Abstract Introduction Risk stratification of indeterminate pulmonary nodules (IPN) is challenging. The Autoantibody Test (AAT) is a commercially available blood test validated for identifying malignant IPNs that warrant expedited evaluation. Use in patients with prior cancer is often discouraged because preexisting autoantibodies may cross-react and increase false positives, yet this population has been minimally studied. We evaluated AAT performance by prior-cancer status to determine the impact on test accuracy. Methods The AAT (the Nodify CDT® test, Biodesix Inc., Louisville, CO) evaluates patients’ plasma for the presence of 7 autoantibodies to tumor antigens. Test positivity was defined as ≥ 1 autoantibody above pre-defined thresholds (Moderate or High Level). For this analysis, patient cohorts evaluated with the AAT were aggregated to create a combined study population (n = 1,601). The specific cohorts were: 1013 Study (n = 268), Fred Hutch Cancer Center (n = 337), PANOPTIC (n = 307), CLARIFY (n = 554), and ORACLE (n = 135). Inclusion criteria were: age ≥40, 4-30mm diameter, and no diagnosis of a non-lung cancer or carcinoid. Specificity, false positive rate (FPR), and positive predictive value (PPV) were calculated for each cohort, and comparisons were made using Fisher’s exact. Results Of 1,601 eligible patients, 1,366 had no prior cancer, and 235 with a prior cancer history. Cancer prevalence was 41% (566/1,366) in the no-history group and 49% (116/235) in the prior-cancer group. Among patients with prior cancer, time since last cancer was 5 years in 18% (43/235), 5 to 10 years in 19% (44/235), 10 years in 27% (64/235), and unknown in 36% (84/235). AAT test positivity and performance were similar between groups: positivity 12% (164/1,366) vs 10% (24/235), p = 0.510; performance: specificity 91% (89-93%) vs 91% (84-95%), p = 0.865 and FPR 9% (7-11%) vs 9% (5-16%), p = 0.865, for participants without prior cancer and with prior cancer history, respectively. Risk reclassification was likewise comparable between groups (Figure 1). Additionally, patients with a post-test pCA 65% had an overall PPV of 73% (63-81%) without prior cancer and a PPV of 62% (41-79%) with prior cancer, p = 0.4261. Conclusions This analysis shows that the AAT maintains high specificity, low false-positive rates, and strong positive predictive value in a large clinical cohort. Performance did not differ by prior cancer history, indicating that prior cancer status does not improve or diminish test accuracy. PPV remained high in both groups, particularly among patients with a post-test cancer probability 65%, supporting broader application of the AAT as a tool for risk-stratifying IPNs in practice. This abstract is funded by: Biodesix
Kammer et al. (Fri,) studied this question.