Abstract Background: NIVO + RELA is approved for unresectable or metastatic melanoma at 480 mg NIVO + 160 mg RELA Q4W based on results from RELATIVITY-047. RELATIVITY-020 (NCT01968109) was a phase 1/2a, dose escalation and cohort expansion open-label study investigating RELA +/− NIVO in advanced solid tumors. Parts A-C investigated dose escalation and expansion into solid tumor types. Methods: See Table for tumor types and abbreviations. Part A was a dose escalation analysis in patients (pts) with immuno-oncology (IO) treatment-naive solid tumors with RELA monotherapy at 20-800 mg with dose expansion studies in IO RF NSCLC and RCC (A1). Part B was NIVO + RELA doseescalation across tumor types plus cervical, ovarian, and colorectal cancer from Q2W 80 + 20 to 240 + 240 mg or Q4W 480 + 160 to 480 + 1440 mg. Primary endpoint in Parts A and B was safety. Part C was expansion at NIVO + RELA 240 + 80 Q2W (with bladder at 480 + 160 Q4W). Endpoints were safety and BICRassessed ORR, disease control rate, and duration of response. Results: Baseline LAG3 expression ≥ 1% was comparable across cohorts (24-41%). Minimum follow-up in mo was: 87-110 (Part A), 44 (B), and 61-84 (C). NIVO + RELA dose escalation demonstrated an acceptable safety profile, with no maximum tolerated dose (MTD) identified. Grade 3-4 treatment-related adverse events occurred in approximately 16% of pts across all Parts, leading to discontinuation in 4-11% of pts. Efficacy was noted across all tumor types tested in Part C (Table). RELA exposure increased doseproportionally as combination or monotherapy. No pharmacokinetic interaction between NIVO and RELA was identified and there was a low incidence of antidrug and neutralizing antibodies for both RELA and NIVO. Conclusions: NIVO + RELA demonstrated preliminary efficacy across several different tumor types. Clinical activity was observed in both IO naive and IO RF pts, with tolerable toxicity observed. Although no MTD was reached, PK data supported the NIVO 480 + RELA 160 mg dose. Citation Format: Paolo A. Ascierto, Carlo Gomez-Roca, Ignacio Melero, Katriina Jalkanen, Rachel E. Sanborn, Juan Martin-Liberal, James Larkin, Rastilav Bahleda, F. Stephen Hodi, Margaret Callahan, Marta Nykas, Giuseppe Curigliano, Sebastian Bauer, Solange Peters, Takekazu Aoyama, Mark Semaan, Sourav Mukherjee, Sonia Dolfi, Satyendra Suryawanshi, Evan Lipson. Dose escalation and expansion of nivolumab plus relatlimab (NIVO + RELA) in solid tumors in RELATIVITY 020 Parts A-C [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(8_Suppl):Abstract nr LB419.

