A Systematic Review Into the Radiologic Features Predicting Local Recurrence After Stereotactic Ablative Body Radiotherapy (SABR) in Patients With Non-Small Cell Lung Cancer (NSCLC).

ABSTRACT

PURPOSE: Posttreatment surveillance for local recurrence (LR) after stereotactic ablative body radiotherapy (SABR) can include both fluorodeoxyglucose-positron emission tomography (FDG-PET) and computed tomography (CT). Radiation-induced lung injury shares a similar appearance to LR after treatment, making the detection of LR on imaging difficult for clinicians. We aimed to summarize radiologic features of CT and FDG-PET predicting LR and to evaluate radiomics as another tool for detecting LR. METHODS AND MATERIALS We searched MEDLINE, EMBASE, and PubMed databases for published studies and Web of Science, Wiley Online, and Science Direct databases for conference abstracts that had patient populations with non-small cell lung cancer and reported post-SABR radiologic features of FDG-PET or CT and radiomics from either FDG-PET or CT. Studies for inclusion were independently reviewed by 2 authors. RESULTS: Across 32 relevant studies, the incidence of LR was 13% (222/1726). On CT, certain gross radiologic appearances and kinetic features of changes in size, diameter, volume, or 3 consecutive rises in volume of masslike consolidation are suggestive of LR. **Particular regard should be made for the presence of any ≥3 high-risk features on CT or the individual high-risk features of enlarging opacity at ≥12 month's post-SABR as being highly suspicious of LR. On FDG-PET a relative reduction of <5% of maximum standardised uptake value (SUVmax) from baseline in the first 12 months or cut-offs of SUVmax >5 and SUVmean >3.44 after 12 months can indicate LR. There is limited evidence available to corroborate radiomic features suggestive of LR. CONCLUSIONS: This research has identified common features of LR compared with radiation-induced lung injury, which may aid in early and accurate detection of LR post-SABR; further research is required to validate these findings.