Development of an Electronic Health Record-Based Algorithm for Predicting Lung Cancer Screening Eligibility in the Population-Based Research to Optimize the Screening Process Lung Research Consortium

PURPOSE

Lung cancer screening (LCS) guidelines in the United States recommend LCS for those age 50-80 years with at least 20 pack-years smoking history who currently smoke or quit within the last 15 years. We tested the performance of simple smoking-related criteria derived from electronic health record (EHR) data and developed and tested the performance of a multivariable model in predicting LCS eligibility.

METHODS

Analyses were completed within the Population-based Research to Optimize the Screening Process Lung Consortium (PROSPR-Lung). In our primary validity analyses, the reference standard LCS eligibility was based on self-reported smoking data collected via survey. Within one PROSPR-Lung health system, we used a training data set and penalized multivariable logistic regression using the Least Absolute Shrinkage and Selection Operator to select EHR-based variables into the prediction model including demographics, smoking history, diagnoses, and prescription medications. A separate test data set assessed model performance. We also conducted external validation analysis in a separate health system and reported AUC, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy metrics associated with the Youden Index.

RESULTS

There were 14,214 individuals with survey data to assess LCS eligibility in primary analyses. The overall performance for assigning LCS eligibility status as measured by the AUC values at the two health systems was 0.940 and 0.938. At the Youden Index cutoff value, performance metrics were as follows: accuracy, 0.855 and 0.895; sensitivity, 0.886 and 0.920; specificity, 0.896 and 0.850; PPV, 0.357 and 0.444; and NPV, 0.988 and 0.992.

CONCLUSION

Our results suggest that health systems can use an EHR-derived multivariable prediction model to aid in the identification of those who may be eligible for LCS.

The Philadelphia Lung Cancer Learning Community: a multi-health-system, citywide approach to lung cancer screening

BACKGROUND

Lung cancer screening uptake for individuals at high risk is generally low across the United States, and reporting of lung cancer screening practices and outcomes is often limited to single hospitals or institutions. We describe a citywide, multicenter analysis of individuals receiving lung cancer screening integrated with geospatial analyses of neighborhood-level lung cancer risk factors.

METHODS

The Philadelphia Lung Cancer Learning Community consists of lung cancer screening clinicians and researchers at the 3 largest health systems in the city. This multidisciplinary, multi-institutional team identified a Philadelphia Lung Cancer Learning Community study cohort that included 11 222 Philadelphia residents who underwent low-dose computed tomography for lung cancer screening from 2014 to 2021 at a Philadelphia Lung Cancer Learning Community health-care system. Individual-level demographic and clinical data were obtained, and lung cancer screening participants were geocoded to their Philadelphia census tract of residence. Neighborhood characteristics were integrated with lung cancer screening counts to generate bivariate choropleth maps.

RESULTS

The combined sample included 37.8% Black adults, 52.4% women, and 56.3% adults who currently smoke. Of 376 residential census tracts in Philadelphia, 358 (95.2%) included 5 or more individuals undergoing lung cancer screening, and the highest counts were geographically clustered around each health system's screening sites. A relatively low percentage of screened adults resided in census tracts with high tobacco retailer density or high smoking prevalence.

CONCLUSIONS

The sociodemographic characteristics of lung cancer screening participants in Philadelphia varied by health system and neighborhood. These results suggest that a multicenter approach to lung cancer screening can identify vulnerable areas for future tailored approaches to improving lung cancer screening uptake. Future directions should use these findings to develop and test collaborative strategies to increase lung cancer screening at the community and regional levels.

Application of team science best practices to the project management of a large, multi-site lung cancer screening research consortium

Research is increasingly conducted through multi-institutional consortia, and best practices for establishing multi-site research collaborations must be employed to ensure efficient, effective, and productive translational research teams. In this manuscript, we describe how the Population-based Research to Optimize the Screening Process Lung Research Center (PROSPR-Lung) utilized evidence-based Science of Team Science (SciTS) best practices to establish the consortium's infrastructure and processes to promote translational research in lung cancer screening. We provide specific, actionable examples of how we: (1) developed and reinforced a shared mission, vision, and goals; (2) maintained a transparent and representative leadership structure; (3) employed strong research support systems; (4) provided efficient and effective data management; (5) promoted interdisciplinary conversations; and (6) built a culture of trust. We offer guidance for managing a multi-site research center and data repository that may be applied to a variety of settings. Finally, we detail specific project management tools and processes used to drive collaboration, efficiency, and scientific productivity.

