Liver transplantation (LT) has become a standard treatment for end-stage liver diseases, and graft injury is intricately associated with poor prognosis. Granzyme B (GZMB) plays a vital role in natural killer (NK) cell biology, but whether NK-derived GZMB affects graft injury remains elusive. Through the analysis of single-cell RNA-sequencing data obtained from human LT grafts and the isolation of lymphocytes from mouse livers following ischemia-reperfusion injury (IRI), we demonstrated that 2NK cells with high expression of GZMB are enriched in patients and mice. Both systemically and liver-targeted depletion of NK cells led to a notable reduction in GZMB⁺ cell infiltration, subsequently resulting in diminished graft injury. Notably, the reconstitution of Il2rg ^-/- Rag2 ^-/- mice with purified Gzmb-KO NK cells demonstrated superior outcomes compared to those with wild-type NK cells. Crucially, global knockout of GZMB and pharmacological inhibition exhibited remarkable improvements in liver function in both mouse IRI and rat LT models. Moreover, a phosphorylated derivative of FDA-approved vidarabine was identified as an effective inhibitor of mouse GZMB activity by molecular dynamics, which could provide a potential avenue for therapeutic intervention. Therefore, targeting NK cell-derived GZMB during the LT process suggests potential therapeutic strategies to improve post-transplant outcomes.

Objective:To investigate the risk factors and their diagnostic efficacy of perioperative lower limb deep vein thrombosis (DVT) in polytrauma patients with predominant severe limb trauma.Methods:A retrospective cohort study was conducted to analyze the clinical data of 155 polytrauma patients with predominant severe trauma who were admitted to Wuxi Ninth People′s Hospital from January 2021 to December 2024, including 64 males and 91 females, aged 13-95 years [(52.1±16.9)years]. Abbreviated injury scale (AIS) was 5-15 points [(7.4±2.1)points] and injury severity score (ISS) was 17-59 points [(21.3±6.5)points]. Based on the occurrence of DVT in the perioperative period, the patients were divided into preoperative DVT group with 17 patients (11.0%) and non-preoperative DVT group with 138 patients (89.0%) as well as postoperative DVT group with 24 patients (15.5%) and non-postoperative DVT group with 131 patients (84.5%). Basic clinical data were collected, including gender, age, body mass index (BMI), underlying diseases (hypertension, diabetes mellitus), hemoglobin level (Hb), platelet count (PLT), D-dimer, ISS, trauma site [cranial and brain trauma, thoracic and abdominal trauma, upper limb trauma, lower limb trauma (femoral fracture, patellar fracture, tibial or fibular fracture, foot fracture, vascular injury), and pelvic fracture], preoperative waiting time for surgery, surgical site (pelvis and lower limb, other areas), surgical protocols (pelvic and lower limb internal fixation, external fixation of lower limb, lower limb amputation), operation duration less or more than 2 hours, amount of intraoperative blood loss, intraoperative blood transfusion requirement, venous thromboembolism (VTE) prophylaxis (pharmacological and mechanical modalities) and length of hospital stay. Univariate analysis and multivariate binary Logistic regression analysis were conducted to investigate the correlation between the aforementioned indicators and incidence of perioperative lower limb DVT in polytrauma patients with predominant severe limb trauma and determine the independent risk factors. Receiver operating characteristic (ROC) curve and area under the curve (AUC) of the relevant risk factors were analyzed to evaluate and compare the diagnostic efficacy of the risk factors for perioperative lower limb DVT in polytrauma patients with predominant severe limb trauma.Results:Univariate analysis results showed that age, history of