AI diagnostics and what clinicians can verify in oncology - Report - MDSpire
Coming Soon: Introducing MDSpire News. Learn more
Conexiant’s news site is now MDSpire News. Learn more

AI diagnostics and what clinicians can verify in oncology

"The more AI changes what we do … the less able we are to check it.”

  • By

  • Lisa Astor

  • October 1, 2026

  • 8 min

Share

Clinical Report: AI diagnostics and what clinicians can verify in oncology

Background

The integration of AI in medical imaging is crucial. Many AI tools remain in research phases and have not yet been widely adopted in clinical practice. Understanding the tiers of AI applications is essential for clinicians to utilize these technologies effectively.

Data Highlights

No numerical data provided in the source material.

Key Findings

  • AI can identify features in medical images that are often missed by human clinicians.
  • Three tiers of AI applications exist: tier one (features visible to humans), tier two (features poorly seen by humans), and tier three (features invisible to humans).
  • Currently, there are no FDA-cleared tier three applications in radiology, although the Artera AI assay is a notable example in oncology.
  • Verifiability of AI applications decreases with higher tiers, complicating their integration into clinical workflows.
  • Data set drift can affect the performance of AI models over time.
  • Foundation models can reduce the need for labeled data.

Clinical Implications

Clinicians should be aware of the varying levels of AI application and their respective challenges in verifiability and integration into practice.

Conclusion

Understanding the dynamics of AI in oncology diagnostics is critical for clinicians.

Related Resources & Content

  1. ASCO AI in Oncology, ASCO, 2026 -- AI Diagnostics and What Clinicians Can Verify in Oncology
  2. ASCO Publications, ASCO, 2016 -- Analytic validation of a clinical circulating tumor DNA assay for patients with solid tumors.
  3. ACR Approves First Practice Parameter for Imaging Artificial Intelligence, ACR, 2026 -- ACR Approves First Practice Parameter for Imaging Artificial Intelligence
  4. Interval cancer, sensitivity, and specificity comparing AI-supported mammography screening with standard double reading, The Lancet, 2026 -- Interval cancer, sensitivity, and specificity comparing AI-supported mammography screening with standard double reading.
  5. ASCO Publications — Analytic validation of a clinical circulating tumor DNA assay for patients with solid tumors.
  6. ASCO Publications — Real-world artificial intelligence use in oncology practice: High clinician verification amid critical governance gaps.
  7. Real-world artificial intelligence use in oncology practice: High clinician verification amid critical governance gaps.
  8. ACR Approves First Practice Parameter for Imaging Artificial Intelligence
  9. Interval cancer, sensitivity, and specificity comparing AI-supported mammography screening with standard double reading without AI in the MASAI study: a randomised, controlled, non-inferiority, single-blinded, population-based, screening-accuracy trial.
  10. Artificial intelligence as a simultaneous second reader in diagnostic radiology: An umbrella review of systematic reviews and meta-analyses - ScienceDirect

Original Source(s)

AI diagnostics and what clinicians can verify in oncology

  • by Lisa Astor

  • October 1, 2026

Related Content

FDA expands bladder cancer therapy approval

The expansion offers another treatment option for patients with high-grade tumors following BCG failure.

Misleading AI summaries may alter recall

Participants exposed to inaccurate postevent summaries were less likely to correctly remember key details from videos they had previously watched.

10 states that support physician work-life balance

These 10 states have paid family or medical leave programs and health, wellness, peer support, or monitoring resources available to physicians or other health professionals.