Applying Risk Prediction Models to Optimize Lung Cancer Screening: Current Knowledge, Challenges, and Future Directions

Abstract

Structured Abstract Purpose of review Risk prediction models may be useful for facilitating effective and high-quality decision-making at critical steps in the lung cancer screening process. This review provides a current overview of published lung cancer risk prediction models and their applications to lung cancer screening and highlights both challenges and strategies for improving their predictive performance and use in clinical practice. Recent findings Since the 2011 publication of the National Lung Screening Trial results, numerous prediction models have been proposed to estimate the probability of developing or dying from lung cancer or the probability that a pulmonary nodule is malignant. Respective models appear to exhibit high discriminatory accuracy in identifying individuals at highest risk of lung cancer or differentiating malignant from benign pulmonary nodules. However, validation and critical comparison of the performance of these models in independent populations are limited. Little is also known about the extent to which risk prediction models are being applied in clinical practice and influencing decision-making processes and outcomes related to lung cancer screening. Summary Current evidence is insufficient to determine which lung cancer risk prediction models are most clinically useful and how to best implement their use to optimize screening effectiveness and quality. To address these knowledge gaps, future research should be directed toward validating and enhancing existing risk prediction models for lung cancer and evaluating the application of model-based risk calculators and its corresponding impact on screening processes and outcomes.

Effects of Personalized Risk Information on Patients Referred for Lung Cancer Screening with Low-Dose CT

“…First, PCRI may help maximize the net benefits of LDCT screening from a population perspective by identifying high-risk individuals who are most likely to benefit and enabling screening to be targeted toward them. 4,[11][12][13][14][15] Secondary analyses of data from the NLST and the US Prostate, Lung, Colorectal and Ovarian Cancer (PLCO) Screening Trial have suggested that risk-based screening would result in fewer individuals being screened, more lung cancers identified, fewer false positives, and fewer false negatives. 16,17 A recent analysis, however, has shown that risk-based screening may not improve the cost-effectiveness of screening because higher-risk patients are more costly to screen and have a lower life expectancy if they survive lung cancer.…”

“…18 Nevertheless, PCRI may also improve LDCT screening from an individual, clinical perspective by enhancing informed and shared decision making. 4,8,9,[11][12][13][19][20][21][22] In theory, PCRI enables patients to determine for themselves whether their own disease risks are sufficiently high to justify undertaking risk-reducing action. 23 Individual patients could compare their PCRI with their own personal risk thresholds to make this determination.…”

“…These concerns have also fueled growing interest in an alternative implementation strategy: ''risk-based,'' 4 ''personalized,'' 5,6 or ''precision'' 7 screening based on personalized cancer risk information (PCRI)-that is, evidence-based estimates of an individual's risk of cancer. 2,[8][9][10] Personalized lung cancer risk information is now obtainable from several validated clinical prediction models and may improve LDCT screening in 2 main ways.…”

Risk prediction models versus simplified selection criteria to determine eligibility for lung cancer screening: an analysis of German federal-wide survey and incidence data

“…Several models have been developed on the basis of an individual's age and detailed smoking history, presence of pulmonary disease (e.g., chronicobstructive-pulmonary-disease (COPD), emphysema), family or personal history of cancer, body-mass-index and socio-economic background indicators. Major models (reviewed in [10][11][12]) have been developed in context of large-scale prospective cohorts and trials, particularly in the USA and the United Kingdom (Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial [PLCO], [CARET], American Association of Retired Persons Study [NIH-AARP], NLST, Liverpool Lung Project [LLP]), and have been externally validated in independent cohorts [12][13][14][15]. Follow-up analyses in the NLST and PLCO trials [12,16] and other population cohorts [14,15] have shown that, at equal numbers of individuals selected as eligible as obtained with concise inclusion criteria, risk models identified more individuals actually developing lung cancer.…”

Lung cancer screening and shared decision making in cancer survivors: the long and winding road

“…Sex, race, prior history of cancer, family history of lung cancer and chronic obstructive lung disease (COPD) are predictors of developing lung cancer. These variables have been incorporated into risk-prediction models (24), and incorporation of these variables into patient selection for screening could lead to more benefits and less harms of LCS compared to the use of current recommendations (25). An argument against the use of risk-prediction models for selecting patients at highest risk for lung cancer is that they may result in the selection of patients who are too sick [because of comorbidities such as prior history of cancer, COPD, cardiovascular disease (CVD)] to benefit from LCS because they are unable to undergo treatment and are at increased risk of dying from a cause other than lung cancer (26).…”