hypertension, D-dimer, thoracic and abdominal trauma, pelvic fracture, preoperative waiting time for surgery, and length of hospital stay were significantly correlated with preoperative of DVT of the lower limbs in the patients ( P<0.05). The results of multivariate binary Logistic regression analysis showed that age ( OR=1.05, 95% CI 1.00, 1.10, P<0.05), pelvic fracture ( OR=5.03, 95% CI 1.09, 23.20, P<0.05), preoperative waiting time for surgery ( OR=1.10, 95% CI 1.00, 1.22, P<0.05) and length of hospital stay ( OR=0.89,95% CI 0.81,0.98, P<0.05) were highly correlated with preoperative DVT of the lower limbs in the patients ( P<0.05). Univariate analysis results showed that age, D-dimer, ISS, foot fracture, and length of hospital stay were significantly correlated with postoperative DVT of the lower limbs in the patients ( P<0.05). The results of multivariate binary Logistic regression analysis showed that age ( OR=1.05, 95% CI 1.01, 1.08, P<0.01), D-dimer ( OR=1.05, 95% CI 1.00, 1.10, P<0.05), ISS ( OR=1.09, 95% CI 1.01, 1.17, P<0.05), and foot fracture ( OR=3.51 , 95% CI 1.25 , 9.87 , P<0.05) were significantly correlated with postoperative DVT of the lower limbs in the patients ( P<0.05). The results of the ROC curve analysis indicated that preoperative waiting time for surgery (AUC=0.83, 95% CI 0.75, 0.91) had the highest diagnostic efficacy for preoperative DVT of the lower limbs in the patients, with the diagnostic efficacies of pelvic fracture (AUC=0.75, 95% CI 0.65, 0.85) and age (AUC=0.70, 95% CI 0.59, 0.82) decreasing successively. For postoperative DVT of the lower limbs in the patients, D-dimer (AUC=0.71, 95% CI 0.61, 0.81) exhibited the highest diagnostic efficacy, followed by age (AUC=0.70, 95% CI 0.59, 0.81), ISS (AUC=0.64, 95% CI 0.51, 0.76) and foot fracture (AUC=0.62, 95% CI 0.49, 0.74), with diagnostic efficacy decreased successively. Conclusions:For polytrauma patients with predominant severe limb trauma, age, pelvic fracture and preoperative waiting time for surgery are independent risk factors for preoperative DVT, while age, D-dimer, ISS and foot fracture are independent risk factors for postoperative DVT. Additionally, preoperative waiting time for surgery has the best diagnostic efficacy for preoperative DVT, followed by pelvic fracture and age. D-dimer has the best diagnostic efficacy for postoperative DVT, followed by age, ISS and foot fracture.

Objective:To analyze the short-term hospital-based transmission characteristics and gene variation of Carbapenem-Resistant Klebsiella pneumoniae (CRKP) by genome-wide technique to provide evidence for transmission control. Methods:The experimental strain was derived from all the CRKP isolated in Affiliated Hospital of North China University of Science and Technology from October 2022 to December 2023. Strain identification and drug susceptibility were tested with VITEK 2-Compact automatic bacterial identification drug susceptibility analyzer or disk method, and the results were interpreted through whole genome sequencing. The ST type, carbapenem resistance gene, virulence factor, and O serotype of the collected strains were analyzed.Results:Among the 115 strains of CRKP, 94 strains were isolated from the intensive care unit (ICU), accounting for 81.7%, and 21 strains were isolated from the non-intensive care unit (NICU), accounting for 18.3%. The 115 strains of CRKP can be divided into 11 ST types, of which ST11 type was the most (54.8%, 63/115), followed by ST15 type (22.6%, 26/115) and ST5492 type (15.7%, 18/115). Type ST5492 was a new clonal group in the region. The 115 strains of CRKP could be divided into 7 O serotypes, most of which were O2a type(32.2%,37/115), followed by O5 type(30.4%,35/115) and O1 type(27.8%,32/115). The resistance genes of carbapenem antibiotics showed that there were 107 strains carrying the blaKPC-2 gene, one