JCO Clinical Cancer Informatics最新文献

Purpose: Emoji are digital images or icons used to express an idea or emotion in electronic communication. The purpose of this study was to develop and evaluate the psychometric properties of two patient-reported scales that incorporate emoji.

Methods: The Emoji Response Scale developed for this study has two parts: the Emoji-Ordinal and Emoji-Mood scales. A pilot study was designed to validate the ordinal nature of the Emoji-Ordinal Scale. Twenty patients were shown all possible pairs of five emoji and asked to select the most positive from each pair. The psychometric ordering was assessed using Coombs unfolding and Thurstone scaling. A separate pilot study was designed to determine which emoji to include in the Emoji-Mood Scale. Ten common feelings experienced by patients with cancer were chosen by the study team. Patients and providers were asked to select the one emoji that best represented each feeling from the selection. The most commonly selected emotions and representative emoji were chosen for the Emoji-Mood Scale. In a randomized study of 294 patients, Spearman correlations, Wilcoxon tests, and Bland-Altman analyses determined the construct validity of the scales compared with Linear Analog Scale Assessments (LASA) and Patient-Reported Outcomes Measurement Information System (PROMIS) scores.

Results: Ninety-five percent of patients selected the same ordering among the ordinal emoji, and Thurstone scaling confirmed the ordinal nature of the response scale. The construct validity of the scales was high with correlations between the Emoji-Ordinal Scale and the LASA scale of 0.70 for emotional well-being, 0.72 for physical well-being, 0.74 for overall quality of life, and -0.81 for fatigue. Emoji-Mood Scale ratings were strongly related to PROMIS global mental, global physical, fatigue, anxiety, sleep disturbance, and social activity scales (P < .0001).

Conclusion: This study provides evidence that scales incorporating emoji are valid for collecting patient-reported outcomes.

Purpose: Modeling passively collected smartphone sensor data (called digital phenotyping) has the potential to detect distress among family caregivers and patients with advanced cancer and could lead to novel clinical models of cancer care. The purpose of this study was to assess the feasibility and acceptability of collecting passive smartphone data from family caregivers and their care recipients with advanced cancer over 24 weeks.

Methods: This was an observational feasibility study. Family caregivers and patients with advanced cancer were recruited through clinic or via social media and downloaded a digital phenotyping application (Beiwe) to their smartphones that passively collected sensor data over 24 weeks. Feasibility was evaluated by quantifying enrollment and retention and the quantity of acquired data. Acceptability was assessed through post-24 week qualitative interviews.

Results: Of 178 caregiver and patient dyads approached, 22.5% of caregivers (n = 40) and 10.1% of patients (n = 18) both consented to the study and successfully downloaded the application, with most recruited through social media (93%). Of 24 weeks (168 days), the median number of days that data were received was 141 days. Interviews yielded three themes: (1) experiences with study procedures were generally positive despite some technical challenges; (2) security and privacy concerns were minimal, mitigated by clear explanations, trust in the health care system, and privacy norms; and (3) a clinical model that used passive smartphone monitoring to automatically trigger assistance could be beneficial but with concern about false alarms.

Conclusion: This pilot study of collecting passive smartphone data found mixed indicators of feasibility, with suboptimal enrollment rates, particularly via clinic, but positive retention and data collection rates for those who did enroll. Participants had generally positive views of passive monitoring.

Purpose: Immune checkpoint inhibitors (ICIs) have demonstrated promise in the treatment of various cancers. Single-drug ICI therapy (immuno-oncology [IO] monotherapy) that targets PD-L1 is the standard of care in patients with advanced non-small cell lung cancer (NSCLC) with PD-L1 expression ≥50%. We sought to find out if a machine learning (ML) algorithm can perform better as a predictive biomarker than PD-L1 alone.

