Browse Articles
Discover research articles across all indexed journals
Antihistamine use and immune checkpoint inhibitor efficacy: An institutional analysis.
e24170 Background: Immune checkpoint inhibitor use has changed the treatment landscape for numerous solid tumor types. Prior work by others has suggested that H1 antihistamine is associated with an improvement in survival among patients receiving immune checkpoint inhibitors for lung cancer. We tested the hypothesis that of patients on a checkpoint inhibitor for any indication, those taking an H1 antihistamine would have longer time on therapy than those who do not. Methods: Deidentified retrospective patient data was collected using Patient Cohort Explorer at University of Mississippi Medical Center. All patients with at least one cycle of pembrolizumab, nivolumab or atezolizumab between 2013 and 2025 were included, as was the subset of patients on one of the aforementioned checkpoint inhibitors plus an oral antihistamine (cetirizine, loratadine, or fexofenadine). Other checkpoint inhibitors were excluded due to insufficient sample size to reasonably maintain anonymity. The primary endpoint was time to treatment discontinuation, measured in cycles, between all those on checkpoint inhibitor and those on checkpoint inhibitor plus an H1 antihistamine. We used this outcome because of its availability in the database. Statistical analysis was performed using unpaired two-tailed T test. Results: Data was collected for 1,182 patients having received a cumulative 9,878 cycles of immune checkpoint inhibitors, including 280 who received concomitant H1 antihistamine. The mean number of immune checkpoint inhibitor cycles delivered to those exposed to H1 antihistamine was significantly higher than those not exposed at 9.9 vs 7.9 (P 0.003, standard deviation 12.1 vs 8.3). Subgroup analysis is presented, with statistical significance demonstrated for pembrolizumab plus antihistamine but not nivolumab or atezolizumab. Sex, race, and insurance status were similar between groups. Conclusions: Our data indicates a positive association between H1 antihistamine use and the number of cycles of immunotherapy a patient receives, suggesting a link affecting the efficacy of immunotherapy. This was demonstrated in a broad population, irrespective of tumor type. Future work will evaluate for associations between H2 antihistamines as well as concurrent chemotherapy administration. All IO w/o AH All IO w/ AH Pembrolizumab w/o AH Pembrolizumab w/ AH Nivolumab w/o AH Nivolumab w/ AH Atezolizumab w/o AH Atezolizumab w/ AH # of patients 901 280 517 155 287 98 97 27 Mean # of cycles* 7.9 +/- 8.9 9.9 +/- 12.1 7.8 +/- 8.6 9.8 +/- 12.1 7.9 +/- 9.3 9.8 +/- 12.1 8.2 +/- 9.7 10.7 +/- 13 Female 371 144 257 92 86 42 28 10 Male 530 136 260 63 201 56 69 17 Median age 62 62 61 61 61 60.5 64 65 White 458 130 235 66 180 52 43 12 Black 422 147 268 88 101 44 53 15 Self pay 72 9 37 4 29 5 6 0 Medicaid 206 73 126 38 53 29 27 6 IO = immunotherapy, AH = antihistamine, * mean +/- standard deviation.
Volumetric analysis of central nervous system hemangioblastomas during belzutifan treatment holiday in von Hippel–Lindau (VHL) disease patients.
e22659 Background: Belzutifan produces durable volumetric responses in von Hippel–Lindau (VHL)–associated central nervous system (CNS) hemangioblastomas. However, data describing tumor behavior during treatment holidays are limited. Methods: We conducted a retrospective lesion-level volumetric analysis of VHL patients with CNS hemangioblastomas treated with belzutifan at Massachusetts General Brigham Cancer Institute between September 2018 and January 2026 who experienced a treatment interruption (holiday) ≥3 months. Longitudinal MRI data were retrospectively reviewed. Lesions were classified at holiday baseline as measurable (≥10 mm), submeasurable (5–9 mm), or non-measurable ( < 5 mm). The primary endpoint was lesion-level volumetric change during treatment holiday; secondary endpoints included ORR, best overall response, and time to response per RECIST 1.1. Lesion-level volumetric analyses were performed for measurable and submeasurable lesions. Treatment holiday was offered after ≥9 months of sustained maximal response. The study was conducted under an institutional review board–approved protocol. Results: Nine patients were included. At treatment initiation baseline imaging, 16 measurable, 11 submeasurable, and 34 non-measurable lesions were identified. At treatment holiday baseline, 7 measurable, 5 submeasurable, and 9 non-measurable lesions were evaluable. Median duration of belzutifan therapy prior to treatment interruption was 34.4 months (range, 2.9–57.7). Median duration of treatment interruption was 12.7 months (IQR, 9.3–15.3). Seven patients had RECIST-evaluable measurable disease. The ORR was 85.7% (6/7 partial responses; 95% CI, 42.1–99.6), with stable disease observed in 14.3% (1/7). No RECIST-defined progression occurred prior to treatment withholding. Median TTR was 5.9 months (range, 1.8–7.9), and responses were ongoing at the time of treatment interruption in all responding patients. From treatment holiday baseline to last eligible MRI, 10/12 evaluable lesions (83.3%) demonstrated volumetric increase ≥10%, while 2/12 lesions (16.7%) decreased in size by ≥10%. Among lesions that grew during treatment holiday, the median volumetric increase was +116.1% (IQR, 187.1), over a median duration of 8.7 months (IQR, 6.0). Conclusions: During belzutifan treatment holiday, most evaluable CNS lesions demonstrated mild volumetric increase over time, although lesion behavior was heterogeneous and occurred over several months rather than immediately after treatment withholding. These findings provide early, lesion-level insight into CNS hemangioblastoma dynamics during belzutifan holidays and support cautious clinical monitoring during treatment interruption. Ongoing follow-up and expanded analyses will further characterize lesion dynamics during belzutifan treatment holidays.
Microbiome associations with chemotherapy-induced nausea and vomiting: A metagenomic shotgun sequencing study.
