Haobo Li
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Haobo Li has a total of 139 co-authors. Here are the top co-authors by number of shared publications:
- Zhenguo Zhai
China-Japan Friendship Hospital
12 shared publications Zhu Zhang
Chinese Academy of Medical Sciences
10 shared publicationsWanmu Xie
National Center for Respiratory Medicine, State Key Laboratory of Respiratory Health and Multimorbidity, National Clinical Research Center for Respiratory Diseases, Institute of Respiratory Medicine, Chinese Academy of Medical Sciences & Peking Union Medical College, Department of Integrative Pulmonary Medicine, Center of Respiratory Medicine, China-Japan Friendship Hospital, Beijing, China (W.X.).
7 shared publicationsPeiran Yang
State Key Laboratory of Respiratory Health and Multimorbidity, Department of Physiology, Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences and School of Basic Medicine, Peking Union Medical College; National Center for Respiratory Medicine; Institute of Respiratory Medicine, Chinese Academy of Medical Sciences; National Clinical Research Center for Respiratory Diseases, Beijing, China
5 shared publicationsShuai Zhang
National Center for Respiratory Medicine; State Key Laboratory of Respiratory Health and Multimorbidity; National Clinical Research Center for Respiratory Diseases; Institute of Respiratory Medicine, Chinese Academy of Medical Sciences; Department of Pulmonary and Critical Care Medicine, Center of Respiratory Medicine, China-Japan Friendship Hospital, Beijing, China
5 shared publicationsYu Zhang
National Center for Respiratory Medicine; State Key Laboratory of Respiratory Health and Multimorbidity; National Clinical Research Center for Respiratory Diseases; Institute of Respiratory Medicine, Chinese Academy of Medical Sciences; Department of Pulmonary and Critical Care Medicine, Center of Respiratory Medicine, China-Japan Friendship Hospital, Beijing, China
5 shared publications
Latest Publications
- Open Access
- 8 November 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16 (21) , 11044
Sparse Point-to-Pseudo-Instance Supervision for Annotation-Efficient Nuclei Segmentation
Accurate nuclei instance segmentation requires pixel-level boundary annotation, which is expensive and time-consuming in histopathology images. We propose a sparse point-to-pseudo-instance supervision framework that learns from limited center-point annotations. The framework first completes sparse nucleus centers through curriculum-based center recovery and then generates instance-level pseudo labels using point-prompted foundation-model proposals. For completed centers without valid proposals, point-derived fallback regions are introduced to recover missing supervision, and all generated regions are assembled into unified pseudo-instance labels. The generated labels are converted into the native training targets of different downstream segmentation models, including HoVerNet and StarDist, while inference relies only on the trained segmentation networks. Experiments on MoNuSeg, CPM17, and CoNSeP evaluate segmentation performance under limited point budgets and analyze the coverage–precision trade-off introduced by fallback recovery. Under a 10% point annotation budget, the framework achieves AJI and PQ scores of 0.6233 and 0.6036 on MoNuSeg and 0.6653 and 0.6626 on CPM17, respectively. Point budgets cover training; fully annotated validation subsets support model selection. Additional validation analyses quantify recovered nuclei and false additions, showing that fallback improves supervision coverage while its effectiveness depends on the proposal coverage and center accuracy. The results demonstrate that sparse point annotations can be effectively converted into transferable pseudo-instance supervision for nuclei instance segmentation.
- 8 November 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16 (21) , 11044
Open AccessSparse Point-to-Pseudo-Instance Supervision for Annotation-Efficient Nuclei Segmentation
Accurate nuclei instance segmentation requires pixel-level boundary annotation, which is expensive and time-consuming in histopathology images. We propose a sparse point-to-pseudo-instance supervision framework that learns from limited center-point annotations. The framework first completes sparse nucleus centers through curriculum-based center recovery and then generates instance-level pseudo labels using point-prompted foundation-model proposals. For completed centers without valid proposals, point-derived fallback regions are introduced to recover missing supervision, and all generated regions are assembled into unified pseudo-instance labels. The generated labels are converted into the native training targets of different downstream segmentation models, including HoVerNet and StarDist, while inference relies only on the trained segmentation networks. Experiments on MoNuSeg, CPM17, and CoNSeP evaluate segmentation performance under limited point budgets and analyze the coverage–precision trade-off introduced by fallback recovery. Under a 10% point annotation budget, the framework achieves AJI and PQ scores of 0.6233 and 0.6036 on MoNuSeg and 0.6653 and 0.6626 on CPM17, respectively. Point budgets cover training; fully annotated validation subsets support model selection. Additional validation analyses quantify recovered nuclei and false additions, showing that fallback improves supervision coverage while its effectiveness depends on the proposal coverage and center accuracy. The results demonstrate that sparse point annotations can be effectively converted into transferable pseudo-instance supervision for nuclei instance segmentation.
