Aide au diagnostic
Bien que le diagnostic soit une façon quotidienne de parler de la performance du dispositif, gardez à l'esprit que ce que le dispositif produit est un représentation interprétative de la distribution des classes possibles de la classification internationale des maladies (CIM) qui pourraient être représentées dans le contenu en pixels de l'image.
En effet, les professionnels de la santé et les organisations peuvent utiliser les données générées par le dispositif pour déterminer un diagnostic, mais ce que le dispositif lui-même génère n'est pas un diagnostic. Ceci est signalé de manière appropriée dans les résultats que génère le dispositif, ce qui est conforme à la norme FHIR lorsqu'on remarque que les résultats générés constituent un DiagnosticReport, avec un statut préliminaire.
Table of contents
Version 2.0
Terminal
https://ai.legit.health/v2/legit_health/predict
Demande du corps
Pour une analyse d'image basique en vue d'obtenir des probabilités de pathologies détectées :
{
"requestId": "unique-request-id",
"data": {
"content": "base64 encoded image"
}
}
Pour plus de précision, vous pouvez inclure des données supplémentaires :
{
"requestId": "unique-request-id",
"data": {
"type": "image",
"modality": "clinical",
"operator": "Practitioner",
"bodySite": "ARM_LEFT",
"subject": {
"identifier": "subject-id",
"gender": "m",
"height": 175,
"weight": 71,
"birthdate": "1986-10-21",
"generalPractitioner": {
"identifier": "practitioner-id"
},
"managingOrganization": {
"identifier": "organization-id",
"display": "Hospital Central"
}
},
"content": "base64 encoded image"
}
}
Voici la description des propriétés pertinentes de la demande du corps :
bodySiteest le code de la zone du corps qui identifie la zone où l'image a été prise. Vous pouvez trouver la liste disponible avec tous les codes admis ici.
Réponse
La réponse fournit des résultats diagnostiques préliminaires, y compris les maladies potentielles, leurs codes associés et la confiance diagnostique (%).
{
"category": [
{
"coding": [
{
"code": "250171008",
"display": "Clinical history and observation finding",
"system": "http://snomed.info/sct"
}
]
}
],
"code": {
"coding": [
{
"code": "10206-1",
"display": "Physical findings of Skin Narrative",
"system": "http://loinc.org"
}
]
},
"conclusions": [
{
"code": {
"code": "DA04.5",
"codeSystem": "ICD-11"
},
"name": "Mucous cyst",
"probability": 87.77
},
{
"code": {
"code": "2F20.2",
"codeSystem": "ICD-11"
},
"name": "Congenital nevus",
"probability": 6.11
},
{
"code": {
"code": "XH4L78",
"codeSystem": "ICD-11"
},
"name": "Nevus",
"probability": 2.16
},
{
"code": {
"code": "ED63.0",
"codeSystem": "ICD-11"
},
"name": "Vitiligo",
"probability": 1.92
}
],
"detectedModality": "Clinical",
"device": {
"deviceName": {
"name": "Legit.Health",
"type": "user-friendly-name"
},
"manufacturer": "AI LABS GROUP SL",
"resourceType": "Device",
"type": {
"code": "string",
"display": "Dermatology picture archiving and communication system application software",
"system": "http://snomed.info/sct"
}
},
"evolution": {},
"explainabilityMedia": {
"content": "",
"modality": "Clinical",
"resourceType": "Media",
"type": "image"
},
"mediaValidity": {
"isValid": true,
"metrics": {
"hasEnoughQuality": true,
"isDermatologyDomain": true
},
"score": 66.0
},
"metrics": {
"category": "calculation",
"resourceType": "DeviceMetric",
"sensitivity": 99.56,
"specificity": 93.99
},
"preliminaryFindings": {
"hasConditionSuspicion": 99.2,
"isMalignantSuspicion": 0,
"isPreMalignantSuspicion": 0.6000000000000001,
"needsSpecialistsAttention": 3.86
},
"resourceType": "DiagnosticReport",
"status": "preliminary",
"time": 1.6485424041748047
}
Version 2.1
Vous pouvez maintenant envoyer plusieurs images de la même structure cutanée.
