Abstract Background Radiotherapy (RT) is a cornerstone of the pediatric cancer care paradigm. However, not all patients are responsive to radiotherapy, with many displaying normal tissue toxicities and irradiation-associated late effects. Despite major advances in genomics and precision medicine, RT dosing continues to be prescribed in a uniform manner irrespective of patient-specific tumor biology. Numerous recent works across cancers have demonstrated the utility of the genomic-adjusted radiation dose (GARD) as a model for determining the expected biological effect of radiation to determine the optimal RT for patients in a personalized manner.[1][2][3] Here, we characterized if GARD could predict RT response and inform biologically personalized treatment strategies in pediatric brain tumors. Methods We analyzed 207 patient samples from the Children’s Brain Tumor Network (CBTN) that received radiotherapy for primary high-grade glioma, ependymoma, or medulloblastoma. GARD was calculated for each patient by integrating tumor-specific radiosensitivity into the classical linear-quadratic model for radiation dosing. Cox proportional hazards models were used to evaluate GARD as a continuous variable, with overall survival (OS) and local recurrence (LC) provided as clinical endpoints. Results We determined broad variability in the expected biological effect of radiotherapy (GARD) in patients regardless of physical dose or tumor type (Range: 0 - 74). GARD was significantly associated with overall survival (HR = 0.96, p = 0.01) and local recurrence (HR = 0.95, p < 0.001), in pooled analyses of patients that received radiotherapy. These results suggest a robust 5% benefit per unit increase in GARD , supporting its efficacy as a predictive metric for patient-specific radiation benefit. Conclusion GARD-based dosing paradigms predict the biological effect of radiotherapy and demonstrate improvements in time to first recurrence and overall survival in pediatric brain tumor patients. This suggests that incorporating GARD can effectively inform decision-making and dosing of radiotherapy in pediatric brain tumors. 1. Scott JG, Berglund A, Schell MJ, et al. Pan-cancer prediction of radiotherapy benefit using genomic-adjusted radiation dose (GARD): a cohort-based pooled analysis. Lancet Oncol. 2021;22(8):1221–1229. doi:10.1016/S1470-2045(21)00347-8. 2. Ho E, De Cecco L, Eschrich SA, et al. Personalized treatment in HPV+ oropharynx cancer using genomic adjusted radiation dose. J Clin Invest. 2025;135(19):e194073. doi:10.1172/JCI194073. 3. Scott JG, Sharifi H, Osborne EM, et al. Personalizing radiotherapy prescription dose using genomic markers of radiosensitivity and normal tissue toxicity in non-small cell lung cancer. J Thorac Oncol. 2021;16(7):1130-1141. doi:10.1016/j.jtho.2020.10.046.

Data modeling in biomedical research often operates in the small-sample regime, where the number of observations is small relative to the data dimensionality; the detrimental effects of limited sample sizes are well documented in cancer studies. Synthetic data offers a potential solution to data shortfalls provided that the data generated is an adequate facsimile of the underlying distribution; the adequacy of such synthetic data remains an open-ended problem. In this work, we evaluate a synthetic generator proposed previously. The generator applies a series of transformations to the observed data to accommodate the small-sample size resulting in an uncoupled representation, where uncorrelated marginal distributions are modeled with optimized univariate kernel density estimation. In this report, (1) we develop a nonparametric method for assessing multivariate similarity based on the Cramér-Wold theorem and random projection testing, (2) investigate when the absence of bivariate correlation approximates independence in a non-normal setting, and (3) evaluate artifacts induced by data compression. The presentation is primarily methodological; low-dimensional data were used so each stage of the generation process could be analyzed explicitly. A formal testing framework was developed by comparing random projection level outcomes with a two-sample test, modeling these outcomes as Bernoulli trials, aggregating replicate outcomes within each projection direction, and pooling outcomes across many directions, yielding a scalable standardized normal test-statistic. The key innovation was decoupling the two-sample test significance level from that governing finalized normal inference. We showed the same projection framework also evaluates the full multivariate covariance structure. The generator produced high-fidelity multivariate synthetic data when the bivariate correlation approximates independence in the non-normal setting; in highly compressed data, residual modes were best modeled as normally distributed regardless of their intrinsic distributional form. Ongoing work includes applying these methods to higher-dimensional, diverse data.

