clinical and radiologic predictors of CK19 expression, and the optimal habitat model score was integrated to build a clinical-radiologic-habitat combined model. The area under the receiver operating characteristic curve (AUC) was used to evaluate model predictive performance. Recurrence-free survival (RFS) was analyzed using the Kaplan-Meier method and the differences in survival curves were compared with the log-rank test. Results:Univariate logistic regression analysis revealed that alpha-fetoprotein (AFP) ( OR=2.629, 95% CI 1.412-4.896, P=0.002), AP enhancement ( OR=3.636, 95% CI 1.642-8.052, P=0.001), AP peritumoral enhancement ( OR=2.219, 95% CI 1.084-4.542, P=0.029), and HBP peritumoral hypointensity ( OR=2.010, 95% CI 1.004-4.021, P=0.049) were potential factors associated with CK19 expression, which were incorporated into the clinical-radiologic model. In the internal and external validation cohorts, the AUC of the clinical-radiologic model was 0.690 (95% CI 0.590-0.778) and 0.650 (95% CI 0.565-0.727), respectively. The habitat radiomics model based on CP images demonstrated the highest performance. It achieved AUC of 0.729 (95% CI 0.622-0.836) and 0.725 (95% CI 0.607-0.842) in the internal and external validation cohorts, respectively. Multivariate analysis identified AFP ( OR=2.494, 95% CI 1.163-5.348, P=0.019), AP enhancement ( OR=5.230, 95% CI 1.868-14.643, P=0.002) and habitat radiomics model score ( OR=4.105, 95% CI 2.643-6.368, P<0.001) as independent predictors of CK19 positivity. Based on these factors, a combined clinical-radiologic-habitat combined model was established. The clinical-radiologic-habitat combined model achieved AUCs of 0.767 (95% CI 0.671-0.846) and 0.730 (95% CI 0.649-0.801) in the internal and external validation cohorts, respectively. Significant differences in RFS were observed between the CK19-positive group (25.1 month) and CK19-negative group (51.0 month) as predicted by the clinical-radiologic-habitat model ( χ2=4.17, P=0.041). Conclusion:The clinical-radiologic-habitat combined model based on Gd-EOB-DTPA-enhanced MRI habitat radiomics demonstrates good predictive performance for CK19 expression in HCC and offers valuable prognostic stratification for clinical practice.

Objective:To explore the value of the deep learning model based on gadolinium ethoxybenzyl diethylenetriamine pentaacetic acid (Gd-EOB-DTPA) enhanced MRI in preoperatively predicting vessels completely encapsulating tumor clusters (VETC) in hepatocellular carcinoma (HCC).Methods:This study adopted a case-control design to retrospectively analyze 420 patients with HCC confirmed by postoperative pathology who underwent Gd-EOB-DTPA enhanced MRI between June 2016 and March 2023. A total of 420 patients were divided into a training set ( n=305) from the First Affiliated Hospital of Soochow University and an external validation set ( n=115) from Affiliated Nantong Hospital 3 of Nantong University. Based on postoperative pathological findings, patients were stratified into VETC-positive and VETC-negative groups. The training set comprised 161 VETC-positive cases and 144 VETC-negative cases, while the external validation set included 55 VETC-positive cases and 60 VETC-negative cases. Tumor regions of interest in arterial, portal venous, and hepatobiliary phases were manually delineated using ITK-SNAP software. Pre-trained Vgg19, Densenet121, and Vision Transformer (ViT) models were employed for transfer learning, extracting deep learning features from each image. Feature data were processed using FAE software, and 12 logistic regression models (arterial phase, portal venous phase, hepatobiliary phase, and combined three-phase models) were constructed to select the optimal deep learning model. Independent predictors in clinical characteristics were identified through univariate and multivariate logistic analyses to establish a clinical model for predicting VETC pattern. Subsequently, a clinical-deep learning fusion model was developed by integrating these clinical predictors with the optimal deep learning features. Model performance in predicting VETC-positive HCC was evaluated using receiver operating characteristic curves, calibration curves, and decision curve analysis (DCA). Results:In the external validation set, the area under the curve (AUC) of the Vgg19 model in the arterial phase, portal venous phase, hepatobiliary phase, and combined three-phase, respectively were 0.799,0.756,0.789,0.821, which were higher than those of Densenet121 (AUC: 0.544,0.581,0.544,0.583) and ViT (AUC: 0.740,0.752,0.785,0.767) model. The three-phase combined Vgg19 model achieved the highest AUC of 0.821 (95% CI 0.746-0.897). Multivariate logistic regression identified alpha-fetoprotein level ( OR=1.826,95% CI 1.069-3.120, P=0.028) and tumor diameter ( OR=1.329,95% CI 1.206-1.466, P<0.001) as independent predictors of VETC-positive HCC, forming the clinical model with an AUC of 0.789 (95% CI 0.703-0.859). The clinical-deep learning fusion model further achieved the AUC of 0.825 (95% CI 0.749-0.900). Calibration curves confirmed high concordance between predicted and actual probabilities for the three-phase Vgg19 model, while DCA revealed greater net clinical benefit for the combined Vgg19 and fusion models compared with the clinical model alone. Conclusions:The deep learning model based on Gd-EOB-DTPA enhanced MRI can be used to predict VETC of HCC preoperatively, among which the three-phase combined Vgg19 model and the clinical-deep learning model provide high predictive value.