br The most common localization of
The most common localization of the recurrence in both groups was mediastinal lymph nodes, followed by lung and brain. All pa-tient characteristics are described in Table 1.
Changes in PD-L1 Expression in Tumors From Patients With or Without Adjuvant Chemotherapy After Primary Surgical Resection
In the cases undergoing adjuvant chemotherapy, 5 (25%) of 20 initial tumor specimens showed PD-L1 AT13387 (> 1%). After platinum-based chemotherapy, 7 (35%) cases showed upregulation of PD-L1 expression at NSCLC recurrence (P ¼ .102) (Figure 1A-D; Table 2). After adjuvant chemotherapy, PD-L1 expression was present in 10 of 20 tumors, corresponding to 50% (Figure 2B).
In cases without adjuvant chemotherapy, initial PD-L1 expres-sion was found in 5 of 16 tumors, corresponding to 31%. In comparison to the group with adjuvant chemotherapy, only 2 (12.5%) of 16 cases showed upregulation in PD-L1 expression at NSCLC recurrence (P ¼ 1.000) (Figures 1E,F and 2A).
Table 1 Patient Characteristics
Adjuvant No Adjuvant
recurrence, d (range)
Abbreviations: ALK ¼ anaplastic lymphoma kinase; EGFR ¼ epidermal growth factor receptor.
Of all 36 cases, 6 patients were PD-L1-negative on initial tumor sample and were positive at NSCLC recurrence, corresponding to 17%.
PD-L1 Expression and Correlation With Oncogenic Driver Alterations
In our study cohort, we assessed the 3 most common oncogenic driver alterations (KRAS, EGFR, and ALK), which also guide therapeutic planning in NSCLC. Of our 36 patients, 10 (28%) patients harbored KRAS mutations, whereas 2 (6%) patients Max Lacour et al
showed an EGFR mutation. None of our cases showed ALK rear-rangements. The most common types of KRAS mutations are G12C, G12D, and G12V.22 Of 10 cases with KRAS mutations in our study cohort, G12C was the most common mutation present in 6 (60%) patients. G12D, G12V, G12A, and G12S were all present in a single case.
Of the 10 patients presenting with KRAS mutation, 7 (70%) patients had PD-L1 expression either on initial tumor or NSCLC recurrence.
The aim of this study was to determine variation in PD-L1 expression from initial diagnosis to recurrence in NSCLC and its relation to chemotherapy. Recently published retrospective studies also investigating the effect of cytotoxic chemotherapy on PD-L1 expression indeed showed a trend towards higher PD-L1 expres-sion after administration of platinum-based chemotherapy.15,17 Interestingly, Sheng et al, as well as Rojko et al, reported a decrease in PD-L1 expression after chemotherapy.14,16 However, the study cohort of Sheng et al included 11 (34.4%) of 32 patients harboring an EGFR mutation, with recent studies demonstrating that PD-L1 expression in EGFR mutated tumor samples are lower than in those of wild type for EGFR.23 Secondly, chemotherapy regimens in the mentioned study included EGFR-tyrosine kinase inhibitors in approximately 25% of the cases, which have been re-ported to down-regulate PD-L1 expression in EGFR-mutated cell lines.24 Furthermore, the study of Rojko et al included samples of not only NSCLC, which may explain the discrepancy in the change in PD-L1 expression compared with our study.
In our study, we demonstrated that platinum-based standard chemotherapy might increase PD-L1 expression in NSCLC. These findings support the use of such chemotherapies in combination with immune checkpoint inhibitors such as pembrolizumab and nivolumab. This new therapeutic approach has been evaluated in the Keynote-021 trial, which demonstrated that the combination of an immune checkpoint inhibitor (pembrolizumab) and a platinum-pemetrexed doublet increases progression-free survival compared with standard chemotherapy alone, even in the PD-L1 low-expression group.25
Moreover, in our study cohort, we have seen that PD-L1 expression might represent a dynamic biomarker, which can be expressed in NSCLC recurrences even if the initial tumor was PD-L1-negative (in 17% of our cases). Accordingly, re-biopsy at NSCLC recurrence should always be considered in order to assess the molecular profile and specific resistance mutations, which further guides therapy.
In recent studies, it was demonstrated that PD-L1 expression might correlate with genetic driver alterations.26 Chen and col-leagues reported that KRAS mutations induce upregulation of PD-L1 expression in lung adenocarcinoma.26 Further, Sumimoto and colleagues showed that KRAS mutation contributes to enhanced PD-L1 expression in human lung cancer.27 Our results show that 70% of KRAS-mutated patients express PD-L1 either in the initial tumor sample or at recurrence. Thus, our data suggests that blockade of the PD-1/PD-L1 pathway might be a promising ther-apeutic option for KRAS-mutated NSCLC. Recent clinical trials demonstrated a benefit for immune checkpoint inhibitors in