LBA9500 Background: NIVO + RELA fixed-dose combination (FDC), compared with NIVO alone, demonstrated a clinically meaningful benefit to progression-free survival (HR 0.79, 95% CI 0.66–0.95) and overall survival (OS; 0.80, 95% CI 0.66–0.99) after a 3-y follow-up, with a manageable safety profile in patients (pts) with untreated advanced melanoma (RELATIVITY-047). To address a current unmet need for more efficacious adjuvant regimens for completely resected melanoma, RELATIVITY-098 was designed to evaluate adjuvant NIVO + RELA FDC vs NIVO alone in pts after complete resection of stage III–IV melanoma (NCT05002569). Methods: In this phase 3, randomized, double-blind study, pts aged ≥ 12 years were stratified by AJCC v8 stage at screening (stage IIIA/IIIB vs IIIC vs IIID/IV) and geographic region (USA/Canada/Australia vs Europe vs rest of the world). Pts were randomized 1:1 to receive NIVO 480 mg + RELA 160 mg FDC or NIVO 480 mg every 4 weeks for a maximum of 1 year or until first recurrence, unacceptable toxicity, or withdrawal of consent. The primary endpoint was recurrence-free survival (RFS) by investigator; secondary endpoints included OS (key), distant metastasis-free survival (DMFS), and safety. Results: Pts randomized to NIVO + RELA (n = 547) vs NIVO alone (n = 546) had stage IIIA/B (38% vs 36%) or IIIC (49% vs 50%) disease; 80% vs 83% had cutaneous nonacral melanoma, 11% vs 10% had cutaneous acral, and 2% vs 1% had mucosal. Median duration of therapy was 11.0 mo for each arm. At a minimum follow-up of 23.4 mo, there was no statistical difference in RFS for NIVO + RELA vs NIVO (Table). The RFS outcome was generally consistent across stratification factors and prespecified subgroups. OS was not tested per the hierarchical testing strategy. There were 148 OS events (48% data maturity). DMFS was similar in both treatment groups (Table). Grade 3/4 treatment-related adverse events (TRAEs) occurred in 19% of pts treated with NIVO + RELA vs 8% with NIVO alone (compared with 22% vs 12% in RELATIVITY-047); any-grade TRAEs led to discontinuation of therapy in 17% vs 9% of pts, respectively. There were 2 treatment-related deaths with NIVO + RELA and 1 with NIVO. Conclusions: NIVO + RELA did not result in significant RFS improvement vs NIVO alone as adjuvant treatment for pts after complete resection of stage III–IV melanoma. The safety profile of NIVO + RELA in this setting was generally consistent with results from RELATIVITY-047. A robust biomarker analysis for the study is currently underway. Clinical trial information: NCT05002569 . NIVO + RELA NIVO RFS 24-mo rate, %(95% CI) 62.0(57.7–66.0) 63.6 (59.4–67.6) Median, mo (95% CI)(events/pts) NR (30.8–NR)(214/547) 33.1 (31.0–NR)(213/546) HR 1.01 (95% CI, 0.83–1.22) DMFS 24-mo rate, %(95% CI) 73.1(68.8–76.9) 76.3(72.3–79.9) Median, mo (95% CI)(events/pts) NR (NR–NR)(133/499) 33.1 (31.5–NR)(129/494) HR 1.07 (95% CI 0.84–1.36) HR, hazard ratio; NR, not reached.

Multiplex immunofluorescence (mIF) is a promising tool for immunotherapy biomarker discovery in melanoma and other solid tumors. mIF captures detailed phenotypic information of immune cells in the tumor microenvironment, as well as spatial data that can reveal biologically relevant interactions among cell types. Given the complexity of mIF data, the development of automated analysis pipelines is crucial for advancing biomarker discovery. In pre-treatment melanoma samples from 50 patients treated with immune checkpoint inhibitors (ICIs), a higher stromal B cell percentage is associated with the clinical benefit of ICI therapy. The automatic detection of B cell aggregates with DBSCAN, a novel application of a computer-aided machine learning algorithm, demonstrates the potential for enhanced accuracy compared to pathologist assessment of lymphoid aggregates. TCF1+ and LAG3- T cell subpopulations are enriched near stromal B cells, suggesting potential functional interactions. These analyses provide a roadmap for the further development of spatial immunotherapy biomarkers in melanoma and other diseases.

Based on RELATIVITY-047, nivolumab plus relatlimab is approved for advanced melanoma. Here, to address a current unmet need for more efficacious adjuvant regimens for completely resected melanoma, the phase 3, double-blind RELATIVITY-098 trial compared adjuvant nivolumab plus relatlimab to nivolumab after complete resection of stage III/IV melanoma. Patients were randomized 1:1 to receive nivolumab 480 mg plus relatlimab 160 mg (n = 547) or nivolumab 480 mg (n = 546) intravenously every 4 weeks for ≤1 year; safety populations totaled 543 and 545 patients, respectively. The primary endpoint was recurrence-free survival (RFS), and the key secondary was overall survival; translational endpoints were exploratory. There was no difference in RFS for nivolumab plus relatlimab versus nivolumab (hazard ratio = 1.01; 95% confidence interval: 0.83-1.22; P = 0.928); therefore, overall survival was not tested. Translational data across trials showed lower circulating LAG-3+ T cells in the adjuvant setting (RELATIVITY-098) versus advanced melanoma (RELATIVITY-047), where LAG-3+ T cells were enriched in tumor versus blood. The absence of macroscopic tumor and reduced peripheral LAG-3+ T cells may explain the lack of added benefit of nivolumab plus relatlimab over nivolumab in resected versus metastatic melanoma. ClinicalTrials.gov identifier: NCT05002569 .

