Osimertinib is a 3rd generation TKI approved for stage IV EGFR+ NSCLC in the first line or post-progression with T790M. The spectrum of osimertinib resistance mutations and clinical outcomes post-osimertinib progression are not well described. Single-center retrospective review of patients with stage IV EGFR+ NSCLC treated with osimertinib was conducted. Resistance mutations were determined via tissue biopsy or circulating tumor DNA (Guardant) prior to and at time of progression on osimertinib. PFS was calculated using Kaplan-Meier method. PFS1 is start of osimertinib to radiographic progression. PFS2 is start of next therapy after osimertinib to next radiographic progression. We identified 95 patients with stage IV EGFR+ lung adenocarcinoma treated with osimertinib detected via NGS (56/95), real-time PCR (29/95), Sanger sequencing (8/95), and other techniques (2/95). Most patients were female (63/95) and never smokers (72/95). Osimertinib resistance and post-progression patterns are shown in Table 1. Potentially targetable mutations were found in 55% (26/47) samples and 14% (6/47) samples had oncogenes targetable with available TKIs. TP53 mutations prior to osimertinib did not significantly influence PFS (36 weeks vs 39 weeks; p = 0.13). MET amplification was only seen in the setting of undetectable T790M or in patients who received first line osimertinib. Median PFS1 for 1st line EGFR TKI (n=17), 2nd line EGFR TKI (n=41), 3rd or greater line EGFR TKI (n=29) was 36, 45 and 39 weeks respectively (p=0.268) with median follow up of 59, 81, and 64 weeks. 10 patients received locally ablative radiotherapy for oligoprogressive disease (defined as ≤3 progressive sites) and continued osimertinib post-progression with median PFS2 of 49 weeks. There is utility to repeat biopsy after progression on osimertinib as targetable oncogenes can be found. Presence of TP53 prior to starting osimertinib did not influence PFS1. Continuing osimertinib and adding radiotherapy for oligoprogressive disease does increase post-progression PFS.

Based on clinical observation, we hypothesized that ROS1 gene-rearranged non-small cell lung cancer (ROS1+ NSCLC) has a higher than expected thromboembolic event (TEE) rate. A multicenter, retrospective cohort study of TEE in advanced ROS1+, KRAS+, ALK+ and EGFR+ NSCLC was conducted. Venous (DVT / PE) and arterial (MI/TIA/CVA) TEE within +/- 365 days of diagnosis of ROS1+, KRAS+, ALK+ or EGFR+ advanced NSCLC at 4 academic centers in USA and China from October 2002 to January 2018 were captured. The primary endpoint was the incidence of TEE in ROS1+ compared to KRAS+ NSCLC as a control group within +/- 90 days of diagnosis. Secondary endpoints compared TEE incidence between ROS1+ and ALK+, and ROS1+ and EGFR+. Fine-Gray Model was used to detect differences in TEE incidence while accounting for death as a competing risk. 105 ROS1+, 101 ALK+, 112 EGFR+, and 114 KRAS+ NSCLC patients were enrolled. Incidence rate of TEE within +/- 90 days of diagnosis was 30.5% (32/105), 12.9% (13/101), 7.1% (8/112), and 12.3% (14/114) in the respective molecular cohorts. Compared to the ROS1+ cohort, the risk of TEE was significantly lower in the three other cohorts (KRAS+ HR 0.334, 95% CI: 0.18-0.62, p=0.001; ALK+ HR 0.357, 95% CI: 0.188-0.68, p=0.002; EGFR+ HR 0.193, 95% CI: 0.089-0.421, p<0.001) (Figure 1). First event TEEs were venous as opposed to arterial in 59.5% (22/37) ROS1+, 87.1% (27/31) ALK+, 80.6% (25/31) EGFR+, and 80% (16/20) KRAS+ cases. The median time (Interquartile Range) to TEE from the time of diagnosis for ROS1+/ALK+/EGFR+/ KRAS+ was 0 days (-6.75 to 7.0), 0 days (-20.0 to 35.0), 0.50 days (-43.7 to 21.3), and 13 days (0.49 to 32.0), respectively. Among common molecular subtypes of NSCLC, ROS1+ oncogene is associated with a significantly higher risk of developing TEE within +/- 90 days of advanced NSCLC diagnosis.

