Accurate prediction of target dose-escalation and organ-at-risk dose levels for non-small cell lung cancer patients

Steven F. Petit*, Wouter van Elmpt

*Corresponding author for this work

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

Purpose/objective: To develop a method to predict feasible organ-at-risk (OAR) and tumour dose levels of non-small cell lung cancer (NSCLC) patients prior to the start of treatment planning. Materials/methods: Included were NSCLC patients treated with volumetric modulated arc therapy according to an institutional isotoxic dose-escalation protocol. A training cohort (N = 50) was used to calculate the average dose inside the OARs as a function of the distance to the planning target volume (PTV). These dose-distance relations were used in a validation cohort (N = 39) to predict dose-volume histograms (DVHs) of OARs and PTV as well as the maximum individualized PTV dose escalation. Results: The validation cohort showed that predicted and achieved MLD were in agreement with each other (difference: -0.1 +/- 1.9 Gy, p = 0.81). The spinal cord was dose limiting in only two patients, which was correctly predicted. The achieved mean PTV dose varied from 52 to 73 Gy and was predicted correctly with an accuracy better than 2 Gy (i.e. 1 fraction) for 79% of the patients. Conclusion: We have shown that the MLD and the prescribed PTV dose could be accurately predicted for NSCLC patients. This method can guide the treatment planner to achieve optimal OAR sparing and tumour dose escalation.
Original languageEnglish
Pages (from-to)453-458
Number of pages6
JournalRadiotherapy and Oncology
Volume117
Issue number3
DOIs
Publication statusPublished - 1 Dec 2015

Keywords

  • Dose prediction
  • Dose-distance relation
  • Knowledge based treatment planning
  • NSCLC dose escalation
  • Treatment planning QA

Fingerprint

Dive into the research topics of 'Accurate prediction of target dose-escalation and organ-at-risk dose levels for non-small cell lung cancer patients'. Together they form a unique fingerprint.

Cite this