2026 Abstracts eFlip Booklet

View 2026 Abstracts eFlip Booklet flipbook.

ABSTRACT BOOKLET Radiation Oncology Research Interactive eFlip-ready edition Click the Table of Contents titles to navigate directly to each abstract.

TABLE OF CONTENTS 1 Automated Generation of Deliverable VMAT Breast Plans from Predicted Dose Distributions: A RayStation Scripting Workflow 2 A Debridement-First Technique for the Management of Vulvar Radionecrosis After Dose-Escalated Interstitial Brachytherapy 3 Addressing Unseen Burdens: Feasibility and Conclusions of Patient-Reported Outcome Collection in Radiation Oncology Clinic 4 Assessing Patient Needs to Inform the Design of CAREConnect, a Patient-Centered Cancer Support Group 5 Identifying Delays in Diagnosis and Treatment of Patients with Stage III NSCLC: Examining Barriers in a Safety-Net Health System 6 The role of PARP inhibition in overcoming NFR2-driven radioresistance in head and neck cancer 7 PATH-RT: Pathway for Accountable Transition and Handoff to Radiation Therapy Interactive PDF: select any abstract title to jump to the abstract.

TABLE OF CONTENTS 8 Predicting Erectile Function After Radiotherapy: Long-Term Evaluation of a Previously Validated 2-year Model out to 10-years 9 Feasibility of a Same-Day OR-to-Outpatient Gynecologic Interstitial HDR Brachytherapy Workflow Without Inpatient Admission 10 Mapping Delays and Surveillance Gaps in Lung SBRT within a Safety-Net Health System 11 Mitochondrial rewiring supports survival of triple negative breast cancer cells after ionizing radiation 12 Resident Integration and Site Efficiency Project (RISE): Improving Site Onboarding in a Multi-Site Radiation Oncology Residency 13 A Quantitative Multiscale Framework for Investigating the Mechanisms of Ultra-High Dose-Rate FLASH Radiotherapy 14 From Consult to Plan Review: Development of a Structured Disease-Site Educational Framework for Radiation Oncology Residents Using Evidence-Based Andragogy Interactive PDF: select any abstract title to jump to the abstract.

Automated Generation of Deliverable VMAT Breast Plans from Predicted Dose Distributions: A RayStation Scripting Workflow Xinru Chen, PhD1, Krystal Morales, MD1, Baozhou Sun, PhD1, Laurence Court, PhD2, and Yao Zhao, PhD2 1. Department of Radiation Oncology, Baylor College of Medicine 2. Department of Radiation Physics, UT MD Anderson Cancer Center Key words: Automated treatment planning; Dose prediction; VMAT; Breast I. INTRODUCTION Radiotherapy planning is time-intensive, and rising patient volumes strain resources in high-throughput centers. Deep learning–based dose prediction can estimate achievable, patient-specific dose distributions from anatomy, improving plan quality and consistency while reducing planning time; multi-center studies show such systems can match manually generated breast plans. However, a predicted dose distribution is not itself deliverable—it lacks the machine parameters needed for delivery and converting it into a deliverable VMAT plan (dose mimicking) is challenging. Most published work has focused on dose predictions alone, leaving a translational gap to clinical implementation. This study developed and tested a fully automated RayStation-scripted workflow that converts predicted dose into a deliverable VMAT breast plan without manual optimization. II. METHODS An automated RayStation-scripted workflow generates a deliverable VMAT plan from a predicted dose distribution: (1) the predicted dose map, PTV/GTV structures, and prescription are acquired; (2) organ-at-risk contours are auto-generated using commercial models; (3) for each organ, optimization constraints are derived—directive-specified OAR constraints are read from achieved values in the predicted dose, otherwise mean dose, maximum dose, EUD, D98%, and D2% are extracted as constraints; (4) isocenter and VMAT arc angles are automatically determined; and (5) the plan is automatically optimized. The workflow was tested on a lumpectomy simultaneous integrated boost (SIB) case, comparing the deliverable plan to the predicted dose. Validation on a wider range of plans are undergoing. III. RESULTS The workflow generated a deliverable plan in approximately 30 minutes. Relative to the prediction, PTV-high V95% rose from 99.88% to 99.99% and PTV-low V95% from 98.05% to 99.80%. Contralateral breast D0.03cc increased from 9.53 to 9.71 Gy and heart dose from 3.83 to 4.03 Gy; ipsilateral lung V4Gy, V8Gy, and V16Gy each increased by less than 1%. IV. CONCLUSION This automated workflow reliably converts predicted dose into a deliverable VMAT breast plan, preserving target coverage and OAR sparing close to prediction. By eliminating manual optimization, it bridges the gap between dose-prediction research and clinical use, reducing planning time and variability in high-volume departments. Validation on larger, more diverse cohorts and comparison with manual plans is recommended.

A Debridement-First Technique for the Management of Vulvar Radionecrosis After Dose-Escalated Interstitial Brachytherapy Authors: Abraham Arenas, MD1; Jake Owen, BS2; Cheryl Claunch, MD, PhD3; Hayden Ansinelli, MD, MSc4; Andrew M. Farach, MD5 1Baylor College of Medicine, Houston, TX; 2Texas A&M School of Medicine, Bryan, TX; 3Department of Radiation Medicine, Oregon Health & Science University, Portland, OR; 4Arizona Oncology, Tucson, AZ; 5Department of Radiation Oncology, Houston Methodist Hospital, Houston, TX ABSTRACT Interstitial brachytherapy achieves highly conformal dose escalation in vulvar cancer, improving local control while sparing adjacent organs at risk. The doses required, however, carry a real risk of late radionecrosis, a severely painful complication for which practical guidance remains scarce. Some radiation oncologists may avoid escalating dose for this reason, even though undertreatment carries the greater risk of local failure. We present our approach to preventing and managing vulvar radionecrosis, illustrated by two patients treated on an identical protocol, only one of whom developed radionecrosis. Our algorithm rests on patient selection, prophylaxis in higher-risk patients, a homogeneous implant, and, once necrosis develops, debridement before pharmacologic therapy.

Addressing Unseen Burdens: Feasibility and Conclusions of Patient-Reported Outcome Collection in Radiation Oncology Clinic Jamie Liang1, Shelly Sharma, MD2 1. Baylor College of Medicine, Medical Student 2. Baylor College of Medicine, Radiation Oncology Introduction: Patient-centered cancer care requires understanding of not only disease but patients’ physical, emotional, social, and functional well-being. Validated patient-reported outcomes (PROs) can capture these dimensions, yet workflow barriers limit routine use. We conducted a feasibility quality-improvement (QI) pilot to evaluate PRO implementation, clinical information yield, and barriers to sustainable integration. Methods: Beginning in August 2025, a feasibility QI pilot incorporated Functional Assessment of Cancer Therapy (FACT) questionnaires into consultation and follow-up visits within an academic breast and gynecologic radiation oncology practice using Plan-Do-Study-Act cycles. Patients completed printed questionnaires before physician evaluation; responses were reviewed and transcribed into structured EPIC SmartTexts. Process measures assessed collection/documentation; outcome measures assessed feasibility and multidimensional information yield; balancing measures included completion time and staff-reported workflow burden. Associations with age, disease stage, and visit type were explored. Results: Feasibility: During limited pilot implementation, PROs were documented in 11.8% of applicable clinic encounters. No patient declined a questionnaire, and completion required 3–5 minutes. Clinical information yield: Among 62 patients with age-linked assessments, 58 had complete FACT-G domains. FACT-G demonstrated substantial variation (mean 88.8±12.7; range 53.8–107). Mean physical, social/family, emotional, and functional scores were 24.7±4.0, 23.9±4.2, 19.2±3.9, and 20.8±5.8, respectively. Age was unrelated to overall FACT-G (r=.008, p=.95) or individual domains; disease stage was not significantly associated with any domain. Compared with follow-up patients, consultation patients reported numerically lower emotional (18.1 vs 20.0; p=.083) and functional well-being (19.1 vs 22.0; p=.070), suggesting potentially greater needs at initial evaluation. Workflow barriers: Printing, scanning completed forms, and secure shredding created nursing-dependent steps that became the principal implementation bottleneck.