IMPORTANCE:While new intrathoracic adenopathy in a patient with cancer can represent progression of disease, the differential diagnosis is broad. Sarcoid-like reactions (SLR) remain an underreported source of lymphadenopathy in patients treated with immune checkpoint inhibitors (ICI), with limited reports in patients with cancers other than melanoma. OBJECTIVE:To characterize SLRs among patients treated with ICI for advanced solid tumors. METHODS:Data were collected on the clinical, pathologic, and radiographic presentation of patients treated with ICI who developed clinical or imaging findings suggestive of an SLR, including the presence of hilar or mediastinal lymphadenopathy, cutaneous/subcutaneous involvement, and/or worsening of existing sarcoidosis on ICI. RESULTS:Twelve patients were identified as having experienced an SLR. While 6 patients had melanoma, SLRs were also observed among patients with lung, gynecologic, and genitourinary cancers, including high-grade serous ovarian carcinoma, and an angiomyolipoma. Median time from initiation of ICI to diagnosis of an SLR was 3.4 months (range: 1.8-9.1). All but one patient (92%) were deemed to have had a radiographic response to ICI. CONCLUSIONS AND RELEVANCE:Clinicians should maintain the awareness of the possibility of SLRs in patients receiving ICI, particularly in patients whose scans show evidence of "mixed" response, with decreases in certain lesions coupled with new/increasing intrathoracic lymphadenopathy and/or other systemic signs of sarcoid.

e16389 Background: Pancreatic cancer is one of the most lethal cancers in the United States (U.S.), with a five-year survival rate of only 12% due to late diagnosis and aggressive progression. End-of-life (EOL) care is often needed for these patients, yet data on EOL care for pancreatic cancer are limited. This study evaluates trends and disparities in EOL care by examining location of death across demographics. Methods: Our population-based retrospective study analyzed death certificate data from the Centers for Disease Control and Prevention’s Wide-Ranging Online Data for Epidemiologic Research (CDC WONDER) from 2003 to 2020 to determine the longitudinal trends of place of death for pancreatic cancer deaths at the age of 35 years and above in the U.S.. We used linear regression to test the proportions of deaths in different locations, and stratified the data by race and ethnicity. Odds ratios were calculated using 2x2 contingency table. Results: From 2003 to 2020, pancreatic cancer accounted for 690,950 deaths in the U.S. Most of these deaths occurred at the deceased's home (48.52%), followed by medical facilities (23.58%), nursing home/long-term care (LTC) (11.49%), hospice facility (10.56%), and other locations (5.85%). Deaths at hospice facilities and homes increased significantly from 0.70% to 11.85% (p < 0.01) and from 46.53% to 58.01% (p < 0.01) respectively. Conversely, nursing home/LTC and medical facilities deaths decreased from 14.32% to 6.84% (p < 0.01) and from 28.81% to 16.93% (p < 0.01) respectively. Hospice facility deaths increased among Whites (0.52% to 9.46%, p < 0.01), African Americans (AA) (0.07% to 1.45%, p < 0.01), Hispanics (0.06% to 0.62%, p < 0.01), and Asians (0.05% to 0.25%, p < 0.01). Home deaths also increased among Whites (39.11% to 44.90%, p < 0.01), AA (4.17% to 5.95%, p < 0.01), Hispanics (2.21% to 4.74%, p < 0.01), and Asians (0.80% to 2.08%, p < 0.01). Meanwhile, medical facility deaths decreased significantly among Whites (21.46% to 11.42%, p < 0.01), AA (4.69% to 3.14%, p < 0.01), and Hispanics (1.66% to 1.45%, p < 0.01). Hispanics, AA, and Asians had lower odds of dying in hospice facility (Odds ratio [OR]: 0.84 [95% CI, 0.82–0.86] vs 0.97 [95% CI, 0.96–0.98] vs and 0.66 [95% CI, 0.64–0.68] and higher odds of dying in medical facilities (OR: 1.35 [95% CI, 1.34-1.36], 1.91[95% CI, 1.90-1.92], and 1.56 [95% CI, 1.54-1.58]) than Whites. Conclusions: From 2003 to 2020, there was a shift toward hospice and home deaths, but racial and ethnic disparities remain. These findings underscore the need to address disparities in EOL care.