12110 Background: Chemotherapy-induced nausea and vomiting (CINV) is common, with four out of every ten patients reporting moderate to severe nausea even after treatment with antiemetics. Existing risk prediction models have focused on demographic, clinical, and patient-reported factors, but predictive accuracy remains low. A missing predictor may be the gut microbiome, or gut microorganisms that engage in reciprocal signaling with the brain via many of the same pathways known to cause nausea. Methods: In this original research study, we conducted whole-metagenomic shotgun sequencing of the gut microbiome to evaluate its association with patient-reported CINV and improve existing risk prediction algorithms by employing a random forest approach. Nausea and vomiting were reported 24 hours after infusion (acute) and 5 days after infusion (delayed) in patients treated with moderately- or highly-emetogenic chemotherapy using the MASCC Antiemesis Tool. Participants were recruited from five ORIEN centers: City of Hope, Moffitt Cancer Center, Ohio State University, Rutgers University, and University of Colorado-Denver. Fecal samples were collected prior to chemotherapy using at-home fecal occult blood test (FOBT) cards. Results: Participants (n=233) were 57 years of age on average (SD=12), and most were female (80%), White (89%), with breast cancer (55%). More than half of participants (65%) had early stage cancer. Any acute and delayed self-reported nausea was reported by 40% and 67% of participants, respectively. Any acute and delayed vomiting was reported by 3% and 12% of participants, respectively; the low rate of acute vomiting precluded additional analyses. No significant difference was observed in microbial diversity and the abundance of dominant bacteria at the phylum level between patients who reported nausea or vomiting and those who did not. Bacillota and Pseudomonadota were two phyla identified as the most significantly enriched in patients who reported nausea or vomiting, at either timepoint (FDR-adjusted p < 0.05). The predictive performance of risk models based on established factors were improved by incorporating the top 15 differentially abundant bacterial species for acute nausea (AUC=69% vs. 58%), delayed nausea (AUC=68% vs. 53%), and delayed vomiting (AUC=77% vs. 54%). Conclusions: Gut microbial diversity did not differ between patients who self-reported acute or delayed nausea and those who did not, but risk prediction models for acute nausea, delayed nausea, and delayed vomiting were all improved by incorporating differentially abundant bacterial species. Results suggest that the gut microbiome is an important factor associated with the development of CINV.
Comparison of different metrics for estimating kidney radiation exposure during treatment for childhood cancer: A report from the St. Jude Lifetime Cohort Study.
10047 Background: Risk for impaired kidney function varies with radiation therapy (RT). We compared different methods for quantifying RT. Methods: Kidney function was graded in 2753 survivors (>10 years post-diagnosis, age ≥18 years) and 806 community controls (CC) per the Kidney Disease International Global Outcomes 2012 Clinical Practice Guideline for the Evaluation and Management of Chronic Kidney Disease (CKD). Multivariable logistic regression was used to evaluate the associations between grades 3-5 CKD and demographics and treatment exposures. RT was expressed as percentage of total kidney volume treated with 5 Gray (Gy) or more (V5), 10 Gy or more (V10) Gray, Children’s Oncology Group (COG) abdominal RT, or total body irradiation (TBI)/whole abdomen RT (WA). Model discrimination was assessed using area under the receiver operating characteristic curve (AUC). Results: 2.1% of survivors had stages 3 to 5 CKD, compared to 0.12% of CC (Odds Ratio (OR): 17.0 (95% confidence interval (CI) 2.9, 682.8), p < 0.001). Significant associations between higher risks of grades 3-5 CKD and demographic and treatment exposures were identified in separate regression models (Table). AUCs were 0.891, 0.888 and 0.896 for V5, V10, and TBI/WA, models respectively. Conclusions: TBI/WA is a more readily available exposure designation than V5 or V10. These results inform identification of survivors at highest risk for CKD. V5 V10 COG TBI/WA Predictors OR (95%CI) p OR (95%CI) p OR (95%CI) p OR (95%CI) p Race (White vs non-White) 1.68(0.81 – 3.30) 0.144 1.71(0.82 – 3.35) 0.133 1.80(0.89 – 3.44) 0.086 1.71(0.85 – 3.28) 0.116 Age at evaluation 1.09(1.04 – 1.13) <0.001 1.09(1.05 – 1.13) <0.001 1.09(1.05 – 1.13) <0.001 1.10(1.05 – 1.14) <0.001 Hypertension 8.21(4.12 – 17.61) <0.001 8.27(4.15 – 17.71) <0.001 8.61(4.41 – 18.02) <0.001 8.61(4.42 – 17.97) <0.001 Nephrectomy (yes/no) 2.44(0.88 – 6.12) 0.069 2.68(0.97 – 6.72) 0.044 4.23(1.81 – 9.48) 0.001 3.42(1.36 – 8.01) 0.006 Ifosfamide (per 1000 mg/m 2 ) 1.04(1.02 – 1.06) <0.001 1.04(1.02 – 1.05) <0.001 1.04(1.03 – 1.06) <0.001 1.04(1.03 – 1.06) <0.001 Cisplatin (per 100 mg/m 2 ) 1.46(1.26 – 1.68) <0.001 1.46(1.26 – 1.67) <0.001 1.47(1.27 – 1.68) <0.001 1.49(1.29 – 1.70) <0.001 Carboplatin (per 100 mg/m 2 ) 1.03(1.00 – 1.06) 0.024 1.03(1.00 – 1.06) 0.027 1.03(1.00 – 1.06) 0.012 1.03(1.00 – 1.06) 0.012 Calcineurin inhibitor (yes/no) 7.29(1.86 – 26.92) 0.003 7.9(11.94 – 30.52) 0.003 14.79(4.32 – 48.28) <0.001 9.31(2.43 – 33.77) 0.001 Doses Abelcet and/or Ambisome 1.02(0.94 – 1.06) 0.516 1.02(0.94 – 1.06) 0.513 1.02(0.95 – 1.06) 0.469 1.02(0.94 – 1.06) 0.555 Doses amphotericin B 1.00(0.95 – 1.03) 0.990 1.00(0.95 – 1.03) 0.998 1.00(0.96 – 1.03) 0.981 1.00(0.95 – 1.03) 0.936 V5 (per 10%) 1.14(1.04 – 1.24) 0.004 V10 (per 10%) 1.12(1.01 – 1.23) 0.021 COG abdominal radiation therapy (yes/no) 1.49(0.76 – 2.85) 0.235 TBI or WA (yes/no) 2.73(1.08 – 6.61) 0.029
Real-world comparative safety and outcomes of ribociclib versus palbociclib in hormone receptor-positive, HER2-negative metastatic breast cancer: A propensity score–matched sensitivity analysis.