strain with the blaNDM-1 gene, and one strain with both the blaKPC-2 and blaNDM-13 genes. Virulence genes were detected in 55 CRKP strains (47.8%, 55/115), among which six strains detected peg-344, iucA, iroB, rmpA, and rmpA2 virulence genes (5.2%, 6/115). Four virulence genes ( peg-344, iucA, rmpA, and rmpA2) were detected in 34 strains (29.6%, 34/115). Three virulence genes ( iucA, iroB and rmpA) were detected in two strains (1.7%, 2/115). Three virulence genes ( peg-344, iucA and rmpA) were detected in one strain (0.8%, 1/115). IucA and rmpA virulence genes were detected in 12 strains (10.4%, 12/115). KPC-2_ST11_O2a, KPC-2_ST15_O1 and KPC-2_ST5492_O5 were dominant clones, and their distribution was mainly in the intensive care unit. The whole genome sequence analysis showed that there were three dominant clones, among which ST11 clones were subdivided into three dominant O serotypes, all of which were mainly in the intensive care unit. Conclusion:The popular strain in the hospital of CRKP is a KPC-2_ST11 clone group carrying iucA, rmpA/rmpA2, with cross-department transmission and mutation. ST5492 is a newly-launched clone type. The intensive care unit of hvKP carrying five virulence genes, including peg-344, should be alert to the epidemic risk of CR-hvKP outbreak.

Objective:To analyze the carriage status, subtype distribution and flanking gene sequence characteristics of oxacillinases (OXA enzyme) in 241 clinical strains of Pseudomonas aeruginosa, and assess their roles in the drug resistance of Pseudomonas aeruginosa and ability to horizontally transfer across species. Methods:Clinical P. aeruginosa isolates were collected from four hospitals in Sanya, Tangshan, Zhangjiakou, and Beijing. The prevalence of oxacillinases and their flanking gene sequences was analyzed by whole-genome sequencing (NGS) and bioinformatic approaches. Results:A total of 241 isolates of P. aeruginosa were gathered, and 35 blaOXA subtypes were identified through screening of 252 blaOXA genes. These genes were classified into three subfamilies: blaOXA-50-like (241, 95.6%), blaOXA-1-like (9, 3.6%) and blaOXA-10-like (2, 0.8%). Among these, 11 subtypes (11, 31.4%) were novel blaOXA subtypes. Nine of these belonged to the blaOXA-50-like subfamily and were designated as blaOXA-1244, blaOXA-1245, blaOXA-1246, blaOXA-1250, blaOXA-1252, blaOXA-1253, blaOXA-1254, blaOXA-1255, and blaOXA-1256. The remaining two belonged to the blaOXA-10-like subfamily and were named blaOXA-1247 and blaOXA-1248. Compared to the amino acid sequence of OXA-10, the newly identified subtype OXA-1247 exhibited a mutation at position 117, where a valine was replaced by a leucine. This change was thought to improve the enzyme′s ability to hydrolyze carbapenems. In the analysis of the flanking sequences of the blaOXA genes, Class I integrons were identified in four bacterial strains. The variable regions of these integrons carried three distinct patterns of resistance gene cassettes: aac( 6′) -Ib-blaOXA-1247-ant( 3′′) -Ia, aac( 6′) -Ib-blaOXA-1248 and aac( 6′) -Ib- blaIMP-45-blaOXA-1-catB3. Among these, the strain BJ2326 carried a class I integron that was connected to the downstream IS CR1 element to form a composite class I integron structure, additionally carrying the resistance gene blaPER-1. Out of the 223 non-wild-type P. aeruginosa strains, 127 strains exhibited non-wild-type profiles to the four beta-lactam antibiotics MEM, CAZ, FEP, and TZP, with the combination of MEM+CAZ+FEP being the most prevalent, representing 57.0% of the total. Conclusions:The blaOXA genes in 241 clinical P. aeruginosa strains showed diversity. Some blaOXA genes had a co-transfer risk with the metallo-β-lactamase resistance gene blaIMP-45. Among the 11 newly discovered blaOXA subtypes, the new subtype OXA-1247 may have carbapenemase activity and potential for horizontal transfer.