Conclusion: PROs efficiently captured substantial multidimensional variation in patients’ lived experience not identified by age or disease stage, with signals of greater emotional and functional burden at consultation. Yet, a 3–5-minute patient assessment generated a disproportionately burdensome paper workflow. Direct electronic PRO capture with automated scoring and actionable response pathways may enable scalable, longitudinal identification of quality-of-life needs and opportunities for intervention. Keywords: Patient-reported outcomes, quality improvement

Assessing Patient Needs to Inform the Design of CAREConnect, a Patient-Centered Cancer Support Group Authors: Avery Franques, 1 Samya Hughes,2 Brian Box, 1• 3 Emaan Lehardy, 1•4 Katie Gonzalez,5 Shelly Sharma 1 • 3 1Baylor College of Medicine, Houston, TX, U.S. 2Texas Southern University, Houston, TX, U.S. 3Department of Radiation Oncology, Dan L Duncan Comprehensive Cancer Center, Baylor College of Medicine, Houston, TX, U.S. 4Department of Medical Oncology, Dan L Duncan Comprehensive Cancer Center, Baylor College of Medicine, Houston, TX, U.S. 5Dan L Duncan Comprehensive Cancer Center, Baylor College of Medicine, Houston, TX, U.S. Introduction/Background Cancer support groups can provide peer support, practical guidance, and patient education. However, support programs may be underutilized when their structure and content do not reflect patient needs. CAREConnect is a quality improvement (QI) initiative designed to develop a feasible, patient-centered support and education program at DLDCCC. As the initial phase of this initiative, we conducted a needs assessment to identify psychosocial and supportive care priorities and patient preferences for program delivery. Methods We conducted a prospectively designed, cross-sectional needs-assessment survey across three oncology clinics (Radiation Oncology, Breast Oncology, and Gynecologic Oncology). The voluntary survey assessed patient interests, supportive care needs, and logistical preferences. Patients could complete the survey on paper or online, with both English and Spanish versions available. A total of 100 responses were collected and analyzed using SPSS. Findings will be used to inform the design of the subsequent CAREConnect QI intervention. Results Among 100 respondents, 87 completed the survey in English and 13 in Spanish; 74 responded on paper, and 26 online. Most respondents (96%) were above the age of 40. Nearly one third (32%) of patients work full-time, 10% part-time, and 57% were not currently employed. The most frequently requested session topics were anxiety (49%), treatment side effects (48%), fatigue (40%), nutrition (45%), and returning to work and normal life (27%), with similar preferences across English- and Spanish- speaking respondents. Nearly half (49%) requested virtual meetings, although 46% of Spanish-speaking respondents reported they would not participate in a support group. Preferred meeting times were weekdays after 5 pm (42%) or before/after their clinic visit (39%), as opposed to meeting on weekends, regardless of employment status. Conclusions and Recommendations This prospective needs assessment identified actionable patient priorities for the CAREConnect QI intervention. Findings support a weekday, predominantly virtual model with language-access services and educational content focused on anxiety, treatment side effects, fatigue, nutrition, and return to normal life. The next phase will implement CAREConnect using these patient-informed design elements and evaluate participation, feasibility, and patient experience. Keywords: Support group, quality improvement, survey

Baylor College of Medicine 10th Annual Radiation Oncology Research Symposium Abstract Title: Identifying Delays in Diagnosis and Treatment of Patients with Stage III NSCLC: Examining Barriers in a Safety-Net Health System Authors and Affiliations: Annabel Merritt, Baylor College of Medicine Madeline Flanagan, M.D., Baylor College of Medicine Alexander Hanania, M.D., MPH, Baylor College of Medicine Abstract: Introduction/Background: Stage III non-small cell lung cancer (NSCLC) can progress rapidly, underscoring the need for timely diagnostic, staging, and multidisciplinary evaluation to initiate definitive treatment. This process requires a radiation oncology consult visit, PET/CT scan, brain MRI, tissue biopsy, CT simulation appointment, and treatment planning, each of which can be difficult to schedule, particularly in safety net systems such as Harris Health, where many patients are uninsured or underinsured. Delays in this pathway may allow disease progression and compromise curative-intent therapy. Methods: We retrospectively identified adults with stage III NSCLC treated within the Harris Health System from 2015 to the present. Patients were included if they had stage III NSCLC at their radiation oncology consultation and were planned for definitive radiotherapy (typically 60-66 Gray in 30-33 fractions). Patients with small cell histology or non-curative-intent radiation were excluded. We anticipate approximately 170 patients will meet inclusion criteria for analysis. We collected data including demographics, comorbidities, smoking history, and ECOG performance status; disease characteristics (histology, tumor size/location, TNM stage per AJCC 8th edition, pretreatment imaging dates); treatment details (radiation dose/fractionation, concurrent chemotherapy, adjuvant immunotherapy); and outcomes (local, regional, and distant control, failure patterns/timing, subsequent therapy, survival, and disease status at last follow-up). We will perform descriptive statistics on each treatment timing interval, compare intervals across patient subgroups, assess trends over time, evaluate associations between delays and clinical outcomes, and identify predictors of delay using multivariable regression. Results: Data collection is ongoing; outcome data is not available at the time of submission. We plan to present preliminary results at the conference.

Conclusions: This ongoing study aims to characterize the diagnostic and treatment pathway for stage III NSCLC within a safety-net health system and identify the largest time intervals contributing to delay. These findings will help target interventions, such as streamlining imaging, referrals, and prior authorizations, to reduce time to definitive treatment and limit opportunity for disease progression, ultimately improving outcomes for an underserved population. (326 words, max 350) Keywords (max of 3): Non-small cell lung cancer, treatment delays, underserved population

