#METABOLOMICS WORKBENCH LMSAC_20250428_202430 DATATRACK_ID:5870 STUDY_ID:ST004047 ANALYSIS_ID:AN006691 PROJECT_ID:PR002537 VERSION 1 CREATED_ON July 17, 2025, 9:42 pm #PROJECT PR:PROJECT_TITLE Patient-centered artificial intelligence platform for endometrial cancer risk PR:PROJECT_TITLE stratification using clinical and molecular multi-omics data PR:PROJECT_SUMMARY Endometrial cancer (ENDOM) prevention remains challenging, creating an urgent PR:PROJECT_SUMMARY need for better risk stratification tools. We developed a patient-centered PR:PROJECT_SUMMARY bimodal multilevel endometrial cancer (2M-EC) predictive platform that PR:PROJECT_SUMMARY integrates clinically accessible data with multi-biofluid multi-omics data. Our PR:PROJECT_SUMMARY study established the MBF-ED cohort (n=531), collecting comprehensive clinical PR:PROJECT_SUMMARY data and multi-dimensional body fluid samples. We processed these using a unique PR:PROJECT_SUMMARY analytical pipeline: (1) simplifying clinical variables through empirical and PR:PROJECT_SUMMARY data-driven methods, (2) extracting ENDOM-specific MS features using machine PR:PROJECT_SUMMARY learning, and (3) developing an innovative bimodal AI architecture that fuses 2D PR:PROJECT_SUMMARY MS omics matrices with 1D clinical vectors. The resulting patient-centered 2M-EC PR:PROJECT_SUMMARY predictive platform provides real-time, interpretable risk stratification PR:PROJECT_SUMMARY through an online interface. The advantages of it include overcoming PR:PROJECT_SUMMARY single-marker limitations via multimodal integration, combining molecular depth PR:PROJECT_SUMMARY with clinical practicality and scalable design adaptable to both PR:PROJECT_SUMMARY resource-limited and advanced healthcare settings. This work demonstrates how AI PR:PROJECT_SUMMARY can bridge cutting-edge molecular profiling with routine clinical practice, PR:PROJECT_SUMMARY offering a new paradigm for patient-centered cancer risk assessment. PR:INSTITUTE Fudan University PR:DEPARTMENT Chemistry department PR:LABORATORY LiangQiao lab PR:LAST_NAME DANDAN PR:FIRST_NAME LI PR:ADDRESS Fudan University PR:EMAIL oceanddl@sina.com PR:PHONE 18061019632 #STUDY ST:STUDY_TITLE Artificial intelligence platform for endometrial cancer risk stratification ST:STUDY_TITLE using clinical and molecular multi-omics data ST:STUDY_SUMMARY Endometrial cancer (ENDOM) prevention remains challenging, creating an urgent ST:STUDY_SUMMARY need for better risk stratification tools. We developed a patient-centered ST:STUDY_SUMMARY bimodal multilevel endometrial cancer (2M-EC) predictive platform that ST:STUDY_SUMMARY integrates clinically accessible data with multi-biofluid multi-omics data. Our ST:STUDY_SUMMARY study established the MBF-ED cohort (n=531), collecting comprehensive clinical ST:STUDY_SUMMARY data and multi-dimensional body fluid samples. We processed these using a unique ST:STUDY_SUMMARY analytical pipeline: (1) simplifying clinical variables through empirical and ST:STUDY_SUMMARY data-driven methods, (2) extracting ENDOM-specific MS features using machine ST:STUDY_SUMMARY learning, and (3) developing an innovative bimodal AI architecture that fuses 2D ST:STUDY_SUMMARY MS omics matrices with 1D clinical vectors. The resulting patient-centered 2M-EC ST:STUDY_SUMMARY predictive platform provides real-time, interpretable risk stratification ST:STUDY_SUMMARY through an online interface. The advantages of it include overcoming ST:STUDY_SUMMARY single-marker limitations via multimodal integration, combining molecular depth ST:STUDY_SUMMARY with clinical practicality and scalable design adaptable to both ST:STUDY_SUMMARY resource-limited and