DICTORA.AI
I built a radiology imaging reader that uses Astra existing image reading capabilities to interpret CT scans of the chest, which include more than 200 slices, generate radiology report findings and impression.
Core trick
Feeds more than 200 CT slices to a vision model to generate radiology findings and impressions, then cross-checks them against a radiologist's report.
Description
I built a radiology imaging reader that uses Astra existing image reading capabilities to interpret CT scans of the chest, which include more than 200 slices, generate radiology report findings and impression. This report was cross-referenced against a report generated by a board-certified radiologist, and it demonstrated superiority of Astra relative to the radiologist without any misses in our five cases, demonstrating Probable AGI at least in the field of radiology and in our five cases. A large blinded academic peer review study is now mandatory as a matter of patient safety to ensure all radiology reports use Astra for interpretation.
README excerpt
# Dictora + Astra — Hackathon review [Watch the one-minute video](https://youtu.be/j_tocTcEGo4) ## Complete source project Download [dictora-hackathon-public.zip](dictora-hackathon-public.zip) and extract it. The full runnable source, tests, and documentation are inside `dictora-hackathon-public/`. The archive preserves the source folder structure; the files are packaged rather than expanded in this repository. ```bash unzip dictora-hackathon-public.zip cd dictora-hackathon-public npm ci npm run build npm start ``` The source build and 13 offline tests passed before upload. API credentials, scans, original dataset reports, and spending ledgers are excluded. Judges supply their own authorized data and API access for a live run. --- # Dictora — Astra radiology hackathon prototype Dictora explores CT report drafting with GPT-6 Astra and gives a radiologist a side-by-side workspace to review AI output against companion CT-RATE reports. ## What judges can review - **CT research:** select a local chest CT, browse axial slices in lung/soft-tissue windows, and generate an image-only research draft from 24 sampled slices. - **Dictation:** turn supplied findings into a report. Completed individual CT reviews automatically transfer the draft and its limitations into Dictation with “CT chest.” - **Batch comparison:** process saved scans sequentially, generate each image-only draft before reading its original report, and identify major omissions, contradictions, and unsupported