Ignite Creation Date:
2025-12-25 @ 2:17 AM
Ignite Modification Date:
2026-01-07 @ 3:21 AM
Study NCT ID:
NCT07004660
Status:
COMPLETED
Last Update Posted:
2025-06-04
First Post:
2025-05-16
Is NOT Gene Therapy:
True
Has Adverse Events:
False
Brief Title:
Performances of Large Language Models in Kidney Allograft Diagnostics
Sponsor:
Organization:
Raw JSON
{'hasResults': False, 'derivedSection': {'miscInfoModule': {'versionHolder': '2025-12-24'}, 'conditionBrowseModule': {'meshes': [{'id': 'D004194', 'term': 'Disease'}], 'ancestors': [{'id': 'D010335', 'term': 'Pathologic Processes'}, {'id': 'D013568', 'term': 'Pathological Conditions, Signs and Symptoms'}]}}, 'protocolSection': {'designModule': {'studyType': 'OBSERVATIONAL', 'designInfo': {'timePerspective': 'PROSPECTIVE', 'observationalModel': 'COHORT'}, 'enrollmentInfo': {'type': 'ACTUAL', 'count': 240}, 'patientRegistry': False}, 'statusModule': {'overallStatus': 'COMPLETED', 'startDateStruct': {'date': '2004-03-01', 'type': 'ACTUAL'}, 'expandedAccessInfo': {'hasExpandedAccess': False}, 'statusVerifiedDate': '2025-05', 'completionDateStruct': {'date': '2023-12-31', 'type': 'ACTUAL'}, 'lastUpdateSubmitDate': '2025-05-26', 'studyFirstSubmitDate': '2025-05-16', 'studyFirstSubmitQcDate': '2025-05-26', 'lastUpdatePostDateStruct': {'date': '2025-06-04', 'type': 'ACTUAL'}, 'studyFirstPostDateStruct': {'date': '2025-06-04', 'type': 'ACTUAL'}, 'primaryCompletionDateStruct': {'date': '2023-12-31', 'type': 'ACTUAL'}}, 'outcomesModule': {'primaryOutcomes': [{'measure': 'Biopsy diagnosis', 'timeFrame': 'The biopsy will be protocol biopsies performed at 3 months and 1 year post transplant, or for-cause biopsies performed at any time post transplant.', 'description': 'The diagnosis of the biopsy will be based on the latest Banff classification (2022). The diagnoses will include i) biopsies with nonspecific lesions or clean (n=40), ii) biopsies with antibody-mediated rejection (AMR) (n=40) among which 14 had acute AMR, 13 chronic active AMR and 13 chronic inactive AMR, iii) biopsies with T cell-mediated rejection (TCMR) (n=40), among which 20 had acute TCMR and 20 had chronic active TCMR, iv) biopsies with borderline for acute TCMR (n=40), v) biopsies with mixed rejection (n=40), and vi) biopsies with microvascular inflammation (MVI) (n=40) among which 20 had probable AMR and 20 had MVI without DSA and without C4d.'}]}, 'oversightModule': {'oversightHasDmc': True, 'isFdaRegulatedDrug': False, 'isFdaRegulatedDevice': False}, 'conditionsModule': {'keywords': ['Large language model', 'Banff classification', 'Diagnosis', 'Kidney transplantation', 'Biopsy'], 'conditions': ['Kidney Transplantation']}, 'descriptionModule': {'briefSummary': 'Kidney allograft rejection diagnosis relies on the complex Banff classification, but its application is limited by variability and workload. Our group previously built a scripted automation system, though it required major expert input. This study assesses whether modern LLMs can achieve similar diagnostic performance using Banff-based prompts, without extensive manual engineering.', 'detailedDescription': 'Kidney allograft rejection remains a leading cause of allograft failure. Histological diagnosis relies on the Banff classification, a complex and evolving rule based framework. While successive Banff working groups refined the guidelines over time, daily interpretation is still hampered by inter and intra pathologist variability and growing demands on renal pathologists. This is why our group previously built a fully scripted Banff automation system. However, this system demanded years of expert curation and bespoke code before reaching acceptable accuracy. Whether modern LLMs, which show high capabilities to generate consistent and transparent reasoning at scale, can match expert pathologists without such resource intensive engineering remains unknown. The present study was therefore designed to benchmark state of the art LLMs against consensus diagnoses from senior renal pathologists on a representative series of kidney allograft biopsies, and to explore whether properly engineered prompts can translate Banff rules into reliable, reproducible diagnostic output.'}, 'eligibilityModule': {'sex': 'ALL', 'stdAges': ['CHILD', 'ADULT', 'OLDER_ADULT'], 'maximumAge': '100 Years', 'minimumAge': '0 Years', 'samplingMethod': 'NON_PROBABILITY_SAMPLE', 'studyPopulation': 'Kidney recipients', 'healthyVolunteers': False, 'eligibilityCriteria': 'Inclusion Criteria:\n\n* Kidney recipients\n\nExclusion Criteria:\n\n* Combined transplant'}, 'identificationModule': {'nctId': 'NCT07004660', 'briefTitle': 'Performances of Large Language Models in Kidney Allograft Diagnostics', 'organization': {'class': 'OTHER', 'fullName': 'Paris Translational Research Center for Organ Transplantation'}, 'officialTitle': 'Performances of Large Language Models in Kidney Allograft Diagnostics', 'orgStudyIdInfo': {'id': 'LLM_vs_Patho_001'}}, 'armsInterventionsModule': {'armGroups': [{'label': 'Necker hospital', 'description': 'Transplant unit from Necker hospital, France'}, {'label': 'Saint-Louis hospital', 'description': 'Transplant unit from Saint-Louis hospital, France'}]}, 'ipdSharingStatementModule': {'ipdSharing': 'UNDECIDED'}, 'sponsorCollaboratorsModule': {'leadSponsor': {'name': 'Paris Translational Research Center for Organ Transplantation', 'class': 'OTHER'}, 'responsibleParty': {'type': 'PRINCIPAL_INVESTIGATOR', 'investigatorTitle': 'PhD', 'investigatorFullName': 'Marc Raynaud', 'investigatorAffiliation': 'Paris Translational Research Center for Organ Transplantation'}}}}