Longitudinal data analysis of the patient’s treatment course is critical to uncovering variables that influence outcomes. However, existing tools have significant limitations in integrating multilayered time-series data, particularly in linking treatment events with survival outcomes. Here, we developed ShinyEvents, a web-based framework for complex longitudinal data analysis. ShinyEvents allows users to upload data and generate interactive timelines of clinical events, enabling cohort-level analyses such as treatment clustering and endpoint assignment. It also provides informative cohort visualizations, such as a Sankey diagram of the treatment line and a Swimmer diagram of the clinical course. Finally, our tool can infer real-world progression-free survival (rwPFS) based on user-defined endpoints and perform Kaplan-Meier and Cox proportional hazards regression analysis. With these features, the tool can then associate treatment lines with clinical outcomes. As a case study, we analyzed Moffitt patients with muscle-invasive bladder cancer treated with neoadjuvant chemotherapy followed by surgery. Patients treated with cisplatin and gemcitabine exhibited more favorable rwPFS and overall survival, which is consistent with prior reports. Altogether, ShinyEvents provides a unified framework for integrating longitudinal real-world data with survival analytics, fostering transparent and reproducible collaboration between clinicians and data scientists. A live demo is available at https://shawlab-moffitt.shinyapps.io/shinyevents/.

Understanding how genetic and phenotypic diversity emerges and evolves within cancer cell populations is a fundamental challenge in cancer biology. CLONEID is a novel framework designed to organize and analyze clone-specific measures as structured time-series data. By integrating and monitoring genotypic and phenotypic experimental data over time, CLONEID facilitates hypothesis-driven and hypothesis-generating research in cancer biology. This article outlines the development, utility, and applications of CLONEID, emphasizing its role in overcoming challenges in data reproducibility, mathematical modeling, and multi-modal data integration. A webportal to the CLONEID database is available at dev.cloneid.org.

CHOICES Decisions Aid (DA) is an evidence-based, interactive, educational decision-making support tool (available in English and Spanish) for patients regarding participation in cancer clinical trials (CCTs). Described here are preliminary data on the first six months following implementation of the CHOICES DA as part of the ACT WONDER2S multi-level intervention aimed at increasing referral and enrollment of diverse patients to CCTs. CHOICES DA has educational content including facts about CCTs, values clarification exercises on CCT participation, and patient narratives. The site was made accessible via the patient portal to new Moffitt patients residing in seven geographically defined study intervention zones within Moffitt’s catchment area. Eligible patients also received an email with details and a link to the CHOICES DA tool. Patients were offered the option to complete a survey when exiting the site, which included questions about subjective CCT knowledge (range 1-7), preparedness to talk with provider about CCT (range 1-7), CCT decision preparedness (range 1-7), and usefulness and satisfaction with the tool (range 1-10). The mean scores of the survey questions were re-scaled to 100% for comparison. CHOICES DA usage metrics were continuously monitored, including return visits and time spent viewing the site. In the first six months following intervention launch (September 17, 2024 – March 16, 2025), 279 of 1,568 eligible patients (17.8%) visited CHOICES DA. Out of the 279 eligible patients who viewed the site, race demographics were: White 57% (n=159); Black/African American 7.2% (n=20); Other race 6.2% (n=17); Not reported 30% (n=83). Regarding ethnicity, 17% (n=47) were Hispanic/Latino, 51% (n=143) were non-Hispanic/Latino, and 43% (n=89) were unknown ethnicity. The number of users who returned to the site at least once was 23 of 279 (8%), and the number of patients who viewed the Spanish version of the site was 34 of 279 (12%). The mean website viewing time was 5 minutes (range 1-77 minutes), with a mean of 10 page views per patient (range 1-50). A total of 17 (6%) patients completed the survey. Mean self-rated CCT knowledge was 78.6%, readiness to decide about CCTs was 80%, and readiness to discuss CCTs with their doctor was 91.4%. Mean website usefulness and satisfaction were 87% and 86%, respectively. Nearly 1 in 5 eligible patients engaged with CHOICES DA. Early findings suggest CHOICES DA is a relatively low resource approach to reach cancer patients with information about clinical trials. Patients who completed the exit survey gave high ratings to their CCT knowledge and their readiness to discuss and make decisions about participation in a CCT. They also found the site useful. Usage data will be continuously monitored to explore ways to increase adoption. CHOICES DA will later be evaluated at the two-year post-implementation time point, and future analyses will also examine the overall impact on CCTs enrollment and retention, and intervention-specific effects. Elliott Tapia-Kwan, Yayi Zhao, Rossybelle Amorrortu, Lindsay Fuzzell, Melany Garcia, Margaret M. Byrne, Steven Eschrich, Guillermo Gonzalez-Calderon, Dana E. Rollison, Susan T. Vadaparampil. Preliminary usage of an interactive, clinical trials educational decision aid for Florida cancer patients [abstract]. In: Proceedings of the 18th AACR Conference on the Science of Cancer Health Disparities; 2025 Sep 18-21; Baltimore, MD. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2025;34(9 Suppl):Abstract nr B091.