1099 Background: Cyclin-dependent kinase 4/6 (CDK4/6) inhibitors are widely used in hormone receptor-positive (HR+) / HER2-negative (HER2-) metastatic breast cancer, yet real-world differences in survival and clinically relevant toxicities across agents remain unclear. We performed a propensity score-matched (PSM) sensitivity analysis comparing ribociclib and palbociclib, focusing on mortality and selected safety endpoints. Methods: Using the TrinetX database, a federated real-world data network, adult patients with HR+/HER2- metastatic breast cancer treated with ribociclib (cohort 1) or palbociclib (cohort 2) were identified. 1:1 propensity score matching was performed on more than 20 baseline characteristics. Outcomes were evaluated using (1) risk analyses excluding patients with the outcome prior to the time window and (2) Kaplan-Meier time-to-event analyses with hazard ratios (HRs). Endpoints included all-cause mortality, prolonged QT, cardiac arrest, liver injury, hospital admission, and emergency department (ED) visits. Results: Patient counts were 2,819 vs 11,871 before matching and 2,696 vs 2,696 after matching. In the matched cohorts, ribociclib was associated with significantly lower all-cause mortality compared with palbociclib (HR 0.65, 95% CI 0.552-0.758, p <0.0001). No statistically significant differences were observed in major cardiovascular safety outcomes. Specifically, rates of cardiac arrest were comparable between groups (HR 0.76, 95% CI 0.384-1.505, p =0.43), as were rates of long QT syndrome (HR 0.837, 95% CI 0.425-1.648, p =0.61). With respect to hepatic and healthcare utilization outcomes, liver injury occurred numerically more often in the ribociclib group, but this difference did not reach statistical significance (HR 1.209, 95% CI 0.896-1.632, p =0.21). Ribociclib was associated with numerically lower rates of hospitalization (HR 0.845, 95% CI 0.700-1.021, p =0.08) and ED visits (HR 0.82, 95% CI 0.659-1.02, p =0.07), though neither outcome met conventional thresholds for statistical significance. Conclusions: In this PSM analysis of HR+/HER2- metastatic breast cancer, ribociclib was associated with significantly lower all-cause mortality compared with palbociclib. Rates of hospitalization, ED visits, prolonged QT, and cardiac arrest were not significantly different. These real-world findings support clinically meaningful heterogeneity across CDK4/6 inhibitors and highlight the importance of individualized agent selection, balancing survival, hepatic toxicity, and healthcare utilization.
Inspiring the next generation of medical oncologists: Passion, purpose, and preventing burnout.
9015 Background: The National Cancer Registry of Ireland projects a doubling of cancer incidence by 2045, with an estimated shortfall of over 100 medical oncologists by 2028. This workforce gap is compounded by rising burnout, with ASCO data showing an increase in oncologists reporting at least one burnout symptom from 34% to 59% over the past decade. Alongside sustained investment in training and recruitment, addressing burnout is essential. This study aimed to identify factors influencing career interest in medical oncology and to explore perceived barriers, informing recruitment strategies and burnout mitigation. Methods: A cross-sectional, anonymous electronic survey assessed interest in medical oncology among medical students, interns, and junior doctors. The 20-item questionnaire examined demographics, career preferences, oncology exposure, and perceived barriers. A separate 28-item follow-up survey evaluated burnout among Irish medical oncologists and trainees. Results: A total of 183 early-career respondents participated (62.3% female). A response rate of 73%, respectively. Most were aged 25–34 years (69.4%), with junior residents comprising the largest group (42.6%). Key motivators for considering oncology included interest in cancer biology and treatment (54.9%), caring for cancer patients (53.2%), managing complex multidisciplinary cases (52.4%), and the dynamic nature of the specialty (51.6%). Personal experience with cancer did not influence career choice for 54%. Deterrents included high patient mortality and emotional burden (71.6%), concerns regarding burnout and work–life balance (57.5%), and prolonged training (36.1%). Oncology exposure was limited, with 38.5% reporting none; exposure most commonly occurred during junior residency (34.3%) and centered on palliative care, systemic therapies, and chemotherapy. Access to mentorship (73.8%) and increased clinical or research opportunities (62.3%) were identified as key factors to enhance recruitment. Forty-one oncologists completed the burnout survey (responses on-going). Most were aged 35–44 years, female, and working in public hospitals, with 65.9% reporting weekly workloads of 50–59 hours. Over half reported burnout at least monthly, and 75.6% reported inadequate work–life balance. Major stressors included high clinical workload, administrative burden, lack of protected academic time, staffing shortages, and moral distress. Despite high professional satisfaction (82.9%), 43.9% had considered reducing hours or leaving practice due to work-related stress. Conclusions: Ireland faces a significant oncology workforce challenge. Early exposure, structured mentorship, and targeted strategies to address workload and burnout are critical to strengthening recruitment and sustaining the oncology workforce.
Immunogenicity profiles in pediatric extra-cranial rhabdoid tumors: A framework for future risk stratification.
10026 Background: Malignant rhabdoid tumors (MRTs) are highly aggressive pediatric cancers driven by SMARCB1 loss. Emerging evidence suggests that the tumor immune microenvironment may influence the biological behavior of tumors. However, data on immune patterns in extra-cranial MRTs remain extremely limited, particularly in low- and middle-income settings. This study characterizes immune marker expression and immune phenotypes in a large single-center cohort to explore their potential relevance for future risk stratification. Methods: Formalin-fixed paraffin-embedded specimens from pediatric extra-cranial MRTs were evaluated for CD4⁺, CD8⁺, and CD68⁺ immune infiltrates using immunohistochemistry. Marker expression was categorized as high or low. Tumor immune phenotypes were subsequently classified based on the balance of infiltrating populations into: (immune-cold (uniformly low), mixed (variable across markers), and inflamed (uniformly high). Associations between immune phenotypes, clinicopathologic features, and survival trends were explored to determine whether immune architecture could serve as a biologically meaningful correlate of clinical behavior. Results: Individually, none of the immune markers (CD4, CD8, CD68) showed statistically significant correlation with overall survival. However, higher CD4⁺ and CD8⁺ T-cell infiltration demonstrated a consistent trend toward improved clinical outcomes. The most informative observation emerged at the phenotype level: tumors with a mixed immune profile exhibited more favorable survival trends and delayed progression compared with immune-cold tumors, whereas fully inflamed tumors were rare. These findings suggest that the equilibrium and interaction of immune infiltrates—rather than absolute immune cell density—may shape clinical behavior. Conclusions: Pediatric extra-cranial MRTs demonstrate distinct immune microenvironment patterns that are readily identifiable using routine immunohistochemistry. Although individual immune markers did not independently predict survival, immune-cold versus mixed phenotypes reflected meaningful biological divergence and aligned with clinical outcomes. Integrating immune phenotyping into diagnostic workflows may provide a pragmatic foundation for future risk stratification and help identify patients most likely to benefit from emerging immuno-epigenetic treatment strategies. Statement of Clinical Significance: Immune profiling using routine immunohistochemistry can classify pediatric extra-cranial rhabdoid tumors into clinically relevant immune phenotypes. Identifying immune-cold versus mixed tumors may support future risk stratification and guide biologically informed treatment selection.
