Improved Lung Cancer Screening Approaches for People Living with HIV

About This Grant

PROJECT SUMMARY Our overall objective is to identify the best strategies for lung cancer screening (LCS) in a representative group of people with HIV (PWH). Lung cancer (LC) is now the leading source of cancer mortality among PWH. Unfortunately, most LCs in PWH are diagnosed at a late stage and have very poor survival. Thus, LCS among PWH is needed to stem this growing source of preventable mortality. Two large trials demonstrated that LCS with low-dose chest computed tomography (LDCT) reduced LC mortality among heavy smokers without HIV (PWoH). We have confirmed the benefits of LCS in well-controlled PWH who have a high-risk smoking history and guidelines recommend LCS for PWH. Many issues regarding the optimal indications for LCS in PWH remain unclear, however. Research in PWoH has shown that LC risk prediction models, such as the PLCOm2012 score, are superior to current eligibility criteria for LCS. PWH develop LC with less tobacco exposure and at earlier ages and existing LC risk models have not been developed or tested in PWH, a population for which several unique LC risk factors (e.g., low CD4/CD8 ratio values) exist. Additionally, false positive lung nodules are common in LCS. Robust algorithms to work-up these nodules are critical to minimize the harms of LCS. However, current nodule follow-up protocols do not consider several important factors affecting PWH like an increased rate of false positive screens, follow-up testing complications and competing risks of death. While LCS is recommended for PWH with well-controlled HIV, several other factors such as increased comorbidities and other HIV-related factors that affect life expectancy and quality of life may impact the optimal regimen for LCS in PWH. Therefore, the VACS index, a validated HIV mortality risk index, may be a convenient tool for further improving the screening decision-making process. Our overall goal is to use simulation modeling to improve LCS regimens for PWH. Our Aims are to: 1) Derive and validate an HIV- specific LC risk prediction model; 2) Compare the impact of LCS eligibility criteria based on USPSTF guidelines, existing risk prediction models for PWoH, and a novel HIV-specific LC risk prediction model on LC mortality reduction in PWH; 3) Identify optimal recommendations for management of screen-detected nodules in PWH; and 4) Determine personalized indications and LCS regimens according to HIV-related prognosis (based on the VACS index) and major comorbidities common among PWH (chronic obstructive pulmonary disease, cardiovascular disease and liver disease) that maximize benefits and minimize harms. In Aim 1, we will apply machine learning methods to data from several of the largest US HIV cohorts to derive and validate an HIV-specific LC risk prediction model. For Aims 2 to 4, we will update and use a well-established simulation modeling framework previously developed by our group. Using the updated model, we will evaluate the best criteria for eligibility for LDCT screening in PWH, identify the most effective strategies to work-up screen- detected nodules and assess personalized LCS indications that consider HIV-related prognosis.