Title: The role of PARP inhibition in overcoming NFR2-driven radioresistance in head and neck cancer Authors: Grace E. Anderson*, Avishi Singh*, Kalil Saab, Pedram Yadollahi, Yang Li, Fonma Essien, Reid T. Powell, Clifford Stephan, Michael J. Bolt, Michael A. Mancini, Mitchell J. Frederick, Vlad C. Sandulache#, Rutulkumar Patel# * These authors contributed equally to this work as co-first authors. # Co-senior authors. Introduction: HPV-independent head and neck squamous cell carcinoma (HNSCC) is a therapeutically challenging malignancy; around 50% of HPV-independent HNSCC do not respond to multimodal treatment regimens, including radiation therapy (RT). Approximately 20% of HPV-independent HNSCC tumors have genetic alterations in the KEAP1/NFE2L2/CUL3 oxidative stress pathway and have been shown to be therapy-resistant. Although loss of KEAP1 function leads to NRF2 (the protein product of the NFE2L2 gene) hyperactivation, NRF2 has not been successfully targeted clinically, perhaps because its role has been viewed as protective rather than as a therapeutic target. Methods and results: This study used murine cancer cell lines from genetically engineered mouse models (GEMMs) of HNSCC with conditional activation of oncogenic PIK3CA and deletion of Trp53, with or without conditional deletion of Keap1. We verified NRF2 activation in Keap1-deficient (Keap1fl/fl) cell lines through RNA-seq analysis, increased antioxidant gene expression, elevated intracellular NADH and glutathione levels, and enhanced radioresistance in clonogenic survival assays. Next, we performed a high-throughput drug screen using FDA-approved oncology drugs and identified poly (ADP-ribose) polymerase (PARP) inhibitors as among the most radiosensitizing agents for Keap1fl/fl cells. Clonogenic assays with olaparib and rucaparib (PARP inhibitors) confirmed in vitro radiosensitization, with dose enhancement factor at 10% survival (DEF10) values consistently higher for Keap1fl/fl cell lines, suggesting genotype-specific radiosensitization. For the in vivo arm, we assessed treatment efficacy in both athymic (immunocompromised) and syngeneic (immunocompetent) models. Athymic mice were given a lingual injection of Keap1fl/fl tumor cells, while syngeneic mice received MOC1 cells. We assessed tumor growth delay and overall survival in each cohort under four conditions: no treatment, RT alone, olaparib alone, or RT and olaparib. Again, combination treatment significantly delayed tumor growth and improved overall survival in both models compared with individual treatments. Conclusions: Our results provide evidence that PARP inhibitor radiosensitization is effective in both GEMM and syngeneic orthotopic tumor models, establishing a mechanistic rationale: PARP inhibition overcomes NRF2-mediated radioresistance in HNSCC by targeting the NRF2 antioxidant shield. Ultimately, our findings are a major step towards clinical evaluation of PARP inhibitors as radiosensitizers for HNSCC.

PATH-RT: Pathway for Accountable Transition and Handoff to Radiation Therapy Authors: Sara Ahmed, Shelly Sharma Affiliation: Radiation Oncology Department, Baylor College of Medicine Background: Hospitalized neurosurgical patients with postoperative brain metastases, spine metastases, or gliomas may require timely outpatient Radiation Oncology (Rad Onc) evaluation following discharge. When evaluation within 2–3 weeks is recommended, failure to reliably place, schedule, or track the referral may delay cancer-directed care. We initiated PATH-RT to characterize process failures during this transition and develop a standardized closed-loop referral pathway. Aim: To systematically identify contributors to delayed or missed post-discharge Rad Onc follow-up and design a closed-loop process targeting 100% timely referral placement and outpatient consultation scheduling among eligible neurosurgical patients. Methods: A structured current-state process analysis was performed from receipt of an inpatient Rad Onc call through scheduling of the recommended outpatient consultation. Process mapping characterized sequential workflow steps, responsible stakeholders, and interservice handoffs. A multidisciplinary root-cause analysis identified candidate failure domains including referral initiation, handoffs between rotating clinicians, longitudinal patient tracking, scheduling and appointment availability, financial clearance, and patient communication. During baseline assessment, eligible patients with delayed or missed follow-up will undergo case-level pathway review. Each case will be reconstructed across key transition points, and all contributing process failures will be classified using predefined failure categories; individual patients may contribute to multiple failure events. Failure-event frequencies will be evaluated using Pareto analysis to identify recurrent, modifiable process vulnerabilities and prioritize targets for intervention. Proposed Intervention: Findings will inform sequential Plan-Do-Study-Act cycles evaluating PATH-RT, a proposed closed-loop pathway incorporating defined referral ownership, standardized resident handoff, longitudinal tracking, and confirmation of outpatient scheduling. Direct referral placement by the Rad Onc team will be evaluated where institutionally feasible.

Measures/Next Steps: Baseline measures will include timely referral placement, consultation scheduling and completion, discharge-to-consultation interval, and frequency and distribution of contributing failure events. Subsequent PDSA cycles will target the highest-frequency modifiable failures and assess process reliability and workflow burden. Conclusion: PATH-RT applies systematic processes and failure-mode analysis to a vulnerable inpatient-to- outpatient oncologic transition. Identifying recurrent failure points before intervention will enable targeted, iterative redesign rather than reliance on nonspecific reminders or additional handoffs.

320 words (plus figure and title) Predicting Erectile Function After Radiotherapy: Long-Term Evaluation of a Previously Validated 2-year Model out to 10-years AUTHORS AND AFFILIATIONS: Kim T, Flanagan MM, Gay H, Michalski J, Patil D, Sandler HM, Hembroff L, Kaplan I, Ciezki J, McLaughlin PW, Patel SA, Sanda MG, Hamstra DA INTRODUCTION/BACKGROUND: The Alemozaffar (JAMA, 2011) model was developed in PROST-QA and externally validated in CaPSURE to predict erectile function 2-years after EBRT or brachytherapy. This analysis evaluates whether the frozen 2011 model remains accurate at 10-years. METHODS AND MATERIALS: 568 PROST-QA participants with complete model covariates (282 EBRT, 286 brachytherapy) were stratified into quartiles from the previous model. The outcome was based on the question on quality of erections, “firm enough for intercourse” or erection sufficient for intercourse (ESI) vs. all other answers. Discrimination was assessed by area under the receiver-operating-characteristic curve (AUC) at each horizon. Time to loss of ESI was analyzed by Cox regression, with baseline sexual score, age, androgen deprivation, and treatment as covariates. RESULTS: ESI by model is shown (figure). The discrimination for EBRT AUC was 0.82 (derivation) and maintained at 0.70 at 10-years; brachytherapy AUC was 0.89 (derivation) and 0.78 at 10- years (P-values <0.001 for all except EBRT at 10-years (P = 0.015)). The Cox model for ESI split at 2-years due to a differential impact of age. Age was not associated with ESI in the first 2-years (HR 1.08, P=0.50) but became the dominant thereafter (HR 1.94, P<0.001). Baseline sexual function predicted ESI across all follow-ups (hazard ratio 1.42 from 0–2 years, 1.40 from 2–15 years; both P<0.01). Androgen deprivation's effect was concentrated early (HR 1.89, P=0.001) and not significant later (HR 0.43, P=0.119). Treatment type was not an independent predictor of ESI (HR 1.34, P=0.072; HR 0.71, P=0.211).

