br The data regarding outcomes resulting from symptomatic
The data regarding outcomes resulting from symptomatic versus nonsymptomatic relapse identification are conflicting. A study evaluating patients enrolled onto the neoadjuvant CRT Intergroup 0160 and 0139 trials demonstrated no survival benefit with more intensive imaging surveillance.15 Additionally, a study of patients managed with complete resection found that relapse detection in asymptomatic patients did not afford any OS benefit16; of note, more than half of the patients in this Vorinostat (SAHA, MK0683) particular study had stage I disease. In contrast, Westeel et al17 examined 192 patients followed with an aggressive postsurgical surveillance regimen of CT imaging and fiberoptic bronchoscopy, and found improved OS in patients with asymptomatic recurrences. Our study found that patients with symptomatic relapses compared to those detected by surveillance imaging had inferior median OS (23 vs. 36 months; P ¼ .013). In our analysis, the definition of relapse required both evidence of progressive disease on imaging as well as a change in oncologic care. Thus, all patients with relapsed disease in our analysis underwent salvage therapy with either systemic therapy or repeat irradiation. The survival curves indicate that detection of first relapse was not dependent on symptomatic presentation (Figure 2A); therefore, the improved OS in our study is likely not inflated by lead-time bias but rather is due to earlier oncologic intervention. r> Among the inherent weaknesses to retrospective analyses, there are additional limitations to our study. For instance, all patients did not undergo invasive mediastinal staging, which could affect the incidence of RRs. Also, the modality of relapse detection was not uniform among all patients, which could bias the occurrence of a relapse event by the imaging quality or interpretation by the radi-ologist. Although a potential confounder, as stated above, a change in oncologic management in addition to imaging findings was needed to define a relapse event in our study. We found that 28%, 57%, and 26% of patients had 18FDG-PET/CT-detected LR, RR, and DR events, respectively, although this method of surveillance was not uniformly used by all physicians and may not accurately reflect detection rates by 18FDG-PET/CT. Previous evidence sug-gests 18FDG-PET/CT has superior sensitivity and specificity in detecting NSCLC recurrences compared to standard CT imaging, especially when relapse is within the mediastinum.18 Although 18FDG-PET/CT may be more robust in detecting relapses, a pro-spective study of 100 NSCLC patients of varying disease stages found that 18FDG-PET/CT imaging 3 months after RT detected more relapses than CT alone, but only 3% of patients had disease amenable to salvage treatment.19 Reddy et al20 also recently re-ported no differences in survival outcomes in patients who experi-enced relapse identified by 18FDG-PET/CT versus CT > 6 months
after completing definitive RT. These findings are reflected in various guidelines, which currently recommend against the use of 18FDG-PET/CT for routine surveillance imaging after curative-intent interventions in NSCLC.
Approximately 15% to 30% of stage III NSCLC patients will have intracranial relapses after definitive therapy.21,22 An analysis of 422 stage III NSCLC patients enrolled onto 4 prospective protocols of combined modality therapy demonstrated that approximately 60% of intracranial relapses occur within 6 months of treatment.23 Further-more, a study of 455 NSCLC patients of differing stages found the rate of symptomatic brain metastases was < 10% over a period of 16 months.24 Intracranial relapses were the most common site of DR in our study (30%) and were mostly identified by symptomatic presen-tation, despite 81% of patients undergoing pretreatment assessment by MRI. Because the majority of patients underwent pretreatment evaluation of the brain, our higher rate of symptomatic brain metastases is likely a reflection of aggressive biology and not profound un-derestimates of the pretreatment intracranial tumor burden.
This study found that most relapses happen within the first year of completing definitive CRT, and just over half of these are identified after symptomatic presentation. Although we found that symptomatic relapses correlated with worse OS, a more intensive surveillance regimen within the first 6 months of completing treatment may help identify more asymptomatic relapses. This would be advantageous because it may provide an opportunity to initiate salvage therapy sooner, possibly in patients with better PS and with a lower burden of disease. Of note, adjuvant therapy with durvalumab after CRT is now the standard of care in stage III disease and has been proven to increase the time to progression.25 Thus, before symptomatic recurrence, administration of durvalu-mab may improve OS when the disease burden is lower or via the synergistic interactions between RT and immunotherapies.26