Medicaid expansion and timely colorectal cancer treatment initiation among young adults in Virginia.
e15679 Background: Colorectal cancer (CRC) remains a leading cause of cancer mortality in the United States, and incidence is rising among younger adults. Treatment within 30 days of CRC diagnosis is associated with higher survival; however, younger survivors, particularly those insured by Medicaid, may face barriers to timely care. We evaluated the association between insurance type and timely receipt of care (TRC) among younger survivors in Virginia. Methods: We conducted a retrospective cohort study using Virginia all-payer claims data (2016–2023). CRC survivors aged 18–54 enrolled in Medicaid or commercial insurance at diagnosis and initial treatment were identified. Survivors with continuous enrollment from diagnosis through 30 days post-diagnosis were included; individuals with any Medicare enrollment or death within 30 days were excluded. Bivariate comparisons (chi-square, t-tests) assessed unadjusted associations between insurance and TRC. Multivariable logistic regression assessed the association between insurance and TRC (treatment ≤30 days) adjusting for demographic and clinical factors (Charlson Comorbidity Index [CCI]). Results: Among 5,211 CRC survivors aged 18–54, most had commercial insurance (74.1% vs 25.3% Medicaid) and resided in urban areas (87.1%). One-half (50.3%) initiated CRC treatment; time to treatment did not differ by payer (mean 71.1 vs 75.9 days, p = 0.593; median 27 vs 26 days, p = 0.866). Overall, 27.4% initiated treatment within 30 days. In adjusted models, Medicaid insurance was associated with 24% lower odds of TRC compared with commercial coverage (aOR 0.76, 95% CI 0.64–0.92; p = 0.004). Adults aged < 40 had higher odds of TRC than those aged 50–54 (aOR 1.27, 95% CI 1.06–1.53); adults aged 40–49 also had higher odds (aOR 1.20, 95% CI 1.04–1.38). Higher comorbidity burden was associated with TRC (aOR 1.18, 95% CI 1.16–1.20). Survivors in large rural areas had lower odds of TRC than urban residents (aOR 0.55, 95% CI 0.38–0.78; p = 0.001). Conclusions: Among insured CRC survivors aged 18–54 in Virginia, Medicaid insurance was associated with lower odds of treatment within 30 days compared with commercial coverage. Younger age and higher comorbidity were associated with timely care. Adjusted odds ratios for timely CRC treatment initiation within 30 days among CRC survivors aged 18–54 (N=5,211). Covariates Adjusted OR 95% CI p-value Medicaid vs Commercial 0.762 0.635–0.918 0.0039 Age group 40–49 vs 50–54 1.199 1.042–1.379 0.0115 <40 vs 50–54 1.271 1.060–1.525 0.0097 Rurality Large rural vs Urban 0.549 0.384–0.784 0.001 Small rural vs Urban 0.868 0.653–1.145 0.3229 Isolated vs Urban 1.058 0.743–1.516 0.7567 CCI (per 1-unit increase) 1.178 1.156–1.200 <0.0001 Note: Models are also adjusted for race, sex, ethnicity; Estimates are not shown due to missingness >10%.
Gliofocus: A global, open-label, randomized phase 3 study comparing niraparib with temozolomide in newly diagnosed MGMT-unmethylated glioblastoma.
TPS2102 Background: Glioblastoma (GBM) remains a fatal malignancy despite multimodality therapy. Approximately 60% of GBMs lack O6-methylguanine-DNA methyltransferase (MGMT) promoter methylation, thus associated with resistance to alkylating chemotherapy and poor survival with median OS of 12.7 months. Poly(ADP-ribose) polymerase (PARP) is a central regulator of DNA damage repair and genomic stability in GBM and represents a rational therapeutic target, particularly in the setting of radiotherapy-induced DNA damage. Niraparib is a selective PARP1/2 inhibitor that has demonstrated excellent central nervous system penetration and favorable tumor pharmacokinetic and pharmacodynamic properties in previously reported early-phase GBM studies. These findings provided the scientific rationale for Gliofocus, a global phase 3 trial designed to evaluate whether replacing temozolomide with niraparib during and after radiotherapy improves outcomes in patients with newly diagnosed, MGMT-unmethylated GBM. Methods: Gliofocus is a phase 3, open-label, randomized, two-arm study enrolling total 450 adults with newly diagnosed, MGMT-unmethylated GBM. Eligibility requires histologic diagnosis of GBM per 2021 WHO classification, Karnofsky performance status ≥70, documented MGMT promoter unmethylated status by validated local testing, and no prior GBM-directed therapy other than biopsy or resection. Patients are randomized 1:1 to receive niraparib or temozolomide concurrently with external-beam radiotherapy (60 Gy in 30 fractions using the ESTRO-EANO single-target volume approach), followed by adjuvant monotherapy of niraparib or temozolomide. Treatment continues until centrally reviewed disease progression per RANO 2.0 guidelines or completion of six cycles of temozolomide in the control arm. The primary endpoint is overall survival (HR = 0.698, 88% power). Secondary endpoints include progression-free survival, objective response rate, health-related quality of life, neurocognitive outcomes, and safety. Integrated correlative assessments include centralized imaging review, longitudinal neurocognitive testing, and patient-reported outcomes. Enrollment began in June 2024. The independent data monitoring committee last reviewed the study in January 2026, and recommended continuation as planned. The trial is sponsored by the Ivy Brain Tumor Center with funding/product support from GSK (NCT06388733). GSK was provided the opportunity to provide a courtesy review of the abstract; however, the authors are solely responsible for final content. Gliofocus study is being conducted across a minimum of 80 sites across 12 countries. Clinical trial information: NCT06388733 .
Racial disparities in utilization of bispecific T-cell engager therapy for relapsed/refractory multiple myeloma: A TriNetX study.
