AI Science Tracker
EN

Medical University of Vienna

3 discoveries from 2020 to 2025

Discoveries

AI imaging measures change during cardiac amyloidosis treatment

Source check: Source checked. Not independent scientific confirmation.

Finding: Automated cardiac segmentation quantified serial SPECT/CT scans in 45 patients receiving transthyretin amyloidosis therapies. Several tracer-uptake markers decreased after treatment, and selected changes were associated with functional measures. These exploratory observational results do not prove treatment causality or removal of amyloid deposits; affected-volume measures did not significantly decline.

Technology: TotalSegmentator; foundational deep-learning CT segmentation

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

AI contribution
A pretrained anatomical segmentation network maps cardiac structures to SPECT, enabling standardized regional tracer measurements over time.
Test and result
Serial baseline/follow-up clinical imaging, automated cardiac-region quantification and exploratory associations with functional measurements.
Limitation / still unknown
Controlled treatment effects, cross-scanner generalizability and clinical decision benefit remain unestablished; the paper has inconsistent aggregate marker counts and small treatment subgroups.
Evidence sources and annotation context

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

Evidence sources: https://link.springer.com/article/10.1007/s00259-025-07673-6

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.

Automated PET measures track immunotherapy thyroid adverse events

Source check: Source checked. Not independent scientific confirmation.

Finding: TotalSegmentator and a tumor-segmentation network quantified organ uptake before and after checkpoint-inhibitor therapy in 64 patients. Larger thyroid uptake changes were associated with recorded thyroid adverse events, particularly hypothyroidism. This small retrospective association does not establish detection before symptoms, causal toxicity prediction or improved management, and thyroid laboratory follow-up was limited.

Technology: TotalSegmentator; nnU-Net; AutoPET-trained lesion segmentation

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

AI contribution
Pretrained neural networks delineate organs and remove tumor uptake, enabling organ-level tracer-change measurements linked to recorded adverse events.
Test and result
Paired clinical PET/CT scans with independently recorded thyroid adverse events and exploratory organ-level uptake comparisons.
Limitation / still unknown
Prospective early-warning utility and causal interpretation remain unestablished; only 11 thyroid adverse events occurred and thyroid-function testing was incomplete.
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/PMC12589324/

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.

PET and MRI learning classifies prostate lesion risk

Source check: Source checked. Not independent scientific confirmation.

Finding: An ensemble of random forests combined PSMA PET and MRI features to classify 121 lesions from 52 surgical prostate-cancer patients as lower or higher risk against whole-mount pathology. Patient-separated repeated internal validation reported AUC 0.86. Additional models explored recurrence and overall risk in smaller subsets. This pilot does not establish biopsy replacement, external performance or clinical benefit.

Technology: Random-forest ensemble

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

AI contribution
An ensemble of nine random forests combines imaging features to classify pathology-defined lesion risk, with patient-level partitioning.
Test and result
Whole-mount pathology reference and repeated Monte Carlo validation that kept all lesions from each patient in one partition.
Limitation / still unknown
External validation and clinical utility remain unestablished; this small surgical cohort, global redundancy filtering and histology-informed lesion localization constrain generalization.
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/PMC8113201/

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.