Artificial Intelligence to Assist the Echocardiographic Identification of Transthyretin Cardiac Amyloidosis

Algalarrondo Vincent15,000 enrolled

Overview

The goal of this study is to develop an algorithm using artificial intelligence (AI) to assist identification of potential ATTR-CM cases using routine transthoracic echocardiography. The main questions it aims to answer are: * is the algorithm able to diagnose ATTR-CM * is the algorithm able to diagnose different types of ATTR-CM (ATTRv, ATTRwt) This is a non interventional study. Participant' echocardiographies will be, after deidentification, used to train, valid and test the algorithm.

Transthyretin (TTR) amyloidosis is a serious systemic disease affecting multiple target organs including the peripheral nervous system, heart, and kidney. In the absence of treatment, the median survival for symptomatic forms with cardiac involvement is 3 to 4 years. In recent years, new treatments have proven their effectiveness in transthyretin amyloidosis, making it possible to slow the progression of neuropathy and cardiac damage. These treatments seem particularly effective when they are initiated at an early stage of the disease. It is therefore necessary to establish the diagnosis as early as possible in order to benefit the most from the treatment. However, during the clinical examination, the electrocardiogram or the routine echocardiography, the signs evoking cardiac amyloidosis are not specific. The initial diagnosis is therefore often difficult, missed or delayed and the median time between the first symptoms and the initiation of treatment is approximately 3 years. It is therefore the initial phase of diagnosis that must be improved in a sufficiently sensitive and specific manner to detect potential cases early while avoiding unnecessary examinations in the event of a low probability. The objective of the study is to develop and validate a tool to assist the screening of cardiac transthyretin amyloidosis, from standard echocardiography, without the need for active participation of the cardiologist in the diagnostic process. This diagnostic contribution will allow the cardiologist to evoke the diagnosis of cardiac amyloidosis and to consider additional explorations.

Study Type

OBSERVATIONAL

Enrollment

15,000

Conditions

Amyloid Cardiomyopathy

Eligibility

Sex: ALLMin age: 18 Years

Medical Language ↔ Plain English

ATTR-CM patients: Inclusion Criteria: * Cardiac transthyretin amyloidosis diagnosed on the classic criteria: 1. Absence of monoclonal immunoglobulin AND 2. Presence of a bisphosphonate scintigraphy with enhancement in the cardiac area OR 2-Presence of a cardiac biopsy showing transthyretin (Congo red positive) cardiac amyloidosis (demonstrated either by immunostaining or by mass spectrometry) OR 3-Presence of a peripheral biopsy showing transthyretin amyloidosis (see above) associated with cardiac infiltration (parietal thickness \>12mm without other cause of cardiac hypertrophy) * No opposition to research Non-inclusion criteria: * Another cause of cardiac amyloidosis: AL AA amyloidosis… * Mixed heart disease with associated presence of non-amyloid heart disease (ischemic heart disease, dilated, etc.) Control patients: Inclusion criteria: * Indication for transthoracic echocardiography as part of cardiological follow-up * Patient affiliated with social security * Patient's agreement to participate in the research and signature of the consent form. * Technical conditions of the examination and echogenicity allowing acquisition of good quality echocardiographic images, allowing post processing Non-inclusion criteria: * Presence of cardiac amyloidosis as defined above * Presence of transthyretin amyloidosis even without demonstrated cardiac involvement * Patient monitored for asymptomatic transthyretin mutation * Minor patient or patient unable to give their consent (unconscious patient, under guardianship)

Outcomes

Primary Outcomes

Building and validating the diagnostic performance metrics curves of the AI algorithm to diagnose ATTR-CM :

To develop and validate a tool using artificial intelligence an algorithm that will improve the automatic detection on routinely acquired echocardiography images of aspects suggestive of transthyretin amyloidosis. A confusion matrix will be built and the following diagnostic performance metrics be computed: * receiver operating characteristic curve (ROC) and area under curve (AUC) of the ROC : AUROC * Precision recall curve (PR) and area under curve (AUC) of the PR curve : AUC-PR

Time frame: year 1

Building and validating the diagnostic performance metrics of the AI algorithm to diagnose ATTR-CM :

To develop and validate a tool using artificial intelligence an algorithm that will improve the automatic detection on routinely acquired echocardiography images of aspects suggestive of transthyretin amyloidosis ATTR. A confusion matrix will be built and the following diagnostic performance metrics be computed: Accuracy, Sensitivity or Recall, Specificity, False positive rate, False Negative Rate, Precision (all are expressed as ratio)

