Bridging clinical informatics and implementation science to improve cancer symptom management in ambulatory oncology practices: experiences from the IMPACT consortium.

Objectives: To report lessons from integrating the methods and perspectives of clinical informatics (CI) and implementation science (IS) in the context of Improving the Management of symPtoms during and following Cancer Treatment (IMPACT) Consortium pragmatic trials. Materials and Methods: IMPACT informaticists, trialists, and implementation scientists met to identify challenges and solutions by examining robust case examples from 3 Research Centers that are deploying systematic symptom assessment and management interventions via electronic health records (EHRs). Investigators discussed data collection and CI challenges, implementation strategies, and lessons learned. Results: CI implementation strategies and EHRs systems were utilized to collect and act upon symptoms and impairments in functioning via electronic patient-reported outcomes (ePRO) captured in ambulatory oncology settings. Limited EHR functionality and data collection capabilities constrained the ability to address IS questions. Collecting ePRO data required significant planning and organizational champions adept at navigating ambiguity. Discussion: Bringing together CI and IS perspectives offers critical opportunities for monitoring and managing cancer symptoms via ePROs. Discussions between CI and IS researchers identified and addressed gaps between applied informatics implementation and theory-based IS trial and evaluation methods. The use of common terminology may foster shared mental models between CI and IS communities to enhance EHR design to more effectively facilitate ePRO implementation and clinical responses. Conclusion: Implementation of ePROs in ambulatory oncology clinics benefits from common understanding of the concepts, lexicon, and incentives between CI implementers and IS researchers to facilitate and measure the results of implementation efforts.

Hypofractionated vs Conventionally Fractionated Postmastectomy Radiation After Implant-Based Reconstruction: A Randomized Clinical Trial.

Importance: Postmastectomy radiation therapy (PMRT) improves local-regional disease control and patient survival. Hypofractionation (HF) regimens have comparable efficacy and complication rates with improved quality of life compared with conventional fractionation (CF) schedules. However, the use of HF after mastectomy in patients undergoing breast reconstruction has not been prospectively examined. Objective: To compare HF and CF PMRT outcomes after implant-based reconstruction. Design, Setting, and Participants: This randomized clinical trial assessed patients 18 years or older undergoing mastectomy and immediate expander or implant reconstruction for breast cancer (Tis, TX, or T1-3) and unilateral PMRT from March 8, 2018, to November 3, 2021 (median [range] follow-up, 40.4 [15.4-63.0] months), at 16 US cancer centers or hospitals. Analyses were conducted between September and December 2023. Interventions: Patients were randomized 1:1 to HF or CF PMRT. Chest wall doses were 4256 cGy for 16 fractions for HF and 5000 cGy for 25 fractions for CF. Chest wall toxic effects were defined as a grade 3 or higher adverse event. Main Outcomes and Measures: The primary outcome was the change in physical well-being (PWB) domain of the Functional Assessment of Cancer Therapy-Breast (FACT-B) quality-of-life assessment tool at 6 months after starting PMRT, controlling for age. Secondary outcomes included toxic effects and cancer recurrence. Results: Of 400 women (201 in the CF arm and 199 in the HF arm; median [range] age, 47 [23-79] years), 330 patients had PWB scores at baseline and at 6 months. There was no difference in the change in PWB between the study arms (estimate, 0.13; 95% CI, -0.86 to 1.11; P = .80), but there was a significant interaction between age group and study arm (P = .03 for interaction). Patients younger than 45 years had higher 6-month absolute PWB scores if treated with HF rather than CF regimens (23.6 [95% CI, 22.7-24.6] vs 22.0 [95% CI, 20.7-23.3]; P = .047) and reported being less bothered by adverse effects (mean [SD], 3.0 [0.9] in the HF arm and 2.6 [1.2] in the CF arm; P = .02) or nausea (mean [SD], 3.8 [0.4] in the HF arm and 3.6 [0.8] in the CF arm; P = .04). In the as-treated cohort, there were 23 distant (11 in the HF arm and 12 in the CF arm) and 2 local-regional (1 in the HF arm and 1 in the CF arm) recurrences. Chest wall toxic effects occurred in 39 patients (20 in the HF arm and 19 in the CF arm) at a median (IQR) of 7.2 (1.8-12.9) months. Fractionation was not associated with chest wall toxic effects on multivariate analysis (HF arm: hazard ratio, 1.02; 95% CI, 0.52-2.00; P = .95). Fewer patients undergoing HF vs CF regimens had a treatment break (5 [2.7%] vs 15 [7.7%]; P = .03) or required unpaid time off from work (17 [8.5%] vs 34 [16.9%]; P = .02). Conclusions and Relevance: In this randomized clinical trial, the HF regimen did not significantly improve change in PWB compared with the CF regimen. These data add to the increasing experience with HF PMRT in patients with implant-based reconstruction. Trial Registration: ClinicalTrials.gov Identifier: NCT03422003.

