Citations

INTRODUCTION: To assess strategies for optimizing participation of underserved minorities in a blood-based early colorectal cancer detection test study (PREEMPT CRC; NCT04369053) at a hospital serving primarily Black individuals. METHODS: Culturally sensitive, racially congruent research staff approached individuals undergoing average-risk screening colonoscopy. Consent/study procedures were synchronized with clinical appointments. Enrolled and not-enrolled patient characteristics were compared. Recruitment was compared with other study sites. RESULTS: In total, 247 of 509 eligible individuals were enrolled; most were identified as Black (88.7%). No baseline characteristics were associated with participation. Recruitment was high compared with other sites (11th centile). DISCUSSION: Recruitment barriers for Black individuals can be overcome when easy, culturally sensitive access is facilitated.

Purpose: This study compares the performance of small, fine-tuned LLMs with the large original models on two radiation oncology clinical tasks: treatment modality selection and ICD-10 code prediction. Methods: We used data of 403 treatment modality selection cases and 377 ICD-10 code prediction cases. The models used were fine-tuned LLaMA 2 7B and Mistral 7B, as well as the larger LLaMA 3.1 70B and 405B models for comparison. We evaluated the accuracy of all models on these two tasks. Results: Both fine-tuned LLaMA 2 7B and Mistral 7B outperformed LLaMA 3.1 70B and 405B in treatment modality selection task. The fine-tuned LLaMA 2 7B and Mistral 7B models achieved accuracy comparable to the larger models in ICD-10 code prediction. Conclusions: Small models, fine-tuned with radiation oncology data, can achieve performance comparable to-or even surpass-that of larger models.

801 Background: SBA is a rare, poorly understood gastrointestinal malignancy. Adenocarcinomas are often grouped with other small bowel cancer histologies in epidemiologic and outcomes reports, which limits the ability to detect SBA-specific trends. Recent studies in European populations show increasing SBA incidence, but there are limited published data in American cohorts. Identifying trends in SBA can improve understanding of the population most at risk for developing SBA. This study describes U.S. trends in SBA incidence and outcomes from 2000 to 2019. Methods: Using the 2022 SEER research-limited dataset, we identified SBA cases and measured the incidence change from 2000-2019. We built linear regression models to evaluate changes in annual incidence by groups, including sex, race, and age of diagnosis, separately. We then compared the overall incidence and trends between subgroups. We investigated overall survival (OS) through the SEER research case listings for SBA patients diagnosed during the study period. Results: An average of 0.67 cases/100,000 SBA cases were reported annually from 2000-2019, with a median diagnosis age of 69 years. SBA incidence has increased annually by 0.3% per year from 2000-2019 (p=0.05). From 2000-2019, men were at a higher risk for SBA compared to women (p=0.02). In men, SBA cases increased at a rate of 0.004 cases/100,000/year from 2000-2019 (p=0.01) and a nonsignificant decrease in women of 0.00012 cases/100,000/year. There were no statistically significant differences in incidence or the change in incidence of SBA by race. The age-adjusted incidence was significantly higher each year between 2000 and 2019 in the > 70 cohort compared to the <70 cohort (p=0.02). Cases of SBA increased at a rate of 0.02 cases/100,000/year in adults >70 years (p=0.02). There were no differences in incidence rates in patients <70 years. There was no significant difference in OS for patients diagnosed with SBA in 2000, 2005, 2010, 2015, and 2019 (p=0.21). Conclusions: The overall annual incidence of SBA is increasing in the U.S. population. The most significant rate increases are in men and adults older than 70. In the SEER population, trends in SBA contrast with the recent rise of colorectal cancer in the young adult population, highlighting the unique nature of SBA compared to other gastrointestinal malignancies. Additionally, survival outcomes in patients diagnosed with SBA have remained stable over the past 20 years, while other malignancies have shown improvement in the same period. The small cohort size and the lack of a universal U.S. cancer registry limited this evaluation. Improved reporting of SBA cases and further research to identify risk factors are urgently needed to reduce the growing impact of SBA.

Background Ischemic stroke remains a leading cause of morbidity and mortality worldwide. Understanding stroke etiology is critical in future stroke prevention. A cardioembolic etiology of stroke makes up a significant proportion of all ischemic stroke. Part of the evaluation for a cardioembolic etiology requires a structural evaluation of the heart to evaluate for sources of thrombus. Traditionally, transthoracic echocardiography (TTE) and transesophageal echocardiography (TEE) have been utilized for this purpose, although both have limitations. Cardiac computed tomography (CCT) has emerged as a potential alternative, offering non‐invasive imaging with high sensitivity and specificity for detecting intracardiac thrombi. This study aimed to evaluate the diagnostic yield of CCT compared to TTE and TEE in detecting high‐risk cardioaortic sources of embolism, including left ventricular (LV) thrombus, left atrial (LA) thrombus, valvular sources of thrombus, and ulcerated aortic plaques, in patients with ischemic stroke. Methods We conducted a retrospective, single‐center cohort study at Mayo Clinic, Rochester, MN, on patients who underwent CCT within 30 days of stroke onset between March 2019 and June 2023. We compared the diagnostic yields of CCT, TTE, and TEE in detecting structural cardiac sources of embolism. Results 426 patients were included in our study. 87 patients underwent both CCT and TEE, 257 underwent CCT and TTE, and 82 only underwent CCT. The overall diagnostic yield of CCT for detecting high‐risk cardioaortic sources was 15.5%. Specifically, CCT identified LV thrombus in 2.3% of patients, LA thrombus in 4.0%, valvular thrombus in 2.8%, and ulcerated aortic plaques in 7.0%. In comparative analysis, CCT demonstrated a higher yield than TTE (16.0% vs. 3.5%) and a comparable or higher yield than TEE (25.3% vs. 20.7%) in detecting these sources. Conclusions CCT proved to be a viable and potentially superior alternative to traditional echocardiography for identifying some high‐risk structural sources of cardioembolism in ischemic stroke patients. Its non‐invasive nature, coupled with the ability to perform imaging soon after stroke onset, underscores its clinical utility. However, further prospective and blinded studies are required to validate these findings and fully establish the efficacy of CCT in this context.