Oncology Clinical Trials Research Articles

Heterogeneity in methods for estimating renal function in anticancer drug development.

e23002 Background: Accurate assessment of renal function is critical in oncology clinical trials. Many anti-cancer agents are renally excreted or cleared, and impaired kidney function can significantly impact drug safety and efficacy. In an effort to modernize renal eligibility criteria the American Society of Clinical Oncology–Friends of Cancer Research Organ Dysfunction Working Group recommended that if renal toxicity and clearance are not of direct treatment-related concern, then patients with impaired renal function should be included in trials (Lichtman, JCO, 2017). However, a diversity of approaches are available to determine renal function. Methods: We systematically searched the clinicaltrials.gov database to review renal eligibility criteria for phase III clinical trials. These trials evaluated systemic anti-cancer therapies in adult patients between October 28th 2013, and October 28th 2023. Studies were excluded if they did not include eligibility criteria related to renal function. The demographics of the trials and methods used for estimating renal function were recorded. Results: 436 trials met criteria for review and 231 trials, enrolling 111,424 patients, met our inclusion criteria. Systemic therapies explored included: immunotherapy (98 trials, 42%), targeted therapy (110 trials, 48%), chemotherapy (17 trials, 7%) or other regimens (6 trials, 3%). Renal function exclusion criteria are shown in the Table. Among studies that estimated renal function with a formula, 140 (61%) studies used Cockcroft-Gault, 3 (1%) used Modification of Diet in Renal Disease, 9 (4%) used an institutional standard, 10 (4%) used a heterogenous combination of estimation equations, and 69 (30%) trials did not specify the estimation approach. Conclusions: In phase III clinical trials involving adult cancer patients, there is substantial heterogeneity in methodologies utilized for determining renal function as related to trial eligibility. Harmonization of approaches to estimate renal function in oncology trials may improve trial accessibility, generalizability, and comparability. [Table: see text]

Cancer Trials Ecosystem in India-Ready for Prime Time?

Executing global clinical trials for cancer is a long, expensive, and complex undertaking. While selecting countries global studies, sponsors must consider several aspects including patient pool, quality of trained investigators, competing trials, availability of infrastructure, and financial investment versus returns. With a large, often treatment-naïve, and diverse patient pool, relatively low cost, good quality health care facilities in urban areas, and a robust and well-trained workforce, India offers several advantages for conducting oncology clinical trials. However, there remains challenges, including a shifting regulatory environment in recent decades. With the implementation of the New Drugs and Clinical Trial Rules in 2019, India's regulatory atmosphere seems to have stabilized. In this article, we present a review of the evolving clinical trial landscape in India, highlight the current regulatory scenario, and discuss the advantages and challenges of selecting India as a potential location for conducting global oncology clinical trials.

The value of blinded independent central review (BICR) in recent trials in non-small cell lung cancer (NSCLC).

e23000 Background: Progression free survival (PFS) is an endpoint that has been increasingly used to obtain approval of new cancer drugs and strategies. However, there are a number of issues that can complicate assessment of PFS. One of these is the potential for investigator bias. This has been addressed by the use of blinded independent central review (BICR). However, BICR increases costs and complexity. Methods: The current study evaluated the results of BICR versus investigator assessment (IA) for phase III randomized clinical trials in thoracic oncology with initial publication from 1.01.2013-12.31.2023. Trials were included if they involved targeted agents, antibody drug conjugates or immunotherapy. Studies with < 100 pts or comparing only cytotoxic chemotherapy or maintenance strategies were excluded. Studies with cytotoxic chemotherapy as part of therapy or within the control arm were included. A search of clinicaltrials.gov with the terms “phase III”, “non-small cell lung cancer” for completed studies was done, supplemented by a bibliographic search. Publications were then reviewed for whether a BICR was utilized and whether both BICR and IA were reported. Hazard Ratios (HR) for PFS between IA vs BICR were compared. Results: 49 randomized phase III trials were identified of which 34 describe BICR of which 21 reported both IA and BICR results. Six studies evaluated immunotherapy and enrolled 3337 pts, 15 evaluated targeted agents and enrolled 6531 pts. All immunotherapy studies were significantly positive for the experimental arm and 14 of 15 of the targeted therapies were positive. PFS data was reported for 17 of the studies (not reached for experimental arms in 2 and not reported in 2 others). HRs were directionally similar in all studies, though were lower (more favorable) in IA vs BICR in 5 of 6 immunotherapy studies and 11 of 15 targeted therapy studies. The mean % difference for immunotherapy studies was 8.9% and the median 10%, range 2-23.5%. For the positive targeted therapies studies the mean/median differences were 14.6%/15.6%, range 1.5-42%. The largest % differences were seen at the lowest HRs (e.g..19 vs .27; 42%). For immunotherapy studies, the median PFS difference was 0.2 and 0.28 months for the control and experimental arms, respectively. For targeted therapy studies, median PFS differences were 0.4 and 0.95 months for the control and experimental arms, respectively. The studies were performed in multiple countries including the United States, European Union, China and Japan. Conclusions: Despite a wide range of agents and settings, no trial for which information is publicly available comparing BICR to IA has demonstrated any qualitative difference in outcome nor more than a marginal numerical difference.