Transforming patient-centered cancer care using telehealth: the MATCHES Center.

Modern cancer care is costly and logistically burdensome for patients and their families despite an expansion of technology and medical advances that create the opportunity for novel approaches to care. Therefore, there is a growing appreciation for the need to leverage these innovations to make cancer care more patient centered and convenient. The Memorial Sloan Kettering Making Telehealth Delivery of Cancer Care at Home Efficient and Safe Telehealth Research Center is a National Cancer Institute-designated and funded Telehealth Research Center of Excellence poised to generate the evidence necessary to inform the appropriate use of telehealth as a strategy to improve access to cancer services that are convenient for patients. The center will evaluate telehealth as a strategy to personalize cancer care delivery to ensure that it is not only safe and effective but also convenient and efficient. In this article, we outline this new center's research strategy, as well as highlight challenges that exist in further integrating telehealth into standard oncology practice based on early experiences.

Optimization of alert notifications in electronic patient-reported outcome (ePRO) remote symptom monitoring systems (AFT-39).

PURPOSE: Clinical benefits result from electronic patient-reported outcome (ePRO) systems that enable remote symptom monitoring. Although clinically useful, real-time alert notifications for severe or worsening symptoms can overburden nurses. Thus, we aimed to algorithmically identify likely non-urgent alerts that could be suppressed. METHODS: We evaluated alerts from the PRO-TECT trial (Alliance AFT-39) in which oncology practices implemented remote symptom monitoring. Patients completed weekly at-home ePRO symptom surveys, and nurses received real-time alert notifications for severe or worsening symptoms. During parts of the trial, patients and nurses each indicated whether alerts were urgent or could wait until the next visit. We developed an algorithm for suppressing alerts based on patient assessment of urgency and model-based predictions of nurse assessment of urgency. RESULTS: 593 patients participated (median age = 64 years, 61% female, 80% white, 10% reported never using computers/tablets/smartphones). Patients completed 91% of expected weekly surveys. 34% of surveys generated an alert, and 59% of alerts prompted immediate nurse actions. Patients considered 10% of alerts urgent. Of the remaining cases, nurses considered alerts urgent more often when patients reported any worsening symptom compared to the prior week (33% of alerts with versus 26% without any worsening symptom, p = 0.009). The algorithm identified 38% of alerts as likely non-urgent that could be suppressed with acceptable discrimination (sensitivity = 80%, 95% CI [76%, 84%]; specificity = 52%, 95% CI [49%, 55%]). CONCLUSION: An algorithm can identify remote symptom monitoring alerts likely to be considered non-urgent by nurses, and may assist in fostering nurse acceptance and implementation feasibility of ePRO systems.

Biomarker Inference and the Timing of Next-Generation Sequencing in a Multi-Institutional, Cross-Cancer Clinicogenomic Data Set.

