Lung cancer screening: Testing the accuracy of risk prediction for Maori

This is a cross-sectional study of people offered one round of lung cancer screening. This study aims to test how an international model for determining the risk of lung cancer (the ‘PLCOm2012’ model) performs in an Aotearoa New Zealand context, specifically with Maori and European populations, and to provide an appropriate ethnic weighting to optimise this model to use locally. The most appropriate ethnic weighting within the PLCOm2012 risk prediction model for lung cancer risk will be determined by comparing risk and outcomes for Maori and European populations. We will not be able to determine the absolute accuracy of the risk predictions from the PLCO model are in the NZ context in this study as we will not have 6 years follow-up. Instead we are using a prevalence outcome which will provide the relative risk associated with Maori ethnicity compared with European ethnicity so that for any given screening threshold (1.7% etc) the algorithm will not underestimate Maori risk. In doing so, we will describe key outcomes that are required to inform a potential national lung cancer screening programme in NZ, and will examine whanau activation and how participants have experienced the lung cancer screening pathway. We will also test the effectiveness of a novel mechanism for inviting participants for lung screening, a voucher provided by friends/whanau. We will be testing the PLCOm2012 model, hypothesising that the PLCOm2012 risk prediction model (used without an ethnic weighting) will under estimate risk for Maori participants. The Prostate Lung Colorectal and Ovarian (PLCOm2012) model is a lung cancer risk prediction model which has been used and validated by multiple lung cancer screening projects internationally. The PLCOm2012 is a logistic regression model based on cancer incidence (over 6 years) that occurred in the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial (PLCO). This model has higher sensitivity, specificity and positive predictive value when compared to the NLST risk prediction criteria (which defines a person as of “high risk” when they have a smoking history of greater than or equal to 30 pack years, have smoked within the past 15 years, and are aged between 55–74). The PLCOm2012 model was developed using U.S. data, and includes variables such as age, education, body mass index, personal history of cancer, family history of lung cancer, COPD, smoking status, tobacco consumption, smoking duration and time since quitting. (ref Tammemagi MC, ten Haag K, Toumazis I, Kong CY, Han SS, Jeon J, et al. Development and validation of a multivariable lung cancer risk prediction model that includes low-dose computed tomography screening results: A secondary analysis of data from the National Lung Cancer Screening Trial. JAMA Netw Open [Internet]. 2019 [cited 2020 Nov 11]; 2(3): e190204) Potentially eligible participants will receive a written invitation to participate in the study, and followed up with a phone call to confirm eligibility and consent to participate. The first stage of participation is to undergo the PLCO risk assessment. The risk assessment will be done by the project research nurse in-person, over the phone or by Zoom/digital technology, depending on participant preference, or self-completed online. Participants who meet risk threshold or criteria-based eligibility for LCS will be offered a Shared Decision-Making process. In this process, participants and the study nurse share the best available evidence when making decisions, and participants are supported to consider options to achieve informed choice. After the shared decision making process, participants who consent will be offered a CT scan of the chest. Results will be followed up and participants informed and offered appropriate follow up, either through their GP or through referral to respiratory / oncology services Estimated time required by participants is as follows: Risk assessment - 10 minutes Shared decision making - 15 minutes Low dose CT scan - 2 minutes (for the scan) plus consenting time and waiting time The study research nurses will collect information during the risk assessment conversation Some participant responses will be double checked at CT scan to audit fidelity Study nurses will interact with the model and if identified as high risk, participants will be offered a CT scan at a community scanning facility. Results will be followed up and participants informed and offered appropriate follow up, either through their GP or through referral to respiratory / oncology services