ABSTRACT:The radiation sensitivity index (RSI) and 12-chemokine gene expression signature (12CK GES) are two gene expression signatures (GES) that were previously developed to predict tumor radiation sensitivity or identify the presence of tertiary lymphoid structures in tumors, respectively. To advance the use of these GESs into clinical trial evaluation, their assays must be assessed within the context of the Clinical Laboratory Improvement Amendments (CLIA) process. Using HG-U133Plus2.0 arrays, we first established CLIA laboratory proficiency. Then the accuracy (limit of detection and macrodissection impact), precision (variability by time and operator), sample type (surgery vs. biopsy), and concordance with a reference laboratory were evaluated. RSI and 12CK GES were reproducible (RSI: 0.01 mean difference, 12CK GES: 0.17 mean difference) and precise with respect to time and operator. Taken together, the reproducibility analysis of the scores indicated a median RSI difference of 0.06 (6.47% of range) across samples and a median 12CK GES difference of 0.92 (12.29% of range). Experiments indicated that the lower limit of input RNA is 5 ng. Reproducibility with a second CLIA laboratory demonstrated reliability with the median RSI score difference of 0.065 (6% of full range) and 12CK GES difference of 0.93 (12% of observed range). Overall, under CLIA, RSI and 12CK GES were demonstrated by the Moffitt Cancer Center Advanced Diagnostic Laboratory to be reproducible GESs for clinical usage. SIGNIFICANCE:The RSI and 12CK GES are two GESs that predict tumor radiation sensitivity or the presence of tertiary lymphoid structures in tumors, respectively. These GESs were assessed within the CLIA process for future clinical use. We established proficiency, reproducibility, and reliability characteristics for both signatures in a controlled setting, indicating these GESs are suitable for validation within future clinical trials.

INTRODUCTION:Non-Hispanic (NH) Black/African American (AA) and Hispanic cancer patients are underrepresented in cancer clinical trials (CCTs) due to patient, physician, and system-level barriers. Therefore, multi-level approaches are critical to address barriers to CCT participation. Presented here are the study design and baseline characteristics of ACT WONDER2S, a multi-level intervention (MLI) aimed to decrease barriers to NH Black/AA and Hispanic patient referral and enrollment in CCTs. METHODS:ACT WONDER2S is an MLI including community outreach and digital interventions for community and Moffitt Cancer Center (MCC) populations. Geospatial analytics were used to identify clusters of census tracts ("priority zones") with high NH Black/AA and Hispanic populations for intervention deployment. Priority zones were then matched on population characteristics and randomized to receive the intervention (n = 7) or to serve as controls (n = 7). Baseline characteristics of the priority zones were described using US Census data and other public sources. RESULTS:Approximately 16.5 % and 35.8 % of the intervention priority zones are NH Black/AA or Hispanic, respectively. There are no statistically significant differences between groups in total population size (paired t-test p-value = 0.63), proportions of NH Black/AA (p = 0.13) and Hispanic populations (p = 0.17), or distance in miles from MCC (p = 0.64). The estimated number of cancer cases and CCT enrollment rates at baseline are also similar between groups. CONCLUSION:If shown to be effective in increasing referral and enrollment of NH Black/AA and Hispanic cancer patients to CCTs, ACT WONDER2)S can be deployed across other geographic settings, thereby reducing disparities to CCT referral and enrollment on a national scale.