PURPOSE: Observational clinicogenomic data sets, consisting of tumor next-generation sequencing (NGS) data linked to clinical records, are commonly used for cancer research. However, in real-world practice, oncologists frequently request NGS in search of treatment options for progressive cancer. The extent and impact of this dynamic on analysis of clinicogenomic research data are not well understood. METHODS: We analyzed clinicogenomic data for patients with non-small cell lung, colorectal, breast, prostate, pancreatic, or urothelial cancers in the American Association for Cancer Research Biopharmaceutical Consortium cohort. Associations between baseline and time-varying clinical characteristics and time from diagnosis to NGS were measured. To explore the impact of informative cohort entry on biomarker inference, statistical interactions between selected biomarkers and time to NGS with respect to overall survival were calculated. RESULTS: Among 7,182 patients, time from diagnosis to NGS varied significantly by clinical factors, including cancer type, calendar year of sequencing, institution, and age and stage at diagnosis. NGS rates also varied significantly by dynamic clinical status variables; in an adjusted model, compared with patients with stable disease at any given time after diagnosis, patients with progressive disease by imaging or oncologist assessment had higher NGS rates (hazard ratio for NGS, 1.61 [95% CI, 1.45 to 1.78] and 2.32 [95% CI, 2.01 to 2.67], respectively). Statistical interactions between selected biomarkers and time to NGS with respect to survival, potentially indicating biased biomarker inference results, were explored. CONCLUSION: To evaluate the appropriateness of a data set for a particular research question, it is crucial to measure associations between dynamic cancer status and the timing of NGS, as well as to evaluate interactions involving biomarkers of interest and NGS timing with respect to survival outcomes.

Identifying Oncology Clinical Trial Candidates Using Artificial Intelligence Predictions of Treatment Change: A Pilot Implementation Study.

PURPOSE: Precision oncology clinical trials often struggle to accrue, partly because it is difficult to find potentially eligible patients at moments when they need new treatment. We piloted deployment of artificial intelligence tools to identify such patients at a large academic cancer center. PATIENTS AND METHODS: Neural networks that process radiology reports to identify patients likely to start new systemic therapy were applied prospectively for patients with solid tumors that had undergone next-generation sequencing at our center. Model output was linked to the MatchMiner tool, which matches patients to trials using tumor genomics. Reports listing genomically matched patients, sorted by probability of treatment change, were provided weekly to an oncology nurse navigator (ONN) coordinating recruitment to nine early-phase trials. The ONN contacted treating oncologists when patients likely to change treatment appeared potentially trial-eligible. RESULTS: Within weekly reports to the ONN, 60,199 patient-trial matches were generated for 2,150 patients on the basis of genomics alone. Of these, 3,168 patient-trial matches (5%) corresponding to 525 patients were flagged for ONN review by our model, representing a 95% reduction in review compared with manual review of all patient-trial matches weekly. After ONN review for potential eligibility, treating oncologists for 74 patients were contacted. Common reasons for not contacting treating oncologists included cases where patients had already decided to continue current treatment (21%); the trial had no slots (14%); or the patient was ineligible on ONN review (12%). Of 74 patients whose oncologists were contacted, 10 (14%) had a consult regarding a trial and five (7%) enrolled. CONCLUSION: This approach facilitated identification of potential patients for clinical trials in real time, but further work to improve accrual must address the many other barriers to trial enrollment in precision oncology research.

Effects of a change in recall period on reporting severe symptoms: an analysis of a pragmatic multisite trial.

BACKGROUND: Optimal methods for deploying electronic patient-reported outcomes to manage symptoms in routine oncologic practice remain uncertain. The electronic symptom management (eSyM) program asks chemotherapy and surgery patients to self-report 12 common symptoms regularly. Feedback from nurses and patients led to changing the recall period from the past 7 days to the past 24 hours. METHODS: Using questionnaires submitted during the 16 weeks surrounding the recall period change, we assessed the likelihood of reporting severe or moderate and severe symptoms across 12 common symptoms and separately for the 5 most prevalent symptoms. Interrupted time-series analyses modeled the effects of the change using generalized linear mixed-effects models. Surgery and chemotherapy cohorts were analyzed separately. Study-wide effects were estimated using a meta-analysis method. RESULTS: In total, 1692 patients from 6 institutions submitted 7823 eSyM assessments during the 16 weeks surrounding the recall period change. Shortening the recall period was associated with lower odds of severe symptom reporting in the surgery cohort (odds ratio = 0.65, 95% confidence interval = 0.46 to 0.93; P = .02) and lower odds of moderate and severe symptom reporting in the chemotherapy cohort (odds ratio = 0.83, 95% confidence interval = 0.71 to 0.97; P = .02). Among the most prevalent symptoms, 24-hour recall was associated with a lower rate of reporting postoperative constipation but no differences in reporting rates for other symptoms. CONCLUSION: A shorter recall period was associated with a reduction in the proportion of patients reporting moderate-severe symptoms. The optimal recall period may vary depending on whether electronic patient-reported outcomes are collected for active symptom management, as a clinical trial endpoint, or another purpose. ClinicalTrials.gov ID NCT03850912.