Methods: Using a real-world, nationwide electronic health record-derived deidentified database of 38,048 patients with advanced NSCLC, we trained binary prediction algorithms to predict likelihood of 12-month progression-free survival (PFS; 12-month PFS) and 12-month overall survival (OS; 12-month OS) from initiation of first-line therapy. We evaluated the algorithms by calculating the AUC on the test set. We plotted Kaplan-Meier curves and fit Cox survival models comparing survival between patients who were classified as low-risk (LR) for 12-month disease progression or 12-month mortality versus those classified as high-risk.

Results: The ML algorithms achieved an AUC of 0.701 (95% CI, 0.689 to 0.714) and 0.718 (95% CI, 0.707 to 0.730) for 12-month PFS and 12-month OS, respectively. Patients in the LR group had lower 12-month disease progression (hazard ratio [HR], 0.47 [95% CI, 0.45 to 0.50]; P < .001) and 12-month all-cause mortality (HR, 0.31 [95% CI, 0.29 to 0.34]; P < .0001) compared with the high-risk group. Patients deemed LR for disease progression and mortality on IO monotherapy were less likely to progress (HR, 0.53 [95% CI, 0.46 to 0.61]; P < .0001) or die (HR, 0.30 [95% CI, 0.24 to 0.37]; P < .001) compared with the high-risk group.

Conclusion: An ML algorithm can more accurately predict response to first-line therapy, including IO monotherapy, in patients with advanced NSCLC, compared with PD-L1 alone. ML may better aid clinical decision making in oncology than a single biomarker.

Purpose: To explore the perceived utility and effect of simplified radiology reports on oncology patients' knowledge and feasibility of large language models (LLMs) to generate such reports.

Materials and methods: This study was approved by the Institute Ethics Committee. In phase I, five state-of-the-art LLMs (Generative Pre-Trained Transformer-4o [GPT-4o], Google Gemini, Claude Opus, Llama-3.1-8B, and Phi-3.5-mini) were tested to simplify 50 oncology computed tomography (CT) report impressions using five distinct prompts with each LLM. The outputs were evaluated quantitatively using readability indices. Five LLM-prompt combinations with best average readability scores were also assessed qualitatively, and the best LLM-prompt combination was selected. In phase II, 100 consecutive oncology patients were randomly assigned into two groups: original report (received original report impression) and simplified report (received LLM-generated simplified versions of their CT report impressions under the supervision of a radiologist). A questionnaire assessed the impact of these reports on patients' knowledge and perceived utility.

Results: In phase I, Claude Opus-Prompt 3 (explain to a 15-year-old) performed slightly better than other LLMs, although scores for GPT-4o, Gemini, Claude Opus, and Llama-3.1 were not significantly different (P > .0033 on Wilcoxon signed-rank test with Bonferroni correction). In phase II, simplified report group patients demonstrated significantly better knowledge of primary site and extent of their disease as well as showed significantly higher confidence and understanding of the report (P < .05 for all). Only three (of 50) simplified reports required corrections by the radiologist.

Conclusion: Simplified radiology reports significantly enhanced patients' understanding and confidence in comprehending their medical condition. LLMs performed very well at this simplification task; therefore, they can be potentially used for this purpose, although there remains a need for human oversight.

Purpose: Lung cancer screening (LCS) has the potential to reduce mortality and detect lung cancer at its early stages, but the high false-positive rate associated with low-dose computed tomography (LDCT) for LCS acts as a barrier to its widespread adoption. This study aims to develop computable phenotype (CP) algorithms on the basis of electronic health records (EHRs) to identify individual's eligibility for LCS, thereby enhancing LCS utilization in real-world settings.

Materials and methods: The study cohort included 5,778 individuals who underwent LDCT for LCS from 2012 to 2022, as recorded in the University of Florida Health Integrated Data Repository. CP rules derived from LCS guidelines were used to identify potential candidates, incorporating both structured EHR and clinical notes analyzed via natural language processing. We then conducted manual reviews of 453 randomly selected charts to refine and validate these rules, assessing CP performance using metrics, for example, F1 score, specificity, and sensitivity.

