Abstract Background: Adjuvant immune checkpoint inhibitors (ICIs) improve recurrence-free survival in high-risk melanoma. Despite this, a subset of patients do not respond to ICIs while other experience immune-related adverse events (irAEs) without therapeutic benefits. Predictive biomarkers remain undefined, and grouping patients without accounting for irAE status may obscure meaningful immune correlates. We hypothesized that comprehensive immune profiling of baseline peripheral blood mononuclear cells (PBMCs) may help identify predictive immune signatures associated with both response and toxicity to adjuvant ICI. Methods: Blood samples were collected from a total of 22 melanoma patients undergoing adjuvant ICI therapies at baseline and on-treatment. PBMCs were isolated using standard density gradient centrifugation and stained using two comprehensive immunophenotyping panels. Data was acquired on FACS symphony A5. FlowJo’s built-in UMAP plugin was used to visualize immune cell clustering and further characterized using Marker Enrichment Modeling. Response to therapy was determined retrospectively. Immune subsets were compared between responders and non-responders at the two time points and was further stratified by occurrence of irAEs. Results: At baseline, responders demonstrated enrichment of central memory T cells and terminally differentiated effector memory T cells, consistent with a primed yet regulated immune state poised for rapid activation. In contrast, non-responders had elevated CD16+ MDSCs, which further expanded on treatment, indicating a suppressive peripheral milieu. Among responders, those developing irAEs demonstrated increased memory-like T cells and CD14+CD62P+ myeloid populations compared to those with no irAEs. Among non-responders, irAE-positive patients displayed higher CD45RA+CD4+CD62L+ memory and CD8+CD57+ effector T cell populations. Within the myeloid compartment, CD62P+ cells were increased in both responders and non-responders with irAEs highlighting platelet-myeloid activation. Non-responders without irAEs showed globally low activation, consistent with an immune quiescent state. No significant treatment induced changes distinguished response categories, underscoring the importance of baseline immune context. Conclusions: Baseline peripheral immune composition reflects distinct states of immune readiness in melanoma patients receiving adjuvant ICI. Responders display a primed immune profile whereas non-responders exhibit myeloid driven suppression. These findings highlight peripheral immune features that may predict therapeutic efficacy and toxicity in melanoma patients in the adjuvant ICI setting. Importantly, stratifying patients by irAE status is essential to reveal true immune correlates of efficacy and toxicity, as pooled analyses risk conflating biologically distinct activation and suppression states. Citation Format: Iluja Gautam, Patricia Rayman, Nickolas Stabellini, Jennifer Powers, Adam Moen, Brian Race, Paul Pavicic, Mona Patel, Alberto J. Montero, C. Marcela Diaz-Montero. Baseline peripheral immune signatures associate with toxicity to immunotherapy in melanoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 6462.

Background: The KEYNOTE (KN)-522 trial established pembrolizumab in combination with chemotherapy as the standard of care for the treatment of eTNBC (clinical stage II-III). This trial demonstrated that the addition of pembrolizumab to standard anthracycline-platinum-taxane neoadjuvant chemotherapy was associated with a clinically and statistically significant improvement in pathological complete response (pCR) rates and event-free survival compared to chemotherapy alone. Pembrolizumab is a programmed cell death protein 1 monoclonal antibody, known to cause irAEs in a significant subset of patients— in the KN-522 trial the rate of any grade and >3 grade irAEs were approximately 34% and 13%, respectively. The addition of pembrolizumab does add significant toxicity to standard chemotherapy, which can lead to early discontinuation of curative intent treatment. Therefore, identification of early predictors of irAEs and ways to mitigate them is very important. We hypothesized that an increase in circulating MDSC levels is an early marker of irAEs associated and that MDSCs differ phenotypically and functionally from tumor driven MDSCs. Accumulation of MDSCs has been previously reported to correlate with stage and inversely correlate with response to systemic therapy. Methods: PBMCs were isolated at baseline, Cycle 1 Day 1, Cycle 1 Day 8, Day 1 Cycles 2-4, and after surgery from n=40 eTNBC patients treated with the KN-522 chemotherapy regimen. Phenotypic characterization of MDSCs was performed by high parameter flow cytometry. Data was concatenated and visualized in UMAPs using FlowJo. Kinetics of MDSC accumulation during treatment was assessed and associated with irAEs defined as: adverse events that associated with exposure to an immune checkpoint inhibitor (ICI) and consistent with an immune phenomenon. Phenotypic characteristics of MDSCs from metastatic TNBC patients at baseline and MDSCs from non-metastatic TNBC at the time of irAEs were compared. Results: Twelve out of 40 patients (30%) on study had one of the following documented iRAEs: hypothyroidism (n=4), hepatitis (n=3), adrenal insufficiency (n=2), colitis (n=1), myocarditis (n=1), and hemophagocytic lymphohistiocytosis (n=1). Levels of MDSCs increased during treatment among non-metastatic patients undergoing ICI, and this surge was more prominent among patients experiencing irAEs. In eTNBC patients with irAE associated MDSCs, we observed marked differences in the PMN-MDSC phenotype when compared to MDSCs from metastatic TNBC patients (tumor derived MDSCs). One key finding was significantly higher CX3CR1 expression, the receptor for fractalkine, among irAE associated MDSCs vs. tumor derived MDSCs. Interestingly, accumulation of CX3CR1+ PMN-MDSCs was also found to be the prominent MDSC subtype in the patient who developed severe grade 4 hemophagocytic lymphohistiocytosis. Conclusions: Our results highlight a potential contribution of a unique subset of PMN-MDSCs to the pathogenesis of irAEs that could involve signaling by fractalkine and provide a potential target of therapeutic intervention. Citation Format: Alberto J. Montero, Naji Mallat, Patricia Rayman, Paul Pavicic, Jennifer Powers, Sherwin DeSouza, Corey Speers, Megan Kruse, Zahraa AlHilli, C. Marcela Diaz-Montero. Unique kinetics of myeloid derived suppressor cell (MDSC) accumulation associated with immune related adverse events (irAEs) in patients with early stage triple negative breast cancer (eTNBC) receiving neoadjuvant chemo-immunotherapy [abstract]. In: Proceedings of the San Antonio Breast Cancer Symposium 2024; 2024 Dec 10-13; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2025;31(12 Suppl):Abstract nr P2-11-19.