Evaluating speaker gender in scientific sessions at ASCO and ASCO GI (2019-2023).

9009 Background: Gender disparities exist in medicine, including academic and medical society leadership, journal editorial boards, and research teams. Women are also underrepresented in clinical trial leadership and authorship, with one study reporting fewer than 10% female corresponding authors for randomized clinical trials in GI Oncology, GU oncology, and Hematology. Clinical trial leadership, subsequent podium presentations and authorship have downstream effects on promotion, reputation, and funding opportunities. We examined the gender of speakers in scientific sessions at the ASCO Annual Meeting (AM) and Gastrointestinal Symposium (ASCO GI) from 2019-2023. Methods: We examined speaker gender, role (session chair, abstract presenter, abstract discussant), speaker region (Asia, Europe, N. America), and presentation topic (colorectal cancer or non-colorectal cancer) for the ASCO AM GI scientific sessions and ASCO GI from 2019-2023. For the ASCO AM, we included GI oral abstracts and plenary sessions; for ASCO GI we included oral and rapid abstracts. Self-identified gender was not available; therefore, gender was assigned by physical appearance via ASCO video or institutional websites. Statistical analyses were performed using the two-proportions z-test with a p value <0.05 considered statistically significant. Results: 208 scientific abstracts were reviewed across both meetings over 5 years. There was a statistically significant difference in the % of men vs. women as abstract presenters (77% vs 23%, p <0.001). This did not differ when looking at ASCO and ASCO GI separately (ASCO 76% vs 24%, ASCO GI 78% vs 22%, p <0.001 for both) and was observed across all 5 years. Similarly, there was a predominance of men as abstract presenters regardless of speaker region (Asia 93%, Europe 74%, N. America 67%, p<0.001 for all) and presentation topic (CRC 76% vs 24%, non-CRC 78% vs. 22%; p <0.001 for both). The percent of men as repeat presenters for scientific abstracts was statistically higher than women (84% vs 16%, p <0.001). Importantly, there was no difference in men vs. women in the invited chair/discussant roles (51% vs 49%, p = 0.89). Conclusions: Our analysis of 5 years of scientific sessions for ASCO and ASCO GI shows a stark difference in the proportion of men vs women as abstract presenters regardless of presentation topic, region of presenter, and meeting year. Our data highlight the need for all trial sponsors to be more intentional in the selection of principal investigators and lead authors as these opportunities are critical to the promotion of women in GI Oncology and to diversifying approaches to clinical trial design and access. Our data is limited by how gender was assigned (i.e., not self-reported). Lastly, we commend ASCO on their efforts to ensure gender equity among the roles that are within their control, and we advocate for them to consider opportunities for self-identification of gender in their membership directory.

Development of an arenavirus-based immunotherapy for treatment of KRAS mutant cancer.