Results: We developed an optimal CP rule that integrates both structured and unstructured data, adhering to the US Preventive Services Task Force 2013 and 2020 guidelines. This rule focuses on age (55-80 years for 2013 and 50-80 years for 2020), smoking status (current, former, and others), and pack-years (≥30 for 2013 and ≥20 for 2020), achieving F1 scores of 0.75 and 0.84 for the respective guidelines. Including unstructured data improved the F1 score performance by up to 9.2% for 2013 and 12.9% for 2020, compared with using structured data alone.

Conclusion: Our findings underscore the critical need for improved documentation of smoking information in EHRs, demonstrate the value of artificial intelligence techniques in enhancing CP performance, and confirm the effectiveness of EHR-based CP in identifying LCS-eligible individuals. This supports its potential to aid clinical decision making and optimize patient care.

Purpose: Undifferentiated pleomorphic sarcomas (UPSs) demonstrate therapy-induced hemosiderin deposition, granulation tissue formation, fibrosis, and calcification. We aimed to determine the treatment-assessment value of morphologic tumoral hemorrhage patterns and first- and high-order radiomic features extracted from contrast-enhanced susceptibility-weighted imaging (CE-SWI).

Materials and methods: This retrospective institutional review board-authorized study included 33 patients with extremity UPS with magnetic resonance imaging and resection performed from February 2021 to May 2023. Volumetric tumor segmentation was obtained at baseline, postsystemic chemotherapy (PC), and postradiation therapy (PRT). The pathology-assessed treatment effect (PATE) in surgical specimens separated patients into responders (R; ≥90%, n = 16), partial responders (PR; 89%-31%, n = 10), and nonresponders (NR; ≤30%, n = 7). RECIST, WHO, and volume were assessed for all time points. CE-SWI T2* morphologic patterns and 107 radiomic features were analyzed.

Results: A Complete-Ring (CR) pattern was observed in PRT in 71.4% of R (P = 7.71 × 10^-6), an Incomplete-Ring pattern in 33.3% of PR (P = .2751), and a Globular pattern in 50% of NR (P = .1562). The first-order radiomic analysis from the CE-SWI intensity histogram outlined the values of the 10th and 90th percentiles and their skewness. R showed a 280% increase in 10th percentile voxels (P = .061) and a 241% increase in skewness (P = .0449) at PC. PR/NR showed a 690% increase in the 90th percentile voxels (P = .03) at PC. Multiple high-order radiomic texture features observed at PRT discriminated better R versus PR/NR than the first-order features.

Conclusion: CE-SWI morphologic patterns strongly correlate with PATE. The CR morphology pattern was the most frequent in R and had the highest statistical association predicting response at PRT, easily recognized by a radiologist not requiring postprocessing software. It can potentially outperform size-based metrics, such as RECIST. The first- and high-order radiomic analysis found several features separating R versus PR/NR.

Purpose: Current image-based long-term risk prediction models do not fully use previous screening mammogram images. Dynamic prediction models have not been investigated for use in routine care.

Methods: We analyzed a prospective WashU clinic-based cohort of 10,099 cancer-free women at entry (between November 3, 2008 and February 2012). Follow-up through 2020 identified 478 pathology-confirmed breast cancers (BCs). The cohort included 27% Black women. An external validation cohort (Emory) included 18,360 women screened from 2013, followed through 2020. This included 42% Black women and 332 pathology-confirmed BC excluding those diagnosed within 6 months of screening. We trained a dynamic model using repeated screening mammograms at WashU to predict 5-year risk. This opportunistic screening service presented a range of mammogram images for each woman. We applied the model to the external validation data to evaluate discrimination performance (AUC) and calibrated to US SEER.

