AI Science Tracker
EN

Massachusetts General Hospital

2 discoveries from 2021

Discoveries

Neural pathology quantification links enlarged brain spaces to vascular amyloid

Source check: Source checked. Not independent scientific confirmation.

Finding: Convolutional models quantified vascular amyloid and related pathology in donated brain tissue. In 19 cerebral-amyloid-angiopathy cases, enlarged perivascular spaces were associated with amyloid burden in overlying vessels after accounting for other pathology. Paired MRI and histology supported anatomical interpretation. Cross-sectional autopsy findings do not establish impaired clearance as the cause or the sequence of changes.

Technology: Aiforia convolutional neural networks; three-dimensional U-Net; SmartInterpol

Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.

AI contribution
Validated CNNs quantify vascular amyloid/fibrin pathology, while learned high-resolution MRI segmentation supports regional anatomical analysis.
Test and result
Cross-sectional autopsy tissue quantification, regional MRI/histology comparisons and mixed-effects association analysis in 19 amyloid-angiopathy cases.
Limitation / still unknown
Causal direction, clearance dynamics and population generalizability remain unestablished; the selected donation cohort was small and automated in-vivo MRI segmentation was not reliable.
Evidence sources and annotation context

Evidence context (descriptions follow the cited sources; test type is a catalog classification):

Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC9047512/

Source checked · editorial details

The source page was checked against this entry. Independent confirmation is not assessed by this label.

Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.

Review date: 21 Nov 2026. Review scope: Primary relevant model Methods, Results and limitations; exact primary identity and current primary PubMed correction/retraction relationships checked..

No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.

Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.

Neural heart measurements reveal a chest-pain risk association

Source check: Source checked. Not independent scientific confirmation.

Finding: Two U-Net models measured heart volume in CT scans from 3,798 PROMISE participants. Smaller body-size-indexed volume was associated with subsequent major adverse events, particularly in patients with non-obstructive coronary disease. This retrospective secondary analysis is hypothesis-generating; it does not establish a causal mechanism, a normal-volume threshold or improved clinical decisions.

Technology: Cascaded U-Net convolutional neural networks

Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.

AI contribution
Two pretrained U-Nets localize and segment the heart, enabling standardized volumetric measurements for clinical event-association analysis.
Test and result
Automated CT measurements linked to independently adjudicated events over median 26-month follow-up in a randomized-trial cohort.
Limitation / still unknown
Independent prognostic validation and clinical utility remain unestablished; patients with known coronary disease or heart failure were excluded and subgroup event counts were limited.
Evidence sources and annotation context

Evidence context (descriptions follow the cited sources; test type is a catalog classification):

Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC8273107/

Source checked · editorial details

The source page was checked against this entry. Independent confirmation is not assessed by this label.

Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.

Review date: 21 Nov 2026. Review scope: Primary relevant model Methods, Results and limitations; exact primary identity and current primary PubMed correction/retraction relationships checked..

No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.

Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.

Open the interactive profile for the timeline, or browse all organizations and model families.