Time frame: year 1

Secondary Outcomes

Building and validating the diagnostic performance metrics of the AI algorithm to diagnose ATTRwt-CM :

To develop and validate a tool using artificial intelligence an algorithm that will improve the automatic detection on routinely acquired echocardiography images of aspects suggestive of transthyretin amyloidosis. A confusion matrix will be built and the following diagnostic performance metrics be computed: Accuracy, Sensitivity or Recall, Specificity, False positive rate, False Negative Rate, Precision (all are expressed as ratio)

Time frame: year 1

Building and validating the diagnostic performance metrics of the AI algorithm to diagnose ATTRv-V122I-CM :

To develop and validate a tool using artificial intelligence an algorithm that will improve the automatic detection on routinely acquired echocardiography images of aspects suggestive of transthyretin amyloidosis. A confusion matrix will be built and the following diagnostic performance metrics be computed: Accuracy, Sensitivity or Recall, Specificity, False positive rate, False Negative Rate, Precision (all are expressed as ratio)

Time frame: year 1

Building and validating the diagnostic performance metrics of the AI algorithm to diagnose ATTRv-CM :

To develop and validate a tool using artificial intelligence an algorithm that will improve the automatic detection on routinely acquired echocardiography images of aspects suggestive of transthyretin amyloidosis. A confusion matrix will be built and the following diagnostic performance metrics be computed: Accuracy, Sensitivity or Recall, Specificity, False positive rate, False Negative Rate, Precision (all are expressed as ratio)

Time frame: year 1

Building and validating the diagnostic performance metrics of the AI algorithm to differentiate ATTR-CM from LV hypertrophy (LVH) :

To develop and validate a tool using artificial intelligence an algorithm that will improve the automatic detection on routinely acquired echocardiography images of aspects suggestive of transthyretin amyloidosis from LVH. A confusion matrix will be built and the following diagnostic performance metrics be computed: Accuracy, Sensitivity or Recall, Specificity, False positive rate, False Negative Rate, Precision (all are expressed as ratio)

Time frame: year 1

Building and validating the diagnostic performance metrics curves of the AI algorithm to diagnose ATTRwt-CM :

To develop and validate a tool using artificial intelligence an algorithm that will improve the automatic detection on routinely acquired echocardiography images of aspects suggestive of transthyretin amyloidosis (ATTRwt subgroup). A confusion matrix will be built and the following diagnostic performance metrics be computed: * receiver operating characteristic curve (ROC) and area under curve (AUC) of the ROC : AUROC * Precision recall curve (PR) and area under curve (AUC) of the PR curve : AUC-PR

Time frame: year 1

Building and validating the diagnostic performance metrics curves of the AI algorithm to diagnose ATTRv-V122I-CM :

To develop and validate a tool using artificial intelligence an algorithm that will improve the automatic detection on routinely acquired echocardiography images of aspects suggestive of transthyretin amyloidosis (ATTRv-V122I subgroup). A confusion matrix will be built and the following diagnostic performance metrics be computed: * receiver operating characteristic curve (ROC) and area under curve (AUC) of the ROC : AUROC * Precision recall curve (PR) and area under curve (AUC) of the PR curve : AUC-PR

Time frame: year 1

Building and validating the diagnostic performance metrics curves of the AI algorithm to diagnose ATTRv-CM :

To develop and validate a tool using artificial intelligence an algorithm that will improve the automatic detection on routinely acquired echocardiography images of aspects suggestive of transthyretin amyloidosis (ATTRv-subgroup). A confusion matrix will be built and the following diagnostic performance metrics be computed: * receiver operating characteristic curve (ROC) and area under curve (AUC) of the ROC : AUROC * Precision recall curve (PR) and area under curve (AUC) of the PR curve : AUC-PR

Time frame: year 1

Building and validating the diagnostic performance metrics curves of the AI algorithm to differentiate ATTR-CM from LV hypertrophy (LVH) :

To develop and validate a tool using artificial intelligence an algorithm that will improve the automatic detection on routinely acquired echocardiography images of aspects suggestive of transthyretin amyloidosis (ATTRv-subgroup) in a subset of patients with LVH. A confusion matrix will be built and the following diagnostic performance metrics be computed: * receiver operating characteristic curve (ROC) and area under curve (AUC) of the ROC : AUROC * Precision recall curve (PR) and area under curve (AUC) of the PR curve : AUC-PR

Time frame: year 1

Artificial Intelligence to Assist the Echocardiographic Identification of Transthyretin Cardiac Amyloidosis

Overview

Conditions

Interventions

Eligibility

Locations (1)

Outcomes

Primary Outcomes

Secondary Outcomes

Central Contacts