Objective:To analyze the short-term hospital-based transmission characteristics and gene variation of Carbapenem-Resistant Klebsiella pneumoniae (CRKP) by genome-wide technique to provide evidence for transmission control. Methods:The experimental strain was derived from all the CRKP isolated in Affiliated Hospital of North China University of Science and Technology from October 2022 to December 2023. Strain identification and drug susceptibility were tested with VITEK 2-Compact automatic bacterial identification drug susceptibility analyzer or disk method, and the results were interpreted through whole genome sequencing. The ST type, carbapenem resistance gene, virulence factor, and O serotype of the collected strains were analyzed.Results:Among the 115 strains of CRKP, 94 strains were isolated from the intensive care unit (ICU), accounting for 81.7%, and 21 strains were isolated from the non-intensive care unit (NICU), accounting for 18.3%. The 115 strains of CRKP can be divided into 11 ST types, of which ST11 type was the most (54.8%, 63/115), followed by ST15 type (22.6%, 26/115) and ST5492 type (15.7%, 18/115). Type ST5492 was a new clonal group in the region. The 115 strains of CRKP could be divided into 7 O serotypes, most of which were O2a type(32.2%,37/115), followed by O5 type(30.4%,35/115) and O1 type(27.8%,32/115). The resistance genes of carbapenem antibiotics showed that there were 107 strains carrying the blaKPC-2 gene, one strain with the blaNDM-1 gene, and one strain with both the blaKPC-2 and blaNDM-13 genes. Virulence genes were detected in 55 CRKP strains (47.8%, 55/115), among which six strains detected peg-344, iucA, iroB, rmpA, and rmpA2 virulence genes (5.2%, 6/115). Four virulence genes ( peg-344, iucA, rmpA, and rmpA2) were detected in 34 strains (29.6%, 34/115). Three virulence genes ( iucA, iroB and rmpA) were detected in two strains (1.7%, 2/115). Three virulence genes ( peg-344, iucA and rmpA) were detected in one strain (0.8%, 1/115). IucA and rmpA virulence genes were detected in 12 strains (10.4%, 12/115). KPC-2_ST11_O2a, KPC-2_ST15_O1 and KPC-2_ST5492_O5 were dominant clones, and their distribution was mainly in the intensive care unit. The whole genome sequence analysis showed that there were three dominant clones, among which ST11 clones were subdivided into three dominant O serotypes, all of which were mainly in the intensive care unit. Conclusion:The popular strain in the hospital of CRKP is a KPC-2_ST11 clone group carrying iucA, rmpA/rmpA2, with cross-department transmission and mutation. ST5492 is a newly-launched clone type. The intensive care unit of hvKP carrying five virulence genes, including peg-344, should be alert to the epidemic risk of CR-hvKP outbreak.