e14672 Background: Oncogenic mutations in KRAS are prevalent in more than 80% of pancreatic ductal adenocarcinomas (PDAC), approximately 30% of colorectal cancer (CRC) and 15-20% of lung adenocarcinoma (LUAD). It has been demonstrated previously that CD8+ T cell responses specific for KRAS mutations can lead to tumor control and regression of metastasis. Replicating arenavirus vectors are currently evaluated in multiple clinical trials in infectious diseases and oncology and have demonstrated potent induction of target antigen specific CD8+ T cell responses of up to almost 50% of the total CD8+ T cell compartment in peripheral blood. To augment KRAS mutation specific CD8+ T cell responses in patients, we are using replicating arenavirus vectors targeting five prevalent KRAS mutations which are found in approximately 95% (PDAC), 74% (CRC), and 83% (LUAD) of the KRAS mutated tumors, representing a promising off-the-shelf immunotherapy for these indications. Here, we systematically investigate immune responses induced by mutant KRAS targeted arenavirus-vector-based cancer vaccines in HLA transgenic mice and characterize the efficiency of target cell killing of activated KRAS mutation specific CD8+ T cells. Methods: A transgene cassette consisting of peptide stretches including KRAS mutations G12D, G12V, G12C, G12R and G13D was generated by in silico aided antigen design. Vectors encoding peptide stretches of single KRAS mutations as well as the corresponding wildtype KRAS sequence served as controls in preclinical experiments. KRAS mutation specific CD8+ T cell expansion was evaluated in HLA transgenic mice and functionality of induced CD8+ T cells was evaluated by assessing CD8+ T cell mediated killing of mutant KRAS peptide loaded target cells in vivo. Results: All treatment regimens were well tolerated, and no mortalities or major adverse events were observed. Following vector administration, HLA I restricted antigen-specific CD8+ T cell responses were detected against 4 encoded mutant KRAS epitopes in HLA-A*11:01 (KRAS G12D and G12V), or HLA-B*07:02 (KRAS G12C and G12R) transgenic mice. These cells did not cross-react with wildtype KRAS epitopes presented in the same HLA I molecules. Specific cytotoxicity against KRAS G12D/V or KRAS G12C/R pulsed target cells was observed in vivo in HLA transgenic animals, indicating functionality of the induced KRAS mutation specific T cells. No specific cytotoxicity towards target cells pulsed with KRAS WT peptides was observed in any of the groups. Conclusions: This study indicates efficient induction of KRAS mutation specific T cell responses in HLA transgenic mice using an arenavirus vector based vaccine. Expanded CD8+ T cells were capable of killing cells loaded with KRAS mutation specific epitopes in vivo indicating functionality of the induced T cell responses. Based on these results, initiation of a Phase I study for the treatment of KRAS mutated cancers is planned.

An Argument for Decentralized Clinical Trials in Gynecologic Oncology

To evaluate factors that may influence clinical trial enrollment, we compared demographics, clinical characteristics, and patient-reported social determinants of health for patients with advanced ovarian cancer who were approached for enrollment in a prehabilitation trial. Although no significant differences were found in the majority of demographic and clinical characteristics, distance to study institution was associated with decision to enroll; 21.9% of enrolled compared with 42.2% of not enrolled patients lived 200 miles or more from the study institution (P< .05). In addition, compared with not enrolled patients, enrolled patients were more likely to be receiving neoadjuvant chemotherapy within the study institution compared with a nonstudy institution clinician (81.2% vs 53.1%, P=.02). These findings highlight barriers to enrollment that can be addressed through decentralization of clinical trials.

Barriers to enrollment in clinical trials in gynecologic oncology.

e13511 Background: While clinical trials improve survival in gynecologic oncology, racial and ethnic minorities remain underrepresented. Our objective was to evaluate factors associated with enrollment among patients considered for trials at a tertiary care center to inform future participant recruitment. Methods: We conducted a retrospective cohort study at the University of Pennsylvania Health System from 1/1/2022-12/31/2022. Our primary objective was the referral rate to clinical trials of patients discussed at a weekly multi-disciplinary tumor board. Patients could be discussed more than once and referred for screening for multiple trials. Each contact was classified as ‘referred/not referred’ for trial(s) screening and ‘enrolled/not-enrolled’ on trial(s). We collected socio-demographic, cancer, trial screening, and enrollment data, including reasons for ineligibility or non-enrollment, from electronic health records. We compared patient factors across 3 subgroups: 1) not-referred, 2) referred not-enrolled, 3) referred and enrolled. We performed descriptive analyses to characterize patient cohorts. Results: In a one-year period, 264 patients—194 unique—were discussed at tumor board. Referral rate for trial screening was 57%, representing 135 trial referrals for 111 unique patients. Patients were referred to screening for 2.2 trials on average. The median age of patients discussed was 65 years (IQR 58-73 years). 65.0% self-identified as White, 24.7% Black, and 10.3% other race. 96% identified English as their preferred language. The majority had Medicare insurance (51.0%), followed by private insurance (32.5%), and Medicaid (14.4%). Of 111 patients referred for trial screening, 12 patients enrolled in a trial for a referral-to-enrollment rate of 10.8% (12/111) and an overall enrollment rate of 6.2%. The majority of enrolled patients were white (10/12) and had Medicare (66%) or private insurance (25%). Cancer type varied by referral status, with most referrals for ovarian (n=87, 64%) and uterine cancer (n=37, 27.4%) and 5 for cervical cancer (3.7%). The non-referred cohort had lower rates of ovarian cancer (49.6%) and higher rates of cervical (12.4%) and other cancers (14.0%). Of patients referred to trial, but not enrolled, only 33.7% had a documented reason for non-enrollment. Documentation for incomplete screening or ineligibility cited the patient opting out of screening (26), medical comorbidity (22), and pursuit of other trial (16). Conclusions: While 57% of tumor board patients were referred for screening for at least one clinical trial, only 10.8% of referred patients and 6.2% of patients overall were enrolled in a trial. Disparities by race and insurance type were apparent between patients enrolled and not enrolled on trial. Improving documentation, educating patients about trials early in their disease course, and less restrictive trial criteria may help improve minority enrollment rates in clinical trials.

