Clinical feasibility of single-source dual-spiral 4D dual-energy computed tomography for proton treatment planning of lung-cancer patients

Purpose/Objective:

Dual-spiral dual-energy CT (DECT) provides additional patient information to improve range accuracy in proton therapy, but is prone to motion between the two consecutively acquired scans. Here, the clinical feasibility of dual-spiral time-resolved DECT (4D-DECT) for proton treatment planning within the thoracic region was evaluated.

Material/Methods:

4D-DECT scans of three non-small-cell lung cancer (NSCLC) patients were acquired during the course of treatment with a Siemens single-source DECT scanner (Fig.1). For temporally averaged datasets and four breathing phases, the geometrical conformity of both 4D-DECT scans before (80kVp/140kVp) and after (58keV/79keV) image post-processing including deformable image registration (DIR) was assessed by normalized cross correlation (NCC).
To evaluate the reliability of dose calculation, clinical treatment plans were recalculated on DECT-derived 79keV MonoCT and 140kVp SECT datasets as reference using a heuristic conversion (HLUT) from CT number to stopping-power ratio (SPR). Dose distributions were compared with gamma analyses (0.1% dose-difference, 1mm distance-to-agreement criterion).
Finally, range differences between HLUT and patient-specific DECT-based SPR prediction were quantified.

Results:

Respiration changes during 4D-DECT acquisition resulted in NCCs>80%, indicating geometrical deviations of (1-2)mm. This was almost completely corrected by DIR leading to a high geometrical conformity with average NCC ± SD = (99.6±0.4)% corresponding to anatomical shifts below 0.2mm (not visually distinguishable). Even the impact of coughing could be corrected by DIR (Fig.2).
Clinical dose distributions on 140kVp and 79keV datasets were similar with average gamma passing rate of 99.9% and maximal dose difference of 0.8%.
Clinically relevant mean range shifts of (2.2±1.2)% were determined between patient-specific DECT-based SPR prediction and HLUT.

Conclusion:

Dual-spiral 4D-DECT is applicable for dose calculation on 79keV MonoCT datasets in NSCLC patients. Patient-specific DECT-based SPR prediction performed properly and revealed its potential for reducing range uncertainty. Even if large motion differences hamper 4D-DECT post-processing, only the 140kVp scan can be used and additional information on respiration variability and robustness is gathered.