Objective:To analyze the carriage status, subtype distribution and flanking gene sequence characteristics of oxacillinases (OXA enzyme) in 241 clinical strains of Pseudomonas aeruginosa, and assess their roles in the drug resistance of Pseudomonas aeruginosa and ability to horizontally transfer across species. Methods:Clinical P. aeruginosa isolates were collected from four hospitals in Sanya, Tangshan, Zhangjiakou, and Beijing. The prevalence of oxacillinases and their flanking gene sequences was analyzed by whole-genome sequencing (NGS) and bioinformatic approaches. Results:A total of 241 isolates of P. aeruginosa were gathered, and 35 blaOXA subtypes were identified through screening of 252 blaOXA genes. These genes were classified into three subfamilies: blaOXA-50-like (241, 95.6%), blaOXA-1-like (9, 3.6%) and blaOXA-10-like (2, 0.8%). Among these, 11 subtypes (11, 31.4%) were novel blaOXA subtypes. Nine of these belonged to the blaOXA-50-like subfamily and were designated as blaOXA-1244, blaOXA-1245, blaOXA-1246, blaOXA-1250, blaOXA-1252, blaOXA-1253, blaOXA-1254, blaOXA-1255, and blaOXA-1256. The remaining two belonged to the blaOXA-10-like subfamily and were named blaOXA-1247 and blaOXA-1248. Compared to the amino acid sequence of OXA-10, the newly identified subtype OXA-1247 exhibited a mutation at position 117, where a valine was replaced by a leucine. This change was thought to improve the enzyme′s ability to hydrolyze carbapenems. In the analysis of the flanking sequences of the blaOXA genes, Class I integrons were identified in four bacterial strains. The variable regions of these integrons carried three distinct patterns of resistance gene cassettes: aac( 6′) -Ib-blaOXA-1247-ant( 3′′) -Ia, aac( 6′) -Ib-blaOXA-1248 and aac( 6′) -Ib- blaIMP-45-blaOXA-1-catB3. Among these, the strain BJ2326 carried a class I integron that was connected to the downstream IS CR1 element to form a composite class I integron structure, additionally carrying the resistance gene blaPER-1. Out of the 223 non-wild-type P. aeruginosa strains, 127 strains exhibited non-wild-type profiles to the four beta-lactam antibiotics MEM, CAZ, FEP, and TZP, with the combination of MEM+CAZ+FEP being the most prevalent, representing 57.0% of the total. Conclusions:The blaOXA genes in 241 clinical P. aeruginosa strains showed diversity. Some blaOXA genes had a co-transfer risk with the metallo-β-lactamase resistance gene blaIMP-45. Among the 11 newly discovered blaOXA subtypes, the new subtype OXA-1247 may have carbapenemase activity and potential for horizontal transfer.

Objective:To investigate the risk factors and their diagnostic efficacy of perioperative lower limb deep vein thrombosis (DVT) in polytrauma patients with predominant severe limb trauma.Methods:A retrospective cohort study was conducted to analyze the clinical data of 155 polytrauma patients with predominant severe trauma who were admitted to Wuxi Ninth People′s Hospital from January 2021 to December 2024, including 64 males and 91 females, aged 13-95 years [(52.1±16.9)years]. Abbreviated injury scale (AIS) was 5-15 points [(7.4±2.1)points] and injury severity score (ISS) was 17-59 points [(21.3±6.5)points]. Based on the occurrence of DVT in the perioperative period, the patients were divided into preoperative DVT group with 17 patients (11.0%) and non-preoperative DVT group with 138 patients (89.0%) as well as postoperative DVT group with 24 patients (15.5%) and non-postoperative DVT group with 131 patients (84.5%). Basic clinical data were collected, including gender, age, body mass index (BMI), underlying diseases (hypertension, diabetes mellitus), hemoglobin level (Hb), platelet count (PLT), D-dimer, ISS, trauma site [cranial and brain trauma, thoracic and abdominal trauma, upper limb trauma, lower limb trauma (femoral fracture, patellar fracture, tibial or fibular fracture, foot fracture, vascular injury), and pelvic fracture], preoperative waiting time for surgery, surgical site (pelvis and lower limb, other areas), surgical protocols (pelvic and lower limb internal fixation, external fixation of lower limb, lower limb amputation), operation duration less or more than 2 hours, amount of intraoperative blood loss, intraoperative blood transfusion requirement, venous thromboembolism (VTE) prophylaxis (pharmacological and mechanical modalities) and length of hospital stay. Univariate analysis and multivariate binary Logistic regression analysis were conducted to investigate the correlation between the aforementioned indicators and incidence of perioperative