7545 Background: There have been remarkable advancements in the treatment of multiple myeloma over the past few decades. With the advent of cell redirecting immunotherapies, such as bispecific T-cell engagers (BiTEs), we have significantly improved rates of progression-free survival. As a result, physicians now prefer early utilization of these therapies in Relapsed/Refractory Multiple Myeloma (RRMM) due to their efficacy. This study aims to investigate the utilization and clinical outcomes of bispecific therapy based on race using a real-world database. Methods: We conducted a retrospective cohort study using the TriNetX US Collaborative Network, including adult patients (≥18 years) with RRMM treated with teclistamab, talquetamab, elranatamab, or linvoseltamab. Outcomes were assessed from 30 days to 5 years following the index diagnosis of RRMM. Propensity score matching (1:1) was performed to balance demographics, comorbidities, prior therapies, transplant status, medications, and laboratory parameters. Kaplan–Meier analyses were used to assess overall survival. Outcomes included hospitalization, overall survival, cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS) rates. Results: A total of 25,225 patients with RRMM were identified. 67.3% were White, 19.5% were Black, and the remainder were classified as other races, including Asian or unknown. Among 1,889 patients treated with bispecific antibodies, 1,256 (66.5%) were White and 403 (21.3%) were Black patients. The median follow-up was 854 days for White patients and 785 days for Black patients. In the unadjusted analysis, mortality was similar between Black and White patients (21.7% vs 23.7%, p = 0.43), with no difference in overall survival (HR 0.98; log-rank p = 0.88). Hospitalization rates (34.6% vs 39.0%), CRS and ICANS were comparable. After propensity score matching, outcomes remained consistent. Mortality was numerically lower among Black patients (20.3% vs 25.4%), though not statistically significant (p = 0.11). Hospitalization, CRS and ICANS did not differ significantly between cohorts. Conclusions: Our findings showed no significant difference in mortality or five-year survival probability between Black and White patients. The incidence of multiple myeloma is more than twice as high among Black vs. White people (14 vs. 6.1 per 100,000). Despite this, White patients are disproportionately more likely to receive BiTE therapy than Black patients. This highlights the impact that lack of access has on myeloma outcomes for Black patients, likely driven by systemic inequities such as socioeconomic status, geographic barriers to access therapy, the requirement of a full-time caregiver and lack of knowledge about newer therapies available. Future efforts should focus on expanding access to BiTE therapy to promote equitable care for Black patients.
Early changes in blood-based immune gene expression as a predictor of outcomes of <sup>177</sup> LuPSMA radioligand therapy in metastatic castration-resistant prostate cancer.
e17066 Background: Metastatic castration-resistant prostate cancer (mCRPC) is characterized by an immunosuppressive tumor immune microenvironment (TIME), limiting the effectiveness of immune checkpoint inhibition. We developed a 38-gene blood-based immune expression panel to characterize systemic immune profiles in prostate cancer treated with ¹⁷⁷Lu-PSMA radioligand therapy (RLT). This study assessed baseline immune profiles across disease states and examined whether early post-RLT transcriptional changes are associated with clinical outcome. Methods: Blood samples were collected from age-matched controls ( n = 60), newly diagnosed PCa ( n = 22), mCRPC ( n = 29) and a subset of mCRPC patients following 1 cycle of 177 Lu-PSMA ( n = 10). Immune gene expression was measured using qPCR and TaqMan primers targeting 38 immune genes and 6 housekeeping genes, normalized by ΔCt. Differential expression was evaluated across cohorts and pre/post-RLT. Clinical outcome analyses compared responders (n = 5; OS 327–1784 days) versus non-responders (n = 5; OS 191–244 days). Statistical analyses included non-parametric testing and Chi-square. Results: Relative to controls, 19 genes were differentially expressed in PCa and 26 in mCRPC, predominantly reflecting immune suppression. Of the 19 genes, 15 were downregulated and 4 upregulated and out of the 26, 19 were downregulated and 7 upregulated. CTLA4 and PD-L1 were both significantly reduced (0.1–0.5-fold, p < 0.05) in PCa and mCRPC compared with controls. Following a single RLT cycle, CTLA4 expression normalized in mCRPC, accompanied by restoration of the associated transcription factor BATF , while PD-L1 remained unchanged. Overall, RLT reversed 79% (15/19; p < 0.05) of downregulated immune genes. Among treated patients, responders demonstrated 2–4-fold increases ( p < 0.05) in nine immune genes, including markers of T-cell infiltration, antigen presentation, cross-presentation, and macrophage activation, compared with patients experiencing early mortality. Conclusions: This blood-based 38-gene immune panel detects systemic immune dysregulation in prostate cancer and captures rapid immunomodulatory effects of ¹⁷⁷Lu-PSMA RLT. Early transcriptional changes—particularly restoration of T-cell and macrophage activation pathways—were associated with improved survival, supporting the potential utility of this assay as a non-invasive pharmacodynamic and response biomarker. These findings suggest a role for immune profiling to identify patients most likely to benefit from RLT and to inform rational combination strategies with immunotherapy.
Efficient I <sup>−</sup> /I <sub>3</sub> <sup>−</sup> /I <sub>2</sub> Conversion and Shuttle‐Suppression in High‐Rate Ah‐Level Zinc‐Iodine Batteries Enabled by Bifunctional Confined‐Catalyst ZrB <sub>2</sub> /AC Host
ABSTRACT Rechargeable aqueous Zn||I 2 batteries have received significant attention for the high energy density, inherent safety, and ecological friendliness. However, the sluggish iodine redox kinetics and the formation of highly soluble polyiodide (I 3 − and I 5 − ) seriously affect the rate capability and cycling stability of Zn||I 2 battery. Herein, ZrB 2 is synthesized as a catalyst host for polyiodide conversion, where unsaturated surface Zr atoms provide Lewis‐acid sites that enable strong interaction with polyiodide species. These sites not only confine polyiodide anions through strong adsorption to suppress polyiodide shuttling, but also catalyze their stepwise conversion from I − to I 3 − and further to I 2 during charging, thereby reducing the energy barrier and improving iodine utilization. Consequently, the Zn||I 2 @ZrB 2 /AC full cell displays remarkable rate performance, retaining 80.5% of the capacity as the current density increases from 1 to 10 A g −1 . Remarkably, the full cells exhibit exceptionally long cycle life of 50,000 cycles with negligible capacity decay. The Zn||I 2 @ZrB 2 /AC pouch cell with a high I 2 mass loading of 33 mg cm −2 maintains a stable capacity of 1.2 Ah over 600 cycles (more than 1400 h).