Background: HPV infection is implicated in approximately half of global penile squamous cell carcinoma (PSCC) cases. Previous studies on HPV DNA and p16INK4a status in PSCC have yielded inconclusive prognostic findings. This meta-analysis aims to elucidate the prognostic role of HPV in PSCC by pooling data on disease-free survival (DFS), disease-specific survival (DSS), and overall survival (OS). Methods: We systematically searched Medline and Embase up to January 2023 for relevant human studies. Data from eligible publications reporting HPV DNA or p16INK4a status, along with and DFS, DSS, or OS outcomes, were extracted. A random-effects meta-analysis model was used to synthesize data, with study weights based on size and significance. The study protocol was registered with PROSPERO (CRD42019131355). Results: Out of 544 studies screened, 34 publications were included, comprising a pooled sample size of 3,944 patients. p16INK4a-positive status was associated with improved OS (hazard ratio [HR], 0.54; 95% CI, 0.39-0.75; I 2=31%), DFS (HR, 0.52; 95% CI, 0.29-0.94; I 2=20%), and DSS (HR, 0.34; 95% CI, 0.23-0.50; I 2=18%). HPV DNA positivity was significantly associated with improved DFS (HR, 0.63; 95% CI, 0.46-0.87; I 2=13%) and DSS (HR, 0.46; 95% CI, 0.29-0.75; I 2=47%) but not OS (HR, 0.92; 95% CI, 0.74-1.11; I 2=0%). Conclusions: This meta-analysis, comprising the largest number of patients with PSCC to date, shows a notable correlation between p16INK4a immunohistochemistry positivity and survival outcomes. These findings support the understanding that penile cancer cases not associated with HPV tend to behave more aggressively. We support p16INK4a immunohistochemistry testing as part of the initial diagnostic evaluation of patients with PSCC.

ABSTRACT Background Treatment decision-making in oropharyngeal squamous cell carcinoma (OPSCC) includes clinical stage, HPV status, and smoking history. Despite improvements in staging with separation of HPV-positive and -negative OPSCC in AJCC 8th edition (AJCC8), patients are largely treated with a uniform approach, with recent efforts focused on de-intensification in low-risk patients. We have previously shown, in a pooled analysis, that the genomic adjusted radiation dose (GARD) is predictive of radiation treatment benefit and can be used to guide RT dose selection. We hypothesize that GARD can be used to predict overall survival (OS) in HPV-positive OPSCC patients treated with radiotherapy (RT). Methods Gene expression profiles (Affymetrix Clariom D) were analyzed for 234 formalin-fixed paraffin-embedded samples from HPV-positive OPSCC patients within an international, multi-institutional, prospective/retrospective observational study including patients with AJCC 7th edition stage III-IVb. GARD, a measure of the treatment effect of RT, was calculated for each patient as previously described. In total, 191 patients received primary RT definitive treatment (chemoradiation or RT alone, and 43 patients received post-operative RT. Two RT dose fractionations were utilized for primary RT cases (70 Gy in 35 fractions or 69.96 Gy in 33 fractions). Median RT dose was 70 Gy (range 50.88-74) for primary RT definitive cases and 66 Gy (range 44-70) for post-operative RT cases. The median follow up was 46.2 months (95% CI, 33.5-63.1). Cox proportional hazards analyses were performed with GARD as both a continuous and dichotomous variable and time-dependent ROC analyses compared the performance of GARD with the NRG clinical nomogram for overall survival. Results Despite uniform radiation dose utilization, GARD showed significant heterogeneity (range 30-110), reflecting the underlying genomic differences in the cohort. On multivariable analysis, each unit increase in GARD was associated with an improvement in OS (HR = 0.951 (0.911, 0.993), p = 0.023) compared to AJCC8 (HR = 1.999 (0.791, 5.047)), p = 0.143). ROC analysis for GARD at 36 months yielded an AUC of 80. 6 (69.4, 91.9) compared with an AUC of 73.6 (55.4, 91.7) for the NRG clinical nomogram. GARD ≥ 64.2 was associated with improved OS (HR = 0.280 (0.100, 0.781), p = 0.015). In a virtual trial, GARD predicts that uniform RT dose de-escalation results in overall inferior OS but proposes two separate genomic strategies where selective RT dose de-escalation in GARD-selected populations results in clinical equipoise. Conclusions In this multi-institutional cohort of patients with HPV-positive OPSCC, GARD predicts OS as a continuous variable, outperforms the NRG nomogram and provides a novel genomic strategy to modern clinical trial design. We propose that GARD, which provides the first opportunity for genomic guided personalization of radiation dose, should be incorporated in the diagnostic workup of HPV-positive OPSCC patients.