lower limb DVT in polytrauma patients with predominant severe limb trauma and determine the independent risk factors. Receiver operating characteristic (ROC) curve and area under the curve (AUC) of the relevant risk factors were analyzed to evaluate and compare the diagnostic efficacy of the risk factors for perioperative lower limb DVT in polytrauma patients with predominant severe limb trauma.Results:Univariate analysis results showed that age, history of hypertension, D-dimer, thoracic and abdominal trauma, pelvic fracture, preoperative waiting time for surgery, and length of hospital stay were significantly correlated with preoperative of DVT of the lower limbs in the patients ( P<0.05). The results of multivariate binary Logistic regression analysis showed that age ( OR=1.05, 95% CI 1.00, 1.10, P<0.05), pelvic fracture ( OR=5.03, 95% CI 1.09, 23.20, P<0.05), preoperative waiting time for surgery ( OR=1.10, 95% CI 1.00, 1.22, P<0.05) and length of hospital stay ( OR=0.89,95% CI 0.81,0.98, P<0.05) were highly correlated with preoperative DVT of the lower limbs in the patients ( P<0.05). Univariate analysis results showed that age, D-dimer, ISS, foot fracture, and length of hospital stay were significantly correlated with postoperative DVT of the lower limbs in the patients ( P<0.05). The results of multivariate binary Logistic regression analysis showed that age ( OR=1.05, 95% CI 1.01, 1.08, P<0.01), D-dimer ( OR=1.05, 95% CI 1.00, 1.10, P<0.05), ISS ( OR=1.09, 95% CI 1.01, 1.17, P<0.05), and foot fracture ( OR=3.51 , 95% CI 1.25 , 9.87 , P<0.05) were significantly correlated with postoperative DVT of the lower limbs in the patients ( P<0.05). The results of the ROC curve analysis indicated that preoperative waiting time for surgery (AUC=0.83, 95% CI 0.75, 0.91) had the highest diagnostic efficacy for preoperative DVT of the lower limbs in the patients, with the diagnostic efficacies of pelvic fracture (AUC=0.75, 95% CI 0.65, 0.85) and age (AUC=0.70, 95% CI 0.59, 0.82) decreasing successively. For postoperative DVT of the lower limbs in the patients, D-dimer (AUC=0.71, 95% CI 0.61, 0.81) exhibited the highest diagnostic efficacy, followed by age (AUC=0.70, 95% CI 0.59, 0.81), ISS (AUC=0.64, 95% CI 0.51, 0.76) and foot fracture (AUC=0.62, 95% CI 0.49, 0.74), with diagnostic efficacy decreased successively. Conclusions:For polytrauma patients with predominant severe limb trauma, age, pelvic fracture and preoperative waiting time for surgery are independent risk factors for preoperative DVT, while age, D-dimer, ISS and foot fracture are independent risk factors for postoperative DVT. Additionally, preoperative waiting time for surgery has the best diagnostic efficacy for preoperative DVT, followed by pelvic fracture and age. D-dimer has the best diagnostic efficacy for postoperative DVT, followed by age, ISS and foot fracture.

Objective To evaluate and compare the success rate and short-term clinical prognosis of transfemoral transcatheter aortic valve replacement(TF-TAVR)for patients with pure aortic regurgitation(PAR)of different annulus sizes.Methods This study is a single center retrospective study,selecting symptomatic PAR patients who received TF-TAVR treatment at Zhongshan Hospital Fudan University from September 2019 to September 2023.Based on preoperative CT results,all patients were divided into three groups:Group A(aortic annulus circumference＜80 mm),Group B(80 mm≤aortic annulus circumference＜85 mm),and Group C(aortic annulus circumference&ge; 85 mm).The primary endpoint was success rate and 30d all-cause mortality,while the secondary endpoint was TAVR related complications.Results A total of 61 PAR patients were included in this study,including 27 in Group A,21 in Group B,and 13 in Group C.The overall success rate is 82.0％,and the 30 d all-cause mortality rate is 3.3％.The success rate of Group C patients was significantly lower(P=0.012),with higher rates of conversion to surgery and valve-in-valve implantation(P=0.022 and P=0.040).In terms of secondary endpoint events,there were no significant differences among the three groups in major bleeding events,major vascular complications,stroke,myocardial infarction,newly developed atrial fibrillation,implantation of new pacemakers,coronary artery occlusion,and postoperative moderate to severe perivalvular leakage(all P＞0.05).Conclusions The circumference of the aortic valve annulus is a key factor affecting the success rate of TF-TAVR in PAR,and PAR patients with an aortic valve annulus circumference less than 85mm may be more suitable for TF-TAVR.

