Assessing Treatment Response after Lung SABR: An Evaluation of the Predictive Value of RECIST Criteria

Abstract

<h3>Purpose/Objective(s)</h3> Response Evaluation Criteria In Solid Tumors (RECIST) – version 1.1 is a commonly used framework to assess treatment response. However, response assessment after lung stereotactic ablative radiotherapy (SABR) can be difficult due to radiation-induced lung changes. The goal of this study was to assess how frequently lung lesions treated with SABR triggered RECIST criteria for recurrence and to correlate RECIST findings with actual treatment outcomes. <h3>Materials/Methods</h3> We reviewed patients who were treated with lung SABR between 2010-2015. Prescribed dose was 54-60 Gy in 3-8 fractions. Both primary lung and metastatic lesions were included. All lesion measurements were based on maximum diameter on CT axial slices. After treatment, progressive disease (PD) by RECIST was based on strict RECIST criteria and defined as: 1) increase of ≥ 20% compared to post-treatment minimum and 2) ≥ 5 mm absolute increase. The final assessment of recurrence was based on subsequent changes in lesions size after PD by RECIST and, when available, PET scan and/or pathology. When assessing subsequent changes in lesion size, stability (≤ 20% change) in size for at least one year or a decrease (> 20%) in size indicated no recurrence, while subsequent growth (> 20%) indicated recurrence. The assessment of lesion response by the treating radiation oncologist (RO) in the clinical record was also recorded when available. <h3>Results</h3> A total of 85 patients, with 88 lesions, met inclusion criteria. 75 of these lesions were lung primaries, the remainder were metastases. Median patient age was 79 years (IQR: 73-85) and mean lesion size was 2.1 ± 0.9 cm. Median follow-up was 52 months (IQR: 33-68). 66% (58/88) of treated lesions met criteria for PD by RECIST. However, on final assessment only 10% (9/88) were found to have recurrence. The positive predictive value (PPV) of RECIST criteria was 0.16. In the subset of patients with primary lung tumors only, 64% (48/75) were classified as PD by RECIST, but only 12% (9/75) were categorized as recurrence on final assessment. PPV was 0.19. The predictive value of RECIST was lower when PD by RECIST was triggered ≤ 12 vs. > 12 months post-treatment (PPV 0.08 vs. PPV 0.21). Median time to PD by RECIST was 19 months (IQR: 12-27) for those with recurrence vs. 15 months (IQR: 8-24) for those without recurrence on final assessment (p=0.13). Assessment by treating RO was only available for 57% (50/88) of cases, but concordance with final study assessment was 0.85, translating to an overall predictive accuracy of 94% (47/50). <h3>Conclusion</h3> Using strict RECIST criteria, two-thirds of patients treated with lung SABR met criteria for PD. However, on further follow-up only a minority of these patients were felt to have recurrence, leading to a poor PPV for RECIST criteria in this setting. Further work is needed to develop validated criteria for designation of recurrence after lung SABR.