PENTAGON (Predicting Clinical Trials in Gynecologic Oncology): A retrospective study assessing study design factors that affect enrollment in gynecologic cancer trials.

e13647 Background: Obstacles to attaining a diverse clinical trial population contribute to poor representation in medicine, increasing health disparities by limiting generalizability and applicability of findings. Our objective is to evaluate trial enrollment trends according to patient and trial demographics for those who have sought care at our multi-center, mixed-setting Gynecologic Oncology practice. Additionally, we aim to determine specific trial criteria that promote or impede enrollment for patients and physicians. Methods: An IRB-approved, retrospective cohort study evaluating all patients who screened positive for a clinical trial through the Gyn Onc practice’s manual screening process was performed. Screening events were new neoplasms, recurrences or progressions, and changes in treatment. Data was collected from July 2022 to December 2023. Demographic information, trials for which each patient was considered, trials patients enrolled on, and trial characteristics were recorded. Results: After adjustment, there were no significant differences between patient factors (age, marital status, race, ethnicity, and number of screening events) and enrollment status. Of trials available, 9% require biopsy and 94.7% are in person. Most studies are interventional (78.9%) and pharmaceutical trials (73.7%). There was a drastic unadjusted estimated increase in enrollment for Medicaid patients compared to other insurances (55.2% vs 32%, p=.031). There was also an increase in enrollment for cervical cancer patients (18.4%, p=.106) compared to other cancer sites (56.5% vs 32.8%, p=.036). There was a positive relationship between distance traveled to the treatment site and clinical trial enrollment. Patients who lived within 20 miles of the clinical trial site had an enrollment percentage of 30.1% (n=93), compared to 46.55% enrollment for those who lived further (n=58). There was a significant increase in enrollment likelihood without an exclusion criterion of previous history of other cancer (50% vs 33.3%, p=.046). Trials that required no history of chemotherapy had an enrollment of 27.3% (compared to 44.1%). Conclusions: Unlike prior reports, we did not demonstrate differences in trial enrollment by race, socioeconomic or insurance status. This suggests that our objective manual screening process enhances equitable clinical trial offerings. It is possible that this objective manual screening process attenuates provider bias and offers clinical trials to historically marginalized and underrepresented groups.

Tumor Size Is Not Everything: Advancing Radiomics as a Precision Medicine Biomarker in Oncology Drug Development and Clinical Care. A Report of a Multidisciplinary Workshop Coordinated by the RECIST Working Group.

Radiomics, the science of extracting quantifiable data from routine medical images, is a powerful tool that has many potential applications in oncology. The Response Evaluation Criteria in Solid Tumors Working Group (RWG) held a workshop in May 2022, which brought together various stakeholders to discuss the potential role of radiomics in oncology drug development and clinical trials, particularly with respect to response assessment. This article summarizes the results of that workshop, reviewing radiomics for the practicing oncologist and highlighting the work that needs to be done to move forward the incorporation of radiomics into clinical trials.

Systematic review of rural interventions addressing barriers and facilitators in recruitment and adherence of oncology clinical trials.

e13675 Background: Rural patients are significantly underrepresented in cancer clinical trials leading to an inequitable distribution of research funding and excluding rural patients from innovative care. Properly understanding the barriers and facilitators to rural patient participation in cancer clinical trials is essential to lessen this disparity in representation. Methods: EBSCO Medline, Embase, Cinahl, Cochrane and PsycInfo were systematically searched for articles related to the primary research objective. The search strategies focused on a combination of keywords pertaining to “rural populations”, “cancer”, and “clinical trials”. Additional search terms included the presence of “barriers” or “facilitators” within the articles. Articles were removed from the study if they were not written in English, were conducted outside the U.S., focused on pediatric populations, or lacked original data. Results: A systematic review of the literature reveals that barriers and facilitators to cancer clinical trial participation by rural populations are present at the patient, physician, and healthcare system levels. Notable barriers found in this review include the burden of travel, lack of patient knowledge about clinical trial participation, insufficient provider communication and awareness of current trials, and strict eligibility criteria. Promising facilitators include caregiver support, close physician support and contact, and programs to increase patient understanding of clinical trials. Potential future interventions include telemedicine, expanding monetary support for clinical trial transportation, utilizing RUCA codes to document disparities and elucidate further interventions, and cultivating stronger partnerships between industry and rural medical facilities, as well as between the NCI, academic institutions, and rural medical facilities. Conclusions: There is a significant and urgent need to strengthen this body of literature as is underscored by the apparent lack of cancer clinical trials in rural counties, specifically in the state of Texas (CPRIT). It is evident that barriers need to be further elucidated to ascertain why rural patients have less participation in clinical trials, and facilitators need to be further explored in order to sufficiently illustrate and thus effectively harness the power of facilitators in the pursuit of lessening disparities and improving cancer outcomes for rural patients.