Differentiating sequencing errors from true variants is a central genomics challenge, calling for error suppression strategies that balance costs and sensitivity. For example, circulating cell-free DNA (ccfDNA) sequencing for cancer monitoring is limited by sparsity of circulating tumor DNA, abundance of genomic material in samples and preanalytical error rates. Whole-genome sequencing (WGS) can overcome the low abundance of ccfDNA by integrating signals across the mutation landscape, but higher costs limit its wide adoption. Here, we applied deep (~120×) lower-cost WGS (Ultima Genomics) for tumor-informed circulating tumor DNA detection within the part-per-million range. We further leveraged lower-cost sequencing by developing duplex error-corrected WGS of ccfDNA, achieving 7.7 × 10-7 error rates, allowing us to assess disease burden in individuals with melanoma and urothelial cancer without matched tumor sequencing. This error-corrected WGS approach will have broad applicability across genomics, allowing for accurate calling of low-abundance variants at efficient cost and enabling deeper mapping of somatic mosaicism as an emerging central aspect of aging and disease.

Recent progress in multiplexed tissue imaging is deepening our understanding of tumor microenvironments related to treatment response and disease progression. However, analyzing whole-slide images with millions of cells remains computationally challenging, and few methods provide a principled approach for integrative analysis across images. Here, we introduce SpatialTopic, a spatial topic model designed to decode high-level spatial tissue architecture from multiplexed images. By integrating both cell type and spatial information, SpatialTopic identifies recurrent spatial patterns, or "topics," that reflect biologically meaningful tissue structures. We benchmarked SpatialTopic across diverse single-cell spatial transcriptomic and proteomic imaging platforms spanning multiple tissue types. We show that SpatialTopic is highly scalable to large-scale images, along with high precision and interpretability. It consistently identifies biologically and clinically significant spatial topics, such as tertiary lymphoid structures, and tracks spatial changes over disease progression. Its computational efficiency and broad applicability will enhance the analysis of large-scale imaging datasets.

Immune checkpoint blockade (ICB) has revolutionized cancer treatment; however, many patients develop therapeutic resistance. We previously identified and validated a pretreatment peripheral blood biomarker, characterized by a high frequency of LAG-3 + lymphocytes, that predicts resistance in patients receiving anti–PD-1 (aPD-1) ICB. To better understand the mechanism of aPD-1 resistance, we identified murine tumor models with a high LAG-3 + lymphocyte frequency (LAG-3 hi ), which were resistant to aPD-1 therapy, and LAG-3 lo murine tumor models that were aPD-1 sensitive, recapitulating the predictive biomarker we previously described in patients. LAG-3 hi tumor-bearing mice were sensitive to aPD-1 + anti–LAG-3 (aLAG-3) therapy, and this benefit was CD8 + T cell dependent. The efficacy of combination therapy was enhanced in LAG-3 hi (but not LAG-3 lo ) mice with depletion of CD4 + T cells. Furthermore, responses to aPD-1 + aLAG-3 correlated with regulatory T cell (T reg ) phenotypic plasticity in LAG-3 hi mice, suggesting a specific role for T regs in response to aPD-1 + aLAG-3 treatment. Using T reg fate–tracking Foxp3 GFP-Cre-ERT2 × ROSA YFP reporter mice, we demonstrated that expanded populations of unstable T regs correlated with improved response to combination therapy in LAG-3 hi mice. Complementing these preclinical data, an increased proportion of unstable T regs also correlated with higher response rate and improved survival after aPD-1 + aLAG-3 therapy in a cohort of patients with metastatic melanoma ( n = 117). These data indicate that T reg phenotypic plasticity affects aPD-1 + aLAG-3 responsiveness, which may represent a biomarker to aid patient selection and a rational therapeutic target for a subset of PD-1–refractory patients.

