引導式測驗 · 中等 等級

人工智慧在肝臟疾病和肝病學測驗的應用

測試肝臟病理學評分、超音波研究、移植風險模型、FDA 資格範圍和專業審查。

問題 1 的 8

FDA 在什麼情況下將 AIM-NASH 授予藥物開發工具資格?

預覽每個問題8 問題

當您參加測驗時,就會出現答案和解釋。

  1. FDA 在什麼情況下將 AIM-NASH 授予藥物開發工具資格?
    • Diagnosing all liver diseases from a blood test
    • Setting transplant priority in every country
    • Scoring liver-biopsy features in MASH clinical trials
    • Replacing liver biopsy with ultrasound
  2. 誰對 AIM-NASH 活檢評分做出最終解釋?
    • The AI tool alone
    • The pathologist after reviewing the whole slide and AI output
    • The trial sponsor
    • A transplant coordinator
  3. 為什麼超音波和活檢人工智慧結果應該被視為不同的任務?
    • Every imaging method measures the same tissue directly
    • A biopsy model automatically validates ultrasound
    • Ultrasound tools never use AI
    • They use different inputs and reference standards
  4. GEMA-AI 研究估計了什麼?
    • A waitlist outcome estimate for transplant candidates using clinical variables
    • A histology grade for steatosis in biopsy images
    • A radiology score for liver ultrasound texture
    • A classification of cancer-signal origin in blood
  5. 研究模型是否會自動成為移植分配規則?
    • Yes, once it has an external test set
    • Yes, when it uses the same lab variables
    • No; allocation requires policy and multidisciplinary review
    • No, because prediction models cannot use lab results
  6. Which liver-biopsy features does AIM-NASH score?
    • Kidney filtration, blood pressure and sodium only
    • Steatosis, inflammation and fibrosis
    • Thyroid shape, margin and echogenicity
    • Tumor stage and medication response
  7. What should a liver imaging team check before using an AI fibrosis score?
    • That the model was trained on any organ image
    • Only whether the report is short
    • Whether a transplant list is available
    • That the model was validated for the relevant imaging task and patient group
  8. What remains central in transplant allocation decisions?
    • Policy, equity, organ availability and transplant-team review
    • The highest model score alone
    • An ultrasound image without other information
    • A model vendor’s ranking