Disparities in gynecologic oncology clinical trial availability in the United States: A state-level analysis.

e23128 Background: Disparities in minority participation in gynecologic cancer clinical trials is well documented. FDA guidance stipulates geographic informed analyses to reduce barriers to trial participation. We performed a geographical analyses of gynecologic cancer clinical trials in the United States relative to racial distribution and social vulnerability indices to identify areas of increased need. Methods: We performed a cross-sectional analysis of NIH ClinicalTrials.gov by retrieving all trials first posted 1/1/2013 through to 1/10/2024. We searched for ovarian, uterine, cervical, endometrial, vaginal/vulvar and gynecological cancer. We reviewed enrollment criteria to exclude non-gynecological cancers (1643 trials) or non-invasive gynecological conditions (224 trials). We aggregated census data, state-level total population size, percent Non-Hispanic White (NHW), and the FEMA Expected Annual Loss per state as a measure of social vulnerability. Using R statistical software (v 4.3.1), we measured the association between these variables and the number of gynecological trials per 100,000 persons using Pearson’s correlation. Results: We identified 3,428 trials, filtered down to 1,561 invasive cancer trials. The most common trials were ovarian (911, 58.3%) and cervical (438, 28.0%), followed by endometrial (385, 24.6%), uterine (158, 10.1%), vulva-vaginal (78, 4.99%), and 7.62% genetic-based. Texas had the highest number of trials (501) followed by California (454) and New York (427). The state with the highest population-adjusted number of trials was South Dakota (8.57 trials per 100,000 persons), followed by Rhode Island at 8.38 per 100,000 persons. California, Mississippi, and Puerto Rico had the lowest number of trials per 100,000 persons with 1.148, 0.979, 0.466 respectively. States/territories with greater than 4 trials per 100,000 were comprised of populations that were > 50% NHW. The correlation between the state-level percentage of NHW in 2020 and the number of trials per 100,000 persons was p = 0.057 indicating that there was a correlation between Whiteness at the state-level and trial availability. We found that states with higher FEMA expected annual loss had lower numbers of gynecological trials per 100,000 persons (p < 0.001). Conclusions: We found a disparity in the geographical distribution of trials available for gynecological cancers that was correlated with racial disparities and social vulnerability. Efforts to increase decentralization of trials and improve access and availability of gynecologic cancer clinical trials are necessary to increase equity for minority representation in clinical trials at the national level.

Navigating the clinical trial maze: Identifying barriers to enrollment among patients with advanced cancer in a community-academic partnership.

e15191 Background: Further research is required to pinpoint barriers responsible for disparities in clinical trial enrollment. In Alabama, a community-academic partnership has been established to broaden clinical trial access and eliminate disparities. This study aims to examine referral patterns for phase 1 and precision oncology clinical trials and identify screening and enrollment barriers. Methods: Between July-December 2023, we conducted weekly virtual meetings for patients (n=115) seen at Infirmary Health Oncology clinic (Mobile, AL) [community site] who had Next Generation Sequencing data and were potentially eligible for a clinical trial. We recorded key sociodemographic variables, retrospectively analyzed genomic data, examined trial enrollment, and explored non-participation reasons. Descriptive statistics were employed for analysis. Results: A total of 115 patients were included in this retrospective study. The median age was 70 years (range: 38-90), 60% were male vs. 40% female, and the majority were Caucasian (63%). The most common tumor types were: non-small cell lung cancer (27.8%), colorectal cancer (13.0%), pancreatic cancer (11.3%), small cell lung cancer (10.4%), and breast cancer (6.9%). We found that genomic alterations commonly included: TP53 (34.7%), KRAS (20%), PTEN (13.9%), NF (11.3%), CDK (7.8%), and PIK3CA (6.1%). Of the patients prescreened, 11.3% did not return to clinic for subsequent follow up visits and 16.5% did not have an available trial. 33% were pre-screened for clinical trial but did not meet trial eligibility criteria because of cancer stage, number of prior therapies therapy, or some other exclusion criteria. Among these, 18.4% were not eligible due to HIV or immunosuppressed status, a second malignancies, brain metastasis or poor performance status. 45 of the 115 patients (39.1%) had a clinical trial available at UAB. However, 26 (57.7%) of these patients were ultimately unable to participate due to: technical factors (n=10, 38.5%) [biopsy not obtainable/QNS/repeat biopsy needed], patient preference (n=9, 34.6%) and physician preference (n=7, 26.9%). Of the 19 (42.22%) remaining patients who were potentially eligible, only 3 (15.8%) patients were eventually enrolled on a clinical trial. All 3 patients went onto disease specific phase 2-3 trials, rather than a precision medicine or phase 1 trial. Conclusions: Our study highlights key factors affecting referrals, screening, and barriers to clinical trial enrollment. Early referrals and improved screening processes are crucial for considering trials, with prior therapy lines and functional status playing significant roles in trial non-enrollment. Our community-academic partnership aims to overcome barriers like travel and trial access, while also leveraging technology (eg. AI and databases) to address technical obstacles in trial participation.

