ePoster Abstract
Interchangeability of a Radiomic Signature Between Conventional and Weekly Cone Beam Computed Tomography Allowing Response Prediction in Non-Small Cell Lung Cancer

https://doi.org/10.1016/j.ijrobp.2016.06.482Get rights and content

Section snippets

Purpose/Objective(s)

A previously validated radiomic signature based on treatment planning CT (pCT) scans showed to have prognostic information for lung cancer patients. Kilovolt cone-beam CT (CBCT) imaging is acquired prior to each fraction and it therefore has great potential to monitor response to treatment. In this study we aimed to assess the interchangeability of the radiomic signature between pCT and pre-radiotherapy CBCT imaging of lung cancer patients and show the potential of CBCT Delta Radiomics (i.e.

Materials/Methods

A total of 122 stage II-III NSCLC patients for which daily CBCTs were available were included in this study. All patients received radiation therapy between January 2012 and January 2014. Overall survival was retrieved for all patients. The radiomic signature is a Cox regression hazards model based on four features1, (I) “Energy,” (II) “Grey Level Nonuniformity,” (III) “HLH Grey Level Nonuniformity,” and (IV) “Compactness.” These features were derived from the pCT and the CBCT prior to the

Results

A linear relationship between pCT and CBCT-FX1 was found for all four features of the signature (slope of I: 1.21, II: 0.70, III: 0.65, IV: 1.01, all R2 >0.98, P < 0.0001). This slope was used to correct the feature values extracted from CBCT-FX1 images to be compared with the pCT scenario. Harrell’s c-index was 0.66 (95% CI = 0.58–0.74, P = 4.7*10-5) for CBCT-FX1 images. A c-index of 0.68 (95% CI = 0.61–0.75, P = 1.9*10-7) was found after adding the Delta volume (week 4 – baseline) in the

Conclusion

The performance of the model on CBCT images shows that similar prognostic information on overall survival can be derived from images of the first fraction of treatment, providing a calibration was performed. Furthermore, these preliminary results show the prognostic value could be improved by including information during treatment (e.g. “Delta volume”): a proof of concept for CBCT Delta radiomics.

References (0)

Cited by (0)

Author Disclosure: J.E. van Timmeren: None. R. Leijenaar: None. W. van Elmpt: None. P. Lambin: None.

View full text