advanced healthcare settings. This work demonstrates how AI ST:STUDY_SUMMARY can bridge cutting-edge molecular profiling with routine clinical practice, ST:STUDY_SUMMARY offering a new paradigm for patient-centered cancer risk assessment. ST:INSTITUTE Fudan University ST:LAST_NAME DANDAN ST:FIRST_NAME LI ST:ADDRESS Fudan University ST:EMAIL oceanddl@sina.com ST:PHONE 18061019632 #SUBJECT SU:SUBJECT_TYPE Human SU:SUBJECT_SPECIES Homo sapiens SU:TAXONOMY_ID 9606 SU:GENDER Female #SUBJECT_SAMPLE_FACTORS: SUBJECT(optional)[tab]SAMPLE[tab]FACTORS(NAME:VALUE pairs separated by |)[tab]Raw file names and additional sample data SUBJECT_SAMPLE_FACTORS - C1 Sample source:cervix | Sample type:cervix Factor=mix 1; RAW_FILE_NAME(Raw file name)=C1.raw SUBJECT_SAMPLE_FACTORS - C2 Sample source:cervix | Sample type:cervix Factor=mix 2; RAW_FILE_NAME(Raw file name)=C2.raw SUBJECT_SAMPLE_FACTORS - C3 Sample source:cervix | Sample type:cervix Factor=mix 3; RAW_FILE_NAME(Raw file name)=C3.raw SUBJECT_SAMPLE_FACTORS - P1 Sample source:blood | Sample type:blood Factor=mix 4; RAW_FILE_NAME(Raw file name)=P1.raw SUBJECT_SAMPLE_FACTORS - P2 Sample source:blood | Sample type:blood Factor=mix 5; RAW_FILE_NAME(Raw file name)=P2.raw SUBJECT_SAMPLE_FACTORS - P3 Sample source:blood | Sample type:blood Factor=mix 6; RAW_FILE_NAME(Raw file name)=P3.raw SUBJECT_SAMPLE_FACTORS - U1 Sample source:uterine cavity | Sample type:uterine cavity Factor=mix 7; RAW_FILE_NAME(Raw file name)=U1.raw SUBJECT_SAMPLE_FACTORS - U2 Sample source:uterine cavity | Sample type:uterine cavity Factor=mix 8; RAW_FILE_NAME(Raw file name)=U2.raw SUBJECT_SAMPLE_FACTORS - U3 Sample source:uterine cavity | Sample type:uterine cavity Factor=mix 9; RAW_FILE_NAME(Raw file name)=U3.raw SUBJECT_SAMPLE_FACTORS - C1-rp Sample source:cervix | Sample type:cervix Factor=mix 10; RAW_FILE_NAME(Raw file name)=C1-rp.raw SUBJECT_SAMPLE_FACTORS - C2-rp Sample source:cervix | Sample type:cervix Factor=mix 11; RAW_FILE_NAME(Raw file name)=C2-rp.raw SUBJECT_SAMPLE_FACTORS - C3-rp Sample source:cervix | Sample type:cervix Factor=mix 12; RAW_FILE_NAME(Raw file name)=C3-rp.raw SUBJECT_SAMPLE_FACTORS - P1-rp Sample source:blood | Sample type:blood Factor=mix 13; RAW_FILE_NAME(Raw file name)=P1-rp.raw SUBJECT_SAMPLE_FACTORS - P2-rp Sample source:blood | Sample type:blood Factor=mix 14; RAW_FILE_NAME(Raw file name)=P2-rp.raw SUBJECT_SAMPLE_FACTORS - P3-rp Sample source:blood | Sample type:blood Factor=mix 15; RAW_FILE_NAME(Raw file name)=P3-rp.raw SUBJECT_SAMPLE_FACTORS - U1-rp Sample source:uterine cavity | Sample type:uterine cavity Factor=mix 16; RAW_FILE_NAME(Raw file name)=U1-rp.raw SUBJECT_SAMPLE_FACTORS - U2-rp Sample source:uterine cavity | Sample type:uterine cavity Factor=mix 17; RAW_FILE_NAME(Raw file name)=U2-rp.raw SUBJECT_SAMPLE_FACTORS - U3-rp Sample source:uterine cavity | Sample type:uterine cavity Factor=mix 18; RAW_FILE_NAME(Raw file name)=U3-rp.raw #COLLECTION CO:COLLECTION_SUMMARY Under sterile conditions in the operating room, endometrial secretions were CO:COLLECTION_SUMMARY collected using a disposable endometrial brush prior to uterine removal. The CO:COLLECTION_SUMMARY brush head was advanced 2 cm through the cervical canal and rotated three times CO:COLLECTION_SUMMARY in each uterine segment. Harvested tissue was transferred to 15 mL tubes CO:COLLECTION_SUMMARY containing 3 mL 80% methanol. Brush heads were detached with sterile forceps and CO:COLLECTION_SUMMARY co-stored. Samples were immediately flash-frozen in laparoscopic specimen liquid CO:COLLECTION_SUMMARY nitrogen containers, transported on dry ice, and stored at -80°C until CO:COLLECTION_SUMMARY metabolomic