Abstract Spatial transcriptomics (ST) is a powerful tool for understanding tissue biology and disease mechanisms. However, the advanced data analysis and programming skills required can hinder researchers from realizing the full potential of ST. To address this, we developed spatialGE, a web application that simplifies the analysis of ST data. The application spatialGE provided a user-friendly interface that guides users without programming expertise through various analysis pipelines, including quality control, normalization, domain detection, phenotyping, and multiple spatial analyses. It also enabled comparative analysis among samples and supported various ST technologies. The utility of spatialGE was demonstrated through its application in studying the tumor microenvironment of two data sets: 10× Visium samples from a cohort of melanoma metastasis and NanoString CosMx fields of vision from a cohort of Merkel cell carcinoma samples. These results support the ability of spatialGE to identify spatial gene expression patterns that provide valuable insights into the tumor microenvironment and highlight its utility in democratizing ST data analysis for the wider scientific community. Significance: The spatialGE web application enables user-friendly exploratory analysis of spatial transcriptomics data by using a point-and-click interface to guide users from data input to discovery of spatial patterns, facilitating hypothesis generation.

Abstract Introduction: Black/African American (AA) and Hispanic cancer patients are underrepresented in cancer clinical trials (CCTs). Disparities stem from multiple factors, thus, comprehensive, multi-level interventions (MLI) are needed. Presented here is a conceptual framework for selecting the appropriate study design for MLIs using ACTWONDER2S, a multi-level intervention (MLI) aimed to increase Black/AA and Hispanic patient participation into CCTs, as an illustrative example. Baseline characteristics of the ACTWONDER2S target populations are also provided. Methods: ACTWONDER2S integrates community health educator and digital tools into an MLI targeting community residents and physicians in the Moffitt Cancer Center (MCC) catchment area, as well as MCC patients, physicians and CCT coordinators. A literature review of the relative strengths and weakness of 5 candidate designs for MLIs was conducted, and a cluster stratified randomized design was selected. Geospatial analytics were used to identify clusters of census tracts (“priority zones”) with high Black/AA and Hispanic populations for intervention deployment. Baseline characteristics of the priority zones were described using US Census data and other sources. Results: 14 priority zones were identified, which were matched on population characteristics and randomized into intervention (n=7) and control (n=7) zones. There were no statistically significant differences between the intervention and control priority zones in total population size (paired t-test p-value=0.63), proportions of Black/AA (p=0.133) and Hispanic populations p=0.17), and distance (in miles) from MCC (p=0.64). Approximately 35.8% and 16.5% of the intervention priority zones were Hispanic or Black/AA, respectively. Average distance to MCC, cancer cases, referral and CCT enrollment rates were also similar between groups. Conclusion: Decision frameworks for MLI study design selection are lacking. The framework provided here for ACTWONDER2S can be applied to other studies seeking to evaluate MLIs. Furthermore, the randomization of the priority zones produced optimal target populations for testing the efficacy of the MIL within ACTWONDER2S. Citation Format: Dana E. Rollison, Rossybelle P. Amorrortu, Lindsay N. Fuzzell, Melany A. Garcia, Elliott S. Tapia-Kwan, Yayi Zhao, Steven A. Eschrich, Bob R. Gore, Brian S. Mittman, Nathanael B. Stanley, Susan T. Vadaparampil. Multi-level intervention to increase minority cancer patient enrollment to clinical treatment trials- Study design considerations and baseline characteristics from the ACTWONDER2S Study [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 4824.