Programmed death ligand-1 (PD-L1) is an inducible protein heterogeneously expressed in melanoma. Assessment of PD-L1 expression is challenging and standard immunohistochemistry (IHC) requires biopsies and cannot capture heterogeneity of expression. Noninvasive imaging methods provide evaluation of expression across lesions in the body. We conducted a prospective pilot trial with PD-L1 PET imaging with [ 18 F]-BMS-986229 as a noninvasive approach to assess PD-L1 expression across lesions, in 10 patients with advanced melanoma, longitudinally during treatment with nivolumab and ipilimumab. PET imaging was performed at baseline and at 6 weeks after-initiation of treatment. We examined the relationship of PD-L1 PET uptake to radiographic clinical response. [ 18 F]-BMS-986229 uptake was variably seen across lesions in patients at baseline. All patients showed positive uptake in lesions at baseline PET with a median SUV max of 3.6 (range: 1.7–8.6). PD-L1 PET SUV max decreased in all but two lesions 6 weeks after treatment initiation. Four of five patients had a mean (SUV max ) greater than or equal to 3.00 in Response Evaluation Criteria in Solid Tumors (RECIST) evaluable lesions at baseline, and all had a RECIST response while all progressors ( n = 3) had baseline PD-L1 mean SUV max less than or equal to 2.60. A higher lesional baseline SUV max was associated with greater individual lesion reduction during treatment. The PD-L1 uptake in lesions showed a low correlation with baseline PD-L1 by IHC. In this small pilot study, PD-L1 PET imaging using [ 18 F]-BMS-986229 showed feasibility in noninvasively assessing lesion uptake and PD-L1 heterogeneity in patients receiving combination immunotherapy. Future exploration of this tracer in larger patient cohorts is necessary to delineate its use in managing immunotherapy treatments.

Distinguishing real biological variation in the form of single-nucleotide variants (SNVs) from errors is a major challenge for genome sequencing technologies. This is particularly true in settings where SNVs are at low frequency such as cancer detection through liquid biopsy, or human somatic mosaicism. State-of-the-art molecular denoising approaches for DNA sequencing rely on duplex sequencing, where both strands of a single DNA molecule are sequenced to discern true variants from errors arising from single stranded DNA damage. However, such duplex approaches typically require massive over-sequencing to overcome low capture rates of duplex molecules. To address these challenges, we introduce paired plus-minus sequencing (ppmSeq) technology, in which both DNA strands are partitioned and clonally amplified on sequencing beads through emulsion PCR. In this reaction, both strands of a double-stranded DNA molecule contribute to a single sequencing read, allowing for a duplex yield that scales linearly with sequencing coverage across a wide range of inputs (1.8-98 ng). We benchmarked ppmSeq against current duplex sequencing technologies, demonstrating superior duplex recovery with ppmSeq, with a rate of 44%±5.5% (compared to ~5-11% for leading duplex technologies). Using both genomic as well as cell-free DNA, we established error rates for ppmSeq, which had residual SNV detection error rates as low as 7.98x10-8 for gDNA (using an end-repair protocol with dideoxy nucleotides) and 3.5x10-7±7.5x10-8 for cell-free DNA. To test the capabilities of ppmSeq for error-corrected whole-genome sequencing (WGS) for clinical application, we assessed circulating tumor DNA (ctDNA) detection for disease monitoring in cancer patients. We demonstrated that ppmSeq enables powerful tumor-informed ctDNA detection at concentrations of 10-5 across most cancers, and up to 10-7 in cancers with high mutation burden. We then leveraged genome-wide trinucleotide mutation patterns characteristic of urothelial (APOBEC3-related and platinum exposure-related signatures) and lung (tobacco-exposure-related signatures) cancers to perform tumor-naive ctDNA detection, showing that ppmSeq can identify a disease-specific signal in plasma cell-free DNA without a matched tumor, and that this signal correlates with imaging-based disease metrics. Altogether, ppmSeq provides an error-corrected, cost-efficient and scalable approach for high-fidelity WGS that can be harnessed for challenging clinical applications and emerging frontiers in human somatic genetics where high accuracy is required for mutation identification.

Melanoma has long been a difficult malignancy to treat with low response rates to standard chemotherapies. In recent years, the use of immune checkpoint inhibitors have demonstrated promising results, paving the way for the use of the rapidly developing novel immune targeting therapies. In this review, we look beyond immune checkpoint inhibitor treatments and summarize several emerging treatment strategies for melanoma, including neoantigen vaccines, conventional antibody drug-conjugates, and bispecific T-cell engager therapies.

