Survival differences between patients with de novo and relapsed/progressed advanced non-small cell lung cancer without epidermal growth factor receptor mutations or anaplastic lymphoma kinase rearrangements

In this population-based study of patients with stage IIIB–IV NSCLC without EGFR or ALK mutations, we found that de novo patients had worse clinical outcomes in terms of OS and TSST than relapsed/progressed patients. This difference remained significant, even after controlling for age, sex, and other factors. The overall treatment patterns for both de novo and relapsed/progressed patients were similar, except that relapsed/progressed patients used more varied therapeutic approaches as first-line therapy.

Although this study targeted patients with NSCLC without EGFR or ALK mutations and included the post-immunotherapy era, our findings were consistent with previous studies targeting patients with different biomarker status or studies conducted in the pre-immunotherapy era [9,10,11,12]. Our results showed that de novo patients had shorter median OS than relapsed/progressed patients (11.0 vs. 11.5 months; HR 1.07) from the initiation of first-line palliative therapy. According to previous studies in Canada that identified patients with metastatic NSCLC in the pre-immunotherapy era regardless of biomarker status, de novo presentation of metastatic NSCLC had HRs ranging from 1.2 to 1.4 in terms of OS than relapsed/progressed presentation [10, 11]. These studies reported the median OS from the date of diagnosis of metastatic disease ranged from 3.7 to 4.7 months for de novo patients and 6.9–8.9 months for relapsed/progressed patients. Similar results were reported in a study using single-center data from the United States, in which the median OS from the date of diagnosis of metastatic disease in patients with KRAS-mutant lung adenocarcinomas was 13 months in de novo patients and 18 months in recurrent patients (HR, 1.41) [9]. These results are in line with our findings that more patients in the de novo group died during first-line therapy than in the relapsed/progressed group (41.8% vs. 39.8%). In addition, fewer patients in the de novo group reached the end of the study period without initiation of second-line therapy or death (16.7% vs. 21.3%).

Other than the OS, we observed TFST and TSST. TFST and TSST reflect the duration of disease and symptom control and incorporate treatment tolerability and patient compliance [22]. Although real-world PFS is used in retrospective studies in oncology, it often requires manual extraction of data from medical charts, potentially slowing research and limiting the number of patients participating in retrospective studies [24]. TFST could be considered a candidate surrogate marker for real-world OS or PFS, although further validation is needed [24,25,26]. TSST could be considered a proxy for time to second objective disease progression or death, as long as the second subsequent therapy is initiated by disease progression rather than the toxicity of the previous therapy [27]. Our results showed that the median TFST was 6.4 months in both groups of patients, but the ongoing treatment probability was higher in relapsed/progressed patients than in de novo patients from 6.4 months since the initiation of first-line treatment (P < 0.001). Median TSST (9.5 vs. 9.9 months, P < 0.001) was shorter in de novo patients than in relapsed/progressed patients. Similar results were reported in a study from Japan, although it was conducted in the pre-immunotherapy era and identified patients with metastatic NSCLC treated with chemotherapy. In the study, de novo patients had a worse median PFS from the initiation of first-line chemotherapy than patients with postoperative recurrence (4.2 vs. 5.5 months, P < 0.01) [12].

To our knowledge, the mechanism underlying the difference in survival between patients with de novo versus relapsed/progressed NSCLC is unknown. The difference may be attributed to the relatively high tumor burden in de novo patients, as reported in previous studies. For instance, Gibson et al. [10] reported that de novo cohort of NSCLC had more extrapulmonary metastatic sites than relapsed cohort (27% vs. 3%, P < 0.001) noting that this could explain the worse prognosis of the de novo cohort. Additionally, Sekine et al. [12] observed that brain and bone metastases were significantly more common in de novo compared to patients with postoperative recurrence, while pulmonary metastases were more frequent in the patients with postoperative recurrence. The relatively limited routine monitoring in de novo patients may contribute to the higher tumor burden in this group. Due to routine monitoring, disease progression is likely to be detected earlier in asymptomatic relapsed/progressed patients, resulting in smaller tumor burden. In contrast, de novo patients are more likely to present with symptoms indicating a more advanced disease stage at the time of detection.