Abstract Patients with lung squamous cell carcinoma (LSCC) require new drug targets and improved biomarkers due to a lack of targetable genomic drivers and low response rates to immune checkpoint blockade. In a previous study, we analyzed a cohort of 108 LSCC patients by integrating DNA copy number variation, somatic mutations, RNA-sequencing, and expression proteomics. The principal discovery was the identification of three proteomic subtypes, with the majority (87%) of tumors comprising two of these subtypes. The "Inflamed" subtype showed enrichment for B-cell-rich tertiary lymphoid structures, while the "Redox" subtype exhibited enrichment for redox pathways and NFE2L2/KEAP1 alterations but had notably lower immune infiltration. We hypothesized that these subtypes would result in distinct metabolic signatures. Using ultra-high-performance liquid chromatographic separation on a HILIC column, followed by analysis on a Q Exactive HF high resolution mass spectrometer, we performed untargeted metabolomics on 87 tumors from the same LSCC proteogenomics cohort. A total of 7,392 features were obtained from this analysis, leading to the identification of 446 metabolites through m/z and retention time matching against an internal reference library. To understand if metabolomics could recapitulate our proteomic subtypes, we applied consensus clustering and non-negative matrix factorization (NMF) and assessed the resulting clusters using a Random Forest (RF) supervised classifier. Area Under the Curve (AUC) values for consensus clustering (5 clusters, AUC = 0.72), NMF (4 clusters, AUC = 0.73), and proteomics subtypes (Stewart et al. Nature communications. 2019:10:3578) (3 clusters, AUC = 0.74) suggest that metabolite abundances do indeed recapitulate the proteomic subtypes. Differential expression between Redox and Inflamed yielded 29 differentially expressed metabolites (p-value < 0.05 and 1.5 fold change). Glutathione, a key redox metabolite, was modestly elevated in the Redox proteomic subtype (0.58 log2 ratio, p = 1.14E-05). Notably, we identified glutathione metabolism (p = 1.29-5) and arginine biosynthesis (p = 5.22-4) as among the most significant pathways among differentially expressed metabolites. Glutathione, a key redox metabolite, was modestly elevated in the Redox proteomic subtype (0.58 log2 ratio, p = 1.14E-05). In conclusion, metabolomics recapitulates the proteomic subtypes, and there are distinct differences between these subtypes at the metabolite level. Ongoing work is developing a novel, network-based analysis framework to integrate these data quantitatively. Citation Format: Isis Y. Narvaez-Bandera, Ashley Lui, Eric Welsh, Dalia Ercan, Vanessa Rubio, Hayley Ackerman, Guohui Li, Lancia Darville, Min Liu, Bin Fang, Steven Eschrich, Brooke Fridley, John Koomen, Eric Haura, Gina M. DeNicola, Elsa Flores, Paul Stewart. Multi-omic landscape of squamous cell lung cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 3495.

Spatial transcriptomics (ST) is a powerful tool for understanding tissue biology and disease mechanisms. However, its potential is often underutilized due to the advanced data analysis and programming skills required. To address this, we present spatialGE, a web application that simplifies the analysis of ST data. The application spatialGE provides a user-friendly interface that guides users without programming expertise through various analysis pipelines, including quality control, normalization, domain detection, phenotyping, and multiple spatial analyses. It also enables comparative analysis among samples and supports various ST technologies. We demonstrate the utility of spatialGE through its application in studying the tumor microenvironment of melanoma brain metastasis and Merkel cell carcinoma. Our results highlight the ability of spatialGE to identify spatial gene expression patterns and enrichments, providing valuable insights into the tumor microenvironment and its utility in democratizing ST data analysis for the wider scientific community.

Penile squamous cell carcinoma (PSCC) is a rare and deadly malignancy. Therapeutic advances have been stifled by a poor understanding of disease biology. Specifically, the immune microenvironment is an underexplored component in PSCC and the activity of immune checkpoint inhibitors observed in a subset of patients suggests immune escape may play an important role in tumorigenesis. Herein, we explored for the first time the immune microenvironment of 57 men with PSCC and how it varies with the presence of human papillomavirus (HPV) infection and across tumor stages using multiplex immunofluorescence of key immune cell markers. We observed an increase in the density of immune effector cells in node-negative tumors and a progressive rise in inhibitory immune players such as type 2 macrophages and upregulation of the PD-L1 checkpoint in men with N1 and N2-3 disease. There were no differences in immune cell densities with HPV status.

