Search

424 Background: Capecitabine (cape), an oral chemotherapy, is the treatment backbone for many GI cancers. Its complex dosing and narrow therapeutic index make medication adherence and toxicity management crucial for quality patient care. Methods: We conducted a feasibility study of “Penny,” an augmented intelligence mobile phone chatbot that leverages algorithmic surveys and natural language processing (NLP) to engage with patients in conversational, bi-directional text messages. Penny provides patients with medication reminders tailored to their prescribed doses and schedules, sends weekly check-in messages, manages low-grade symptoms in real time, and escalates high-grade symptoms for resolution by the clinical team. Patients ≥18 years old receiving cape for the treatment of a GI cancer were accrued in sequential cohorts of 20 for participation over a three-month period. Feasibility was assessed during planned interim analyses and was predefined as the completion of a 20-patient cohort without a safety event, defined as the communication of incorrect medication or symptom management recommendations as ascertained by two independent clinician reviewers (Κ = 0.89). Secondary outcomes included patient-reported adherence and engagement with the chatbot’s weekly check-in messages. At study completion, all patients were invited to participate in structured interviews to provide feedback on the platform. Results: The first cohort of 20 patients was enrolled from 8/2021 to 4/2022; the median age was 57 years, and patients were primarily female (55%), white (65%), commercially insured (55%), and had colorectal cancer (55%). Chemotherapy regimens included cape with oxaliplatin (50%), concurrent RT (30%), temozolomide (5%), and monotherapy (15%). A total of 2,149 text messaging exchanges were reviewed with 150 (7%) medication-related and 9 (0.4%) symptom-related safety events identified. Most medication-related safety events were due to misalignment with prescribed chemotherapy schedules (55%) and doses (32%). Symptom-related safety events were primarily due to the misinterpretation of patient messages by Penny’s NLP functionality (89%). Average patient-reported adherence was 67% (SD 27%), and patients engaged with 27% (SD 24%) of the chatbot’s weekly check-in messages. In post-study interviews with 12 patients, participants reported that the medication reminders were reliable and user-friendly, whereas the symptom management tool was too simplistic to be helpful. Conclusions: Although Penny has not yet met its feasibility endpoint, the lessons learned from this first cohort have informed further refinements to the platform. Ongoing efforts aim to integrate Penny with the electronic health record and further train the chatbot’s NLP functionality to minimize medication- and symptom-related safety events, respectively. Clinical trial information: NCT05113264.

362 Background: Despite current guidelines, less than 50% of patients with metastatic (m) non-squamous (NSq) NSCLC undergo comprehensive molecular genotyping. At our institution, based on improved comprehensiveness of genotyping with use of concurrent tissue (T) and plasma (P) next generation sequencing (NGS), we designed an electronic medical record (EMR)-based “nudge intervention” to auto-generate an order for P NGS at the time of initial consultation, while T NGS was carried out reflexively based on institutional pathways. Methods: A prospective study was conducted at the Abramson Cancer Center and 2 community sites within the University of Pennsylvania Health System after IRB approval. A provider team-focused EMR-based “nudge intervention” was designed to order P NGS at the time of new patient consultation. Eligible patients for the nudge were identified using an EMR based checklist, that included 3 criteria i. newly diagnosed, ii. treatment naïve, iii. mNSq NSCLC. Results from the intervention period (4/2021-12/2021) were compared to baseline data from similar patients treated at our institution between 01/2019 and 03/2021. Categories of NCCN guideline recommended molecular genotyping were defined as: i) comprehensive: EGFR, ALK, BRAF, ROS1, MET, RET, NTRK testing, and ii) incomplete or no testing performed. The proportion of patients with comprehensive molecular genotyping prior to 1st-line therapy were compared in the pre- and post-intervention groups using the chi-square test. Results: 526 patients with mNSq NSCLC were included in this analysis: 381 in the pre-intervention cohort, 145 in the post-intervention cohort. After implementation of the EMR-based nudge, a higher proportion of patients underwent concurrent T+P testing resulting in improved comprehensive molecular genotyping. In addition, a greater proportion of patients had comprehensive genotyping available prior to 1st-line therapy in the post-intervention vs pre-intervention cohort (Table). Conclusions: Across 3 practice sites, a provider team-focused EMR-based nudge intervention was associated with a significantly higher proportion of patients with mNSq NSCLC undergoing comprehensive molecular genotyping, both overall and prior to 1st-line therapy. These findings demonstrate that behavioral, EMR-based nudges can promote guideline concordant diagnostic testing at both community and academic sites and should be studied further as a tool to improve rates of molecular testing in NSCLC.[Table: see text]

