Abstract 32P
Background
Immune checkpoint inhibitors like pembrolizumab have changed the treatment paradigm of advanced non-small cell lung cancer (aNSCLC), especially for tumors with high PD-L1 expression. However, responses to pembrolizumab vary, highlighting the need for effective and easy-to-implement biomarkers. Although numerous scores assessing systemic inflammation have been published, most of them lack external validation, impeding their clinical implementation. This study aims to validate the prognostic utility of the Scottish Inflammatory Prognostic Score (SIPS) in an independent cohort.
Methods
We included 257 aNSCLC patients with PD-L1 expression ≥50% who were treated with first-line pembrolizumab at The Christie between June 2017 and May 2023. Data on demographics, histology, PD-L1 status, and laboratory metrics were collected. The SIPS was determined by allocating 1 point each for a baseline albumin <35 g/l and a neutrophil count >7.5 × 10ˆ9/l, resulting in a three-tier classification: 0 for good prognosis, 1 for intermediate prognosis, and 2 for poor prognosis.
Results
The cohort had a median age of 70, with 48% being female. Baseline characteristics stratified by SIPS are highlighted in the table. SIPS effectively stratified PFS: 14.0 months (m) for SIPS 0, 7.9 m for SIPS 1, and a mere 2.5 m for SIPS 2 (p < 0.001). Similarly, SIPS stratified OS: 23.4 m for SIPS 0, 16.4 m for SIPS 1, and 3.7 m for SIPS 2 (p = 0.010). SIPS was predictive of both PFS and OS regardless of PD-L1 expression. The combination of PD-L1 expression ≥90% and SIPS 0 identified a subgroup comprising 23% (n=60) of all patients, with the most favorable PFS (18.9 m) and OS (28.2 m). Table: 32P
Baseline characteristics stratified by SIPS
| Characteristic | Total | SIPS 0 (n=144) | SIPS 1 (n=93) | SIPS 2 (n=20) | p-value |
| Age | 0.021 | ||||
| <65 | 85 (33%) | 41 (9%) | 39 (42%) | 5 (25%) | |
| 65-74 | 104 (40%) | 55 (38%) | 37 (40%) | 12 (60%) | |
| ≥75 | 68 (26%) | 48 (33%) | 17 (18%) | 3 (15%) | |
| ECOG PS | 0.065 | ||||
| 0 | 36 (14%) | 26 (18%) | 10 (11%) | 0 (0%) | |
| 1 | 191 (74%) | 103 (72%) | 73 (79%) | 15 (75%) | |
| 2 | 30 (12%) | 15 (10%) | 10 (11%) | 5 (25%) | |
| Histology | 0.147 | ||||
| Squamous | 76 (30%) | 44 (31%) | 20 (22%) | 12 (60%) | |
| Non-squamous | 181 (70%) | 100 (69%) | 73 (79%) | 8 (40%) | |
| PD-L1 | 0.951 | ||||
| <90% | 148 (58%) | 84 (58%) | 53 (57%) | 11 (55%) | |
| ≥90% | 109 (42%) | 60 (42%) | 40 (43%) | 9 (45%) | |
| Brain metastases | 0.036 | ||||
| Absent | 225 (88%) | 131 (91%) | 75 (81%) | 19 (95%) | |
| Present | 32 (12%) | 13 (9%) | 18 (19%) | 1 (5%) |
Conclusions
SIPS is a valuable prognostic tool for aNSCLC patients treated with pembrolizumab in the context of high PD-L1 expression. It offers a simple and clinically implementable method to stratify patients, potentially guiding treatment decisions in aNSCLC.
Clinical trial identification
Editorial acknowledgement
Legal entity responsible for the study
The authors.
Funding
Has not received any funding.
Disclosure
All authors have declared no conflicts of interest.