Radiomic phenotyping of the lung parenchyma in a lung cancer screening cohort

High-throughput extraction of radiomic features from low-dose CT scans can characterize the heterogeneity of the lung parenchyma and potentially aid in identifying subpopulations that may have higher risk of lung diseases, such as COPD, and lung cancer due to inflammation or obstruction of the airways. We aim to determine the feasibility of a lung radiomics phenotyping approach in a lung cancer screening cohort, while quantifying the effect of different CT reconstruction algorithms on phenotype robustness. We identified low-dose CT scans (n = 308) acquired with Siemens Healthineers scanners from patients who completed low-dose CT within our lung cancer screening program between 2015 and 2018 and had two different sets of image reconstructions kernel available (i.e., medium (I30f.), sharp (I50f.)) for the same acquisition. Following segmentation of the lung field, a total of 26 radiomic features were extracted from the entire 3D lung-field using a previously validated fully-automated lattice-based software pipeline, adapted for low-dose CT scans. The lattice in-house software was used to extract features including gray-level histogram, co-occurrence, and run-length descriptors. The lattice approach uses non-overlapping windows for traversing along pixels of images and calculates different features. Each feature was averaged for each scan within a range of lattice window sizes (W) of 4, 8 and 20 mm. The extracted imaging features from both datasets were harmonized to correct for differences in image acquisition parameters. Subsequently, unsupervised hierarchical clustering was applied on the extracted features to identify distinct phenotypic patterns of the lung parenchyma, where consensus clustering was used to identify the optimal number of clusters (K = 2). Differences between phenotypes for demographic and clinical covariates including sex, age, BMI, pack-years of smoking, Lung-RADS and cancer diagnosis were assessed for each phenotype cluster, and then compared across clusters for the two different CT reconstruction algorithms using the cluster entanglement metric, where a lower entanglement coefficient corresponds to good cluster alignment. Furthermore, an independent set of low-dose CT scans (n = 88) from patients with available pulmonary function data on lung obstruction were analyzed using the identified optimal clusters to assess associations to lung obstruction and validate the lung phenotyping paradigm. Heatmaps generated by radiomic features identified two distinct lung parenchymal phenotype patterns across different feature extraction window sizes, for both reconstruction algorithms (P < 0.05 with K = 2). Associations of radiomic-based clusters with clinical covariates showed significant differences for BMI and pack-years of smoking (P < 0.05) for both reconstruction kernels. Radiomic phenotype patterns were more similar across the two reconstructed kernels, when smaller window sizes (W = 4 and 8 mm) were used for radiomic feature extraction, as deemed by their entanglement coefficient. Validation of clustering approaches using cluster mapping for the independent sample with lung obstruction also showed two statistically significant phenotypes (P < 0.05) with significant difference for BMI and smoking pack-years. Radiomic analysis can be used to characterize lung parenchymal phenotypes from low-dose CT scans, which appear reproducible for different reconstruction kernels. Further work should seek to evaluate the effect of additional CT acquisition parameters and validate these phenotypes in characterizing lung cancer screening populations, to potentially better stratify disease patterns and cancer risk.

Effect of Financial Incentives on Patient Use of Mailed Colorectal Cancer Screening Tests: A Randomized Clinical Trial

Importance

Mailing fecal immunochemical test (FIT) kits to patients' homes has been shown to boost colorectal cancer screening rates, but response rates remain limited, and organized programs typically require repeated outreach attempts. Behavioral economics has shown that offering salient financial incentives to patients may increase participation in preventive health.

Objective

To compare the impact of different financial incentives for mailed FIT outreach.