- 15 October 2026
- Published by Wolters Kluwer Health in Arteriosclerosis, Thrombosis, and Vascular Biology
Prior Venous Thromboembolism Defines Distinct Genetic and Molecular Phenotypes in Chronic Thromboembolic Pulmonary Hypertension
BACKGROUND: Chronic thromboembolic pulmonary hypertension (CTEPH) has traditionally been considered a long-term complication of pulmonary embolism. However, a substantial proportion of patients have no history of venous thromboembolism (VTE), suggesting underlying biological heterogeneity. METHODS: We analyzed clinical and imaging characteristics of 712 patients with CTEPH stratified by prior VTE history. Whole-genome sequencing was performed in all patients and 2952 healthy controls to assess common variants. Bulk RNA sequencing of pulmonary endarterectomy tissues and plasma protein quantification were integrated to identify and validate subgroup-specific molecular signatures, with further validation in tissue and primary endothelial cells. RESULTS: Patients with CTEPH and prior VTE (n=576, 80.9%) exhibited a distinct thrombotic phenotype with more proximal disease distribution (92.5% versus 66.2%) and greater likelihood of undergoing pulmonary endarterectomy (22.6% versus 3.7%). Genome-wide association study in CTEPH-VTE(+) patients versus controls identified coagulation-related loci at ABO (rs676996, P =1.1×10 –9 ) and FGG (rs7681423, P =2.5×10 –8 ), with high genetic correlation with pulmonary embolism (rg=0.52, P =0.028). In contrast, no consistent genomic signal was identified in patients without prior VTE. Integrative multiomics (transcriptome-wide association study, MAGMA, and bulk RNA sequencing differential expression analysis) convergently prioritized CTSS (cathepsin S) as a key molecular candidate in CTEPH-VTE(+). CTSS was consistently elevated in plasma, pulmonary endarterectomy tissue, and primary endothelial cells from CTEPH-VTE(+) patients and promoted endothelial proliferation via ERK signaling, which was attenuated by the selective CTSS inhibitor LY3000328. CONCLUSIONS: Stratification by prior VTE history reveals biologically distinct CTEPH phenotypes with divergent genetic architectures and identifies cathepsin S as a genetically nominated therapeutic target for thrombosis-associated pulmonary vascular remodeling.
- 15 October 2026
- Published by Wolters Kluwer Health in Arteriosclerosis, Thrombosis, and Vascular Biology
Prior Venous Thromboembolism Defines Distinct Genetic and Molecular Phenotypes in Chronic Thromboembolic Pulmonary Hypertension
BACKGROUND: Chronic thromboembolic pulmonary hypertension (CTEPH) has traditionally been considered a long-term complication of pulmonary embolism. However, a substantial proportion of patients have no history of venous thromboembolism (VTE), suggesting underlying biological heterogeneity. METHODS: We analyzed clinical and imaging characteristics of 712 patients with CTEPH stratified by prior VTE history. Whole-genome sequencing was performed in all patients and 2952 healthy controls to assess common variants. Bulk RNA sequencing of pulmonary endarterectomy tissues and plasma protein quantification were integrated to identify and validate subgroup-specific molecular signatures, with further validation in tissue and primary endothelial cells. RESULTS: Patients with CTEPH and prior VTE (n=576, 80.9%) exhibited a distinct thrombotic phenotype with more proximal disease distribution (92.5% versus 66.2%) and greater likelihood of undergoing pulmonary endarterectomy (22.6% versus 3.7%). Genome-wide association study in CTEPH-VTE(+) patients versus controls identified coagulation-related loci at ABO (rs676996, P =1.1×10 –9 ) and FGG (rs7681423, P =2.5×10 –8 ), with high genetic correlation with pulmonary embolism (rg=0.52, P =0.028). In contrast, no consistent genomic signal was identified in patients without prior VTE. Integrative multiomics (transcriptome-wide association study, MAGMA, and bulk RNA sequencing differential expression analysis) convergently prioritized CTSS (cathepsin S) as a key molecular candidate in CTEPH-VTE(+). CTSS was consistently elevated in plasma, pulmonary endarterectomy tissue, and primary endothelial cells from CTEPH-VTE(+) patients and promoted endothelial proliferation via ERK signaling, which was attenuated by the selective CTSS inhibitor LY3000328. CONCLUSIONS: Stratification by prior VTE history reveals biologically distinct CTEPH phenotypes with divergent genetic architectures and identifies cathepsin S as a genetically nominated therapeutic target for thrombosis-associated pulmonary vascular remodeling.