4569 Background: Although PD-1 blockade is active in varied mUC treatment settings, predicting response and overcoming resistance remain unmet needs. We hypothesized that immunosuppressive M-MDSCs (CD14+Lin-/HLA-DRlow/-) and TCR dynamics in tumors and blood may correlate with outcomes from NIVO in pts with platinum-refractory mUC (NCT02553642). Methods: 69 pts with mUC were treated with NIVO 240mg intravenously every 2 weeks (wks). Pre-treatment peripheral M-MDSCs among lineage-negative CD14+ monocytes (%) were estimated by flow cytometry using an algorithm that calculates MDSC frequencies based on HLA-DR mean fluorescence intensity (Kitano et al. 2014). High throughput DNA sequencing of the CDR3 region of the TCR beta chain was performed on baseline tumors (N = 57) and pre and post-treatment (wk 2) peripheral blood mononuclear cells using Adaptive Immunosequencing (Adaptive Biotechnologies). M-MDSC levels and TCR metrics were correlated with clinical outcomes: complete or partial response (CR, PR) vs. stable or progressive disease (SD, PD) by RECIST 1.1, clinical benefit (CR + PR + SD), and progression-free and overall survival (PFS, OS). Groups were compared using Wilcoxon signed rank (paired), Wilcoxon rank sum tests (unpaired), and log-rank tests (time-to-event). Cox proportional hazards and logistic regression models were used to analyze binary and time-to-event outcomes, respectively. Results: Higher pre-treatment M-MDSCs were associated with worse PFS and OS univariably (PFS HR 1.07; 95% CI, 1.01-1.14; p = 0.025; OS HR 1.07; 95% CI, 1.00-1.14; p = 0.038), and after adjustment for liver disease and PD-L1 at baseline (PFS HR 1.12; 95% CI, 1.04-1.21, p = 0.004; OS HR 1.11; 95% CI, 1.03-1.20, p = 0.010). Baseline M-MDSCs did not significantly differ between responders and non-responders, but were significantly lower in patients with clinical benefit (median 12.8 [IQR: 10.7-15.9] vs. median 15.6 [IQR: 13.5-18.3]; p = 0.039), and remained significantly associated after adjustment for PD-L1 score in a multivariable model (OR 0.82; 95% CI, 0.67-0.97; p = 0.037). A higher number of tumor-associated clones in the blood (BTACs) at baseline was associated with response (p = 0.049). Overall, BTACs significantly increased at wk 2 (p < 0.001), and wk 2 values were associated with response (p = 0.003). PFS also significantly differed between BTAC high (high: > median 2,012 clones) and low groups at wk 2, with improved PFS in the high group (log-rank p = 0.026). OS rate for the BTAC high group was 34% vs. 16% for the low group at wk 2, but this did not achieve significance (p = 0.098). Conclusions: Peripheral M-MDSCs may promote resistance to PD-1 blockade in pts with mUC. NIVO stimulated tumor-specific TCR clones in the blood, which correlated with improved response and outcomes. Clinical trial information: NCT02553642 .

IntroductionAlthough checkpoint inhibitors (CPIs) have improved outcomes for patients with metastatic melanoma, those progressing on CPIs have limited therapeutic options. To address this unmet need and overcome CPI resistance mechanisms, novel immunotherapies, such as T-cell engaging agents, are being developed. The use of these agents has sometimes been limited by the immune response mounted against them in the form of anti-drug antibodies (ADAs), which is challenging to predict preclinically and can lead to neutralization of the drug and loss of efficacy.MethodsTYRP1-TCB (RO7293583; RG6232) is a T-cell engaging bispecific (TCB) antibody that targets tyrosinase-related protein 1 (TYRP1), which is expressed in many melanomas, thereby directing T cells to kill TYRP1-expressing tumor cells. Preclinical studies show TYRP1-TCB to have potent anti-tumor activity. This first-in-human (FIH) phase 1 dose-escalation study characterized the safety, tolerability, maximum tolerated dose/optimal biological dose, and pharmacokinetics (PK) of TYRP1-TCB in patients with metastatic melanoma (NCT04551352).ResultsTwenty participants with cutaneous, uveal, or mucosal TYRP1-positive melanoma received TYRP1-TCB in escalating doses (0.045 to 0.4 mg). All participants experienced ≥1 treatment-related adverse event (TRAE); two participants experienced grade 3 TRAEs. The most common toxicities were grade 1–2 cytokine release syndrome (CRS) and rash. Fractionated dosing mitigated CRS and was associated with lower levels of interleukin-6 and tumor necrosis factor-alpha. Measurement of active drug (dual TYPR1- and CD3-binding) PK rapidly identified loss of active drug exposure in all participants treated with 0.4 mg in a flat dosing schedule for ≥3 cycles. Loss of exposure was associated with development of ADAs towards both the TYRP1 and CD3 domains. A total drug PK assay, measuring free and ADA-bound forms, demonstrated that TYRP1-TCB-ADA immune complexes were present in participant samples, but showed no drug activity in vitro.DiscussionThis study provides important insights into how the use of active drug PK assays, coupled with mechanistic follow-up, can inform and enable ongoing benefit/risk assessment for individuals participating in FIH dose-escalation trials. Translational studies that lead to a better understanding of the underlying biology of cognate T- and B-cell interactions, ultimately resulting in ADA development to novel biotherapeutics, are needed.

