Molecular profiling of metastatic lung squamous cell carcinoma (mLUSC) to identify patients with differential response to immune checkpoint inhibitor (ICI) therapy.
Bibliographic record
Abstract
e20555 Background: Molecular profiling is inconsistently performed in patients with mLUSC as it rarely provides actionable information. Whether it can aid in the selection of ICI based therapy is unknown. Methods: Patients with mLUSC (n=963) who received ICI therapy at 9 academic institutions or as part of 3 clinical trials were identified. Overall response rate (ORR) by RECIST 1.1 and clinical progression-free survival (PFS) were the primary outcomes. Propensity matching on age, smoking history, ECOG, sex and PDL1 among previously untreated patients was performed to compare outcomes in patients treated with ICIs alone (ICI-mono) or in combination with chemotherapy (ICI-chemo). Among patients with complete clinico-genomic annotation, including coverage of KEAP1 (n=248), we trained logistic regression (LR), random forest (RF), and gradient boosting (GB) models to predict response. An 80/20 split was used to separate the training and testing cohorts and hyperparameters were optimized for area under the receiver operating characteristic curve (AUROC). Coefficients of the LR model were interrogated to identify predictors of response. The AACR GENIE LUSC cohort (n=2618) was analyzed to assess the prognostic effects of genomic biomarkers. Results: Among propensity matched patients, there was no difference in ORR (ORR: 29 vs 29, p=1.0) or PFS (HR=0.90, p=0.72) between patients who received ICI-chemo versus ICI-mono. Integrating clinical and molecular data improved on PDL1 in predicting response to ICI therapy (AUROC, LR: 0.73, RF: 0.78, GB: 0.69, LR with PDL1 only: 0.70). Top features predictive of response were ICI-chemo, PDL1, sex, and KDM6A , NFE2L2 / KEAP1 , and TP53 alterations; top features predictive of non-response were prior treatment, KRAS and DNMT3A alterations. Among previously untreated patients, improved outcomes were observed in those with vs without KDM6A (n: 13 vs 185; ORR: 85 vs 42, p=0.0032; PFS: HR=0.50, p=0.064) and NFE2L2 / KEAP1 alterations (n: 45 vs 145; ORR: 60 vs 39, p=0.017; PFS: HR=0.60, p=0.0084), while those with vs without KRAS alterations trended towards decreased response rates and PFS (n: 31 vs 274; ORR: 26 vs 43, p=0.083; PFS: HR=1.40, p=0.088). There was no difference in outcomes among patients with KRAS alterations who received ICI-mono vs ICI-chemo. In the GENIE cohort, 4.5% of patients harbored KRAS mutations, which was associated with decreased overall survival (HR=1.5, p=0.0015) as compared to wild-type KRAS . Conclusions: Molecular profiling identified predictors of response to ICI therapy in patients with mLUSC. KDM6A and KEAP1 / NFE2L2 alterations were predictive of response to ICI therapy and may identify patients who could be spared chemotherapy, while KRAS alterations were predictive of non-response and poor prognosis and may identify patients who would benefit from novel treatment strategies.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".