AI in Pathology Quiz

Answers and explanations appear as you take the quiz.

1. What is a 'whole-slide image' in digital pathology?
   • A gigapixel digital scan of an entire tissue slide
   • A single cell photographed with a phone
   • An X-ray of a patient's chest
   • A handwritten lab report
2. Why do AI models process slides in small tiles or patches?
   • Because color is unimportant in pathology
   • Because the full gigapixel image is too large to feed in at once
   • Because tissue is always square
   • To make the image black and white
3. Multiple-instance learning helps when:
   • There are no labels at all and no goal
   • The model needs to translate text
   • Only a slide-level label is known, not the exact tumor location
   • The slide is a normal photograph
4. What regulatory pathway did FDA use for Paige Prostate?
   • It issued a blanket approval for all pathology AI
   • It granted a drug marketing authorization
   • It declared that pathologists are no longer needed
   • It granted a De Novo classification for the specific device and intended use
5. What does an AI Ki-67 count measure in a tumor?
   • How quickly the tumor cells are proliferating
   • The patient's blood pressure
   • The size of the hospital
   • The slide's thickness in millimeters
6. What can an AI system provide in a pathology workflow, according to the scoped examples?
   • A definitive cancer diagnosis without a pathologist
   • A decision-support output that a pathologist reviews in context
   • An automatic treatment prescription
   • A replacement for tissue sampling
7. What is a key idea in CLAM for whole-slide classification?
   • Every patch must have a manually annotated pixel mask
   • The model analyzes only patient age
   • Attention-based multiple-instance learning can identify informative image patches from slide-level labels
   • It replaces the slide image with a text report
8. What did Paige Prostate receive De Novo authorization to assist pathologists with?
   • Automatically selecting cancer treatment
   • Replacing biopsies with blood tests
   • Diagnosing every cancer type from MRI
   • Identifying suspicious regions on digitized prostate biopsy slides