Abstract Brain tumour classification from Magnetic Resonance Imaging (MRI) requires both high diagnostic accuracy and computationally efficient model adaptation. Although deep transfer learning has improved automated tumour recognition, most existing approaches remain limited by dependence on generic ImageNet-pretrained weights, prolonged fine-tuning time, and weak institution-specific adaptability under small local medical datasets. To address these limitations, this paper proposes an Adaptive Incremental Domain Pretraining and Frozen-Weight Local Rapid Adaptation framework built on the ResNet18 backbone. Unlike conventional one-step transfer learning, the proposed method recursively preserves and updates tumour-specialised parameter states across multiple same-domain MRI repositories, progressively constructing a domain-adapted diagnostic backbone. The final inherited model is then subjected to frozen-weight local rapid adaptation for institution-specific MRI customisation. Experimental evaluation shows that direct ImageNet-based ResNet18 transfer learning achieves 93.27. Similar content being viewed by others Subjects Introduction Brain tumours constitute one of the most critical neurological disorders because of their direct influence on the central nervous system and their potentially life-threatening progression. Uncontrolled abnormal growth of intracranial or spinal cells can severely impair physiological regulation, cognitive activity, and neural communication, thereby resulting in substantial morbidity and mortality. Clinically, brain tumours are broadly categorised as benign or malignant, with benign tumours generally slow-growing and localised, whereas malignant tumours are highly invasive, aggressive, and rapidly progressive1. According to the World Health Organization (WHO) classification, low-grade tumours (Grades I and II) exhibit comparatively better prognosis, whereas high-grade tumours (Grades III and IV) are associated with extremely poor survival outcomes, thereby making early-stage identification and precise grading an indispensable requirement for effective treatment planning and prognosis estimation2,3. Among the various diagnostic modalities available for tumour identification, including computed tomography (CT), electroencephalography (EEG), and invasive biopsy-assisted examinations, Magnetic Resonance Imaging (MRI) has emerged as the most reliable and widely preferred non-invasive clinical tool for brain tumour assessment