This project investigates how local growth mechanisms at cranial sutures and synchondroses translate into large-scale craniofacial morphology. While craniofacial growth is often described using age-based or purely morphometric models, this work seeks to integrate biological and developmental information into shape analysis.

By combining multimodal imaging, tissue-level characterization, and 3D morphometrics, the project aims to clarify how local growth potential and maturation patterns influence global phenotype. A key objective is to move toward biologically informed models of craniofacial growth that better reflect developmental processes.

• Research fellows: N Kogane (PhD candidate), E Jourdain (M2)

• Collaborators: E Dambroise, KH Khonsari, S Laporte, L Legeai-Mallet, K Rajabizadeh

• Imagine, AP-HP, i-biomec ENSAM

Craniofacial growth has historically been described through multiple theoretical frameworks, often based on implicit or partially incompatible assumptions. This project aims to translate these theories into formal ontological and symbolic models, making their assumptions explicit and comparable.

By structuring growth theories as formal models, we seek to evaluate their internal coherence and to assess how well they align with clinical and imaging data. In a second stage, these formalized frameworks are confronted with real datasets using data-driven and machine-learning approaches.

The broader goal is to bridge conceptual models of development with empirical observations and to clarify which theoretical frameworks are most compatible with observed craniofacial growth patterns.

• Research fellow: V Franchetti-Romero (PhD candidate)

• Collaborators: J Baxter, P Jannin, RH Khonsari

• Imagine, AP-HP, Rennes Univ., CHU Nantes

Mandibular growth is shaped not only by developmental programs but also by functional and mechanical constraints. This project explores how muscular forces and loading patterns interact with mandibular growth and temporomandibular joint (TMJ) development.

Using imaging-derived segmentations of bone and muscle structures, biomechanical models are developed to estimate force distributions and functional environments. Initial work focuses on static bite models, with future extensions toward dynamic modelling frameworks.

By superimposing functional and morphological spaces, this research aims to clarify how mechanical environments may influence growth trajectories, particularly in pathological situations where altered function and altered growth interact.

• Research fellow: A Aram (MD thesis)

• Collaborators: H Dutel, RH Khonsari, L Van de Lande

• Imagine, AP-HP, Bordeaux Univ., EMC Rotterdam

A growing number of pipelines exist for 3D morphometric analysis and statistical shape modelling, yet their assumptions, sensitivities, and domains of validity are rarely compared systematically. As a result, methodological choices can strongly influence biological or clinical interpretations.

This project aims to benchmark commonly used morphometric and shape modelling pipelines on shared datasets, evaluating their robustness, reproducibility, and interpretability. Rather than identifying a single “best” method, the goal is to clarify under which conditions specific approaches are more reliable.

This work contributes to promoting methodological transparency and informed method selection in craniofacial research and beyond.

• Research fellow: S Benhamouche (M2)

• Collaborators: J Feydy, R Magnet

• Imagine, INRIA

Most statistical shape models are designed for cross-sectional data, whereas clinical growth studies often rely on sparse, irregular, and heterogeneous longitudinal datasets. This mismatch limits our ability to describe and interpret developmental trajectories.

This project investigates how longitudinal statistical shape modelling frameworks can be adapted to real clinical contexts. Particular attention is given to disentangling true growth-related changes from inter-individual variability, cohort effects, and measurement noise.

Beyond technical development, the project also examines the conceptual limits of longitudinal shape models, asking what kinds of biological or clinical questions these models can realistically answer.

• Research fellow: Open position

• Collaborators: J Feydy, R Magnet, O Lienhard

• Imagine, INRIA, AIDY

Reliable segmentation is a critical prerequisite for quantitative craniofacial shape analysis, yet it remains labor intensive, and a source of variability across studies. This project focuses on developing and refining tools that enable reproducible segmentation of craniofacial structures from CT data.

BounTi4Slicer is an open module for 3D Slicer adapted from the original BounTi tool, allowing semi-automatic segmentation of cranial bones using iterative gray-level approaches. The tool is particularly suited for pediatric and research datasets where anatomical variability is high.

Current work also documents the practical limitations of gray-level–based segmentation, especially in partially fused or low-contrast structures, thereby clarifying where more advanced approaches will be needed in the future.

• Collaborators: M Didziokas, N Kogane, K Rajabizadeh

• Imagine, UCL, Aarhus Univ.

Three-dimensional facial photogrammetry is increasingly used in craniofacial research and clinical follow-up, yet acquisition protocols and analytical pipelines remain highly heterogeneous across centers. This variability limits data comparability and the ability to conduct robust multicenter studies.

Within the ERN Cranio 3D Working Group, this project contributes to the development of harmonized acquisition protocols and analysis pipelines for 3D facial data. The goal is to improve reproducibility, reduce methodological bias, and enable meaningful cross-center comparisons in rare craniofacial disorders.

Beyond technical harmonization, the project also addresses how standardized 3D phenotyping can support longitudinal follow-up and clinical decision-making.

• Collaborators: T Abdel-Alim, RH Khonsari, I Mathijssen, L Smith, M-L van Veelen

• Imagine, AP-HP, ERN CRANIO, EMC Rotterdam, GOSH

Craniofacial morphology plays a major role in upper airway function and in the risk of obstructive sleep apnea (OSAS), particularly in pediatric and syndromic populations such as FGFR-related craniosynostoses. This project investigates how variations in both bony and soft-tissue morphology relate to airway structure and respiratory function.

One axis of the project focuses on the morphological characterization of the upper airways in both typical and FGFR-related craniofacial development. By combining 3D imaging and morphometric analyses, this work aims to clarify how anatomical variation may contribute to airway obstruction and functional impairment.

A second axis examines how external mechanical constraints, such as prolonged use of non-invasive ventilation masks, may influence craniofacial morphology during growth. This allows the study of how externally applied forces interact with developmental and functional processes.

Together, these approaches contribute to a broader understanding of morphology–function–forces interactions in the craniofacial system. Future developments may include airflow and fluid dynamics modelling to better link anatomical structure and respiratory function.

• Research fellow: M Ntirubuza (MD thesis)

• Collaborators: B Fauroux, S Khirani, RH Khonsari, S Laporte, M Moazen, Y Heuzé

• Imagine, AP-HP, i-biomec ENSAM, UCL, Bordeaux Univ.