CONCLUSION: Analysis after 10-years of follow up shows us that the 2011 model continues to accurately predict patient’s ESI after EBRT and brachytherapy. However, the relative weight of its predictors shifts over time: baseline sexual function remains stable, and androgen deprivation's influence is front-loaded, and age becomes increasingly important beyond 2- years. Keywords: Erectile function, Prediction model, 10-year update

Title: Feasibility of a Same-Day OR-to-Outpatient Gynecologic Interstitial HDR Brachytherapy Workflow Without Inpatient Admission Authors: Ansh Purohit1, Ryan LaFratta1, Ren Tong1, Jingzhu Xu1, Snehal Desai1, Jan S. Sunde2, Daniel Hamstra1, Shelly Sharma1 Affiliations: 1Department of Radiation Oncology, Dan L. Duncan Comprehensive Cancer Center, Baylor College of Medicine, Houston, TX; 2Department of Obstetrics and Gynecology, Division of Gynecologic Oncology, Baylor College of Medicine, Houston, TX Introduction/Background: Interstitial high-dose-rate (HDR) brachytherapy is essential for locally advanced gynecologic cancers when intracavitary applicators cannot adequately cover the target, yet it is traditionally delivered as an inpatient procedure with prolonged immobilization and hospitalization, a burden that contributes to declining brachytherapy utilization. We implemented a fully outpatient pathway in which applicator placement, CT simulation, planning, HDR delivery, applicator removal, and discharge occur within a single calendar day, and evaluated its feasibility, efficiency, dosimetric quality, and safety. Methods: We reviewed all consecutive patients treated on the same-day pathway from March 2024 through April 2026. Applicators (median 8 needles) were placed under general anesthesia in a day-surgery operating room, stabilized with a custom external fixation belt, and transferred via air mattress for CT-based planning and HDR delivery, with the implant re-created at every fraction. The primary endpoint was time from operating room entry to applicator removal. Cumulative EQD2 was compared with EMBRACE-II planning aims. Safety endpoints included bleeding, procedural complications, and unplanned admissions or emergency visits within 30 days. Results: Twenty-two patients (18 cervical, 3 vaginal, 1 endometrial; median age 50) underwent 68 implant sessions, most commonly 8 Gy × 3. Half were Hispanic and 50% lived >20 miles away. All 68 sessions ended in same-day discharge with no overnight admissions. Mean operating room entry to applicator removal decreased from 458 minutes in the first five fractions to 337 in the most recent five (−27%); total encounter time fell from 478 to 385 minutes. In definitive-intent cervical patients (n = 15), median HR-CTV D90 was 90.1 Gy, and median bladder, rectum, sigmoid, and bowel D2cm³ were 74.0, 62.4, 60.2, and 56.2 Gy, all within EMBRACE-II planning aims. One minor bleeding event was managed with in-clinic packing; there were no unplanned admissions or emergency visits. Crude local control among definitive-intent cervical patients was 94.1% (16/17) at a median follow-up of 5.3 months. Conclusions and Recommendations: Same-day OR-to-outpatient interstitial HDR brachytherapy is feasible, safe, and dosimetrically sound, with efficiency improving as the workflow matures. Eliminating admission may lower the barrier to offering interstitial brachytherapy, particularly in safety-net settings. Planning time is the primary target for further optimization. Keywords: brachytherapy; interstitial; outpatient workflow

Title: Mapping Delays and Surveillance Gaps in Lung SBRT within a Safety-Net Health System Authors: Ansh Purohit1, Ashley Chavana1, Sarah Kim2, Quillan Huang3, Pavan Mukesh Jhaveri1, Babith Mankidy4, Ravi Ghanta5, Aparna Chakravarti Jotwani3, Alexander Nicola Hanania1 Affiliations: 1Department of Radiation Oncology, Dan L. Duncan Comprehensive Cancer Center, Baylor College of Medicine, Houston, TX; 2Department of Internal Medicine, Washington University School of Medicine in St. Louis, St. Louis, MO; 3Department of Medical Oncology, Baylor College of Medicine, Houston, TX; 4Department of Interventional Pulmonology, Baylor College of Medicine, Houston, TX; 5Department of Thoracic Surgery, Baylor College of Medicine, Houston, TX Introduction/Background: Stereotactic body radiation therapy (SBRT) is the standard curative option for medically inoperable early-stage lung cancer, yet treatment timeliness and post-treatment surveillance remain challenging in underserved populations. We performed a process-level audit of the SBRT care continuum at a large urban safety-net health system to identify high-yield quality improvement (QI) targets. Methods: We retrospectively reviewed 106 patients treated with definitive lung SBRT from 2014 to 2022. The interval from suspicious imaging to first fraction was partitioned into four sequential segments: imaging–referral, referral–consult, consult–simulation, and simulation–first fraction. Factors associated with treatment delay and surveillance non-adherence (failure to complete NCCN-concordant imaging) were evaluated by descriptive comparison and multivariable logistic regression. The yield of post-SBRT recurrence workup was assessed. Results: The cohort was 56% Black, 14% Hispanic, 6% Asian, and 24% White; 33% were uninsured. Among 104 patients evaluable for timing, median imaging-to-treatment interval was 6.4 months [IQR 4.4–12.3], and only 12.5% were treated within 12 weeks. The imaging-to-referral interval was the dominant contributor (median 4.5 months [IQR 2.2–10.0]), whereas post-referral radiation oncology intervals were short: referral-to-consult 2.6 weeks, consult-to-simulation 1.7 weeks, and simulation-to-first fraction 2.6 weeks. Median time-to-treatment was longer for Hispanic (8.0 months) and Black (6.6 months) than White (5.6 months) patients; pre-treatment biopsy and Black/Hispanic race/ethnicity were associated with numerically higher odds of delay without reaching statistical significance. Of 100 patients evaluable for surveillance, 41% completed NCCN-concordant follow-up; loss to follow-up (33%) and temporary care discontinuation (14%) were the leading drivers of non-adherence. Among 28 patients with post-SBRT imaging concerning for recurrence, 11 underwent successful biopsy and only 4 (14%) had confirmed recurrence; PET avidity did not reliably distinguish recurrence from post-treatment change (2 of 7 PET-avid lesions). Conclusions and Recommendations: Delays in lung SBRT in this safety-net cohort arose predominantly in the pre-referral interval rather than the radiation oncology workflow, with numerically longer delays among Black and Hispanic patients and those undergoing biopsy. Surveillance adherence was low, and most imaging concerns for recurrence proved benign. We

recommend four QI targets: systematic imaging-to-referral tracking, expedited biopsy or empiric-treatment pathways for selected patients, navigation-based surveillance support, and standardized multidisciplinary review of post-SBRT imaging changes. Keywords: stereotactic body radiation therapy; lung cancer; health disparities