Efficient site selection for clinical trials using simulated annealing.

1558 Background: Identifying sites with eligible patients for multi-site oncology clinical trials is challenging given increasingly complex inclusion/exclusion criteria, limited site resources, and the need to ensure diversity in trials. This is a complex systems problem with many factors that present nonlinear behaviors. Given manifold combinations of sites, research capabilities, and study characteristics, artificial intelligence solutions may help research sites and sponsors optimize site selection to achieve enrollment, speed, and diversity of goals. We report on one of the first deployments of a multi-AI model solution to optimize site selection for an ongoing clinical trial. Methods: The clinical trials were translated into a standardized digital form expressed as rules fully encompassing the inclusion and exclusion criteria. A patient cohort was created from the criteria using the ConcertAI oncology research database, including site affiliations. The sites had numerical features specific to the protocol indication, such as physician count and projected patient count. Each site was scored by its features with the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) to create higher alignment to the best matched sites. Feature weights allowed TOPSIS to emphasize desirable attributes in an interpretable way. A simulated annealing algorithm, adept with nonlinear objectives, explored site combinations, yielding a high-scoring, constraint satisfying solution. Independent simulations measured convergence and provided a distribution of solutions. Shapley analysis clarified site contributions to score and constraints. The implementation was compiled and parallelized. Results: Utilizing a Multiple Myeloma clinical trial protocol with 7 criteria, we formed a cohort of 15,881 patients across 1,246 US clinical trial sites. Each site had 12 features. The weights optimized patient diversity while enrolling at least 10 patients per month across all sites to achieve the targeted study accrual timeline. The result was a set of 3 sites enrolling 12 patients per month with 51% diversity, in agreement with the optimal solution obtained through detailed, multi-week, manual analysis. As constraints increased, and exact verification became impractical, convergence was gauged by decreasing variance in independent solutions. Conclusions: Our study introduced a robust, efficient, and interpretable approach to clinical trial site selection. It addressed the challenges posed by increasingly complicated protocols, requirements for study participant diversity, and heterogeneous site features. The multi-AI model scalability underscored its practical utility in streamlining the site selection process for clinical research networks, academic centers, industry sponsored trials, cooperative group studies, and investigator-initiated studies.

Human-AI teams to improve accuracy and timeliness of oncology trial prescreening: Preplanned interim analysis of a randomized trial.