extraction. During preoperative examination, cervical secretion were CO:COLLECTION_SUMMARY collected using HPV brushes prior to bimanual palpation. Brush heads were CO:COLLECTION_SUMMARY directly immersed in 15 mL tubes with 80% methanol (Thermo Fisher Scientific, CO:COLLECTION_SUMMARY USA). Peripheral blood (8 mL) was collected preoperatively into EDTA tubes (BD CO:COLLECTION_SUMMARY Biosciences, USA). CO:SAMPLE_TYPE Uterine secretion, cervical secretion, whole blood #TREATMENT TR:TREATMENT_SUMMARY no treatment #SAMPLEPREP SP:SAMPLEPREP_SUMMARY In blood metabolomics, 50 μL whole blood was mixed with 200 μL ice-cold SP:SAMPLEPREP_SUMMARY methanol:acetonitrile (1:1, v/v; Thermo Fisher Scientific, USA). The mixture was SP:SAMPLEPREP_SUMMARY vortexed, sonicated for 2 min, and incubated at -20°C for 30 min. Following SP:SAMPLEPREP_SUMMARY centrifugation at 16,000×g (4°C, 20 min), 100 μL supernatant was lyophilized SP:SAMPLEPREP_SUMMARY and reconstituted in 20 μL methanol:acetonitrile:water (1:1:2, v/v/v). Samples SP:SAMPLEPREP_SUMMARY were thawed at 4°C and homogenized by gentle vortex mixing. After SP:SAMPLEPREP_SUMMARY centrifugation at 12,000 rpm (4°C, 5 min), the supernatant was carefully SP:SAMPLEPREP_SUMMARY collected and aliquoted into two replicates. The supernatant was concentrated SP:SAMPLEPREP_SUMMARY using a vacuum centrifugal concentrator (approximately 2 h) until lyophilized SP:SAMPLEPREP_SUMMARY powder was obtained. The dried metabolites were stored at −80°C prior to SP:SAMPLEPREP_SUMMARY downstream analysis. After reconstitution, samples were centrifuged at 16,000×g SP:SAMPLEPREP_SUMMARY (4°C, 30 min) and the supernatant was analyzed. For plasma processing, 100 μL SP:SAMPLEPREP_SUMMARY plasma was mixed with 400 μL (4× volume) of ice-cold methanol/acetonitrile SP:SAMPLEPREP_SUMMARY (1:1, v/v), vortexed for 30 s, and sonicated for 2 min. The mixture was SP:SAMPLEPREP_SUMMARY incubated at -20°C for 30 min for protein precipitation, then centrifuged at SP:SAMPLEPREP_SUMMARY 16,000×g (4°C, 20 min). 300 μL supernatant was collected, concentrated to SP:SAMPLEPREP_SUMMARY dryness by centrifugation, and stored at -80°C. After reconstitution, samples SP:SAMPLEPREP_SUMMARY were centrifuged at 16,000×g (4°C, 30 min) and the supernatant was subjected SP:SAMPLEPREP_SUMMARY to instrumental analysis. #CHROMATOGRAPHY CH:CHROMATOGRAPHY_SUMMARY RP (LC/MS) CH:CHROMATOGRAPHY_TYPE Reversed phase CH:INSTRUMENT_NAME Thermo Vanquish CH:COLUMN_NAME Phenomenex Kinetex C18 (100 x 2.1mm,2.6um) CH:SOLVENT_A 100% water; 0.01% acetic acid CH:SOLVENT_B 50% acetonitrile/50% toluene dicarboxylic acid CH:FLOW_GRADIENT Linear gradient from 99% A to 1% A over 8 minutes, followed by a return to 99% A CH:FLOW_GRADIENT at 9.1 minutes, with mobile phase B inversely following the same pattern. CH:FLOW_RATE 0.3 mL/min CH:COLUMN_TEMPERATURE 25 #ANALYSIS AN:ANALYSIS_TYPE MS #MS MS:INSTRUMENT_NAME Thermo Orbitrap Exploris 480 MS:INSTRUMENT_TYPE Orbitrap MS:MS_TYPE ESI MS:ION_MODE UNSPECIFIED MS:MS_COMMENTS The electrospray ionization mass spectra were acquired in polarity switching. MS:MS_COMMENTS Data dependent acquisition (DDA) was used to collect full scan MS and MSMS MS:MS_COMMENTS information simultaneously. The ion spray voltage was set to 3,500 V for MS:MS_COMMENTS positive mode and 2,800 V for negative mode. The survey of full scan MS spectra MS:MS_COMMENTS (m/z 70-1200) was acquired in the Orbitrap with 60,000 resolutions. The MS:MS_COMMENTS normalized automatic gain control (AGC) target at 100% and the maximum injection MS:MS_COMMENTS time MS:MS_RESULTS_FILE ST004047_AN006691_Results.txt UNITS:mg/L Has m/z:Yes Has RT:No RT units:No RT data #END