Unlike the differences in OS and TSST, treatment patterns for both de novo and relapsed/progressed patients were similar, except that treatment regimens for relapsed/progressed patients were more varied in first-line therapy. Among patients who received first-line palliative therapy, platinum-based chemotherapy was the most prevalent in both groups, consistent with the results of previous studies [17, 18]. De novo patients used paclitaxel/pemetrexed/gemcitabine + platinum as first-line therapy more frequently than relapsed/progressed patients (92.0% vs. 63.9%). Previous use of the platinum-based regimen as adjuvant therapy may have affected the treatment pattern of the relapsed/progressed group. Although recent studies in the United States reported high use of immunotherapy in the first-line setting [15, 28], a direct comparison is inappropriate because, in our study, immunotherapies were not reimbursable for first-line therapy during the study period, resulting in low use of immunotherapy.

Therapeutic approaches varied in second-line therapy, with no regimen accounting for > 20%. Similarly, previous studies conducted before immunotherapies became prevalent in the United States showed that various chemotherapies were used in advanced NSCLC [16, 29]. In our study, docetaxel was the preferred second-line therapy for both groups of patients. However, previous studies conducted in patients with metastatic NSCLC without EGFR or ALK mutations and in the post-immunotherapy era showed that immunotherapy was the most common second-line therapy. For example, Simeone et al. [17] reported that nivolumab was the most frequent regimen, accounting for 31% of second-line therapies among patients with metastatic NSCLC, using Flatiron health data from January 2013 to January 2017. Similarly, in a study using Flatiron health data from 2018 to 2019, most patients (50.7%) with metastatic NSCLC used second-line therapy containing immunotherapy [28]. Compared with previous studies, the patients included in this study were treated with chemotherapy more than immunotherapy. The difference might be due to limited patient access to immunotherapy, considering that the study period included the era before immunotherapy reimbursement for the second and subsequent line of therapy, which has been effective since 2017.

This study represents a large-scale, multi-year analysis of prognosis and treatment patterns between patients with de novo and relapsed/progressed NSCLC. In this study, both inpatient and outpatient prescriptions were confirmed as part of the national health insurance system in South Korea based on a fee-for-service delivery system. The results of our study are representative of patients with stage IIIB–IV NSCLC in South Korea, as the database covers nearly the entire Korean population. Considering the impact of the disease stage at the time of the initial diagnosis on OS, our findings highlight the importance of screening for the early detection of NSCLC. In addition, these findings suggest that the stage of the disease at the time of the initial diagnosis should be considered in observational studies and clinical trials as a prognostic factor. To the best of our knowledge, this is the first study to represent the treatment pattern of patients with NSCLC without EGFR or ALK mutations, divided into de novo and relapsed/progressed patients.

This study had several limitations. First, we could not obtain genomic information as the HIRA database does not provide them. Therefore, EGFR mutation and ALK translocation statuses were inferred based on the use of EGFR and ALK tyrosine kinase inhibitors, not by molecular testing. In addition, patients with EGFR mutations or ALK translocations may not have been treated with targeted therapies. However, EGFR and ALK tyrosine kinase inhibitors are the most commonly recommended regimens for patients with EGFR and ALK mutations [30]. Similarly, patients with SCLC were excluded based on first-line treatment, which was used as an alternative for biopsy results. Also, we were unable to identify PD-L1 status in the HIRA database. However, patients receiving immunotherapies may have a certain level of PD-L1 expression, given that PD-L1 expression levels are used to determine reimbursement for immunotherapies in South Korea. Second, it was impossible to identify the actual status of the cancer stage in the HIRA database. However, we used an operational definition from a previous study that identified patients with stage IIIB–IV NSCLC using the HIRA claims data [20], and the approach was discussed with clinical experts. Third, there is a possibility of misclassifying palliative therapy initiated within 6 months of thoracic surgery as adjuvant therapy. This could underestimate the OS observed in the relapsed/progressed group and misclassify second-line therapy as first-line therapy. However, the difference in OS remained significant when we changed the operational definition to 5 months. In addition, the operational definition was based on a previous study and the results of consultation with clinical experts that setting a narrower gap can misclassify adjuvant therapy as first-line palliative therapy [20]. Finally, there could be potential confounders, and some prognostic variables were not available, such as squamous cell histology and smoking history.

留言 (0)

沒有登入
gif