En effet, cette version prend en charge l'envoi de plusieurs images pour des résultats plus solides.
Cela ne doit pas être la même image avec des transformations, mais des clichés différents de la même chose.
- Éléments recommandés : 3 images.
- Maximum : 5 images.
Terminal
https://ai.legit.health/v2/legit_health/diagnosis_support
Demande du corps
{
"requestId": "unique-request-id",
"data": {
"content": ["base64 image 1", "base64 image 2", "base64 image 3"]
}
}
Comme vous pouvez le voir, le contenu est un tableau dans lequel vous pouvez envoyer jusqu'à 5 images de la même structure cutanée. Cela ne doit pas être la même image avec des transformations, mais des clichés différents de la même chose.
Pour plus de précision, des données supplémentaires, similaires à la version 2.0, peuvent également être incluses.
{
"requestId": "unique-request-id",
"data": {
"type": "image",
"modality": "clinical",
"operator": "Practitioner",
"bodySite": "ARM_LEFT",
"subject": {
"identifier": "subject-id",
"gender": "m",
"height": 175,
"weight": 71,
"birthdate": "1986-10-21",
"generalPractitioner": {
"identifier": "practitioner-id"
},
"managingOrganization": {
"identifier": "organization-id",
"display": "Hospital Central"
}
},
"content": ["base64 image 1", "base64 image 2", "base64 image 3"]
}
}
Réponse
La réponse dans la version 2.1 a une clé supplémentaire, imagingStudySeries, qui fournit des résultats individuels pour chaque image.
{
"category": [
{
"coding": [
{
"code": "250171008",
"display": "Clinical history and observation finding",
"system": "http://snomed.info/sct"
}
]
}
],
"code": {
"coding": [
{
"code": "10206-1",
"display": "Physical findings of Skin Narrative",
"system": "http://loinc.org"
}
]
},
"conclusions": [
{
"code": {
"code": "DA04.5",
"codeSystem": "ICD-11"
},
"name": "Mucous cyst",
"probability": 87.77
},
{
"code": {
"code": "2F20.2",
"codeSystem": "ICD-11"
},
"name": "Congenital nevus",
"probability": 6.11
},
{
"code": {
"code": "XH4L78",
"codeSystem": "ICD-11"
},
"name": "Nevus",
"probability": 2.16
},
{
"code": {
"code": "ED63.0",
"codeSystem": "ICD-11"
},
"name": "Vitiligo",
"probability": 1.92
}
],
"device": {
"deviceName": {
"name": "Legit.Health",
"type": "user-friendly-name"
},
"manufacturer": "AI LABS GROUP SL",
"resourceType": "Device",
"type": {
"code": "string",
"display": "Dermatology picture archiving and communication system application software",
"system": "http://snomed.info/sct"
}
},
"explainabilityMedia": {
"content": "",
"modality": ["Clinical", "Clinical", "Clinical"],
"resourceType": "Media",
"type": "image"
},
"mediaValidity": {
"isValid": true,
"metrics": {
"hasEnoughQuality": true,
"isDermatologyDomain": true
},
"score": 66.0
},
"metrics": {
"category": "calculation",
"resourceType": "DeviceMetric",
"sensitivity": 99.56,
"specificity": 93.99
},
"preliminaryFindings": {
"hasConditionSuspicion": 99.2,
"isMalignantSuspicion": 0,
"isPreMalignantSuspicion": 0.6000000000000001,
"needsSpecialistsAttention": 3.86
},
"imagingStudySeries": [ {
