Application of Multimodal Large Language Model in HFpEF

Recruiting

• Conditions: Heart Failure With Preserved Ejection Fraction
• Interventions: Diagnostic Test: Multimodal Large Language Model Diagnosis; Diagnostic Test: Routine diagnostic and therapeutic procedure

• Registration Number: NCT06486649
• Lead Sponsor: Peking University Third Hospital

Brief Summary

This study will validate the effectiveness of a multimodal large language model to screen for heart failure with preserved ejection fraction (HFpEF), comparing it with the traditional clinical standardized assessment process.

Detailed Description

Heart failure is a major complication of various heart diseases and is the leading lethal cause of cardiovascular death worldwide. Based on the left ventricular ejection fraction (LVEF), heart failure can be divided into heart failure with reduced ejection fraction (HFrEF), heart failure with preserved ejection fraction (HFpEF) and heart failure with mildly reduced ejection fraction (HFmrEF). Heart failure rehospitalization rates and in-hospital complications did not differ between HFrEF and HFpEF. However, over the past two decades, the survival rate of HFrEF has improved significantly, whereas HFpEF has remained stagnant. One of the major reasons for this is that the diagnostic process of HFpEF is complicated, and it is easy to cause missed diagnosis in the clinic, resulting in delayed treatment.

Multimodal large language models are capable of integrating and analyzing medical data from different sources, including textual data (e.g., medical records, medical literature), image data (e.g., electrocardiograms, CT scan images), and audio data (e.g., symptoms narrated by patients). This multimodal data integration capability is crucial for understanding complex medical scenarios, as it provides a more comprehensive view of the condition than a single data source.

The diagnosis of HFpEF faces many challenges and requires clinicians to make judgments on multi-dimensional data, which can easily lead to the underdiagnosis and misdiagnosis of the disease. As a generative artificial intelligence tool, a large language model is able to integrate and analyze data from different sources and has the ability to learn and evolve from existing clinical evidence. Based on this, this study intends to evaluate the effectiveness of multimodal large language model for screening for heart failure with preserved ejection fraction (HFpEF), comparing it with the traditional clinical standard assessment process.

Basic Information
|
Recruitment & Eligibility
|
Study & Design
|
Trial Locations

Study Timeline

• Study Start: 2023-12-20
• Study First Submitted: 2023-12-22
• Status Verified: 2024-06
• Study First Submitted QC: 2024-06-27
• Last Update Submitted: 2024-06-27
• Study First Posted: 2024-07-03
• Last Update Posted: 2024-07-03
• Primary Completion: 2024-12-20
• Study Completion: 2024-12-20

Recruitment & Eligibility

• Status: RECRUITING
• Sex: All
• Target Recruitment: 80

Inclusion Criteria

1. Age 18-80 years, male or female;
2. Cardiology inpatients with suspected heart failure with preserved ejection fraction (cardiac ultrasound suggestive of LVEF ≥50% with at least 1 of the following: 1, left ventricular hypertrophy and/or left atrial enlargement; and 2, abnormal diastolic cardiac function);
3. Current or previous at least one symptom of heart failure, including dyspnea (including exertional dyspnea, nocturnal paroxysmal dyspnea, and telangiectasia), malaise, nausea, and bilateral lower extremity edema;
4. Voluntary participation and signed informed consent.

Read More

Exclusion Criteria

1. Acute heart failure or acute worsening of chronic heart failure;
2. Severe coronary stenosis (≥75% stenosis) without revascularization;
3. Patients who are unable to perform exercise stress echocardiography or have contraindications to the test;
4. are participating in other clinical trials;
5. Those with severe organic pathologies of the liver, kidney, or hematologic system or those with chronic diseases;
6. Those who are unable to follow the trial procedures;
7. Those who refuse to sign the informed consent.

Read More

Study & Design

• Study Type: OBSERVATIONAL
• Study Design: Not specified

Arm && Interventions

Group	Intervention	Description
single group	Multimodal Large Language Model Diagnosis	The routine consultation process was performed first: according to the process recommended by the 2023 edition of the Chinese Expert Consensus on the Diagnosis and Treatment of Heart Failure with Preserved Ejection Fraction, the attending cardiologist completed the subject's clinical criteria assessment and performed the HFpEF diagnosis (yes/no). During the attending physician's checkup visit, the multimodal large language model screening system (MedGuide-72B) collected routine visit data, recorded relevant data and indicators during the patient's communication with MedGuide-72B and made the diagnosis.
single group	Routine diagnostic and therapeutic procedure	The routine consultation process was performed first: according to the process recommended by the 2023 edition of the Chinese Expert Consensus on the Diagnosis and Treatment of Heart Failure with Preserved Ejection Fraction, the attending cardiologist completed the subject's clinical criteria assessment and performed the HFpEF diagnosis (yes/no). During the attending physician's checkup visit, the multimodal large language model screening system (MedGuide-72B) collected routine visit data, recorded relevant data and indicators during the patient's communication with MedGuide-72B and made the diagnosis.

Primary Outcome Measures

Name	Time	Method
dignostic sensitivity	through study completion, an average of 8 months	dianostic sensitivity comparison between routine diagnosis and therapy and large language model diagnosis
dignostic specificity	through study completion, an average of 8 months	dianostic specificity comparison between routine diagnosis and therapy and large language model diagnosis

Secondary Outcome Measures

Name	Time	Method
consistency rate	through study completion, an average of 8 months	consistency rate between routine diagnosis and therapy and large language model diagnosis
economic cost analysis	through study completion, an average of 8 months	comparison of economic cost between routine diagnosis and therapy and large language model diagnosis by the total cost of treatment
time spent on diagnosis	through study completion, an average of 8 months	comparison of time spent on diagnosis between routine diagnosis and therapy and large language model diagnosis
patient satisfaction	through study completion, an average of 8 months	comparison of patient satisfaction between routine diagnosis and therapy and large language model diagnosis by questionnaire
false discovery rate	through study completion, an average of 8 months	comparison of false discovery rate between routine diagnosis and therapy and large language model diagnosis
diagnosis efficiency	through study completion, an average of 8 months	The probability of accuracy compared to the final diagnosis of the patient's visit
physician workload assessment	through study completion, an average of 8 months	comparison of physician workload between routine diagnosis and therapy and large language model diagnosis according to counting the number of participants with treatment-related

Trial Locations

Locations (1)

Peking UniversityThird Hospital; 🇨🇳 Beijing, Beijing, China