Genomic Characterization and Clinical Outcomes of Patients with Peritoneal Metastases from the AACR GENIE Biopharma Collaborative Colorectal Cancer Registry.

Peritoneal metastases (PM) are common in metastatic colorectal cancer (mCRC). We aimed to characterize patients with mCRC and PM from a clinical and molecular perspective using the American Association of Cancer Research Genomics Evidence Neoplasia Information Exchange (GENIE) Biopharma Collaborative (BPC) registry. Patients' tumor samples underwent targeted next-generation sequencing. Clinical characteristics and treatment outcomes were collected retrospectively. Overall survival (OS) from advanced disease and progression-free survival (PFS) from start of cancer-directed drug regimen were estimated and adjusted for the left truncation bias. A total of 1,281 patients were analyzed, 244 (19%) had PM at time of advanced disease. PM were associated with female sex [OR: 1.67; 95% confidence interval (CI): 1.11-2.54; P = 0.014] and higher histologic grade (OR: 1.72; 95% CI: 1.08-2.71; P = 0.022), while rectal primary tumors were less frequent in patients with PM (OR: 0.51; 95% CI: 0.29-0.88; P < 0.001). APC occurred less frequently in patients with PM (N = 151, 64% vs. N = 788, 79%) while MED12 alterations occurred more frequently in patients with PM (N = 20, 10% vs. N = 32, 4%); differences in MED12 were not significant when restricting to oncogenic and likely oncogenic variants according to OncoKB. Patients with PM had worse OS (HR: 1.45; 95% CI: 1.16-1.81) after adjustment for independently significant clinical and genomic predictors. PFS from initiation of first-line treatment did not differ by presence of PM. In conclusion, PM were more frequent in females and right-sided primary tumors. Differences in frequencies of MED12 and APC alterations were identified between patients with and without PM. PM were associated with shorter OS but not with PFS from first-line treatment. SIGNIFICANCE: Utilizing the GENIE BPC registry, this study found that PM in patients with colorectal cancer occur more frequently in females and right-sided primary tumors and are associated with worse OS. In addition, we found a lower frequency of APC alterations and a higher frequency in MED12 alterations in patients with PM.

Strategies for Implementing an Electronic Patient-Reported Outcomes-Based Symptom Management Program Across Six Cancer Centers.

Background: Electronic patient-reported outcome (ePRO)-based symptom management improves cancer patients' outcomes. However, implementation of ePROs is challenging, requiring technical resources for integration into clinical systems, substantial buy-in from clinicians and patients, novel workflows to support between-visit symptom management, and institutional investment. Methods: The SIMPRO Research Consortium developed eSyM, an electronic health record-integrated, ePRO-based symptom management program for medical oncology and surgery patients and deployed it at six cancer centers between August 2019 and April 2022 in a type II hybrid effectiveness-implementation cluster randomized stepped-wedge study. Sites documented implementation strategies monthly using REDCap, itemized them using the Expert Recommendations for Implementation Change (ERIC) list and mapped their target barriers using the Consolidated Framework for Implementation Research (CFIR) to inform eSyM program enhancement, facilitate inter-consortium knowledge sharing and guide future deployment efforts. Results: We documented 226 implementation strategies: 35 'foundational' strategies were applied consortium-wide by the coordinating center and 191 other strategies were developed by individual sites. We consolidated these 191 site-developed strategies into 64 unique strategies (i.e., removed duplicates) and classified the remainder as either 'universal', consistently used by multiple sites (N=29), or 'adaptive', used only by individual sites (N=35). Universal strategies were perceived as having the highest impact; they addressed eSyM clinical preparation, training, engagement of patients/clinicians, and program evaluation. Across all documented SIMPRO strategies, 44 of the 73 ERIC strategies were addressed and all 5 CFIR barriers were addressed. Conclusion: Methodical collection of theory-based implementation strategies fostered the identification of universal, high-impact strategies that facilitated adoption of a novel care-delivery intervention by patients, clinicians, and institutions. Attention to the high-impact strategies identified in this project could support implementation of ePROs as a component of routine cancer care at other institutions.