1524 Background: Identifying eligible patients for oncology clinical trials (“prescreening”) relies on manual chart review by clinical research coordinators (CRCs), which is time-consuming and often inaccurate. Consequently, 70% of patients with cancer who meet trial eligibility criteria are not offered participation. Natural language processing (NLP) may improve the accuracy and timeliness of prescreening. Methods: This was a preplanned interim analysis of a paired-design noninferiority trial, powered to assess timeliness. We adapted NLP algorithms to identify 13 common eligibility criteria related to cancer type/stage, prior systemic therapy, actionable biomarkers, and response criteria. NLP systems performed optical character recognition, entity and relationship extraction, and natural language inference through deep learning and symbolic AI techniques. Deidentified unstructured electronic health records (EHRs) from real-world patients with non-small cell lung (NSCLC) or colorectal (CrCa) cancer were presented to CRCs via a secure platform. Two CRCs were randomized 1:1 to view blocks of 20 charts each with (Human+AI) or without (Human-alone) NLP annotations. In pre-trial assessment, the CRCs had 86% inter-rater agreement. The primary outcome was overall chart-level accuracy, defined as percent of the 13 CRC-coded eligibility items matching a gold standard set, determined by 2-3 clinicians blinded to experimental arms. Paired t-tests compared overall accuracy (noninferiority margin 5%) and criteria-specific accuracy. Significant differences were assessed for superiority using two-sided paired t-tests. Mann-Whitney tests compared the secondary outcome of timeliness (time per chart review). Results: Among 74 (40 NSCLC; 34 CrCa) of a planned 400 patients, overall accuracy for Human+AI was noninferior to Human-alone (78.7% vs. 76.7%, mean difference 2.0%, p<0.001 rejecting inferiority); both were greater than AI-alone (63.5%). Superiority for Human+AI was demonstrated for RECIST response, but not for overall nor other criteria-specific accuracy (Table). Median time per review was lower for Human+AI than Human-alone (34.1 vs 43.9 min, adjusted p=0.05). Conclusions: Human+AI teams can improve timeliness of trial prescreening with noninferior accuracy. This platform is being used for eligibility assessment in an ongoing clinical trial. [Table: see text]

Chemotherapy-Induced Peripheral Neuropathy and Falls in Cancer Survivors Relate to Digital Balance and Gait Impairments.

Chemotherapy-induced peripheral neuropathy (CIPN) and falls can be persistent side effects of cancer treatment. Standing postural sway and gait tests with body-worn, inertial sensors provide objective digital balance and gait measures that represent several different domains controlling mobility. Specific domains of balance and gait that related to neuropathy and falls are unknown. The aim of this study was to determine which domains of balance and gait differed between cancer survivors who report (1) CIPN symptoms versus no symptoms, (2) a history of falls in the past 6 months versus no falls, and (3) prospective falls over 12 months versus no falls. Postural sway during 30 seconds of quiet standing and gait characteristics from a 7-m timed up and go test were recorded with six synchronized inertial sensors (Opals by APDM Wearable Technologies, a Clario Company) in 425 older, female cancer survivors (age: 62 ± 6 years). A principal component analysis (PCA) approach was used to identify independent domains of mobility from 15 balance and gait measures. PCA analysis revealed five independent domains (PC1 = sway amplitude, PC2 = gait pace, PC3 = sway frequency, PC4 = gait spatial-temporal, and PC5 = turning) that accounted for 81% of the variance of performance. Cancer survivors who reported CIPN symptoms had significantly higher sway frequency (PC3) than asymptomatic survivors. Past fallers had significantly larger sway area (PC1) and slower gait pace (PC2) than nonfallers. Prospective fallers showed a significantly smaller stride length (PC4) than nonfallers. Digital balance and gait measures using wearable sensors during brief standing and walking tests provide objective metrics of CIPN-related mobility impairment and fall risk that could be useful for oncology clinical trials.

Precision oncology clinical trials: A systematic review of phase II clinical trials with biomarker-driven, adaptive design.

e23005 Background: Novel clinical trial designs are conducted in the precision medicine era. This study aimed to evaluate biomarker-driven, adaptive phase II trials in precision oncology, focusing on infrastructure, efficacy, and safety. Methods: We systematically reviewed and analyzed the target studies. EMBASE and PubMed searches from 2015 to 2023 generated 29 eligible trials. Data extraction included infrastructure, biomarker screening methodologies, efficacy, and safety profiles. Results: Government agencies, cancer hospitals, and academic societies with accumulated experiences led investigator-initiated precision oncology clinical trials (IIPOCTs), which later guided sponsor-initiated precision oncology clinical trials (SIPOCTs). Most SIPOCTs were international studies with basket design. IIPOCTs primarily used the central laboratory for biomarker screening, but SIPOCTs used both central and local laboratories. Most of the studies adapted next-generation sequencing and/or immunohistochemistry for biomarker screening. Fifteen studies included an independent central review committee for outcome investigation. Efficacy assessments predominantly featured objective response rate as the primary endpoint, with varying results. Nine eligible studies contributed to the United States Food and Drug Administration’s marketing authorization. Safety monitoring was rigorous, but reporting formats lacked uniformity. Health-related quality of life and patient-reported outcomes were described in some protocols but rarely reported. Table. Precision Oncology Trials led to the FDA approval. Conclusions: Our results reveal that precision oncology trials with adaptive design rapidly and efficiently evaluate anticancer drugs’ efficacy and safety, particularly in specified biomarker-driven cohorts. The evolution from IIPOCT to SIPOCT has facilitated fast regulatory approval, providing valuable insights into the precision oncology landscape. [Table: see text]