Objective:To investigate the effects of PD-1 monoclonal antibody therapy on the peripheral and local immune microenvironment of patients with microsatellite instability-high(MSI-H)rectal cancer.Methods:Samples of peripheral blood and tumor biopsy were collected from a patient with MSI-H rectal cancer before and after PD-1 monoclonal antibody treatment.The samples were dissociated into single-cell suspensions using a combination of enzymatic and mechanical methods.Immune-related marker expression on peripheral and tumor-infiltrating im-mune cells was analyzed using single-cell mass cytometry(CyTOF).Results:According to the results of CyTOF analysis,CD45+immune cells in the peripheral blood and tumor tissues were categorized into 39 and 34 cell subsets,respectively,before and after PD-1 monoclonal antibody treatment(the correlation is unclear and ambiguous).After PD-1 monoclonal antibody treatment,differences were observed in the relative abundance of immune cell subsets:B cells significantly decreased in the peripheral blood,while B cells and γδT cells significantly increased in the tumor tissue;neutrophils significantly decreased,and the proportion of CD4+TEM cells in T cell subsets significantly increased,whereas CD4+Treg cells significantly decreased.Additionally,there were differences in the expression of immune-related markers in multiple immune cell subsets in both peripheral blood and tumor tissues,with CCR6 showing a significant increase in expression across all subsets,while ICOS and PD-1 expressions in T cell subsets were significantly reduced(the specific tissues for these cells or factors are unclear).Conclusion:After PD-1 monoclonal antibody treatment in MSI-H rectal cancer,changes occurred in the composition of immune cells and the expression of immune-related markers in both peripheral blood and tumor tissues.This study reveals the dynamic adjustment of the immune microenvironment and provides important evidence for understanding the therapeutic mechanism of PD-1 monoclonal antibodies.

Objective To investigate the clinical characteristics of female migraine patients with patent foramen ovale(PFO)and design a risk prediction model for PFO in female migraine patients(migraineur patients PFO risk prediction model,MPRPM).Methods Female migraine patients who visited Zhongshan Hospital,Fudan University from Jun 1,2019 to Dec 31,2022 were included.Preoperative information and follow-up results after discontinuation of medication were collected.Patients were divided into PFO-positive and PFO-negative groups based on transesophageal echocardiography results.A multivariate Logistic regression model and a random forest model were constructed,and the random forest model was validated multidimensionally.Key features were selected based on the mean decrease accuracy(MDA)to construct MPRPM.Results A total of 305 female patients were included in the study,with 204 patients in the PFO-positive group and 101 patients in the PFO-negative group.Multivariate Logistic regression analysis showed that age at migraine onset,attack frequency,severe impact on life during attacks,exercise-related headaches,menstruation-induced headaches,aura migraines,and a history of cryptogenic stroke were predictive factors for PFO positivity.The random forest model effectively predicted the incidence of PFO in female migraine patients,with an AUC of 0.895(95%CI:0.847-0.943).MPRPM demonstrated a sensitivity of 71.6%and specificity of 91.1%(AUC:0.862,95%CI:0.818-0.906,P<0.001).The optimal cut-off value was 2.5 points.Patients correctly classified by the model showed a higher rate of symptom improvement compared to incorrectly classified patients(94.3%vs.82.0%,P=0.023).Conclusion We identified predictive factors for PFO in migraine patients.MPRPM can provide guidance in the diagnostic process and therapeutic decision-making for female migraine patients,assist in patient triage,and reduce the healthcare burden.