A Scalable Perovskite Platform With Multi‐State Photoresponsivity for In‐Sensor Saliency Detection
ABSTRACT Artificial vision systems are increasingly central to edge intelligence, yet they often suffer from high data latency and energy consumption due to sensor‐processor separation. In‐sensor computing (ISC) provides a promising solution by integrating sensing and computation. However, current ISC devices remain constrained by scalability, uniformity, and processability. Here, we address these limitations via a reconfigurable perovskite‐photovoltaic platform that can be facilely processed from solutions. This architecture allows precise, reconfigurable photoresponsivity tuning with ultra‐low variability and supports fabrication on both rigid and flexible substrates. The device exhibits up to ±1120 mA W − 1 photoresponsivity and 1000 programmable states, with excellent air stability (30 days) and thermal reliability (80°C). The scalability of these devices is demonstrated via a proof of concept 32 × 32 array. The excellent uniformity and programmability of the array are utilized in energy‐efficient face detection applications (achieving 95.2% sensitivity and 4.51 × speedup for subsequent computation) in addition to image feature extraction and MNIST digit recognition tasks (96.97% accuracy). Compared to previous ISC implementations, our system offers enhanced tunability, fabrication scalability, and functional stability. These results establish a practical perovskite‐based ISC platform, offering new avenues for intelligent computing systems in robotics, wearable electronics, and neuromorphic vision.
Survival trends and subsite variation in mucosal melanoma: A SEER database analysis.
e21606 Background: Mucosal melanomas are rare neoplasms, representing approximately 2% of all melanoma cases. They often present at more advanced stages, and carcinogenesis is poorly understood. This study aimed to characterize current survival trends in mucosal melanoma across primary subsites and compare outcomes with its cutaneous counterpart using the Surveillance, Epidemiology, and End Results (SEER) Database. Methods: SEER Research Plus Data (17 Registries, 2000-2022) was used to extract all melanoma cases with ICD codes 8720-8790. Mucosal cases were subdivided into oral/pharyngeal (OP), airway/sinonasal (AS), gastrointestinal (GI), genitourinary (GU), renal/urinary, and lacrimal sites using ICD-O-3 primary site codes. Overall survival (OS) was estimated using Kaplan-Meier methods and compared using the log-rank test. Multivariable Cox models estimated hazard ratios HR) for OS, adjusting for age, sex, race, diagnostic era, and SEER summary stage. Results: Among 464,736 melanoma tumor cases, 447,387 were cutaneous and 4,910 were mucosal. Median age at diagnosis was 70-74 years. Mucosal melanoma was most common in Caucasians (73.5%), followed by Hispanic (12.2%), Asian (8.0%), Black (5.3%), and American Indian/Alaska Native (0.7%) patients. GU sites comprised 39% of mucosal cases, followed by AS (26.8%), GI (24.0%), and OP sites (7.8%). Renal/urinary and lacrimal primaries were rare ( < 3%). In unadjusted analyses, median OS was shorter for mucosal compared with cutaneous melanoma (24 vs 240 months; log-rank p < 0.001). After adjustment, mucosal melanoma remained independently associated with worse OS compared with cutaneous melanoma (HR 2.33, 95% CI 2.21-2.45; p < 0.001). Comparing mucosal sites, GU had the longest median OS (36 months; 95% CI 31-41), followed by OP (28; 95% CI 22-39), AS (21; 95% CI 19-24), and GI (16; 95% CI 15-18). In multivariable analysis, GI were independently associated with worse OS compared with OP (HR 1.35; 95% CI 1.17-1.56), while AS and GU were not significantly different. Conclusions: This analysis reaffirms the rarity of mucosal melanoma and highlights its markedly poorer prognosis compared with cutaneous melanoma. Survival differed significantly by primary subsite, with GU and OP primaries demonstrating the most favorable outcomes and GI primaries the poorest. Despite its rarity, persistently poor survival underscores the need for further investigation into mucosal melanoma pathogenesis and site-specific treatment strategies. Overall survival by primary mucosal melanoma subsite. Group / Subsite N patients Median OS, months (95% CI) 5-year OS, % (95% CI) Cutaneous melanoma 447,387 240 (238-242) 80.0 (79.8-80.1) Mucosal melanoma (overall) 4,910 24 (23-26) 29.9 (28.4-31.4) Oral/Pharyngeal 416 28 (22-39) 31.7 (26.7-37.7) Airway/Sinonasal 1296 21 (19-24) 25.9 (23.3-28.7) Gastrointestinal 1172 16 (15-18) 18.8 (16.4-21.6) Genitourinary 1909 36 (31-41) 39.0 (36.6-41.6)
Population-based survival and demographic disparities in acute megakaryoblastic leukemia with t(1;22)(p13;q13); RBM15–MKL1: A SEER analysis (2000–2021).
e18550 Background: Acute megakaryoblastic leukemia (AMKL) with t(1;22)(p13;q13); RBM15–MKL1 is a rare, genetically defined AML subtype, predominantly affecting infants and young children. Prior literature consists largely of case reports and small series, with no population-based studies describing survival patterns or demographic disparities. Understanding national-level outcomes is critical to guide treatment strategies and improve health equity. Methods: SEER data (2000–2021) were queried for patients with AMKL harboring t(1;22); RBM15–MKL1. Overall survival (OS) was estimated using Kaplan–Meier methods. A multivariable Cox proportional hazards model assessed the impact of age at diagnosis, sex, race, year of diagnosis, and receipt of chemotherapy on OS. Results: Among 79 patients, 55 deaths occurred during follow-up. Increasing age was associated with inferior OS: patients aged 50–79 years had significantly higher mortality compared with those <50 years (HR 3.01; 95% CI, 1.14–7.92); p = 0.026. Patients ≥80 years showed an elevated risk (HR 1.93) that did not reach statistical significance (p = 0.3), likely due to small sample size. Receipt of chemotherapy was associated with markedly improved survival, reducing the hazard of death by approximately 83% (HR 0.17; 95% CI, 0.07–0.43); p < 0.001 (Table). No significant differences were observed by sex (HR 1.00; p > 0.9), race/ethnicity, or year of diagnosis, although power was limited for subgroup analyses. Conclusions: This population-based analysis provides the first national-level survival estimates for AMKL with t(1;22); RBM15–MKL1. Age at diagnosis is a key prognostic factor, while chemotherapy confers substantial survival benefit. No clear disparities by sex or race/ethnicity were observed, but small numbers limit interpretation. These findings reinforce the importance of systemic therapy and provide a benchmark for future clinical and translational studies in AMKL. Multivariable Cox regression for overall survival in t(1;22) acute megakaryoblastic leukemia. Characteristic Hazard ratio (95% CI) P value Age at diagnosis (ref: <50 years) 50–79 years 3.01 (1.14–7.92) 0.026 ≥80 years 1.93 (0.52–7.14) 0.30 Sex (ref: female) Male 1.00 (0.55–1.81) >0.90 Year of diagnosis (continuous) 1.00 (0.93–1.09) >0.90 Chemotherapy (ref: no/unknown) Yes 0.17 (0.07–0.43) <0.001
Pancreatic cancer incidence and survival in lean and obese adults with type 2 diabetes: A propensity matched analysis.
