AI in Ophthalmology Quiz

Answers and explanations appear as you take the quiz.

1. Why are retinal images used as inputs in ophthalmic deep-learning research?
   • They contain visible retinal structures that can be analyzed for specific research tasks
   • They provide a complete diagnosis of every systemic disease
   • They replace all clinical examinations
   • They guarantee equal performance across populations
2. What regulatory action did the FDA take for IDx-DR in 2018?
   • Cleared it through a 510(k) as a generic camera
   • Granted a De Novo request and classified the device
   • Approved it as a drug
   • Granted an emergency-use authorization
3. What is OCT, the imaging method central to many retinal AI tools?
   • A blood test for the eye
   • A type of contact lens
   • An optical scan giving micron-resolution cross-sections of the retina's layers
   • A method of measuring eye pressure
4. What did Poplin et al. report predicting from retinal fundus photographs?
   • A patient’s complete genome sequence
   • Exact medication adherence for the next year
   • Every future eye disease with certainty
   • Age, sex, smoking status, blood pressure and cardiovascular outcomes
5. What did the RETFound paper evaluate?
   • A retinal-image foundation model pretrained on fundus photographs and OCT
   • A model trained only on chest radiographs
   • A surgical robot for retinal operations
   • A device that measures blood pressure from sound
6. In the Moorfields-DeepMind paper, what were the two main network stages?
   • A speech recognizer followed by a language translator
   • A segmentation network followed by a classification network that gives diagnoses and referral recommendations
   • A tumor detector followed by a robot surgeon
   • A gene sequencer followed by a medication planner
7. Which imaging modality did the Moorfields-DeepMind model analyze to recommend referrals for retinal disease?
   • Chest X-rays
   • Ultrasound images of the heart
   • Optical coherence tomography (OCT) scans
   • Skin photographs
8. Which neural-network family did the Poplin retinal-photo study use?
   • A finite-state transducer
   • A nearest-neighbor graph search
   • A recurrent speech recognizer
   • A deep convolutional neural network