Serial 4-Dimensional (4D) Computed Tomography/4D Positron Emission Tomography Imaging to Predict and Monitor Response for Locally Advanced Non-Small Cell Lung Cancer Radiation Therapy

Comparing serial 3DPET imaging acquired pre-, post-, and during chemo-radiation therapy (CRT) for locally-advanced non-small cell lung cancer (LA-NSCLC) patients has demonstrated that FDG-PET image characteristics correlate with treatment outcome. However, the literature disagrees on which image characteristic has optimal prognostic potential and on the timing of optimal response assessment; furthermore, characteristics for nodal disease has been largely ignored. Since respiratory-correlated (4DPET) has shown impact on response assessment, we compared seven image-based characteristics, identified from the literature, extracted from 4DCT/4DPET studies acquired at multiple time points before and during CRT. We hypothesized that the prognostic power of these images can be improved by identifying which of these seven characteristic is optimal, as well as the optimal timing for image-based response assessment. Consecutive consenting patients with LA-NSCLC receiving curative intent CRT were enrolled. Radiotherapy was delivered using IMRT plans developed from the planning 4DCT/4DPET scan. Additional 4DPET/4DCT scans were acquired 2, 4, and 7 weeks during treatment. Clinical outcome data (progression-free and overall survival, local and locoregional control) was collected two years after CRT. Image-based characteristics were extracted and compared for each time point. For each characteristic, individual patient time trends were grouped according to treatment outcome, and the average trends for each group compared for statistically significant difference (P < 0.05) using Student statistics. Univariate and multivariate analyses were performed on tumor and nodal image characteristics, and also on the rate of change of these characteristics as treatment progressed. Thirty-two patients were recruited, 27 completing all scans. Image characteristics with the highest significance for loco-regional failure were the tumor (LRC; P = 0.013) and node (PFS; P = 0.03) volumes defined by the SUV ≥3, and GTVCT, tumor (LRC; P = 0.017). When considering rate of change, multiple characteristics with prognostic potential were identified: when the volumes described by SUV ≥3 on 4DPET or by the GTV contoured on 4DCT are stable late in the treatment, there is a mean progression-free survival advantage [13 (P = 0.013) and 11 months (P = 0.03), respectively]. While characteristics obtained from 4DCT/4DPET images obtained prior to CRT can be predictive of outcome, the rate of change of these characteristics also shows strong correlations. Typically, 4DPET image characteristics that change little between weeks 4 and 7 of treatment are correlated with improved outcomes.