Parkinson disease (PD) entails progressive motor and non-motor symptoms that reduce quality of life for people with PD (PwPD) and increase caregiver burden. Static functionalities of current assistive systems (AS) are poorly aligned with changing needs. Pro-adaptive AS using digital twins appear to address these problems. This study sought to identify PD-specific requirement clusters, derive user-centered key requirements, and assess their technical feasibility and availability to inform development of adaptive AS. We compiled relevant needs from ICD-10, scientific literature and German care-level criteria, then filtered by potential addressability through AS and measurability of AS effectiveness. Remaining items were grouped into 16 heuristic requirement clusters, from which one open-ended question per cluster was derived to form the interview guide, conducted with PwPD and informal caregivers. Three technical experts independently rated each cluster on 0-4 scales for technical feasibility and for existence/availability of solutions. Our identified most important requirements were physical symptoms and memory impairments. To this end, we propose a pro-adaptive, AI-based digital twin model to detect PD-related symptoms and monitor disease progression and tracking disease progression using a digital twin model.