"conclusions": [
{
"code": {
"code": "DA04.5",
"codeSystem": "ICD-11"
},
"name": "Mucous cyst",
"probability": 87.77
},
{
"code": {
"code": "2F20.2",
"codeSystem": "ICD-11"
},
"name": "Congenital nevus",
"probability": 6.11
},
{
"code": {
"code": "XH4L78",
"codeSystem": "ICD-11"
},
"name": "Nevus",
"probability": 2.16
},
{
"code": {
"code": "ED63.0",
"codeSystem": "ICD-11"
},
"name": "Vitiligo",
"probability": 1.92
}
],
"detectedModality": "Clinical",
"explainabilityMedia": {
"content": "",
"modality": "Clinical",
"resourceType": "Media",
"type": "image"
},
"mediaValidity": {
"isValid": true,
"metrics": {
"hasEnoughQuality": true,
"isDermatologyDomain": true
},
"score": 66.0
},
"metrics": {
"category": "calculation",
"resourceType": "DeviceMetric",
"sensitivity": 99.56,
"specificity": 93.99
},
"preliminaryFindings": {
"hasConditionSuspicion": 99.2,
"isMalignantSuspicion": 0,
"isPreMalignantSuspicion": 0.6000000000000001,
"needsSpecialistsAttention": 3.86
}
},
"conclusions": [
{
"code": {
"code": "DA04.5",
"codeSystem": "ICD-11"
},
"name": "Mucous cyst",
"probability": 87.77
},
{
"code": {
"code": "2F20.2",
"codeSystem": "ICD-11"
},
"name": "Congenital nevus",
"probability": 6.11
},
{
"code": {
"code": "XH4L78",
"codeSystem": "ICD-11"
},
"name": "Nevus",
"probability": 2.16
},
{
"code": {
"code": "ED63.0",
"codeSystem": "ICD-11"
},
"name": "Vitiligo",
"probability": 1.92
}
],
"detectedModality": "Clinical",
"explainabilityMedia": {
"content": "",
"modality": "Clinical",
"resourceType": "Media",
"type": "image"
},
"mediaValidity": {
"isValid": true,
"metrics": {
"hasEnoughQuality": true,
"isDermatologyDomain": true
},
"score": 66.0
},
"metrics": {
"category": "calculation",
"resourceType": "DeviceMetric",
"sensitivity": 99.56,
"specificity": 93.99
},
"preliminaryFindings": {
"hasConditionSuspicion": 99.2,
"isMalignantSuspicion": 0,
"isPreMalignantSuspicion": 0.6000000000000001,
"needsSpecialistsAttention": 3.86
}
},
"conclusions": [
{
"code": {
"code": "DA04.5",
"codeSystem": "ICD-11"
},
"name": "Mucous cyst",
"probability": 87.77
},
{
"code": {
"code": "2F20.2",
"codeSystem": "ICD-11"
},
"name": "Congenital nevus",
"probability": 6.11
},
{
"code": {
"code": "XH4L78",
"codeSystem": "ICD-11"
},
"name": "Nevus",
"probability": 2.16
},
{
"code": {
"code": "ED63.0",
"codeSystem": "ICD-11"
},
"name": "Vitiligo",
"probability": 1.92
}
],
"detectedModality": "Clinical",
"explainabilityMedia": {
"content": "",
"modality": "Clinical",
"resourceType": "Media",
"type": "image"
},
"mediaValidity": {
"isValid": true,
"metrics": {
"hasEnoughQuality": true,
"isDermatologyDomain": true
},
"score": 66.0
},
"metrics": {
"category": "calculation",
"resourceType": "DeviceMetric",
"sensitivity": 99.56,
"specificity": 93.99
},
"preliminaryFindings": {
"hasConditionSuspicion": 99.2,
"isMalignantSuspicion": 0,
"isPreMalignantSuspicion": 0.6000000000000001,
"needsSpecialistsAttention": 3.86
}
}
],
"resourceType": "DiagnosticReport",
"status": "preliminary",
"time": 1.6485424041748047
}