1506 Background: Oncologists suboptimally assess patient symptoms and functional status, possibly leading to poor symptom management or over-treatment. Remote patient-reported symptoms and passive activity monitoring may provide objective measures of symptoms and functional status to improve patient-clinician communication and symptom understanding. We assessed the impact of a clinician-centered dashboard of longitudinal patient-reported symptoms and step counts on patient-clinician communication regarding symptoms and functional status. Methods: This randomized trial enrolled 108 patients with incurable GI or lung cancers treated with chemotherapy at a large academic health center. Patients were randomized to either of Arms A) control, B) weekly patient-reported symptoms via text message + step tracking from a wearable activity monitor, with summary dashboards given to clinicians at each visit, or C) arm B plus text message-based prompts to patients encouraging discussion of symptoms and functional status prior to each visit. We used Kruskal Wallis tests to compare co-primary outcomes (patient-reported perceptions of clinician symptom and functional status understanding at 6 months after enrollment) between control (A) and intervention (B+C) arms on a 5-point scale (1 = Not at all; 2 = Slightly; 3 = Moderately; 4 = Considerably; 5 = Completely). Results: 33, 37, and 38 patients were enrolled in arms A, B, and C, respectively. Patients were 54.6% male, mean age was 58.9 years, 77% had GI cancer, and 23% had lung cancer. At six months, there was no difference between control and intervention arms in patient perception of clinician understanding of symptoms (Arm A: 4.5, Arm B/C: 4.5, p = 0.85) or functional status (Arm A: 4.5, Arm B/C: 4.3, p = 0.59). Patients reported that their oncology team seldom discussed PROstep data during appointments (mean 2.3 on 5-point scale where 2 = seldom). Hospitalization rates were 42% and 45% for Arms A and B/C (p = 0.8), respectively, and new palliative care referrals were 9% and 10% (p = 0.8), respectively. Mean adherence to weekly patient reports and Fitbit data (at least 4 of 7 days in a week) was 64% and 53%, respectively. Net promoter score was 8.3 on a 10-point scale. Conclusions: Clinician and patient-directed dashboards based on patient-generated health data did not lead to higher patient-perceived clinician understanding of symptoms and functional status, although this was limited by moderate adherence to remote symptom and step count collection and low frequency of clinician discussion of PROStep data with patients, highlighting challenges to clinical application of these data sources. Further efforts are needed to improve patient-clinician communication about symptoms and functional status. Clinical trial information: NCT04616768.

9022 Background: Current guidelines recommend comprehensive molecular genotyping for newly diagnosed patients (pts) with metastatic non-squamous (non-Sq) NSCLC. We have previously demonstrated that concurrent plasma (P) and tissue (T) based next-generation sequencing (NGS) improves detection of clinically actionable mutations in pts with advanced NSCLC. We analyzed the impact of concurrent T+P NGS on comprehensiveness of molecular genotyping and on overall survival (OS). Methods: A retrospective cohort study of pts with newly diagnosed stage IV non-Sq NSCLC who received therapy at our institution between 01/2019 and 12/2020 was performed. Categories of NCCN guideline testing were defined, i) comprehensive: EGFR, ALK, BRAF, ROS1, MET, RET, and NTRK testing, ii) incomplete: 2-6 genes tested, and iii) no testing performed. The proportion of pts with comprehensive molecular testing performed, prior to 1st-line therapy and by detection modality (T NGS vs. T+P NGS), were compared using Fisher’s exact test. Median OS was estimated using Kaplan-Meier methodology from diagnosis to death or censored at most recent follow-up. Results: 335 patients were included in this analysis, 98.5% (330/335) underwent molecular testing: either comprehensive: n = 291 (86.9%), incomplete testing: n = 39 (11.6%); or no testing n = 5 (1.5%). Testing with T NGS was completed in 32.7% (108/330); 67.2% (222/330) underwent concurrent T+P NGS. These groups were well balanced for baseline characteristics, with the exception of a higher number of never smokers in T+P vs. T NGS (30.2% vs. 14.8%, p < 0.0001). Proportion of pts with comprehensive molecular testing was higher among pts with T+P NGS: 99.5% (221/222) vs. T NGS: 64.8% (70/108), p < 0.0001. All pts with T+P NGS testing had results available prior to 1st line therapy; 100% (204/204) compared to 60.7% (51/84) for T NGS, p < 0.0001. With median follow up of 20.5 months (mos, range 0.3 - 33.1), median OS was 18.6 mos. Median OS for pts tested with T+P NGS vs T alone was numerically longer at 23.2 vs. 14.1 mos, but not statistically significant (p = 0.078). However, regardless of testing modality, patients with comprehensive molecular genotyping had superior OS compared to those with incomplete or no testing (22.1 mos vs. 11.6 mos, p = 0.017). The institution of oral targeted therapy had no bearing on this difference in OS (test for interaction, p = 0.6509). Conclusions: Performance of concurrent T+P NGS testing was associated with a higher likelihood of comprehensive molecular genotyping, as well as improved availability of results, including prior to first line therapy. Patients with comprehensive genotyping have improved OS compared to patients with incomplete or no testing. These results support implementation of a concurrent T+P NGS approach upon initial diagnosis of metastatic non-Sq NSCLC.

