Hennocq Quentin, Bongibault Thomas, Marlin Sandrine, Amiel Jeanne, Attie-Bitach Tania, Baujat Geneviève, Boutaud Lucile, Carpentier Georges, Corre Pierre, Denoyelle Françoise, Djate Delbrah François, Douillet Maxime, Galliani Eva, Kamolvisit Wuttichart, Lyonnet Stanislas, Milea Dan, Pingault Véronique, Porntaveetus Thantrira, Touzet-Roumazeille Sandrine, Willems Marjolaine, Picard Arnaud, Rio Marlène, Garcelon Nicolas, Khonsari Roman H
Imagine Institute, INSERM UMR1163, Paris, France.
Service de Chirurgie Maxillo-Faciale et Chirurgie Plastique, Hôpital Necker-Enfants Malades, Assistance Publique-Hôpitaux de Paris, Centre de Référence des Malformations Rares de la Face et de la Cavité Buccale MAFACE, Filière Maladies Rares TeteCou, Faculté de Médecine, Université de Paris Cité, Paris, France.
Front Pediatr. 2023 Aug 17;11:1171277. doi: 10.3389/fped.2023.1171277. eCollection 2023.
Mandibulo-Facial Dysostosis with Microcephaly (MFDM) is a rare disease with a broad spectrum of symptoms, characterized by zygomatic and mandibular hypoplasia, microcephaly, and ear abnormalities. Here, we aimed at describing the external ear phenotype of MFDM patients, and train an Artificial Intelligence (AI)-based model to differentiate MFDM ears from non-syndromic control ears (binary classification), and from ears of the main differential diagnoses of this condition (multi-class classification): Treacher Collins (TC), Nager (NAFD) and CHARGE syndromes.
The training set contained 1,592 ear photographs, corresponding to 550 patients. We extracted 48 patients completely independent of the training set, with only one photograph per ear per patient. After a CNN-(Convolutional Neural Network) based ear detection, the images were automatically landmarked. Generalized Procrustes Analysis was then performed, along with a dimension reduction using PCA (Principal Component Analysis). The principal components were used as inputs in an eXtreme Gradient Boosting (XGBoost) model, optimized using a 5-fold cross-validation. Finally, the model was tested on an independent validation set.
We trained the model on 1,592 ear photographs, corresponding to 1,296 control ears, 105 MFDM, 33 NAFD, 70 TC and 88 CHARGE syndrome ears. The model detected MFDM with an accuracy of 0.969 [0.838-0.999] ( < 0.001) and an AUC (Area Under the Curve) of 0.975 within controls (binary classification). Balanced accuracies were 0.811 [0.648-0.920] ( = 0.002) in a first multiclass design (MFDM vs. controls and differential diagnoses) and 0.813 [0.544-0.960] ( = 0.003) in a second multiclass design (MFDM vs. differential diagnoses).
This is the first AI-based syndrome detection model in dysmorphology based on the external ear, opening promising clinical applications both for local care and referral, and for expert centers.
伴有小头畸形的下颌面部发育不全（MFDM）是一种罕见疾病，症状范围广泛，其特征为颧骨和下颌骨发育不全、小头畸形以及耳部异常。在此，我们旨在描述MFDM患者的外耳表型，并训练一个基于人工智能（AI）的模型，以区分MFDM患者的耳朵与非综合征对照耳朵（二元分类），以及区分MFDM患者的耳朵与该病症主要鉴别诊断的耳朵（多分类）：特雷彻·柯林斯综合征（TC）、纳杰尔综合征（NAFD）和CHARGE综合征。
训练集包含1592张耳部照片，对应550名患者。我们提取了48名完全独立于训练集的患者，每位患者每只耳朵仅有一张照片。在基于卷积神经网络（CNN）的耳部检测之后，图像被自动标记地标。然后进行广义普罗克拉斯提斯分析，并使用主成分分析（PCA）进行降维。主成分被用作极端梯度提升（XGBoost）模型的输入，该模型使用五折交叉验证进行优化。最后，该模型在一个独立的验证集上进行测试。
我们在1592张耳部照片上训练该模型，这些照片对应1296只对照耳朵、105只MFDM耳朵、33只NAFD耳朵、70只TC耳朵和88只CHARGE综合征耳朵。在对照中（二元分类），该模型检测MFDM的准确率为0.969 [0.838 - 0.999]（<0.001），曲线下面积（AUC）为0.975。在第一个多分类设计（MFDM与对照及鉴别诊断）中，平衡准确率为0.811 [0.648 - 0.920]（=0.002），在第二个多分类设计（MFDM与鉴别诊断）中，平衡准确率为0.813 [0.544 - 0.960]（=0.003）。
这是首个基于外耳的畸形学中基于AI的综合征检测模型，为本地护理和转诊以及专家中心开启了有前景的临床应用。