Mitochondrial rewiring supports survival of triple negative breast cancer cells after ionizing radiation Steven W. Wall1, Karen Wang1, Allison Greer1, Mokryun L. Baek1, Audra Lane1, Mariah J. Berner1, Jonathan T. Lei1,2,3, Ángel A. Garcés1,4, Sofia Khoury5, Emily G. Caggiano6,7,8, Matthew D. Meyer9, Sarah Latka10, Hugo Villanueva10, Emil Schüler7, Cullen M. Taniguchi*6, Rutulkumar Patel2,12, Gloria V. Echeverria1,2,3,12,13 1Lester and Sue Smith Breast Center and Department of Medicine, Baylor College of Medicine, Houston, TX, USA 2Dan L Duncan Comprehensive Cancer Center, Baylor College of Medicine, Houston, TX, USA 3Department of Molecular and Cellular Biology, Baylor College of Medicine, Houston, TX, USA 4Medical Scientist Training Program, Baylor College of Medicine, Houston, TX, USA 5Department of Biosciences, Rice University, Houston, TX, USA 6Department of Experimental Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA 7Department of Radiation Physics, The University of Texas MD Anderson Cancer Center, Houston, TX, USA 8The University of Texas MD Anderson Cancer Center UTHealth Houston Graduate School of Biomedical Sciences, Houston TX, USA 9Shared Equipment Authority, Rice University, Houston, TX, USA 10Advanced Technology Cores, Baylor College of Medicine, Houston, TX, USA 11Department of Otolaryngology – Head and Neck Surgery, Baylor College of Medicine, Houston, TX, USA 12Department of Radiation Oncology, Baylor College of Medicine, Houston, TX, USA 13Department of Medicine, Baylor College of Medicine, Houston, TX, USA *Posthumous Keywords: mitochondria, breast cancer, therapy resistance Triple negative breast cancer (TNBC) is an aggressive form of breast cancer and has limited therapy options. Conventional treatments include neoadjuvant chemo-immunotherapy followed by surgical resection and may include further adjuvant immunotherapy and/or radiotherapy of the tumor bed and lymph nodes. Despite these intensive regimens, many TNBC patients with residual disease at the time of surgery have a rapid metastatic recurrence and thus improved therapies are critically needed. Mitochondrial adaptations have been established as functional, targetable mediators of TNBC chemoresistance. We thus investigated whether mitochondrial adaptations support the metabolic rewiring and survival of residual TNBC cells following ionizing radiation (IR). We established longitudinal in vitro models of post-IR human TNBC, characterized by cellular die-off (i.e., regression) to a residual phenotypic state (the cell population nadir), then eventual cell repopulation (i.e., regrowth). These population dynamics were accompanied by plastic adoption of unique metabolic, proteomic, and morphologic changes that largely reverted when cells regrew. Following IR, residual cells exhibited extensive mitochondrial rewiring, including elevated mitochondrial copy number, oxidative phosphorylation (oxphos) rates, inner membrane structure, and mitochondrial metabolite levels. Concomitantly, levels of the short protein isoform of the mitochondrial inner membrane protein optic atrophy 1 (OPA1) were significantly elevated in residual cells, and OPA1 knockout ablated mitochondrial adaptations induced by IR. Further, OPA1 genetic or pharmacologic perturbations led to improved cellular responses to radiation. Metabolomic and proteomic analyses of radio-residual cells uncovered a coordinated program of antioxidant and redox capacity elevation with mitochondrial metabolism, which was corroborated by analysis of external datasets.

Together, these findings provide evidence that TNBC cells surviving radiotherapy adopt an OPA1-dependent program of mitochondrial reorganization that supports their survival and regrowth, and that genetic and pharmacologic disruption of OPA1 improved radiation responses of TNBC cells, thereby positioning OPA1 as a therapeutic dependency that could improve radiotherapy efficacy and reduce TNBC recurrence.

Title: Resident Integration and Site Efficiency Project (RISE): Improving Site Onboarding in a Multi-Site Radiation Oncology Residency Authors and Affiliations: Brian Box, MD; Shelly Sharma, MD; Department of Radiation Oncology, Baylor College of Medicine, Houston, TX. Introduction/Background: Radiation oncology residents at our institution rotate across three clinical sites with distinct workflows, software platforms, documentation processes, and team structures. Repeated re-orientation creates onboarding friction that diverts time from education and patient care. During the 2026-2027 academic year, RISE aimed to increase the proportion of residents achieving operational comfort within 0-1 day of starting a new site rotation from 14% at baseline to ≥ 50%. Methods: Using sequential Plan-Do-Study-Act (PDSA) cycles, we surveyed residents pre- intervention (June 2026) with 5-point Likert items and free-text prompts. Thematic review identified workflow, documentation, and EMR navigation as key modifiable barriers, informing site-specific guides. Guides were distributed electronically before the next rotation. Post- intervention surveys (September 2026) assessed receipt and review (process), preparedness and time to operational comfort (outcome), and unintended effects (balancing). Surveys were anonymous and unpaired; data were analyzed descriptively. This abstract reports Cycle 1. Results: Response rates were 78% (7/9) pre- and 89% (8/9) post-intervention. At baseline, 1/7 residents (14%) reached operational comfort within 0-1 day; EMR navigation (6/7) and simulation (5/7) were common barriers. Post-intervention, 8/8 (100%) received the guide before rotation and 7/8 (88%) reviewed it in week one. Reported benefits included preparedness and documentation clarity (each 6/8, 75%), workflow confidence (5/8, 63%), and reduced logistical time and transition stress (each 4/8, 50%). Six of 8 (75%) each rated it concise and accurate, endorsed continued use, and reported no unintended effects; one cited excessive density and outdated material. Operational comfort within 0-1 day improved to 3/8 (38%), short of the 50% target; 3/8 (38%) required 5 or more days. Conclusions and Recommendations: PDSA Cycle 1 achieved near-universal adoption but did not meet the 0-1 day aim; gains concentrated among residents with prior site exposure. Findings are descriptive; the small, unpaired sample precludes causal attribution. Cycle 2 will condense guides into topic-specific sections, add contact lists, EMR access instructions, and target residents without prior site exposure. Keywords: Quality Improvement; Graduate Medical Education; Radiation Oncology

Title A Quantitative Multiscale Framework for Investigating the Mechanisms of Ultra-High Dose-Rate FLASH Radiotherapy Authors and Affiliations Yiding Han, PhD¹; Sethi A. Ammar¹; Baozhou Sun, PhD¹; Daniel Allan Hamstra, MD PhD¹; Ramiro Pino, PhD¹; Zaid A. Siddiqui, MD¹; ¹Department of Radiation Oncology, Baylor College of Medicine, Houston, TX, USA Introduction/Background Ultra-high dose-rate FLASH radiotherapy can reduce normal-tissue toxicity while maintaining tumor response compared with conventional dose-rate (CDR) irradiation. However, no established quantitative framework links radiation delivery, radical kinetics, molecular damage, and biological lesion processing across experiments. We developed a mechanistically interpretable framework to test whether proposed FLASH mechanisms can simultaneously explain published observations within a physically and biologically consistent parameter space. Methods A coupled two-subsystem model was developed. The physicochemical subsystem describes radiation-induced radical production, first-order reactions, nonlinear radical- radical interactions, molecular reaction products, and spatial diffusion. The downstream biological subsystem describes DNA lesion formation from direct radiation damage, radical-mediated indirect damage, and lesion repair. Peak lesion burden was used as the damage metric. Parameter-sensitivity analyses compared conventional and FLASH dose rates to identify processes capable of producing either comparable effects or FLASH- associated sparing. Results Model sensitivity analysis demonstrated that differences in lesion repair rate alone cannot straightforwardly explain simultaneous normal-tissue sparing and preserved tumor response. Depending on the repair-rate regime, increasing repair can alter or even reverse the relative difference between FLASH and CDR, while a separate region exists in which the FLASH/CDR difference becomes relatively insensitive to repair. Direct radiation-induced damage depends primarily on total dose rather than dose rate and therefore does not independently explain differential FLASH responses. In contrast, parameters governing

radical production, radical consumption/interactions, and radical-mediated lesion generation can produce qualitatively distinct regimes ranging from comparable FLASH/CDR effects to substantial FLASH sparing. Importantly, multiple parameter combinations can reproduce an individual experimental phenotype, demonstrating substantial parameter non-uniqueness. Conclusions and Recommendations The coupled radical-kinetic and lesion-processing framework provides a systematic approach for testing the quantitative compatibility of proposed FLASH mechanisms. Preliminary results suggest that the FLASH effect is unlikely to be uniquely explained by repair kinetics or any single isolated mechanism, but may emerge from nonlinear interactions between physicochemical and biological processes. Future work will follow two complementary directions: simultaneous fitting across multiple published datasets to identify experimentally consistent and biologically plausible parameter regions, and quantum-mechanical calculations to determine whether additional chemical pathways are required to explain FLASH–CDR differences. Cross-validation between macroscopic modeling and microscopic chemistry may provide a more complete mechanistic explanation of the FLASH effect. Keywords FLASH radiotherapy; Radiation chemistry; Mechanistic modeling