- 7 July 2026
- Published by Wolters Kluwer Health in Journal of the American Heart Association
- Vol. 15 (13) , e047624
Evaluation of Pulmonary Artery Flow Reserve in Patients With Chronic Thromboembolic Pulmonary Hypertension: A Pilot Study
Background: Pressure‐wire fractional flow reserve (FFR) can guide balloon pulmonary angioplasty in chronic thromboembolic pulmonary hypertension, but lesion crossing with a distal wire increases cost and procedural risk. We developed an angiography‐derived physiologic index—pulmonary artery flow reserve (PFR)—computed from selective pulmonary angiography using fluid‐dynamic modeling without distal pressure‐wire advancement. Methods: In this prospective observational study, selective pulmonary angiography was acquired in 2 projections for type A (ring‐like), type B (web lesion) and type C (subtotal) pulmonary artery lesions. Proximal pressure (Pa) was measured at the target vessel ostium using a pressure‐sensor catheter; flow velocity was estimated by frame counting; pressure drop (ΔP) was computed by a proprietary computational fluid dynamics approach, and PFR was derived as (Pa − ΔP)/Pa. Pressure‐wire FFR (Pd/Pa) served as the reference standard (FFR r =0.970, P r =0.967; all P <0.001). Conclusions: Angiography‐derived PFR demonstrated strong agreement with pressure‐wire FFR for physiologic assessment of pulmonary artery stenoses in chronic thromboembolic pulmonary hypertension and may provide a simpler alternative to distal pressure‐wire measurements to support physiology‐guided balloon pulmonary angioplasty.
- 7 July 2026
- Published by Wolters Kluwer Health in Journal of the American Heart Association
- Vol. 15 (13) , e047624
Evaluation of Pulmonary Artery Flow Reserve in Patients With Chronic Thromboembolic Pulmonary Hypertension: A Pilot Study
Background: Pressure‐wire fractional flow reserve (FFR) can guide balloon pulmonary angioplasty in chronic thromboembolic pulmonary hypertension, but lesion crossing with a distal wire increases cost and procedural risk. We developed an angiography‐derived physiologic index—pulmonary artery flow reserve (PFR)—computed from selective pulmonary angiography using fluid‐dynamic modeling without distal pressure‐wire advancement. Methods: In this prospective observational study, selective pulmonary angiography was acquired in 2 projections for type A (ring‐like), type B (web lesion) and type C (subtotal) pulmonary artery lesions. Proximal pressure (Pa) was measured at the target vessel ostium using a pressure‐sensor catheter; flow velocity was estimated by frame counting; pressure drop (ΔP) was computed by a proprietary computational fluid dynamics approach, and PFR was derived as (Pa − ΔP)/Pa. Pressure‐wire FFR (Pd/Pa) served as the reference standard (FFR r =0.970, P r =0.967; all P <0.001). Conclusions: Angiography‐derived PFR demonstrated strong agreement with pressure‐wire FFR for physiologic assessment of pulmonary artery stenoses in chronic thromboembolic pulmonary hypertension and may provide a simpler alternative to distal pressure‐wire measurements to support physiology‐guided balloon pulmonary angioplasty.