Quantitative assessment of multiplex immunofluorescence (mIF) data represents a powerful tool for immunotherapy biomarker discovery in melanoma and other solid tumors. In addition to providing detailed phenotypic information of immune cells of the tumor microenvironment, these datasets contain spatial information that can reveal biologically relevant interactions among cell types. To assess quantitative mIF analysis as a platform for biomarker discovery, we used a 12-plex mIF panel to characterize tumor samples collected from 50 patients with melanoma prior to treatment with immune checkpoint inhibitors (ICI). Consistent with prior studies, we identified a strong association between stromal B cell percentage and response to ICI therapy. We then compared pathologist assessment of lymphoid aggregates with a density based clustering algorithm, DBSCAN, to both automatically detect B cell aggregates and quantify their size, morphology, and distance to tumor. Spatial neighborhood analysis identified TCF1+ and LAG3- T cell subpopulations enriched near stromal B cells. These analyses provide a roadmap for the further development and validation of spatial immunotherapy biomarkers in melanoma and other diseases. ### Competing Interest Statement J.W.S. Research funding: IO Biotech (Inst), Regeneron (Inst), Daiichi Sankyo (Inst); Consulting or advisory role: IO Biotech; X.P. No disclosures; F.D.E. No disclosures; A.P.M. No disclosures; M.Y. No disclosures; C.M. No disclosures; N.A. No disclosures; R.V. No disclosures M.Z. No disclosures; J.L. No disclosures; M.B. No disclosures. Y.L.: Employment: Bristol Myers Squibb; M.A.P. Consulting or Advisory Role: Bristol Myers Squibb; Cancer Expert Now; Chugai Pharma; Eisai; Erasca, Inc; Intellisphere; Merck; MJH Associates; Nektar; Novartis; Pfizer; WebMD; Research Funding: Array BioPharma (Inst); Bristol Myers Squibb (Inst); Infinity Pharmaceuticals (Inst); Merck (Inst); Novartis (Inst); Rgenix (Inst); K.S.P. Stock ownership in 23and Me, Vincerx, Eyepoint, & Kyverna T.J.H. Employment: Bristol Myers Squibb; M.K.C. Bristol Myers Squibb: Research support (Inst), advisory role/consulting; Medimmune: advisory role/consulting; Immunocore: advisory role/consulting; Merus: family member employee; R.S. No disclosures ### Funding Statement This work was supported in part by the Memorial Sloan Kettering Cancer Center (MSK) NCI Core Grant P30 CA008748 and grant support from the V Foundation (T2021-007, M.K.C and R.S.). ### Author Declarations I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained. Yes The details of the IRB/oversight body that provided approval or exemption for the research described are given below: Memorial Sloan Kettering Cancer Center Institutional Review Board gave ethical approval for this work I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals. Yes I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance). Yes I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable. Yes All code describing cell phenotyping pipeline, application of DBSCAN for quantification of B cell aggregates, and B cell neighborhood analysis are available on GitHub. Raw mIF data files and de-identified clinical data are available by request to the corresponding author.