Flash

Classic Radiosensitivity Mechanism repair of DNA damage (R1) redistribution of cells in the cycle (R2) re-population (R3) re-oxygenation of hypoxic regions (R4). radiosensitivity defined by the intrinsic response of different tissues and cells (R5) Today’s radiotherapy teratments: all cancer cells within the same tumour will respond similarly to radiation-> Higher dose more kill (Dose rate independent) reactivation defined by extrinsic elements making tumour cells more sensitive to radiation (that is, oxygen, drugs and the immune system) (R6) reinforcement by extrinsic factors from the microenvironment (R7) However, when we increase the the dose rate from 0.2Gy/s to 10^6Gy/s, we observe some interesting things

Previous Trials and Studies • Similar result on overall tumour control compared with conventional dose rate radiotherapy • Induce minimal toxicity to normal tissue • Potential1: Flash can introduce more dose to the target with maintain similar toxicity to the normal tissue : sarcomas and other common carcinomas (60-70 Gys with largely relapse) • Potential2:Effectively treating recurrent: Treat->bad outcome- >Use flash treat again withtout normal tissue toxicity • Potential3: Improve overall treatment outcome without damaging remain tissue/ surgically removing organ (rectal cancer) • But the problem is: WHY the damage effect different for tumor and healthy tissue

Flash: Less ROS Conventional: More ROS Modification of the dose rate did not modify the DNA damage induced in plasmids, and this was maintained under biochemical conditions that mimic some aspects of the tumour microenvironment, such as hypoxia and high Fe2+ levels. Physicochemical Hypothesis

Flash: Less ROS Conventional: More ROS The expression of oxylipins, bioactive lipid hydroperoxides derived from the oxygenation of polyunsaturated fatty acids, finding lower levels in the lungs of mice after FLASH irradiation compared with conventional dose rate radiotherapy. Modification in the protein backbone Samples irradiated at conventional dose rates had increased β-sheet content relative to the α-helical secondary structures, whereas samples irradiated FLASH retained a secondary structure profile similar to the non- irradiated controls. However, modification of the dose rate did not modify the DNA damage induced in plasmids, and this was maintained under biochemical conditions that mimic some aspects of the tumour microenvironment, such as hypoxia and high Fe2+ levels. Failed mechanism to explain difference between tumor and normal tissue, lower ROS level should protect both kinds of tissue at the same time. Physicochemical Hypothesis

Healthy tissues are resisetant to FLASH When zebrafish embryos are irradiated with electron or proton beam FLASH radiotherapy at 4 h post- fertilization, an apoptotic peak occurs 24 h later and is accompanied by a high rate of proliferation. A 5 days post-irradiation, only FLASH-irradiated zebrafish embryos presented a spared morphogenesis, suggesting that FLASH radiotherapy preserves the repair potential of the embryo. Interestingly, the relative resistance of healthy tissues to FLASH radiotherapy is also maintained with age. With different age-stage animal test with whole-brain radiation: conventional dose rate irradiation at any age induces irreversible cognitive deficits, with juvenile brains being the most sensitive, while FLASH irradiation maintain a relative cognitive acuity.

Tumours are sensitive to FLASH repair of DNA damage (R1) redistribution of cells in the cycle (R2) re-population (R3) re-oxygenation of hypoxic regions (R4). radiosensitivity defined by the intrinsic response of different tissues and cells (R5) Today’s radiotherapy teratments: all cancer cells within the same tumour will respond similarly to radiation-> Higher dose more kill reactivation defined by extrinsic elements making tumour cells more sensitive to radiation (that is, oxygen, drugs and the immune system) (R6) reinforcement by extrinsic factors from the microenvironment (R7) R1(DNA damge) is identical after FLASH and conventional dose rate Clonogenic survival, gold standard for R5 (intrinsic radidation sensitivity), identical for FLASH and conventional dose rate. However, T-ALL xenograted in immunocompromised mice show difference between FLASH and conventional dose: 2/3 more responsive to FLASH 1/3 unresponsive to FLASH but controlled by conventional radiotherapy Difference mechanism?

Tumours are sensitive to FLASH: new Mechanism FLASH-sensitive T-ALL overecpressed Gadd45B (R2 and R3), a gene involved in the inhibition of the G2/M transition. Under hypoxic condition (R4), FLASH uniquiely retained antitumour efficacy associated with high expression of Gadd45B. -> FlASH overcome hypoxia One hot but wrong explanation: oxygen depletion: ultrahigh dose rate could deplete oxygen more efficiently than conventional dose. In fact, from xogyen tension performed in bitro and in situ as well as in xenografted studies, local oxygen depletion rate is too low to affect tumour response. Another popular hypothesis: FLASH enhance immunogenicity (R6 and R7) monitoring of circulating immune cells in rats with orthotopic glioblastoma found identicial for FLASH and conventional radiation. More studies show FLASH and conventional radiation show similar immune response for both immunocompetent and immunodeficient animals. No TGFβ production changed However, FLASH response in normal tissues (Lung and skin) of rodents and canines associated with less TGFβ production. Proton-based FLASH radiotherapy but not proton-based conventional radiotherapy was able to sensitize medulloblastoma to GD2 chimeric antigen receptor-T cell therapy through FLASH radiotherapy-induced effects

Tumours are sensitive to FLASH: new Mechanism FLASH radiation led to less vasculature collapse and decreased expression of phosphorylated myosin light chain, a protein involved in the control of vascular smooth muscle cells contraction, vascular tone and resistance, as well as blood pressure and tissue perfusion, compared with conventional dose rate radiotherapy. This preserved vascularization in tumours could promote the influx of tumoricidal immune cells; however, it was not associated with an enhanced antitumour effect. In the healthy brain, where radiotherapy at a conventional dose rate is known to induce persistent vasculature abnormalities by disrupting the cells of the blood–brain barrier by altering the tight junction proteins. FLASH presever the expression of tight junction protein such as occuldin and claudin-5 -> preserve of cognition

Current research Gap • Current researches are focus on “hypothesis”, expect one statement can explain the deviation between the FLASH and CDR • No established quantitative framework • Radiation delivery → radical kinetics → molecular damage → lesion formation and repair • Whether a physically consistent set of mechanisms and parameters can explain the observed dose-rate dependence across different experiments. • “An important outcome of this study was the recognition of inconsistent biological endpoints and reporting standards across studies, which presents a major barrier to quantitative comparison and meta- analysis. ” Isabella C, et.al., Physics and Imaging in Radiation Oncology, Volume 37, 2026.