Assessing Patient Readiness for an Electronic Patient-Reported Outcome-Based Symptom Management Intervention in a Multisite Study.

PURPOSE: While the use of electronic patient-reported outcomes (ePROs) in routine clinical practice is increasing, barriers to patient engagement limit adoption. Studies have focused on technology access as a key barrier, yet other characteristics may also confound readiness to use ePROs including patients' confidence in using technology and confidence in asking clinicians questions. METHODS: To assess readiness to use ePROs, adult patients from six US-based health systems who started a new oncology treatment or underwent a cancer-directed surgery were invited to complete a survey that assessed access to and confidence in the use of technology, ease of asking clinicians questions about health, and symptom management self-efficacy. Multivariable ordinal logistic regression models were fit to assess the association between technology confidence, ease of asking questions, and symptom management self-efficacy. RESULTS: We contacted 3,212 individuals, and 1,043 (33%) responded. The median age was 63 years, 68% were female, and 75% reported having access to patient portals. Over 80% had two or more electronic devices. Most patients reported high technology confidence, higher ease of asking clinicians questions, and high symptom management self-efficacy (n = 692; 66%). Patients with high technology confidence also reported higher ease of asking nurses about their health (adjusted odds ratio [AOR], 4.58 [95% CI, 2.36 to 8.87]; P ≤ .001). Those who reported higher ease of asking nurses questions were more likely to report higher confidence in managing symptoms (AOR, 30.54 [95% CI, 12.91 to 72.30]; P ≤ .001). CONCLUSION: Patient readiness to use ePROs likely depends on multiple factors, including technology and communication confidence, and symptom management self-efficacy. Future studies should assess interventions to address these factors.

Effects of KRAS Genetic Interactions on Outcomes in Cancers of the Lung, Pancreas, and Colorectum.

BACKGROUND: KRAS is among the most commonly mutated oncogenes in cancer, and previous studies have shown associations with survival in many cancer contexts. Evidence from both clinical observations and mouse experiments further suggests that these associations are allele- and tissue-specific. These findings motivate using clinical data to understand gene interactions and clinical covariates within different alleles and tissues. METHODS: We analyze genomic and clinical data from the AACR Project GENIE Biopharma Collaborative for samples from lung, colorectal, and pancreatic cancers. For each of these cancer types, we report epidemiological associations for different KRAS alleles, apply principal component analysis (PCA) to discover groups of genes co-mutated with KRAS, and identify distinct clusters of patient profiles with implications for survival. RESULTS: KRAS mutations were associated with inferior survival in lung, colon, and pancreas, although the specific mutations implicated varied by disease. Tissue- and allele-specific associations with smoking, sex, age, and race were found. Tissue-specific genetic interactions with KRAS were identified by PCA, which were clustered to produce five, four, and two patient profiles in lung, colon, and pancreas. Membership in these profiles was associated with survival in all three cancer types. CONCLUSIONS: KRAS mutations have tissue- and allele-specific associations with inferior survival, clinical covariates, and genetic interactions. IMPACT: Our results provide greater insight into the tissue- and allele-specific associations with KRAS mutations and identify clusters of patients that are associated with survival and clinical attributes from combinations of genetic interactions with KRAS mutations.

Disparities in electronic health record portal access and use among patients with cancer.