- 7 May 2026
- Published by Elsevier in Journal of Thrombosis and Haemostasis
- Vol. 24 (10) , 3544-3552
Ambient temperature, air pollutants, and pulmonary embolism onset: a 17-year nationwide study in China
- 7 May 2026
- Published by Elsevier in Journal of Thrombosis and Haemostasis
- Vol. 24 (10) , 3544-3552
Ambient temperature, air pollutants, and pulmonary embolism onset: a 17-year nationwide study in China
- Open Access
- 22 December 2025
- Published by Elsevier in The Lancet Regional Health - Western Pacific
- Vol. 66
Population-specific ABO haplotypes reveal distinct venous thromboembolism risk in East Asians: insights from a large-scale genetic study
- 22 December 2025
- Published by Elsevier in The Lancet Regional Health - Western Pacific
- Vol. 66
Open AccessPopulation-specific ABO haplotypes reveal distinct venous thromboembolism risk in East Asians: insights from a large-scale genetic study
- Open Access
- 12 April 2025
- Published by Springer Nature in Communications Biology
- Vol. 8 (1) , 1-9
Whole genome sequencing identifies pathogenic genetic variants in Han Chinese patients with familial venous thromboembolism
Genetic factors play a pivotal role in determining venous thromboembolism (VTE) risk, particularly in cases of unprovoked early-onset VTE and those with a family history. While genome-wide association studies (GWAS) has advanced our understanding, high-quality whole-genome sequencing (WGS) from family-based studies is essential to elucidate the role of rare variants. In this study, we performed WGS on 216 individuals from 35 Han Chinese VTE pedigrees and validated findings in 99 high-heritability VTE cases using whole-exome sequencing. Functional impact was assessed via qPCR and Western Blot in HEK293T cells. Classical genes explained partial familial inheritance (20/35), while non-classical genes showed comparable effects on VTE recurrence and CTEPH. From 36 rare variants, 34 pedigrees (97%) were interpreted, with 29 variants reported for the first time. Notably, three novel variants, GP6 (c.G1094A:p.R365H), TET2 (c.G3451T:p.E1151X), and JAK2 (c.G380A:p.G127D), shared in two unrelated pedigrees each and are classified as low frequency in East Asians. Functional analyses revealed significant changes in GP6 and TET2 expression compared to the wild type. These findings provide novel insights into the genetic architecture of VTE and highlight GP6, TET2, and JAK2 as potential risk factors in East Asian populations, underscoring the clinical relevance of rare variants in VTE pathogenesis.
- 12 April 2025
- Published by Springer Nature in Communications Biology
- Vol. 8 (1) , 1-9
Open AccessWhole genome sequencing identifies pathogenic genetic variants in Han Chinese patients with familial venous thromboembolism
Genetic factors play a pivotal role in determining venous thromboembolism (VTE) risk, particularly in cases of unprovoked early-onset VTE and those with a family history. While genome-wide association studies (GWAS) has advanced our understanding, high-quality whole-genome sequencing (WGS) from family-based studies is essential to elucidate the role of rare variants. In this study, we performed WGS on 216 individuals from 35 Han Chinese VTE pedigrees and validated findings in 99 high-heritability VTE cases using whole-exome sequencing. Functional impact was assessed via qPCR and Western Blot in HEK293T cells. Classical genes explained partial familial inheritance (20/35), while non-classical genes showed comparable effects on VTE recurrence and CTEPH. From 36 rare variants, 34 pedigrees (97%) were interpreted, with 29 variants reported for the first time. Notably, three novel variants, GP6 (c.G1094A:p.R365H), TET2 (c.G3451T:p.E1151X), and JAK2 (c.G380A:p.G127D), shared in two unrelated pedigrees each and are classified as low frequency in East Asians. Functional analyses revealed significant changes in GP6 and TET2 expression compared to the wild type. These findings provide novel insights into the genetic architecture of VTE and highlight GP6, TET2, and JAK2 as potential risk factors in East Asian populations, underscoring the clinical relevance of rare variants in VTE pathogenesis.
- 12 March 2025
- Published by Springer Nature in Phenomics
- Vol. 5 (4) , 347-360
A Self-Adapting Polygenic Risk Score Model Improves Risk Prediction of Venous Thromboembolism in Han Chinese Cohorts
Most genome-wide association studies (GWAS) of Venous Thromboembolism (VTE) have used data from individuals of European descent, however, genetic factors for VTE have not been fully identified in Chinese populations, which causes the limited use of existing polygenic risk scores (PRS) to identify subpopulations at high risk of VTE for prevention. We, therefore, aimed to curate all the potential VTE-related single-nucleotide polymorphisms (SNPs) for the construction of a new improved PRS model based on the self-adapting method, and then evaluate its utility and effectiveness in the stratification of VTE risk in Chinese populations. We comprehensively analyzed the mutation spectrum of VTE-associated SNPs in the Chinese cohort, and ranked their individual risk effects independently using risk ratio, logistic regression coefficient, and penalty regression coefficient as evaluation criteria. By integrating various algorithms and evaluating their performance, we trained the optimal prediction model of VTE risk in the Chinese population with the least SNP features, established an adaptive PRS model with progressive SNP overlay, and tested it on an independent Chinese population cohort. Self-adaptive polygenic risk score model based on all 318 SNPs or on the 44 most strongly associated SNPs performed similarly (areas under receiver-operating characteristic curves (AUCs) of 0.739 and 0.709, respectively) on the testing dataset of the Chinese VTE cohort, and that achieve the overall best level of the AUC from a conventional PRS model based on known genetic risk factors (0.620–0.718). In addition, we observed the self-adaptive PRS model was an independent effective risk stratification indicator beyond other clinical characteristics including age and smoking status. Our data revealed that only 44 SNPs-derived PRS model can be effectively used in discriminating subpopulations at high risk of VTE. To become clinically useful, our model could benefit from a practically feasible VTE screening program for precision prevention in Chinese populations.