9 Background: Penile squamous cell carcinoma (PSCC) is a rare and aggressive malignancy with limited treatment options in the advanced or recurrent settings. Immunotherapy can yield clinical responses in PSCC, but less than 20% of patients benefit. Herein, using multiplex immunofluorescence (MIF), we analyzed the composition of the tumor immune microenvironment as it varies with human papillomavirus (HPV) status and across advancing disease stages with an aim to further our understanding of how immune composition may impact clinical outcome, potential role in disease progression and response to immune therapies. Methods: Formalin-fixed and paraffin-embedded tissue samples from men with PSCC treated at Moffitt cancer center (Tampa, FL) were immunostained for CD3, CD4, CD8, CD68, PD-1, PD-L1, CD163 and CD 206 using OPAL TM 7 kit (AKOYA Biosciences) for MIF. Densities of various cell phenotypes (expressed as cells/mm 2 ) were quantified using an automated quantitative image analysis system (InForm). Immune cell phenotype densities were stratified by HPV status and by tumor stage. We will apply the nearest neighbor cell analysis to identify the spatial orientation of immune cells in the tumor microenvironment. Results: 57 PSCC patients (median age, 60 years [IQR (interquartile range) 31-92]; 60% HPV negative) had MIF analysis. More than half (31/57) had lymph nodes involved (6 N1, 25 N2-3). The immune composition did not differ significantly with HPV status. When investigated by individual clinical stage, we observed an increase in median density of total (CD3+), helper (CD3+CD4+), cytotoxic (CD3+CD8+) and CD3+PD-L1+ T cells from N0 to N1 stage (pN0: CD3+ 44.23, CD3+CD4+ 14.77, CD3+CD8+ 16.15 cells/mm 2 ; pN1: CD3+ 188.36, CD3+CD4+ 71.69, CD3+CD8+ 40.22 cells/mm 2 ) followed by a decrease in N2-N3 stage (CD3+CD4+ 18.45 and CD3+CD8+ 28.22 cells/mm 2 ).The median density of total macrophages increased with stage (pN0: 306.38; pN1: 502.47; pN2-3: 648.27 cells/mm 2 ). While the activated M1 macrophages increased from N0 to N1 and decreased in the more advanced N2-3 stage, M2 macrophages steadily increased across stages and became the dominant type in N2-3. Conclusions: This study describes the interplay of T cells and macrophages across disease stages using MIF. We observed an initial T cell and myeloid immune response in the early locoregional stage, followed by the emergence of immune exhaustion, marked by a decline in the density of cytotoxic T cells, rising PD-L1 expression, and the progressive replacement of anti-tumor M1 macrophages with pro-tumorigenic M2 macrophages across stages. Our findings can inform treatments utilizing immune manipulation. Geospatial analysis exploring the proximity relationships of different immune cell types to each other and to the tumor is ongoing and will be presented at the meeting.

Herein, we report the characterization of four cohorts of breast cancer patients including (1) non-Hispanic Whites in Florida, (2) non-Hispanic Blacks in Florida, (3) Hispanics in Florida, and (4) Hispanics in Puerto Rico. Data from female breast cancer patients were collected from cancer registry (n = 9361) and self-reported patient questionnaires (n = 4324). Several statistical tests were applied to identify significant group differences. Breast cancer patients from Puerto Rico were least frequently employed and had the lowest rates of college education among the groups. They also reported more live births and less breastfeeding. Both Hispanic groups reported a higher fraction experiencing menstruation at age 11 or younger (Floridian Hispanics [38