BACKGROUND: Electronic health record-linked portals may improve health-care quality for patients with cancer. Barriers to portal access and use undermine interventions that rely on portals to reduce cancer care disparities. This study examined portal access and persistence of portal use and associations with patient and structural factors before the implementation of 3 portal-based interventions within the Improving the Management of symPtoms during And following Cancer Treatment (IMPACT) Consortium. METHODS: Portal use data were extracted from electronic health records for the 12 months preceding intervention implementation. Sociodemographic factors, mode of accessing portals (web vs mobile), and number of clinical encounters before intervention implementation were also extracted. Rurality was derived using rural-urban commuting area codes. Broadband access was estimated using the 2015-2019 American Community Survey. Multiple logistic regression models tested the associations of these factors with portal access (ever accessed or never accessed) and persistence of portal use (accessed the portal ≤20 weeks vs ≥21 weeks in the 35-week study period). RESULTS: Of 28 942 eligible patients, 10 061 (35%) never accessed the portal. Male sex, membership in a racial and ethnic minority group, rural dwelling, not working, and limited broadband access were associated with lower odds of portal access. Younger age and more clinical encounters were associated with higher odds of portal access. Of those with portal access, 25% were persistent users. Using multiple modalities for portal access, being middle-aged, and having more clinical encounters were associated with persistent portal use. CONCLUSION: Patient and structural factors affect portal access and use and may exacerbate disparities in electronic health record-based cancer symptom surveillance and management.

Systematic symptom management in the IMPACT Consortium: rationale and design for 3 effectiveness-implementation trials.

Cancer and its treatment produce deleterious symptoms across the phases of care. Poorly controlled symptoms negatively affect quality of life and result in increased health-care needs and hospitalization. The Improving the Management of symPtoms during And following Cancer Treatment (IMPACT) Consortium was created to develop 3 large-scale, systematic symptom management systems, deployed through electronic health record platforms, and to test them in pragmatic, randomized, hybrid effectiveness and implementation trials. Here, we describe the IMPACT Consortium's conceptual framework, its organizational components, and plans for evaluation. The study designs and lessons learned are highlighted in the context of disruptions related to the COVID-19 pandemic.

Feasibility of Weekly Electronic Health Record-Embedded Patient-Reported Outcomes for Patients Starting Oral Cancer-Directed Therapy.

PURPOSE: To examine the feasibility of integrating a symptom management platform into the electronic health record (EHR) using electronic patient-reported outcomes (ePROs) during oral cancer-directed therapy (OCDT) and explore the impact of prompting oncology nurse navigators (ONNs) to respond to severe symptomatic adverse events (SAEs). MATERIALS AND METHODS: Adults prescribed OCDT at Dana-Farber Cancer Institute were consecutively invited to participate. Participants received weekly messages to complete ePROs. The first half enrolled in a passive (P) group where ePROs responses could be viewed anytime, but outreach was not expected. The second half enrolled in an active (A) group where severe SAEs prompted emails to ONNs for outreach within 1 business day. Feasibility was the proportion of participants completing ≥2 ePROs during the first 30 days. Participants were followed for up to 90 days. RESULTS: From June 25, 2019, to August 18, 2021, 100 participants enrolled, and 96 remained enrolled for at least 30 days. Overall, average age was 59 years, 80% female, and 9% used the platform in Spanish. Twenty-two A (45%) and 27 P (57%) participants met the feasibility threshold (P = .26). ePROs returned at 30 days were similar (P = .50): 0 ePROs 17 A, 13 P; 1 ePRO 10 A, 7 P; 2 ePROs 3 A, 5 P; 3 ePROs 1 A, 4 P; 4 ePROs 7 A, 8 P; and 5 ePROs 11 A, 10 P. Documented telephone encounters at 30 days were similar (109 A, 101 P; P = .86). CONCLUSION: EHR-embedded ePROs administered weekly for people on OCDT was feasible, although many went incomplete. ePRO completion was not clearly affected by nursing calls for severe SAEs. Future efforts will investigate improving engagement and addressing symptoms proactively.

Empirical evaluation of language modeling to ascertain cancer outcomes from clinical text reports.