- 12 March 2025
- Published by Springer Nature in Phenomics
- Vol. 5 (4) , 347-360
A Self-Adapting Polygenic Risk Score Model Improves Risk Prediction of Venous Thromboembolism in Han Chinese Cohorts
Most genome-wide association studies (GWAS) of Venous Thromboembolism (VTE) have used data from individuals of European descent, however, genetic factors for VTE have not been fully identified in Chinese populations, which causes the limited use of existing polygenic risk scores (PRS) to identify subpopulations at high risk of VTE for prevention. We, therefore, aimed to curate all the potential VTE-related single-nucleotide polymorphisms (SNPs) for the construction of a new improved PRS model based on the self-adapting method, and then evaluate its utility and effectiveness in the stratification of VTE risk in Chinese populations. We comprehensively analyzed the mutation spectrum of VTE-associated SNPs in the Chinese cohort, and ranked their individual risk effects independently using risk ratio, logistic regression coefficient, and penalty regression coefficient as evaluation criteria. By integrating various algorithms and evaluating their performance, we trained the optimal prediction model of VTE risk in the Chinese population with the least SNP features, established an adaptive PRS model with progressive SNP overlay, and tested it on an independent Chinese population cohort. Self-adaptive polygenic risk score model based on all 318 SNPs or on the 44 most strongly associated SNPs performed similarly (areas under receiver-operating characteristic curves (AUCs) of 0.739 and 0.709, respectively) on the testing dataset of the Chinese VTE cohort, and that achieve the overall best level of the AUC from a conventional PRS model based on known genetic risk factors (0.620–0.718). In addition, we observed the self-adaptive PRS model was an independent effective risk stratification indicator beyond other clinical characteristics including age and smoking status. Our data revealed that only 44 SNPs-derived PRS model can be effectively used in discriminating subpopulations at high risk of VTE. To become clinically useful, our model could benefit from a practically feasible VTE screening program for precision prevention in Chinese populations.
- 14 September 2024
- Published by European Respiratory Society (ERS) in European Respiratory Journal
- Vol. 64 (suppl 68) , OA5557
Genetic Insights through Chinese VTE Pedigree Interpretation Using the ACMG Method: Unraveling Potencial Genetic Risk Factors
Aim: The study aims to bridge existing gaps in understanding VTE genetics by focusing on familial units with unexplained VTE occurrences. Methodology: Participant Selection: Identify and recruit individuals from 9 Chinese pedigrees with a clear family history of VTE. Employ high-throughput whole-genome sequencing (WGS) technology for comprehensive genetic analysis. Utilize the American College of Medical Genetics and Genomics (ACMG) scoring system for variant interpretation and familial analysis. Result: We successfully identified and interpreted 10 variants directly linked to thrombosis through harmful mutation filtering and VTE-specific ACMG rating, which enabled the interpretation of thrombotic risk in 8 pedigree, shedding light on the elusive genetic underpinnings of VTE within these cohorts. Moreover, our investigation revealed notable risk mutations within the ADAMTSs gene family associated with thrombosis susceptibility across multiple families, completing the interpretation of all pedigrees. Conclusion: By applying the ACMG scoring system and evidence-based medicine principles, the study seeks to identify and characterize potential genetic factors associated with VTE, ultimately providing valuable insights for clinical practice and future research in VTE genetics.