Our research Motivation • Can the currently proposed physicochemical and biological mechanisms, when placed into a common quantitative framework, simultaneously explain the existing experimental observations of FLASH radiotherapy? • Our Hypothesis: • The FLASH effect is a summation of multiple effects from chemical physics, biomolecular reactions and biological processing/repair • Dose-rate dependence originates from nonlinear competition among radiation-induced radical production, biomolecular reactions, radical– radical interactions, and spatial transport, while the differential response of normal tissue and tumor emerges downstream from differences in biological processing and repair of the resulting lesions.

Our research: Dose rate dependent model • We separate the model into two subsystem • Chemical physics: Radical subsystem • Biological damage: DNA lesion

Our research: Dose rate dependent model • Chemical physics: Radical subsystem 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 = 𝑔𝑔̇𝐷𝐷 − 𝑘𝑘1𝑅𝑅 − 2𝑘𝑘2𝑅𝑅2 − 𝑘𝑘3𝑅𝑅𝑅𝑅 + 𝐷𝐷𝜕𝜕𝛻𝛻2𝑅𝑅 With 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 = 𝑘𝑘2𝑅𝑅2 Buxton GV, et.al. Journal of Physical and Chemical Reference Data. 1988;17(2):513–886. Wardman P. Radiation Research. 2020;194(6):607–617. Labarbe R, et.al. Radiotherapy and Oncology. 2020;153:303–310.

Our research: Dose rate dependent model • Chemical physics: Radical subsystem 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 = 𝑔𝑔̇𝐷𝐷 − 𝑘𝑘1𝑅𝑅 − 2𝑘𝑘2𝑅𝑅2 − 𝑘𝑘3𝑅𝑅𝑅𝑅 + 𝐷𝐷𝜕𝜕𝛻𝛻2𝑅𝑅 With 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 = 𝑘𝑘2𝑅𝑅2 Abolfath R, et.al. Med Phys. 2020 Dec;47(12):6551-6561. Herskind, et.al, radical reactions, and the role of oxygen. Radiat Oncol 21, 111 (2026).

Our research: Dose rate dependent model • Chemical physics: Radical subsystem 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 = 𝑔𝑔̇𝐷𝐷 − 𝑘𝑘1𝑅𝑅 − 2𝑘𝑘2𝑅𝑅2 − 𝑘𝑘3𝑅𝑅𝑅𝑅 + 𝐷𝐷𝜕𝜕𝛻𝛻2𝑅𝑅 With 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 = 𝑘𝑘2𝑅𝑅2

Our Model vs Microscopic result Abolfath R, et.al. Med Phys. 2020 Dec;47(12):6551-6561.

Our research: Dose rate dependent model • Biological damage: DNA lesion 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 = 𝑌𝑌𝑑𝑑𝑑𝑑𝑑𝑑̇𝐷𝐷 + 𝛽𝛽𝑅𝑅 − 𝑘𝑘𝑑𝑑𝑟𝑟𝑟𝑟𝑑𝑑

Our research: Dose rate dependent model • Biological damage: DNA lesion 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 = 𝑌𝑌𝑑𝑑𝑑𝑑𝑑𝑑̇𝐷𝐷 + 𝛽𝛽𝑅𝑅 − 𝑘𝑘𝑑𝑑𝑟𝑟𝑟𝑟𝑑𝑑 Ward JF. Progress in Nucleic Acid Research and Molecular Biology. 1988;35:95–125. Sachs RK, Hahnfeldt P, Brenner DJ. Journal of Radiation Biology. 1997;72(4):351–374.

Our research: Dose rate dependent model • Biological damage: DNA lesion 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 = 𝑌𝑌𝑑𝑑𝑑𝑑𝑑𝑑̇𝐷𝐷 + 𝛽𝛽𝑅𝑅 − 𝑘𝑘𝑑𝑑𝑟𝑟𝑟𝑟𝑑𝑑 Curtis SB. Radiation Research. 1986;106(2):252–270.

The limitation of K_rep • Can repair ability difference between tissue and tumor explain the tissue sparing with similar tumor effect? • Bad repair ability => More damage on CDR • Good repair ability=> More damage on FLASH • It conflicts with the observation. Is model wrong?

K_rep effect for constant environment setting There is a region where the Krep ability is not sensitive to the damage difference between FLASH and CDR Unsensitive region

Our research: Dose rate dependent model • Biological damage: DNA lesion 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 = 𝑌𝑌𝑑𝑑𝑑𝑑𝑑𝑑̇𝐷𝐷 + 𝛽𝛽𝑅𝑅 − 𝑘𝑘𝑑𝑑𝑟𝑟𝑟𝑟𝑑𝑑 It would conflict the observation when krep located at sensitive range It indicates that Krep should located at insensitive range for radiation therapy

Our research: Dose rate dependent model • Biological damage: DNA lesion 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 = 𝑌𝑌𝑑𝑑𝑑𝑑𝑑𝑑̇𝐷𝐷 + 𝛽𝛽𝑅𝑅 − 𝑘𝑘𝑑𝑑𝑟𝑟𝑟𝑟𝑑𝑑 Depend on total dose instead of dose rate Equally effect on both tissue and tumor Cannot be used to explain the deviation between tissue sparing and tumor effect

Our research: Dose rate dependent model • Biological damage: DNA lesion 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 = 𝑌𝑌𝑑𝑑𝑑𝑑𝑑𝑑̇𝐷𝐷 + 𝛽𝛽𝑅𝑅 − 𝑘𝑘𝑑𝑑𝑟𝑟𝑟𝑟𝑑𝑑 Logically, the free radical part would dominate the effect deviation between tissue and tumor ONLY IF the classic biological model is correct In other word, if the free radical physics model cannot/show conflict result while explaining the experiments, it indicates the necessary of the new biological model.

Beta effect Comparable effect Sparing Effect • Biological damage: DNA lesion 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 = 𝑌𝑌𝑑𝑑𝑑𝑑𝑑𝑑̇𝐷𝐷 + 𝛽𝛽𝑅𝑅 − 𝑘𝑘𝑑𝑑𝑟𝑟𝑟𝑟𝑑𝑑 Whether tissue with higher beta Tumor with lower beta ? Oxygen role?

Free radical variables effect • Chemical physics: Radical subsystem 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 = 𝑔𝑔̇𝐷𝐷 − 𝑘𝑘1𝑅𝑅 − 2𝑘𝑘2𝑅𝑅2 − 𝑘𝑘3𝑅𝑅𝑅𝑅 + 𝐷𝐷𝜕𝜕𝛻𝛻2𝑅𝑅 With 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 = 𝑘𝑘2𝑅𝑅2

Free radical variables effect • Chemical physics: Radical subsystem 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 = 𝑔𝑔̇𝐷𝐷 − 𝑘𝑘1𝑅𝑅 − 2𝑘𝑘2𝑅𝑅2 − 𝑘𝑘3𝑅𝑅𝑅𝑅 + 𝐷𝐷𝜕𝜕𝛻𝛻2𝑅𝑅 With 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 = 𝑘𝑘2𝑅𝑅2 Comparable effect Sparing effect

Free radical variables effect • Chemical physics: Radical subsystem 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 = 𝑔𝑔̇𝐷𝐷 − 𝑘𝑘1𝑅𝑅 − 2𝑘𝑘2𝑅𝑅2 − 𝑘𝑘3𝑅𝑅𝑅𝑅 + 𝐷𝐷𝜕𝜕𝛻𝛻2𝑅𝑅 With 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 = 𝑘𝑘2𝑅𝑅2 Sparing effect Comparable effect

The issue and our solution • Challenge — parameter non-uniqueness: A wide range of parameter combinations can reproduce either FLASH sparing or comparable FLASH/conventional effects in individual experiments. Therefore, fitting a single experiment does not uniquely constrain the underlying mechanism, highlighting the need for a quantitative cross-experiment framework. • Our solution — cross-experiment constraints: For each experiment, identify the parameter region that provides an acceptable fit. Then compare these regions across experiments and exclude parameter sets that are biologically implausible or mutually inconsistent. The key test: Can a single physically and biologically plausible parameter space explain the published experimental observations simultaneously?