109 Background: Early serious illness conversations (SICs) between oncology clinicians and patients are associated with improved mood, quality of life, and quality of end-of-life (EOL) care. Yet, most patients with cancer die without a documented SIC. We report on pre-specified 40-week SIC and EOL outcomes from a stepped-wedge randomized clinical trial (NCT03984773) testing the impact of clinician-directed behavioral nudges to prompt SICs among patients with cancer at high risk of mortality based on a machine learning algorithm. Methods: Our sample consisted of patients with cancer receiving care at one of 9 tertiary or community-based medical oncology clinics between June 2019 and April 2020. We identified high-risk patients using a prospectively validated electronic health record machine learning algorithm to predict 6-month mortality. The intervention consisted of: (1) Weekly emails comparing individual oncologists’ SIC rate relative to peers; (2) Weekly lists of forthcoming encounters with high-risk patients; and (3) Opt-out text messages to prompt SICs before high-risk patient encounters. Clinics were randomized in stepped-wedge fashion to receive the intervention in 4-week intervals through week 16, when all clinics received the intervention. Patients were followed through week 40. The primary outcome was SIC rates for all and high-risk patients. EOL outcomes among decedents were based on ASCO/NQF guidelines and included death in the hospital, intensive care unit admission within 30 days of death, receipt of systemic therapy within 14 days of death, hospice enrollment prior to death, and hospice length of stay. Intention-to-treat analyses were adjusted for clinic and wedge fixed effects and clustered at the oncologist-level. Results: The sample consisted of 20,506 patients and 41,021 encounters. 1,324 (6.5%) patients died by the end of follow-up. Among high-risk patients, the unadjusted SIC rate was 3.4% (59/1754) in the control period and 13.5% (510/3765) in the intervention period and remained >12% throughout follow-up. In adjusted analyses, the intervention was associated with an increase in SICs (adjusted odds ratio 2.09, 95% CI 1.53-2.87, p

528 Background: ‘Synthetic’ control arms (SCAs), created using electronic health records (EHRs), have immense potential to augment clinical trial findings and provide a rich context of real-world evidence (RWE) while reducing patient (pt) and sponsor burden (Gottlieb 2019). PICI0002 (PRINCE) is a ph1b/2 study evaluating APX005M with gemcitabine (gem) and nab-paclitaxel (NP) ± nivolumab for the treatment of metastatic pancreatic adenocarcinoma (mPDAC; NCT03214250). PRINCE pts were enrolled from select US academic cancer centers and a global, ph3 study was utilized for an historical reference control (Von Hoff 2013). To address the perceived limitations of this design, we explored the feasibility and utility of developing a contemporary SCA using real-world data (RWD). Methods: The SCA was derived using retrospective pt data from the two highest enrolling participating centers on PRINCE. Pts meeting key PRINCE eligibility criteria, who received gem/NP in the two years preceding the trial start date, were identified using an electronic phenotyping algorithm applied to cancer registry and EHR data, followed by manual review. Baseline characteristics, treatment exposure, efficacy and survival data were extracted electronically and via manual chart abstraction. Data were stored in a REDCap database built and housed by the Parker Institute for Cancer Immunotherapy. SCA pt characteristics were compared with PRINCE and overall survival (OS; time from initiation of gem/NP to death) was compared to historical reference controls (Table). Results: N=68 pts treated with gem/NP meeting PRINCE eligibility criteria were identified. All pts were deceased at the time of analysis. SCA pts had comparable baseline characteristics to PRINCE pts; key differences included inferior performance status and a higher proportion of pts presenting with a de novo mPDAC diagnosis. Median time on gem/NP was 4.8 months (mos; range 0-39). Median OS was 11.5 mos (95% CI 9.0-13.6) and 1-year OS was 43% (95% CI 31-55), in line with historical controls (Table). Conclusions: This study confirms the feasibility and utility of generating a control arm via a semi-automated approach. Current limitations entail manual oversight requirements as well as the known constraints of RWD, including associated biases and lack of available RECIST data. These limitations stand to evolve alongside EHR technologies. SCAs using RWD may help inform the value of prospective data by providing a contemporary reference of RWE. In some circumstances, SCAs may also serve as an alternative to traditional control arms, particularly for well-characterized standard therapies.[Table: see text]