Leptomeningeal disease (LMD) is a devastating complication of melanoma with a dismal prognosis. We present the case of a young man with stage IV BRAF V600E mutant melanoma with lung, lymph node, and brain metastases initially treated with ipilimumab and nivolumab, who subsequently developed LMD. Upon change to BRAF/MEK targeted therapy with nivolumab, a durable complete response was achieved and remains ongoing, off treatment, 7 years from diagnosis. Management of symptomatic LMD remains a critical unmet clinical challenge, with limited clinical trial data. This exceptional case is instructive, as the first published case of the use of the triplet, and the first durable response with therapy discontinuation, in melanoma LMD. The triple-drug regimen may be considered a viable option in fit patients. This case highlights the potential for long-term disease control and the critical and urgent need to develop clinical trials inclusive of patients with LMD to define the best treatment strategies. We present the case of a young man with stage IV BRAF V600E mutant melanoma with lung, lymph node, and brain metastases initially treated with ipilimumab and nivolumab, who subsequently developed leptomeningeal disease (LMD). Upon change to dabrafenib and trametinib with nivolumab, a durable complete response was achieved and remains ongoing, off-treatment, 7 years from diagnosis. Management of symptomatic LMD remains a critical unmet clinical challenge, with limited clinical trial data. This exceptional case is instructive, as the first published case of the use of the triplet, and the first durable response with therapy discontinuation, in melanoma LMD (Lochrin et al. (2024). Pigment Cell & Melanoma Research. ).image

The multiplexed immunofluorescence (mIF) platform enables biomarker discovery through the simultaneous detection of multiple markers on a single tissue slide, offering detailed insights into intratumor heterogeneity and the tumor-immune microenvironment at spatially resolved single cell resolution. However, current mIF image analyses are labor-intensive, requiring specialized pathology expertise which limits their scalability and clinical application. To address this challenge, we developed CellGate, a deep-learning (DL) computational pipeline that provides streamlined, end-to-end whole-slide mIF image analysis including nuclei detection, cell segmentation, cell classification, and combined immuno-phenotyping across stacked images. The model was trained on over 750,000 single cell images from 34 melanomas in a retrospective cohort of patients using whole tissue sections stained for CD3, CD8, CD68, CK-SOX10, PD-1, PD-L1, and FOXP3 with manual gating and extensive pathology review. When tested on new whole mIF slides, the model demonstrated high precision-recall AUC. Further validation on whole-slide mIF images of 9 primary melanomas from an independent cohort confirmed that CellGate can reproduce expert pathology analysis with high accuracy. We show that spatial immuno-phenotyping results using CellGate provide deep insights into the immune cell topography and differences in T cell functional states and interactions with tumor cells in patients with distinct histopathology and clinical characteristics. This pipeline offers a fully automated and parallelizable computing process with substantially improved consistency for cell type classification across images, potentially enabling high throughput whole-slide mIF tissue image analysis for large-scale clinical and research applications.

In solid tumor oncology, circulating tumor DNA (ctDNA) is poised to transform care through accurate assessment of minimal residual disease (MRD) and therapeutic response monitoring. To overcome the sparsity of ctDNA fragments in low tumor fraction (TF) settings and increase MRD sensitivity, we previously leveraged genome-wide mutational integration through plasma whole-genome sequencing (WGS). Here we now introduce MRD-EDGE, a machine-learning-guided WGS ctDNA single-nucleotide variant (SNV) and copy-number variant (CNV) detection platform designed to increase signal enrichment. MRD-EDGESNV uses deep learning and a ctDNA-specific feature space to increase SNV signal-to-noise enrichment in WGS by ~300× compared to previous WGS error suppression. MRD-EDGECNV also reduces the degree of aneuploidy needed for ultrasensitive CNV detection through WGS from 1 Gb to 200 Mb, vastly expanding its applicability within solid tumors. We harness the improved performance to identify MRD following surgery in multiple cancer types, track changes in TF in response to neoadjuvant immunotherapy in lung cancer and demonstrate ctDNA shedding in precancerous colorectal adenomas. Finally, the radical signal-to-noise enrichment in MRD-EDGESNV enables plasma-only (non-tumor-informed) disease monitoring in advanced melanoma and lung cancer, yielding clinically informative TF monitoring for patients on immune-checkpoint inhibition. Detection of circulating tumor DNA using MRD-EDGE, a machine-learning-guided single-nucleotide variant and copy-number variant detection platform for signal enrichment, enables monitoring of minimal residual disease and immunotherapy response in settings of low tumor burden.