Eligibility criteria amongst patient cohort.

Pre-treatment core needle biopsy sites in patients with evaluated tumors

Hallmark and KEGG full gene set enrichment data in TCGA.

MSS and PMN-MDSC differences by HRD and Non-HRD status.

gPV categories broken up by HRD, Non-HRD, and MMR.

MSS and PMN-MDSC differences by known and putative melanoma variants.

AbstractPurpose: Malignant melanoma represents the most lethal skin cancer with germline predispositions thought to comprise 10% to 15% of all melanoma cases. No studies to date examine the immunologic features that may differentiate survival differences between germline pathogenic variant (gPV)–positive patients with melanoma from gPV-negative patients with melanoma. Experimental Design: Adult patients with melanoma and clinical characteristics suggesting hereditary predisposition to cancer were prospectively recruited to undergo germline testing and flow cytometric analysis of peripheral immune suppressor cells. Results: In this cohort, gPV-positive patients (n = 72) had a significantly improved melanoma-specific survival (MSS) compared with gPV-negative patients (n = 411; HRadj, 0.32; 95% CI, 0.13–0.82; P = 0.01). These survival improvements among gPV-positive patients were most apparent among cutaneous melanoma subtypes (HRadj, 0.12; 95% CI, 0.016–0.86; P = 0.03) and numerically improved in later-stage (IIB–IV) patients (HRadj, 0.34; 95% CI, 0.10–1.11; P = 0.06). Further, gPV-positive patients had a significantly lower level of total circulating PMN-MDSC compared with gPV-negative patients (P = 0.01), which was most apparent in those diagnosed with later stages (IIB–IV) of melanoma (P = 0.009). Finally, a significant upregulation of inflammatory transcriptome signatures in later-stage gPV-positive patients (n = 21) was observed in comparison with gPV-negative patients (n = 173) in the cutaneous melanoma cohort (SKCM) of The Cancer Genome Atlas (TCGA). Conclusions: gPV-positive patients with melanoma exhibit improved MSS in addition to reduced peripheral PMN-MDSC and an enhanced inflammatory microenvironment.

Heat map of Reactome Pathway Enrichment (Top 25, Bottom 25, and selected pathways) by pairwise comparison of radiographic response category (CR, PR, SD, PD)

Selected Reactome Pathways for Figure 2A

Genes associated with response to CNPs

Cox proportional hazards analysis of MSS.

Germline variants in TCGA vs manuscript cohort.

Differential IGHG1 expression between the IGHG1high (highest tertile) versus IGHG1low (lower tertiles) groups

Representativeness of study participants.

Representativeness of Study Participants

AbstractPurpose: A single arm, phase II trial of carboplatin, nab-paclitaxel, and pembrolizumab (CNP) in metastatic triple-negative breast cancer (mTNBC) was designed to evaluate overall response rate (ORR), progression-free survival (PFS), duration of response (DOR), safety/tolerability, overall survival (OS), and identify pathologic and transcriptomic correlates of response to therapy. Patients and Methods: Patients with ≤2 prior therapies for metastatic disease were treated with CNP regardless of tumor programmed cell death-ligand 1 status. Core tissue biopsies were obtained prior to treatment initiation. ORR was assessed using a binomial distribution. Survival was analyzed via the Kaplan–Meier method. Bulk RNA sequencing was employed for correlative studies. Results: Thirty patients were enrolled. The ORR was 48.0%: 2 (7%) complete responses (CR), 11 (41%) partial responses (PR), and 8 (30%) stable disease (SD). The median DOR for patients with CR or PR was 6.4 months [95% confidence interval (CI), 4–8.5 months]. For patients with CR, DOR was >24 months. Overall median PFS and OS were 5.8 (95% CI, 4.7–8.5 months) and 13.4 months (8.9–17.3 months), respectively. We identified unique transcriptomic landscapes associated with each RECIST category of radiographic treatment response. In CR and durable PR, IGHG1 expression was enriched. IGHG1high tumors were associated with improved OS (P = 0.045) and were concurrently enriched with B cells and follicular helper T cells, indicating IGHG1 as a promising marker for lymphocytic infiltration and robust response to chemo-immunotherapy. Conclusions: Pretreatment tissue sampling in mTNBC treated with CNP reveals transcriptomic signatures that may predict radiographic responses to chemo-immunotherapy.