- 14 September 2024
- Published by European Respiratory Society (ERS) in European Respiratory Journal
- Vol. 64 (suppl 68) , OA5557
Genetic Insights through Chinese VTE Pedigree Interpretation Using the ACMG Method: Unraveling Potencial Genetic Risk Factors
Aim: The study aims to bridge existing gaps in understanding VTE genetics by focusing on familial units with unexplained VTE occurrences. Methodology: Participant Selection: Identify and recruit individuals from 9 Chinese pedigrees with a clear family history of VTE. Employ high-throughput whole-genome sequencing (WGS) technology for comprehensive genetic analysis. Utilize the American College of Medical Genetics and Genomics (ACMG) scoring system for variant interpretation and familial analysis. Result: We successfully identified and interpreted 10 variants directly linked to thrombosis through harmful mutation filtering and VTE-specific ACMG rating, which enabled the interpretation of thrombotic risk in 8 pedigree, shedding light on the elusive genetic underpinnings of VTE within these cohorts. Moreover, our investigation revealed notable risk mutations within the ADAMTSs gene family associated with thrombosis susceptibility across multiple families, completing the interpretation of all pedigrees. Conclusion: By applying the ACMG scoring system and evidence-based medicine principles, the study seeks to identify and characterize potential genetic factors associated with VTE, ultimately providing valuable insights for clinical practice and future research in VTE genetics.
- 14 September 2024
- Published by European Respiratory Society (ERS) in European Respiratory Journal
- Vol. 64 (suppl 68) , OA5551
Effect of genetic variants on bleeding with rivaroxaban in pulmonary embolism patients: A multicenter prospective cohort study
Background: The genetic feature is a critical factor for the efficacy and safety of rivaroxaban by affecting plasma concentration during extended treatment of pulmonary embolism (PE). However, the existing bleeding risk scores do not include genetic variants. The study aimed to explore whether incorporating genetic factors in the current model improved the accuracy. Methods: PE patients receiving therapeutic doses of rivaroxaban were enrolled. The primary outcome, bleeding events, was monitored over a one-year follow-up period. A genome-wide association study and candidate gene association analyses were conducted. Genetic variants were integrated into four established clinical bleeding risk scores to develop combined clinical-genetic risk assessments. Findings: This study included 3,498 patients treated with rivaroxaban. The bleeding rate during extended treatment was 4.3%. Risk factors identified for bleeding included stroke, abnormal liver function, and hypertension. In 510 patients underwent genotyping, one single-nucleotide polymorphism (SNP) ABLIM2 rs6819260 was significantly related to bleeding (OR = 8.91, 95% CI: 4.08–19.47, p = 4.22 × 10−8). Incorporating ABLIM2 rs6819260 into existing risk models significantly improved the AUC by more than 10.0%. Interpretation: Genetic variations significantly influence bleeding risk in Chinese PE patients treated with rivaroxaban during extended anticoagulation. Integrating the ABLIM2 rs6819260 SNP into bleeding risk models substantially enhances their predictive accuracy and facilitates the use of these models to guide individualized treatment decisions in clinical practice.
- 14 September 2024
- Published by European Respiratory Society (ERS) in European Respiratory Journal
- Vol. 64 (suppl 68) , OA5551
Effect of genetic variants on bleeding with rivaroxaban in pulmonary embolism patients: A multicenter prospective cohort study
Background: The genetic feature is a critical factor for the efficacy and safety of rivaroxaban by affecting plasma concentration during extended treatment of pulmonary embolism (PE). However, the existing bleeding risk scores do not include genetic variants. The study aimed to explore whether incorporating genetic factors in the current model improved the accuracy. Methods: PE patients receiving therapeutic doses of rivaroxaban were enrolled. The primary outcome, bleeding events, was monitored over a one-year follow-up period. A genome-wide association study and candidate gene association analyses were conducted. Genetic variants were integrated into four established clinical bleeding risk scores to develop combined clinical-genetic risk assessments. Findings: This study included 3,498 patients treated with rivaroxaban. The bleeding rate during extended treatment was 4.3%. Risk factors identified for bleeding included stroke, abnormal liver function, and hypertension. In 510 patients underwent genotyping, one single-nucleotide polymorphism (SNP) ABLIM2 rs6819260 was significantly related to bleeding (OR = 8.91, 95% CI: 4.08–19.47, p = 4.22 × 10−8). Incorporating ABLIM2 rs6819260 into existing risk models significantly improved the AUC by more than 10.0%. Interpretation: Genetic variations significantly influence bleeding risk in Chinese PE patients treated with rivaroxaban during extended anticoagulation. Integrating the ABLIM2 rs6819260 SNP into bleeding risk models substantially enhances their predictive accuracy and facilitates the use of these models to guide individualized treatment decisions in clinical practice.