AI-Based CT Image Recognition With Med-Gemini-3D in the Diagnosis of a Rare Craniofacial Condition: A Catlin Mark Skull

Artificial intelligence (AI) is increasingly applied in diagnostic imaging to enhance pattern recognition and support clinical decision-making. In 2024, Google introduced Med-Gemini-3D, a multimodal platform capable of interpreting 3-dimensional computed tomography scans and generating radiologist-level reports. Although not yet approved for independent clinical use, such systems may assist in identifying rare conditions that are unfamiliar to clinicians. The authors describe a 22-month-old girl who presented with persistent bilateral parietal skull defects and global developmental delay. Computed tomography demonstrated symmetric ossification defects adjacent to the sagittal suture that were not initially recognized by the treating physician. The patient's mother used a smartphone application powered by Med-Gemini-3D to analyze a 3D-CT reconstruction image, which suggested a diagnosis of "Catlin mark skull," a historical term for Enlarged Parietal Foramina (EPF). This prompted genetic evaluation and identification of a CDC42BPB variant associated with Chilton-Okur-Chung neurodevelopmental syndrome-a finding not previously reported in association with EPF. Establishing the diagnosis facilitated earlier therapeutic interventions and informed long-term management. This case underscores the potential role of AI-assisted tools in recognizing rare craniofacial anomalies. While such technologies cannot replace clinical expertise and remain limited by variable accuracy, they may help expand differential diagnoses, expedite referrals, and improve outcomes through earlier intervention. Continued research is needed to validate their reliability and to define their optimal integration into clinical practice.