From Consult to Plan Review: Development of a structured Disease-site educational framework for Radiation Oncology Residents using Evidence-Based Andragogy to Improve Efficiency and Reinforce Disease-Site Mastery Author: Dr. Jayna Reichert, MD (PGY-4), Department of Radiation Oncology, Baylor College of Medicine Research Advisor(s): Dr. Shelly Sharma, MD, Dr. Zaid Siddiqui, MD Partners: Dr. Stephen Finn, Director, Center for Faculty Excellence, United States Military Academy, West Point, NY Background. Radiation oncology residents must integrate rapidly evolving information from multiple fragmented resources — clinical guidelines, contouring atlases, treatment protocols, and institution-specific resources across dozens of disease sites. Trying to consolidate this vast clinical information while simultaneously optimizing oncologic patient care coordination, maximizing workflow efficiency, and improving facility in radiation treatment planning can result in suboptimal learning when under high-cognitive load conditions without adequate educational frameworks given the propensity to overload working-memory capacity (Young et al., Medical Teacher, 2014). Cognitive load theory holds that instructional "scaffolding" improves learning by reducing extraneous load and titrating intrinsic load (van Merriënboer and Sweller, Medical Education, 2010). Standardized checklists act as layered defenses in the Swiss cheese model of error prevention (Wachter, JAMA Internal Medicine, 2024), while graded spaced retrieval practice using common resources durably strengthens knowledge retention (Wlodarczyk and Dhaliwal, JAMA Network Open, 2022). A comparable radiation oncology curricular redesign raised mean TXIT clinical percentile from the 17th to the 59th nationally over two years (Khan et al., Advances in Radiation Oncology, 2024). We developed a structured, disease-site educational framework designed to scaffold clinical reasoning while providing residents with practical tools that mirror the longitudinal workflow of radiation oncology care.

Methods. Using principles of cognitive load theory and instructional scaffolding this resident-led, faculty-mentored educational intervention was developed using prostate cancer as the initial disease-site model. Content was organized around sequential clinical competencies: diagnostic workup and risk stratification, treatment selection, simulation, target and organ-at-risk contouring, prescription and dose constraints, and treatment-plan evaluation. The framework integrates evidence-based resources into standardized “consult-to-plan-review” tools, including consultation templates, CT-simulation guidance, treatment directives, contouring resources, and plan-review checklists. Content was developed with faculty mentorship and designed to reinforce clinical reasoning and oral-board-level decision making while providing point-of-care resources applicable to routine resident workflow. Results: A prostate cancer educational framework spanning the clinical pathway from consultation through treatment-plan review was developed as the initial disease-site module. The intervention consolidates previously distributed educational and clinical resources into a sequential framework aligned with resident workflow. Based on this initial development phase, a prospective evaluation strategy was designed to assess objective learning rather than resource utilization alone. Planned measures include case- based knowledge assessment, resident confidence across defined competencies, and standardized contouring and treatment-plan evaluation exercises administered before and after curriculum exposure. A delayed assessment is planned to evaluate retention. Resident feedback will identify components requiring refinement prior to adaptation to additional disease sites. We project ≥25% gains in workflow efficiency consistency, improved resident confidence, and a durable centralized resource library, with TXIT and mock-oral scores as longitudinal outcome measures. Conclusions. A structured consult-to-plan-review framework provides a feasible approach for organizing complex disease-site education around the actual clinical workflow of radiation oncology residents. Prostate cancer will serve as the pilot disease site for prospective evaluation of knowledge and applied clinical skills. Demonstrating educational effectiveness within a focused disease-site model may provide a foundation Commented [JR1]: Each resident is paired with a mentor who recently proctored oral boards and assigned two "bread-and-butter" sites, spanning 16 sites across eight disease groups. Pairs build standardized "cradle-to-grave" templates — Epic consult dot-phrases, CT-simulation notes, treatment directives, and plan-review checklists — embedding oral-board-level clinical reasoning, then deploy them to the residency shared drive by. Measures include time-to-completion, attending feedback, pre/post confidence surveys, TXIT scores, and mock oral board performance (Peters et al., BMC Medical Education, 2020). Commented [JR2]: (Sites in progress: Rectal, Sarcoma, Breast, Meningioma, Laryngeal, thyroid, vestibular schwannoma, brain mets, and more.. Commented [JR3]: Coupling workflow standardization with structured, board-aligned template development translates learning science into a scalable intervention that provides concrete intellectual scaffolding for incoming residents.

for subsequent expansion across additional disease sites and future evaluation of workflow efficiency, documentation consistency, and longitudinal educational outcomes. References: 1. Young JQ, Van Merrienboer J, Durning S, Ten Cate O. Cognitive load theory: implications for medical education: AMEE Guide No. 86. Med Teach. 2014;36(5):371-384. 2. Seshia SS, Young GB, Makhinson M, et al. Gating the holes in the Swiss cheese (part I): expanding Professor Reason's model for patient safety. J Eval Clin Pract. 2018;24(1):187-194. 3. Wlodarczyk S, Dhaliwal G. The learning sciences meet the learning health system. JAMA Netw Open. 2022;5(12):e2245972. 4. Wachter RM. COVID-19 and patient safety—lessons from 2 efforts to keep people safe. JAMA Intern Med. 2024;184(2):111-112. 5. Khan Q, Bowar B, Myers B, et al. Advancing resident education: experiential success in the creation of a comprehensive clinical didactic series in radiation oncology. Adv Radiat Oncol. 2024;9(5):101455. 6. van Merriënboer JJ, Sweller J. Cognitive load theory in health professional education: design principles and strategies. Med Educ. 2010;44(1):85-93. 7. Peters GW, Decker RH, Park HS, Yu JB, Evans SB. Resident attitudes and benefits of mock oral board examinations in radiation oncology. BMC Med Educ. 2020;20(1):200. 8. A Practical Guide to QI Data Analysis: Run and Statistical Process Control Charts. Hosp Pediatr. 2024;14(1):e1-e8. 9. Raman A, Ahmed AT, Sanghvi G, et al. Innovative approaches to training in evidence-based neurology. J Eval Clin Pract. 2025;31(1):XX-XX. 10. Thammasitboon S, Brand PLP. The physiology of learning: strategies clinical teachers can adopt to facilitate learning. Eur J Pediatr. 2022;181(8):2783-2790. 11. Taylor DC, Hamdy H. Adult learning theories: implications for learning and teaching in medical education: AMEE Guide No. 83. Med Teach. 2013;35(11):e1561-e1572. 12. Jimenez RB, et al. The impact of an introductory radiation oncology curriculum (IROC) for radiation oncology trainees across the United States and Canada. Int J Radiat Oncol Biol Phys. 2020;107(3):408-416.