Abstract Background: Our group has previously developed the radiosensitivity index (RSI) using a multigene expression model that is directly proportional to tumor radioresistance (high RSI = increased radioresistance). RSI has been previously validated in two datasets of patients with triple negative breast cancer (TNBC). In this study, we will run a selective dose personalization study in TNBC patients undergoing breast conservation therapy (BCT). Based on patients RSI scores, they will either receive a radiation therapy (RT) boost of 10 Gy to the tumor cavity or not. Given our data in two independent datasets, the current study will reveal the feasibility and benefit of selective genomic dose personalization in TNBC following BCT. Trial Design: The study is designed as a prospective, nonrandomized, phase II trial of genomically guided RT in the management of TNBC undergoing BCT. Patients will be allocated to one of two groups based on their RSI determination from fresh frozen tissue collected by biopsy or at the time of BCT. These groups will be Group A, RSI optimized whole breast radiotherapy alone or with a 10 Gy boost or Group B, RSI not optimized whole breast radiotherapy with a boost of 10 Gy to tumor cavity. Patients will receive standard of care chemotherapy, neoadjuvant or adjuvant. Eligibility: TNBC patients undergoing BCT. Specific Aims: To determine the three-year local control following genomically guided dose personalization in the management of TNBC following BCT. Secondary objectives include determination of overall survival (OS), progression free survival (PFS), and quality of life (QOL) following genomically guided dose personalization. Statistical Methods: The primary hypothesis is the three-year local control rates differ for groups A and B, against the null hypothesis that the two rates are identical. Patients will allocate approximately 78% in group A and 22% in group B and local control rates are expected to be 96% and 75%, respectively. Assuming 80% power and 10% type I error for a log-rank test, 86 patients are needed. An interim analysis will be completed when 4 disease progression events are observed. There will be approximately 43 patients at the time of interim analysis. Patient Accrual: This study is open with 1 patient enrolled at the time of submission. A total of 86 patients will be enrolled. Contact Information: Kamran A. Ahmed MD, Moffitt Cancer Center, email: kamran.ahmed@moffitt.org, Clinical trial information: NCT05115474. Funding: Moffitt and Morton Plant Mease Foundations. Citation Format: Kamran Ahmed, Iman Washington, Matthew Mills, Michelle DeJesus, Youngchul Kim, Ronica Nanda, Javier Torres-Roca, Steven Eschrich, Janis De La Iglesia, John Puskas, Marilin Rosa, Jason Wilson, Paula Lundgren, Negar Golesorkhi, Nazanin Khakpour, Susan Hoover, Marie Lee, John Kiluk, Melissa Mallory, Christine Laronga, Laura Kruper, Brian Czerniecki, Roberto Diaz. Phase II Study of Genomically Guided Radiation Dose Personalization in the Management of Triple Negative Breast Cancer [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO3-20-05.

Abstract Background Soft tissue sarcomas (STS), have significant inter- and intra-tumoral heterogeneity, with poor response to standard neoadjuvant radiotherapy (RT). Achieving a favorable pathologic response (FPR ≥ 95%) from RT is associated with improved patient outcome. Genomic adjusted radiation dose (GARD), a radiation-specific metric that quantifies the expected RT treatment effect as a function of tumor dose and genomics, proposed that STS is significantly underdosed. STS have significant radiomic heterogeneity, where radiomic habitats can delineate regions of intra-tumoral hypoxia and radioresistance. We designed a novel clinical trial, Habitat Escalated Adaptive Therapy (HEAT), utilizing radiomic habitats to identify areas of radioresistance within the tumor and targeting them with GARD-optimized doses, to improve FPR in high-grade STS. Methods Phase 2 non-randomized single-arm clinical trial includes non-metastatic, resectable high-grade STS patients. Pre-treatment multiparametric MRIs (mpMRI) delineate three distinct intra-tumoral habitats based on apparent diffusion coefficient (ADC) and dynamic contrast enhanced (DCE) sequences. GARD estimates that simultaneous integrated boost (SIB) doses of 70 and 60 Gy in 25 fractions to the highest and intermediate radioresistant habitats, while the remaining volume receives standard 50 Gy, would lead to a > 3 fold FPR increase to 24%. Pre-treatment CT guided biopsies of each habitat along with clip placement will be performed for pathologic evaluation, future genomic studies, and response assessment. An mpMRI taken between weeks two and three of treatment will be used for biological plan adaptation to account for tumor response, in addition to an mpMRI after the completion of radiotherapy in addition to pathologic response, toxicity, radiomic response, disease control, and survival will be evaluated as secondary endpoints. Furthermore, liquid biopsy will be performed with mpMRI for future ancillary studies. Discussion This is the first clinical trial to test a novel genomic-based RT dose optimization (GARD) and to utilize radiomic habitats to identify and target radioresistance regions, as a strategy to improve the outcome of RT-treated STS patients. Its success could usher in a new phase in radiation oncology, integrating genomic and radiomic insights into clinical practice and trial designs, and may reveal new radiomic and genomic biomarkers, refining personalized treatment strategies for STS. Trial registration NCT05301283. Trial status The trial started recruitment on March 17, 2022.