We have established a real-world Advanced Non-Small Cell Lung Holistic Registry (ANCHoR) to assess how immunotherapy impacts treatment choice, clinical outcomes, and patient-reported outcomes (PROs) of aNSCLC. Our aim in this analysis was to assess the ability of the MDASI-LC to differentiate between patients who are responding or who are progressing during treatment. Between May 2017 and December 2018, patients with aNSCLC at a single institution were enrolled in ANCHoR and completed the MDASI-LC prior to therapy (PTT) and at routine clinic visits. The MDASI-LC consists of 16 symptom severity and 6 interference items rated on 0-10 scales (0 = no symptom or interference, 10 = worst imaginable symptom or complete interference). MDASI-LC scores from PTT to first recorded response determination (FRD) were compared by response group using linear mixed modeling (LMM). One hundred one patients completed the MDASI-LC PTT and at FRD. Mean patient age was 63.8 years (standard deviation = 10.29) and 55% were males. Fifty percent of patients received chemotherapy (CTX), 22% immunotherapy (IM), 19% CTX+IM or angiogenesis inhibitor, and 9% targeted therapy. Median time from PTT to FRD was 105 days (lower quartile = 63, upper quartile = 224). Forty-six percent of patients had a complete or partial response (RECIST criteria CR, PR), 14% had stable disease (RECIST SD), and 41% progressed (RECIST PD). LMM showed progressing patients had significantly more fatigue (estimated effect [est] =1.39; p = 0.031), sleep disturbance (est=1.37; p = 0.046), and drowsiness (est=1.33; p = 0.037) and reported significantly more interference with work (est=1.67; p = 0.016) over time than responding patients. The MDASI-LC differentiated the symptom burden of patients with responding disease from that of patients with progressive disease. Patients with progressive disease had more fatigue, disturbed sleep, drowsiness, and greater interference with work than those with responsive disease. Further research is needed to determine if the MDASI-LC can predict response to therapy in patients and may be useful in delineating treatment benefit.

STK11/LKB1 alterations are found in 20-30% of NSCLC and used to co-occur with KRAS mutations. Because LKB1 activates AMPK, many of the best known functions of LKB1 are attributed to its ability to control metabolic alterations in cells. Our laboratory have previously reported that loss of LKB1 promotes enhanced glycolysis and elevated lactate production and more recently we demonstrated that STK11/LKB1 mutations are the strongest predictors of de novo resistance to immunotherapy in NSCLC. Prior studies have revealed an association between alterations in the LKB1/AMPK pathway and worse clinical outcomes in NSCLC and in patients treated with chemotherapy and bevacizumab. Given the roles of LKB1 in the regulation of cell metabolism and resistance to immunotherapy, it is feasible that LKB1 also impacts on the response to anti-angiogenic therapies. Xenograft mouse models were established by subcutaneous injection of H460 cells (LKB1-deficient) and H460 LKB1-expressing in nude mice and LKR10 (KRASG12D) LKB1 wild-type (K) or LKB1- knockout (KL) into 129Svmice. Mice were randomized to vehicle or B20-4.1.1 anti-VEGF antibody. Glycolytic activity of LKB1-intact and -deficient NSCLC cells was measured by Seahorse assay. We analyzed gene expression of SLC16A3 (MCT4) by qPCR and Western blot. Genetic disruption of MCT4 in the K and KL cell lines was done using CRISPR-Cas9 and mouse models were established by subcutaneous injection into mice. Mice bearing LKB1-expressing H460 xenografts treated with anti-VEGF antibody showed a significant decrease in tumor volume (p<0.05) compared with their vehicle-treated counterparts. However, mice bearing LKB1-deficient H460 xenografts showed markedly reduced efficacy of anti-VEGF therapy compared with that in LKB1-expressing xenografts. Anti-VEGF therapy significantly reduced growth of LKR10 K tumors (p<0.001) but not in LKR10 KL tumors. Microvascular density was not increased in KL tumors following anti-VEGF treatment compared to K. Human isogenic LKB1-deficient cells showed a significantly increased rate of glycolysis and lactate secretion compared with cells expressing LKB1. Human and murine LKB1-deficient cells also had increased MCT4 expression compared to K cells. Immunofluorescence and RPPA analysis of tumor samples from the K and KL mouse models showed that KL tumors upregulated MCT4 protein expression compared with K tumors (p<0.0001). The genetic disruption of MCT4 KL tumors significantly improved tumor volume reduction to anti-VEGF therapies in vivo (p<0.001). LKB1 loss is associated with increased lactate secretion and resistance to VEGF inhibition in NSCLC. The targeting of the lactate transporter MCT4 enhance the sensitivity of LKB1-deficient NSCLC to anti-VEGF therapy.

