Background <p>Dementia is frequently invoked in medico-legal contexts where questions of legal capacity, guardianship, testamentary validity, transactional validity, and criminal responsibility arise. Although subtype differentiation and cognitive assessment are central to clinical dementia evaluation, it remains unclear to what extent such clinically meaningful detail is visible in appellate judicial reasoning. This study examined the visibility of dementia subtype specification and cognitive evidence in decisions of the Turkish Court of Cassation (2005–2025), and examined the ethical implications of these textual patterns for transparency, accountability, and procedural fairness in judicial decision-making.</p> Methods <p>We conducted a retrospective, cross-sectional, document-based analysis of publicly accessible Court of Cassation decisions. Using keyword-based searches (“demans,” “bunama,” and “Alzheimer”), 1,180 decisions were identified; after screening for legal relevance, 1,031 decisions were included. Each decision was coded across 15 variables, including dementia subtype specification, cognitive evidence visibility, legal context, and medical report source. Descriptive statistics were calculated for all variables. Temporal trends in dementia subtype visibility were analysed using logistic regression, supplemented by year-by-year distributional assessment and multivariable sensitivity modelling. A parallel logistic regression was also conducted for the broader binary outcome of any visible cognitive evidence (specific or non-specific; <i>n</i> = 104 events).</p> Results <p>Dementia subtype information was explicitly specified in 46.4% of decisions (478/1,031; 95% CI, 43.3–49.5%), whereas 53.6% used only general or unspecified terminology. Alzheimer’s disease was the most frequently reported subtype, appearing in 451 decisions (43.7% of the full sample; 94.4% of subtype-specified decisions). Subtype visibility increased significantly over time (OR = 1.062 per year; 95% CI, 1.034–1.091; <i>p</i> &lt; 0.001). In contrast, visible cognitive evidence remained limited: 89.9% of decisions contained no reference to cognitive assessment, 9.8% included only non-specific references, and only 0.3% explicitly named the Mini-Mental State Examination (MMSE; 95% CI, 0.1–0.8%). No decision referred to the Montreal Cognitive Assessment (MoCA). Any visible cognitive evidence was present in 10.1% of decisions (104/1,031; 95% CI, 8.3–12.1%) and was highest in guardianship/restriction cases (19.7%). Temporal modelling of any visible cognitive evidence was associated with a statistically significant decrease over the study period (OR = 0.918 per year; 95% CI, 0.880–0.958; <i>p</i> &lt; 0.001).</p> Conclusions <p>Dementia subtype specification became moderately more visible in appellate judicial reasoning over time, whereas the textual visibility of cognitive evidence remained limited. These findings concern the traceability of clinically relevant information within decision summaries rather than the adequacy of the underlying evidentiary process itself. Strengthening the structured reporting of diagnostic basis, relevant cognitive findings, and functionally relevant capacity-related findings in medico-legal documentation may support improved transparency and accountability in the translation of clinical evidence into legal reasoning.</p>