Central nervous system (CNS) metastases in lung cancer are a frequent cause of morbidity and mortality. There are conflicting data on the incidence of CNS metastases in ROS1+ NSCLC at diagnosis and rate of CNS progression on crizotinib. Retrospective review of 579 patients with stage IV NSCLC between June 2008 to December 2017 was performed. We captured presence of brain metastases and oncogene status. We measured progression free survival (PFS) and time to CNS progression in ROS1+ and ALK+ patients on crizotinib. We identified 33 ROS1+ and 115 ALK+ patients with advanced NSCLC. The incidence of brain metastases for treatment-naïve ROS1+ and ALK+ NSCLC was 36% (12/33) and 34% (39/115) respectively. There were no statistically significant differences in incidence of brain metastases across all oncogene sub-groups. Complete survival data was available for 19 ROS1+ and 83 ALK+ patients. Median PFS for the ROS1+ and ALK+ cohort was 11 and 8 months (p = 0.304). The CNS was the first site of progression for 52% (10/19) ROS1+ NSCLC and 43% (36/83) ALK+ NSCLC with no significant differences between the groups (p = 0.610). Among patients without CNS metastases prior to crizotinib therapy, 50% of ROS1+ and ALK+ patients developed CNS metastases as only site of progression at 24 and 21 months respectively. Brain metastases are common in treatment-naïve stage IV ROS1+ NSCLC, though incidence does not differ from other oncogene cohorts. The CNS is a common first site of progression in patients with ROS1+ NSCLC on crizotinib. This study reinforces the need to develop CNS-penetrant TKIs for patients with ROS1+ NSCLC, similar to ALK+ NSCLC.

Indoleamine 2,3-dioxygenase-1 (IDO-1) is a cytosolic enzyme involved in the catabolism of tryptophan; IDO-1-related immune suppression is due to decreased tryptophan availability and to the generation of tryptophan metabolites, culminating in substantial suppression of T-lymphocytes. Here we investigate IDO-1 expression in a cohort of non-small-cell lung cancer (NSCLC) specimens, both in tumor cells and in immune infiltrate, with correlation of IDO-1 to PD-L1 expression, clinical patient demographics and outcomes. A cohort of 1.200 NSCLC samples were obtained from 437 patients who underwent surgical lung resections at Austin Health, Melbourne, Australia. IDO-1 expression was evaluated by immunohistochemistry. Correlations were assessed using Spearman and Kendall tests. A Cox proportional hazards (PH) model was used to assess if overall survival (OS) was associated with IDO-1 positivity in univariate and multivariable settings. Samples from 437 patients were analyzed for IDO-1 expression, with 111 (25.4%) determined as positive (H-Score ≥ 1) and 326 patients (74.6%) as negative (H-Score: 0). IDO-1 expression was determined to be greater in tumor immune infiltrate, with 406 patients (93.8%) determined as positive, while just 27 (6.2%) were IDO-1 negative. There was a significant positive correlation between IDO-1 positive tumor cells and immune cells (0.2167, p < 0.001). Both continuous and binary versions of tumor H-Score showed a significant positive correlation with the amount of tumor immune infiltrate (0.1806 and 0.1698, p < 0.0001, respectively). None of the analyzed variables (age, sex, histology, stage, EGFR, KRAS and PD-L1 status) were found to display a significant correlation with IDO-1 positivity in tumor and immune cells. IDO-1 positivity in tumor cells was found to be significantly associated with OS in the univariate setting and in the multivariable model where variables age, sex, histology, stage, EGFR, KRAS and PD-L1 status were included [P-value = 0.009 and 0.021, respectively; HR: 0.72 (95% CI: 0.55-0.95)]. IDO-1 positivity in immune cells was found to be significantly associated with OS in the univariate setting and was borderline significant in the multivariable model [P-value = 0.006 and 0.053, respectively; HR: 0.798 (95% CI: 0.635-1.003)]. To our knowledge, this is the most extensive analysis of IDO-1 expression in NSCLC patients reported in the literature. Our results suggest the possible prognostic role of IDO-1 expression in tumor and immune cells, highlighting the relevance of IDO-1 detection in tumor tissue. Since new compounds targeting IDO-1 are actually under investigation, the identification of potential prognostic and predictive biomarkers will be needed.