KEY POINTS
- This retrospective cohort included 155 patients with oncologic emergencies treated from 2022–2025, including superior vena cava syndrome, metastatic spinal cord compression, brain metastases with neurological deficits, tumor bleeding and painful non-spinal bone metastases.
- The integrated CT-linac workflow combined CT simulation, target delineation, automated planning, plan review and first treatment without moving the patient between separate systems. Median total workflow time was only 24.66 minutes.
- The largest time savings came from automated segmentation. In 108 patients with available models, contouring took a median 5.40 minutes automatically versus 17.47 minutes manually, a median reduction of 11.70 minutes (p<0.001).
- Plan quality remained generally acceptable despite the compressed workflow. Median target coverage was 95.0%, 95.48% of plans met predefined target goals, and the median organ-at-risk constraint compliance rate was 100%.
- Delivery verification was also reassuring: median 2D in-vivo gamma passing rate was 97.35%, median offline 3D reconstructed rate was 92.26%, and phantom-based verification was essentially 100%.
- Acute symptom relief occurred in 89.68% of patients. Median ECOG performance status improved from 2 to 1; pain scores fell from 3 to 1; and among bleeding patients, 86.67% improved and 53.33% achieved complete hemostasis.
- Three-month local control was 96.77%, with no grade ≥3 treatment-related toxicity. Median overall survival was 5.37 months, consistent with a population presenting with advanced, often life-threatening disease.
CLINICAL TAKEAWAY
For oncologic emergencies, the clinically interesting result is not simply automation: simulation, planning and first treatment were compressed into roughly 25 minutes while maintaining acceptable dosimetry and high symptom-response rates. The workflow is promising for centres with integrated CT-linac infrastructure, but the evidence remains retrospective and single-centre.