10576 Background: Type 2 diabetes mellitus (T2DM) and excess adiposity are independent risk factors for pancreatic cancer. Pancreatic cancer risk within T2DM across body-mass index (BMI) strata is heterogeneous in real-world cohorts and lean T2DM individuals may represent a distinct high risk phenotype. We compared pancreatic cancer incidence and post-diagnosis survival across BMI strata. Methods: We performed a retrospective comparative outcomes analysis using the TriNetX network. Adults with T2DM (HbA1c ≥6.5%) were categorized as obese (BMI ≥30 kg/m²) or lean (BMI <25 kg/m²). The lean cohort additionally excluded any patient with prior BMI of >25 kg/m² to control for cancer associated weight loss. The index event was first T2DM diagnosis, and outcomes were evaluated after a 365-day window. Patients with pancreatic cancer prior to the outcome window were excluded. Propensity score matching was done using the following parameters: age, sex, smoking status, pancreatitis, pancreatic cysts, chronic kidney disease, gallstones, family history of pancreatic cancer, and HbA1c. Outcomes assessed were incident pancreatic cancer over 10 years and all-cause mortality among patients with pancreatic cancer and T2DM. Results: Total number of patients in the obese group was 1,759,075 with mean age of 57 ± 13.4 years and in the lean cohort were 205,193 with a mean age of 61.8 ± 16.6 years. After propensity score matching, there were 204,086 patients in each group. Mean HbA1c levels were 7.4 ± 1.8% in the obese group and 7.9 ± 2.2% in the lean BMI group. The incidence of pancreatic cancer was 0.16% (n = 335) in obese patients as compared to 0.23% in lean (n = 478) patients with a risk difference of 0.07% (p <0.001). Obesity was associated with lower pancreatic cancer risk (RR 0.692, 95% CI 0.603–0.797; p < 0.001). In the subgroup analysis of patients diagnosed with pancreatic cancer, obese patients had longer median survival (1,109 vs 651 days) with HR 0.853 (95% CI 0.793 - 0.918; p <0.001). Conclusions: Lean adults with T2DM had a higher incidence of pancreatic cancer and worse post-diagnosis survival than obese patients. These results identify lean T2DM as a distinct high-risk phenotype and demonstrate that pancreatic cancer risk within T2DM is heterogeneous and not solely attributable to excess adiposity. Further investigation is needed to elucidate the biological mechanisms underlying this association and its implications for clinical risk assessment. Pancreatic cancer incidence and associated risk estimates among obese and lean adults with type 2 diabetes mellitus. Cohort Pancreatic cancer incidence RD RR Obese T2DM(n = 204,086) 0.16% (n = 335) 0.07% (95% CI: 0.04 - 0.10), p < 0.001 1.44 (95% CI: 1.25 - 1.66) Lean T2DM(n = 204,086) 0.23% (n = 478) Abbreviations: CI, confidence interval; RD, risk difference; RR, relative risk; T2DM, type 2 diabetes mellitus.
Artificial intelligence–based predictive models for recurrence and recurrence-free survival in gastrointestinal stromal tumors: A contemporary systematic review.
11536 Background: Recurrence after complete resection of GISTs is still a major contributing factor to long-term survival. Traditional stratification models for risk depend on clinicopathologic variables and have limited precision. Recent improvements in artificial intelligence (AI), such as machine learning (ML), deep learning (DL), and multimodal methodologies, might enhance personalized prediction of recurrence and recurrence-free survival (RFS). We performed a contemporary systematic review to evaluate the performance and clinical utility of AI-based prognostic models in resected GIST. Methods: PRISMA 2020 guidelines were followed for conducting a systematic review. PubMed, Scopus, and Web of Science were analyzed to search published studies published during 2019–2025 that evaluated AI-, ML-, or DL-based models predicting recurrence or RFS following complete surgical resection of localized primary GISTs. Eligible studies included retrospective or multicenter cohorts with radiologic, histopathologic, genomic, or multimodal data. Data extracted included model type, input modalities, validation strategy, and performance metrics (area under the curve [AUC] and concordance index [C−index]). Results: Eight studies encompassing approximately 4,000 patients met inclusion criteria. Both deep learning and multimodal fusion models showed the highest prognostic accuracy (C-index values 0.86–0.96, AUC up to 0.995). Radiomics based ML models using CT, MRI, or ultrasound yielded AUCs between 0.85 and 0.92 which were consistently superior to traditional clinicopathologic methodologies. Genomic ML models fine-tuned recurrence risk stratification beyond conventional benchmarks into molecularly distinct prognostic subgroups. DL-based histopathology estimators both predicted for RFS and for key driver mutations (KIT, PDGFRA) by bridging morphologic and molecular features. External validation cohorts showed stable performance (AUC 0.87–0.96), with calibration and decision curve analyses supporting clinical utility. Conclusions: AI-based predictive models demonstrate strong and reproducible performance for recurrence and RFS prediction following GIST resection, consistently outperforming traditional risk stratification systems. Multimodal and deep learning approaches integrating radiologic, pathologic, and genomic data appear most promising for precision prognostication and personalized adjuvant therapy selection. Prospective validation and explainable AI integration are needed prior to routine clinical adoption.
