Cerebral small vessel disease lesion segmentation methods: A systematic review
Bibliographic record
Abstract
• Systematic review of segmentation methods for cerebral small vessel disease lesions • Good evidence for validated methods for segmenting white matter hyperintensity • Fewer methods for microbleeds, perivascular spaces, and lacunes Cerebral small vessel disease (CSVD) can manifest as brain lesions visible on magnetic resonance imaging, including white matter hyperintensities (WMH), cerebral microbleeds (CMB), perivascular spaces (PVS), lacunes, and recent small subcortical infarcts (RSSI). Detection and segmentation of these imaging markers can provide valuable information on brain health, including prevention and treatment of dementia. However, manual segmentation is cumbersome, especially for large cohorts in research studies. There has been extensive research into the development of automated tools using machine learning to increase accuracy and efficiency in lesion segmentation. This systematic review aimed to summarize novel automated methods developed over the last 10 years that segment CSVD lesion types and have been validated on a population with or at risk for CSVD ( e.g., older adults, those with cognitive disorders, or those with vascular risk factors). A search on Web of Science and PubMed yielded 2764 studies, of which 89 were included after screening and full text review. 59 of these methods segmented WMH, 23 detected or classified CMB, 6 detected or segmented PVS, 5 detected, classified, or segmented lacunes, and 2 segmented RSSI. Of these, 30 studies (23 for WMH, 5 for CMB, 1 for PVS, and 1 for lacunes) included links to download code or pre-trained models, including one commercial tool, and one that relied on a commercial tool for input. Overall, this review found good evidence for high quality tools available for WMH segmentation, with fewer tools available to accurately segment other CSVD lesion types.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.018 | 0.090 |
| Meta-epidemiology (narrow) | 0.003 | 0.002 |
| Meta-epidemiology (broad) | 0.011 | 0.013 |
| Bibliometrics | 0.015 | 0.013 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.004 | 0.004 |
| Open science | 0.003 | 0.003 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.008 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".