Creating and establishing a sustainable culture of clinical and translational research at an academic cancer center.

e23220 Background / Objectives: At the Ellis Fischel Cancer Center (EFCC) at the Univ. of Missouri (MU) School of Medicine (SOM), our recent (2016 - 2020) accurals to interventional oncology clinical trials (CT) have been suboptimal. Hence, the EFCC research leadership undertook a number of initiatives since 2021 to improve the function and productivity of the oncology CT enterprise. The objective of our retrospective review was to determine the impact of our initiatives including improving enrollments to clinical trials. Methods: These major programmatic initiatives have included the following: (1) CT portfolio analysis; (2) faculty survey to assess concerns/ needs related to enrollment; (3) Development of a "One-Pager" list of open CT; (4) Working with faculty leader "champions"; (5) Increasing the number of clinical research staff; (6) Initiation of a monthly "Clinical Trials Working Group (CTWG)" to monitor enrollment to clinical trials, troubleshooting obstacles to accruals and on-going CT portfolio analysis and adjustments; (7) Development of multi-disciplinary "Disease Working Groups" (DWGs) in Breast, GU, GYN-Onc., Lung, GI and CNS cancers to establish a Protocol Review & Monitoring System (PRMS) with the objective of aligning our clinical trials with our patient population; (9) Initiation of a monthly “Investigator Education Workshop” (IEW) to educate our faculty on the conduct of investigator initiated CT and pharma sponsored CT; (10) Initiation of a "Top Enrolling Ceremony" to honor the top enrollers to clinical trials. Results: The analysis of our CT portfolio revealed that in both December 2020 and May 2021, more than half [i.e., 52% (n = 17/33 and n = 13/25, respectively] of our CT had never enrolled a patient in the lifetime of the protocol. Since the implementation of the PRMS in late 2021, we have observed a continual decrease in the status of those clinical trials with “never” any enrollments from a historical average of 52% to 34% in January 2022, to 32% in January 2023 and to 28% in October 2023, p = 0.0173. In addition, implementation of the PRMS also had an immediate significant impact such that only 53% of all of the protocols reviewed by the Protocol Review and Monitoring Committee were approved in 2021 vs. a historical average of 75% from 2018 - 2020. Moreover, between 2021 and 2023, an average total of 82 patients were enrolled to interventional oncology clinical trials vs. the historical average of 42 patients per year from 2016- 2020, representing an overall increase of 95% p = 0.0004. Conclusions: The development and establishment of the PRMS within the last 3 years has had a very positive impact on our clinical trial enterprise at the EFCC by reducing active clinical trials that are not suitable for our patient population and significantly increasing our enrollment to interventional oncology clinical trials. For 2024, we will continue our implementation of the PRMS at the EFCC.

Seeing the Trees From the Forest: Challenges in Subgroup Analysis-Based Guidelines in Oncology

As clinical trials in oncology require substantial efforts, maximizing the insights gained from them by conducting subgroup analyses is often attempted. The goal of these analyses is to identify subgroups of patients who are likely to benefit, as well as the subgroups of patients who are unlikely to benefit from the studied intervention. International guidelines occasionally include or exclude novel medications and technologies for specific subpopulations based on such analyses of pivotal trials without requiring confirmatory trials. This Perspective discusses the importance of providing a complete dataset of clinical information when reporting subgroup analyses and explains why such transparency is key for better clinical interpretation of the results and the appropriate application to clinical care, by providing examples of transparent reporting of clinical studies and examples of incomplete reporting of clinical studies.

Use of Closed System Transfer Devices in Oncology Clinical Trials: Establishing Processes and Overcoming Challenges

Use of Closed System Transfer Devices in Oncology Clinical Trials: Establishing Processes and Overcoming Challenges

Avoiding Delays in Reporting Time-to-Event Randomized Trials: Calendar Backstops and Other Approaches.

New oncology therapies that extend patients' lives beyond initial expectations and improving later-line treatments can lead to complications in clinical trial design and conduct. In particular, for trials with event-based analyses, the time to observe all the protocol-specified events can exceed the finite follow-up of a clinical trial or can lead to much delayed release of outcome data. With the advent of multiple classes of oncology therapies leading to much longer survival than in the past, this issue in clinical trial design and conduct has become increasingly important in recent years. We propose a straightforward prespecified backstop rule for trials with a time-to-event analysis and evaluate the impact of the rule with both simulated and real-world trial data. We then provide recommendations for implementing the rule across a range of oncology clinical trial settings.