Mortality trends involving colon cancer and atrial fibrillation in the United States,1999-2023: A CDC WONDER analysis.
e24021 Background: Colon cancer is a leading cause of cancer-related mortality in the United States, while atrial fibrillation (AF) is prevalent among older, medically complex patients. AF may complicate cancer management through shared risk factors, treatment-related cardio toxicity, and competing mortality risks. However national long-term mortality trends involving colon cancer and AF remain poorly defined. Methods: U.S. mortality data was analyzed from 1999-2023 using the CDC WONDER database. Deaths involving colon cancer (ICD-10: C18) and AF (ICD-10: I48) were identified. Age-adjusted mortality rates were stratified by ten-year age group, sex, race, census region, place of death and urbanization. Joinpoint regression assessed temporal trends for ten-year age group, sex, census region, and urbanization, reporting annual percentage change (APC) and average annual percentage change (AAPC). Race and place of death were analyzed descriptively. Results: Overall mortality involving colon cancer and AF accelerated around 2015-2016. The steepest increases occurred in adults aged 55-64, with deaths rising (APC 2.6%) annually from (1999-2016), followed by post-2016 acceleration of (APC 12.2% per year), resulting in an upward trend (AAPC 5.3%, p < 0.001). Adults aged 65-74 also showed post-2015 increases (APC 6.7% per year).Trends among older age groups were more gradual. Post-2016 increases in mortality were observed for both males and females (APC 4-5% per year), indicating no sex-specific differences.The Southern U.S. showed acceleration compared to other regions (AAPC 2.5%, p < 0.001). Urbanization analysis (1999-2020) showed increasing mortality in nonmetropolitan areas (AAPC 2.3%, p < 0.001). Deaths predominated among white individuals and in inpatient hospital settings. Conclusions: Mortality involving colon cancer and AF has risen substantially in the U.S. since 2016, shifting towards middle-aged adults and nonmetropolitan populations. This highlights a widening "cardio-oncology" gap underscoring the need for earlier risk recognition, improved cancer screening, integrated care, and targeted resource allocation in high-burden regions.
Real-time liquid biopsy guided assessment of radiation response: Longitudinal circulating free DNA kinetics during precision radiotherapy across solid tumors from India.
e15056 Background: Precision radiotherapy has improved dose delivery, but biological response assessment is delayed, especially in high-volume Indian settings where early recognition of radio-resistance could alter outcomes. Circulating free DNA (cfDNA) reflects dynamic tumor burden and may offer a real-time biomarker. We assessed cfDNA kinetics during precision radiotherapy to predict response across diverse solid tumors. Methods: In this prospective, single-centre study we enrolled 22 adults undergoing image-guided radiotherapy for lung, esophageal, cervical, gastric or breast cancers. Peripheral blood was drawn at baseline, mid-course and completion. cfDNA was isolated from plasma via standardized EDTA-centrifugation protocol and quantified fluorometrically. Changes were calculated relative to baseline. End-of-treatment cfDNA below baseline defined a responder; mid-course decline defined early decline. Radiologic response was evaluated at 6–8 weeks post-RT using RECIST 1.1. Results: Baseline cfDNA values were heterogeneous. During treatment, cfDNA decreased in 8 patients and increased in 14. At completion, 10 patients (45%) had cfDNA decrease and were radiologic responders, while 12 (55%) had an increase and were non-responders. Early decline identified 5 of 10 final responders and early rise identified 9 of 12 final non-responders, so the early cfDNA trend matched the final trend in 14 of 22 cases (64%). cfDNA response agreed with imaging in 19 of 22 patients (90% concordance). In the breast-cancer subset (14 patients), 7 experienced cfDNA decline by the end of therapy. Conclusions: Real-time cfDNA monitoring during precision radiotherapy is feasible and provides an early indicator of treatment response across heterogeneous Indian cancers. Early cfDNA rise predicted resistance, whereas decline correlated with radiologic response. Though limited by sample size, this study shows that cfDNA trends are simple to measure and could be integrated into weekly clinic visits to guide adaptive dosing, prompt systemic therapy initiation, and engage patients. Further multicenter studies are needed to validate thresholds, interpret dynamics, and integrate cfDNA into personalized radiation oncology.
Impact of the 2022 NCCN Guideline Update on germline testing in early-onset colorectal cancer: Insights from a large tertiary referral center.
e15741 Background: In 2022, the NCCN expanded guidelines to recommend universal germline genetic testing for patients diagnosed with early-onset colorectal cancer (EOCRC), independent of family history or tumor-based criteria. This change was prompted by evidence that approximately one in six patients with EOCRC harbor pathogenic germline variants. Real-world utilization and timeliness of germline testing following this guideline expansion remain incompletely characterized. We hypothesized that EOCRC patients diagnosed after the 2022 NCCN update would have higher rates of genetic counseling, germline testing completion, and shorter time to testing compared with patients diagnosed previously. Methods: We conducted a retrospective cohort study of patients with EOCRC who presented to Moffitt Cancer Center between 2016-2025 (n = 140). Patients were stratified by diagnosis date relative to the date the expanded NCCN guidelines were published (pre-4/2022 vs post-4/2022). The primary endpoint was completion of germline genetic counseling, and key secondary endpoint was time from diagnosis to testing completion. Comparisons between cohorts were performed using nonparametric statistical tests, with p < 0.05 considered significant. Results: Of the 140 patients with EOCRC, genetic counseling was completed in 49/101 (48.5%) diagnosed before 4/2022 compared with 29/39 (74.4%) diagnosed after 4/2022 (p = 0.0029). Restricting the pre-guideline cohort to 1/2019–4/2022 yielded similar results (55.6% vs 74.4%, p = 0.0316). Median time from diagnosis to testing completion decreased following guideline expansion (5.00 months vs 2.07 months, p = 0.0047). In the 1/2019-4/2022 cohort, median time to testing was 3.63 months, significantly longer than post-4/2022 (p = 0.043). The rate of genetic counseling rose significantly following the publication of updated NCCN guidelines.from 56.4% in 2019-2022 to 80.6% in 2023-2025 (p = 0.013). Conclusions: Our findings demonstrate improved utilization and timeliness of germline genetic testing following the 2022 NCCN guideline update. Rate of genetic counseling increased, and time from diagnosis to testing was shorter. Pandemic-related disruptions may have contributed to delays, but universal testing was not achieved even post-COVID. Documented barriers included transitions to hospice, missed appointments, and incomplete multi-step testing workflows. Timely germline testing is critical as results inform early management decisions and enable cascade testing for at-risk relatives. Future efforts should focus on system-level implementation strategies such as opt-out testing, point-of-care consent and sample collection, automated referral workflows, and remote genetic counseling models to improve equitable and timely germline testing in EOCRC.