Results: Using 3 years of previous mammogram images available at the current screening visit, we obtained a 5-year AUC of 0.80 (95% CI, 0.78 to 0.83) in the external validation. This represents a significant improvement over the current visit mammogram AUC 0.74 (95% CI, 0.71 to 0.77; P < .01) in the same women. When calibrated, a risk ratio of 21.1 was observed comparing high (>4%) to very low (<0.3%) 5-year risk. The dynamic model classified 16% of the cohort as high risk among whom 61% of all BCs were diagnosed. The dynamic model performed comparably in Black and White women.

Conclusion: Adding previous screening mammogram images improves 5-year BC risk prediction beyond static models. It can identify women at high risk who might benefit from supplemental screening or risk-reduction strategies.

Purpose: Breast cancer relapses are rarely collected by cancer registries because of logistical and financial constraints. Hence, we investigated natural language processing (NLP), enhanced with state-of-the-art deep learning transformer tools and large language models, to automate relapse identification in the text of computed tomography (CT) reports.

Methods: We analyzed follow-up CT reports from patients diagnosed with breast cancer between January 1, 2005, and December 31, 2014. The reports were curated and annotated for the presence or absence of local, regional, and distant breast cancer relapses. We performed 10-fold cross-validation to evaluate models identifying different types of relapses in CT reports. Model performance was assessed with classification metrics, reported with 95% confidence intervals.

Results: In our data set of 1,445 CT reports, 799 (55.3%) described any relapse, 72 (5.0%) local relapses, 97 (6.7%) regional relapses, and 743 (51.4%) distant relapses. The any-relapse model achieved an accuracy of 89.6% (87.8-91.1), with a sensitivity of 93.2% (91.4-94.9) and a specificity of 84.2% (80.9-87.1). The local relapse model achieved an accuracy of 94.6% (93.3-95.7), a sensitivity of 44.4% (32.8-56.3), and a specificity of 97.2% (96.2-98.0). The regional relapse model showed an accuracy of 93.6% (92.3-94.9), a sensitivity of 70.1% (60.0-79.1), and a specificity of 95.3% (94.2-96.5). Finally, the distant relapse model demonstrated an accuracy of 88.1% (86.2-89.7), a sensitivity of 91.8% (89.9-93.8), and a specificity of 83.7% (80.5-86.4).

Conclusion: We developed NLP models to identify local, regional, and distant breast cancer relapses from CT reports. Automating the identification of breast cancer relapses can enhance data collection about patient outcomes.

Purpose: Artificial intelligence (AI) applications in radiotherapy (RT) are expected to save time and improve quality, but implementation remains limited. Therefore, we used implementation science to develop a format for designing an implementation strategy for AI. This study aimed to (1) apply this format to develop an AI implementation strategy for our center; (2) identify insights gained to enhance AI implementation using this format; and (3) assess the feasibility and acceptability of this format to design a center-specific implementation strategy for departments aiming to implement AI.

Methods: We created an AI-implementation strategy for our own center using implementation science methods. This included a stakeholder analysis, literature review, and interviews to identify facilitators and barriers, and designed strategies to overcome the barriers. These methods were subsequently used in a workshop with teams from seven Dutch RT centers to develop their own AI-implementation plans. The applicability, appropriateness, and feasibility were evaluated by the workshop participants, and relevant insights for AI implementation were summarized.

Results: The stakeholder analysis identified internal (physicians, physicists, RT technicians, information technology, and education) and external (patients and representatives) stakeholders. Barriers and facilitators included concerns about opacity, privacy, data quality, legal aspects, knowledge, trust, stakeholder involvement, ethics, and multidisciplinary collaboration, all integrated into our implementation strategy. The workshop evaluation showed high acceptability (18 participants [90%]), appropriateness (17 participants [85%]), and feasibility (15 participants [75%]) of the implementation strategy. Sixteen participants fully agreed with the format.

Conclusion: Our study highlights the need for a collaborative approach to implement AI in RT. We designed a strategy to overcome organizational challenges, improve AI integration, and enhance patient care. Workshop feedback indicates the proposed methods are useful for multiple RT centers. Insights gained by applying the methods highlight the importance of multidisciplinary collaboration in the development and implementation of AI.