Design, Setting, and Participants

This 4-parallel-arm randomized clinical trial included patients aged 50 to 75 years who had an established primary care clinician, at least 2 visits in the prior 2 years, and were eligible for colorectal cancer screening and not up-to-date. This study was conducted at urban primary care practices at an academic health system from December 2015 to February 2018. Data analysis was conducted from March 2018 to September 2018.

Interventions

Eligible patients received a letter from their primary care clinician that included a mailed FIT kit and instructions for use. They were randomized in a 1:1:1:1 ratio to receive (1) no financial incentive; (2) an unconditional $10 incentive included with the mailing; (3) a $10 incentive conditional on FIT completion; or (4) a conditional lottery with a 1-in-10 chance of winning $100 after FIT completion.

Main Outcomes and Measures

Fecal immunochemical test completion within 2 and 6 months of initial outreach.

Results

A total of 897 participants were randomized, with a median age of 57 years (interquartile range, 52-62 years); 56% were women, and 69% were black. The overall completion rate across all arms was 23.5% at 2 months. The completion rate at 2 months was 26.0% (95% CI, 20.4%-32.3%) in the no incentive arm, 27.2% (95% CI, 21.5%-33.6%) in the unconditional incentive arm, 23.2% (95% CI, 17.9%-29.3%) in the conditional incentive arm, and 17.7% (95% CI, 13.0%-23.3%) in the lottery incentive arm. None of the arms with an incentive were statistically superior to the arm without incentive. The overall FIT completion rate across all arms was 28.9% at 6 months, and there was also no difference by arm. The completion rate at 6 months was 32.7% (95% CI, 26.6%-39.3%) in the no incentive arm, 31.7% (95% CI, 25.7%-38.2%) in the unconditional incentive arm, 26.8% (95% CI, 21.1%-33.1%) in the conditional incentive arm, and 24.3% (95% CI, 18.9%-30.5%) in the lottery incentive arm.

Conclusions and Relevance

Mailed FIT resulted in high colorectal cancer screening response rates in this population, but different forms of financial incentives of the same expected value ($10) did not incrementally increase FIT completion rates. The incentive value may have been too small or financial incentives may not be effective in this context.

Trial Registration

ClinicalTrials.gov Identifier: NCT02594150.

Modifiable Failures in the Colorectal Cancer Screening Process and Their Association With Risk of Death

BACKGROUND & AIMS

Colorectal cancer (CRC) deaths occur when patients do not receive screening or have inadequate follow-up of abnormal results or when the screening test fails. We have few data on the contribution of each to CRC-associated deaths or factors associated with these events.

METHODS

We performed a retrospective cohort study of patients in the Kaiser Permanente Northern and Southern California systems (55-90 years old) who died of CRC from 2006 through 2012 and had ≥5 years of enrollment before diagnosis. We compared data from patients with those from a matched cohort of cancer-free patients in the same system. Receipt, results, indications, and follow-up of CRC tests in the 10-year period before diagnosis were obtained from electronic databases and chart audits.

RESULTS

Of 1750 CRC deaths, 75.9% (n = 1328) occurred in patients who were not up to date in screening and 24.1% (n = 422) occurred in patients who were up to date. Failure to screen was associated with fewer visits to primary care physicians. Of 3486 cancer-free patients, 44.6% were up to date in their screening. Patients who were up to date in their screening had a lower risk of CRC death (odds ratio, 0.38; 95% confidence interval, 0.33-0.44). Failure to screen, or failure to screen at appropriate intervals, occurred in a 67.8% of patients who died of CRC vs 53.2% of cancer-free patients; failure to follow-up on abnormal results occurred in 8.1% of patients who died of CRC vs 2.2% of cancer-free patients. CRC death was associated with higher odds of failure to screen or failure to screen at appropriate intervals (odds ratio, 2.40; 95% confidence interval, 2.07-2.77) and failure to follow-up on abnormal results (odds ratio, 7.26; 95% confidence interval, 5.26-10.03).

CONCLUSIONS

Being up to date on screening substantially decreases the risk of CRC death. In 2 health care systems with high rates of screening, most people who died of CRC had failures in the screening process that could be rectified, such as failure to follow-up on abnormal findings; these significantly increased the risk for CRC death.