BACKGROUND: Longitudinal data on key cancer outcomes for clinical research, such as response to treatment and disease progression, are not captured in standard cancer registry reporting. Manual extraction of such outcomes from unstructured electronic health records is a slow, resource-intensive process. Natural language processing (NLP) methods can accelerate outcome annotation, but they require substantial labeled data. Transfer learning based on language modeling, particularly using the Transformer architecture, has achieved improvements in NLP performance. However, there has been no systematic evaluation of NLP model training strategies on the extraction of cancer outcomes from unstructured text. RESULTS: We evaluated the performance of nine NLP models at the two tasks of identifying cancer response and cancer progression within imaging reports at a single academic center among patients with non-small cell lung cancer. We trained the classification models under different conditions, including training sample size, classification architecture, and language model pre-training. The training involved a labeled dataset of 14,218 imaging reports for 1112 patients with lung cancer. A subset of models was based on a pre-trained language model, DFCI-ImagingBERT, created by further pre-training a BERT-based model using an unlabeled dataset of 662,579 reports from 27,483 patients with cancer from our center. A classifier based on our DFCI-ImagingBERT, trained on more than 200 patients, achieved the best results in most experiments; however, these results were marginally better than simpler "bag of words" or convolutional neural network models. CONCLUSION: When developing AI models to extract outcomes from imaging reports for clinical cancer research, if computational resources are plentiful but labeled training data are limited, large language models can be used for zero- or few-shot learning to achieve reasonable performance. When computational resources are more limited but labeled training data are readily available, even simple machine learning architectures can achieve good performance for such tasks.

Financial Toxicity Monitoring in a Randomized Controlled Trial of Patient-Reported Outcomes During Cancer Treatment (Alliance AFT-39).

PURPOSE: Financial toxicity (FT) affects 20% of cancer survivors and is associated with poor clinical outcomes. No large-scale programs have been implemented to mitigate FT. We evaluated the effect of monthly FT screening as part of a larger patient-reported outcomes (PROs) digital monitoring intervention. METHODS: PRO-TECT (AFT-39) is a cluster-randomized trial of patients undergoing systemic therapy for metastatic cancer. Practices were randomly assigned 1:1 to digital symptom monitoring (PRO practices) or usual care (control practices). Digital monitoring consisted of between-visit online or automated telephone patient surveys about symptoms, functioning, and FT (single-item screening question from Functional Assessment of Chronic Illness Therapy-COmprehensive Score for financial Toxicity) for up to 1 year, with automated alerts sent to practice nurses for concerning survey scores. Clinical team actions in response to alerts were not mandated. The primary outcome of this planned secondary analysis was development or worsening of financial difficulties, assessed via the European Organisation for Research and Treatment of Cancer QLQ-C30 financial difficulties measure, at any time compared with baseline. A randomly selected subset of patients and nurses were interviewed about their experiences with the intervention. RESULTS: One thousand one hundred ninety-one patients were enrolled (593 PRO; 598 control) at 52 US community oncology practices. Overall, 30.2% of patients treated at practices that received the FT screening intervention developed, or experienced worsening of, financial difficulties, compared with 39.0% treated at control practices (P = .004). Patients and nurses interviewed stated that FT screening identified patients for financial counseling who otherwise would be reluctant to seek, or unaware of the availability of, assistance. CONCLUSION: In this report of a secondary outcome from a randomized clinical trial, FT screening as part of routine digital patient monitoring with PROs reduced the development, or worsening, of financial difficulties among patients undergoing systemic cancer therapy.

Machine learning for genetics-based classification and treatment response prediction in cancer of unknown primary.

Cancer of unknown primary (CUP) is a type of cancer that cannot be traced back to its primary site and accounts for 3-5% of all cancers. Established targeted therapies are lacking for CUP, leading to generally poor outcomes. We developed OncoNPC, a machine-learning classifier trained on targeted next-generation sequencing (NGS) data from 36,445 tumors across 22 cancer types from three institutions. Oncology NGS-based primary cancer-type classifier (OncoNPC) achieved a weighted F1 score of 0.942 for high confidence predictions ([Formula: see text]) on held-out tumor samples, which made up 65.2% of all the held-out samples. When applied to 971 CUP tumors collected at the Dana-Farber Cancer Institute, OncoNPC predicted primary cancer types with high confidence in 41.2% of the tumors. OncoNPC also identified CUP subgroups with significantly higher polygenic germline risk for the predicted cancer types and with significantly different survival outcomes. Notably, patients with CUP who received first palliative intent treatments concordant with their OncoNPC-predicted cancers had significantly better outcomes (hazard ratio (HR) = 0.348; 95% confidence interval (CI) = 0.210-0.570; P = [Formula: see text]). Furthermore, OncoNPC enabled a 2.2-fold increase in patients with CUP who could have received genomically guided therapies. OncoNPC thus provides evidence of distinct CUP subgroups and offers the potential for clinical decision support for managing patients with CUP.