Recent success using immune checkpoint blockade (ICB) in the metastatic setting has raised the need to understand the immune microenvironment (IME) in early-stage disease. Moreover, pre-clinical evidence suggests that cytotoxic agents can modulate this IME. A recent study conducted by our group showed that non-small cell lung cancer (NSCLC) patients who received neoadjuvant chemotherapy followed by surgery (NCT), as compared to patients who received upfront surgery (US), had higher densities of CD3+ lymphocytes and CD68+ tumor-associated macrophages (TAMs). CD3+CD4+ lymphocytes and TAMs also correlated with better clinical outcomes. In this study, we explored the relationships between NCT and the IME by harvesting tumor samples of multiple surgical NSCLC cohorts. The PROSPECT microarray database was queried in NCT (n=45) and US (n=200) patients to investigate differentially expressed genes related to immunogenic cell death (ICD), susceptibility to CD8+ T cell and NK cell cytotoxicity, priming of antigen presenting cells, immunosuppressive enzymes and intra-tumoral cytokines. Available data from the ImmunogenomiC prOfiling of NSCLC (ICON) and other surgical NSCLC cohorts was evaluated to determine: 1) differential immune profiling using FACS (NCT=17; US=39) and multiplex IHC imaging (NCT=10; US=72); 2) plasma circulating cytokines (NCT=18; US=73); 3) tumor mutational burden (TMB) (NCT=40; US=61). Participants who received NCT or US were excluded according to these criteria: 1) concurrent treatment in addition to NCT; 2) sarcomatoid and small cell histologies; 3) clinical or pathological TNM Stage 4 disease; 4) synchronous malignancies other than lung. PROSPECT NCT patients expressed increased damage-associated molecular pattern (DAMP) genes (HSPA2, HSPA4, HSPE1, and S100A2; p<0.05) and T cell-related chemotaxis and antigen presentation genes (CXCR7, CD1A; p<0.05). Concordantly, the ICON cohort FACS results showed that NCT patients display increases in: 1) infiltration of CD8+ T cells (p=0.004); 2) proliferating Ki67+CD8+ T cells (p=0.02); 3) tissue resident memory CD8+CD103+ (p=0.02) and CD4+CD103+ non-Treg cells (p=0.01). Trends from the ICON multiplex IHC also highlighted increases in CD8+ T cells (p=0.09), CD20+ cells (p=0.08), as well as PD-L1+ malignant cells (p=0.08) and PD-L1+ TAMs (p=0.08) in NCT patients, the latter finding being supported by increased circulating MCP-1 (p=0.03). TMB was similar between NCT and US groups (p=0.912). Our data provides the first evidence of ICD (i.e., increased DAMP gene expression) following NCT in human early-stage NSCLC. Furthermore, our data highlights the association of NCT with a favorable IME (i.e., increased T cell infiltration), supporting the rationale of NCT and ICB combinations in localized NSCLC.

Studies from our group and others have shown that bacteria and viruses present in the tumor may impact therapeutic responses. In the specific context of non-small cell lung cancer (NSCLC), intra-tumoral viral DNA and bacteria have been reported previously to be linked to therapeutic outcomes. However, the interplay between intra-tumoral microorganisms and the host immune response in NSCLC remains unknown. Moreover, the prognostic and predictive therapeutic value of localized NSCLC-specific microbial composition has yet to be defined. RNA-sequencing (RNA-seq) (n=82) and whole exome sequencing (WES) (n=80) was performed on surgically resected (pTNM I-III) tumors from lung cancer patients enrolled in the ImmunogenomiC prOfiling of NSCLC (ICON) project. Intra-tumoral bacteria, viruses and fungi were queried with MetaPhlAn2, a bioinformatical analysis pipeline which employs unique clade-specific marker genes, using reads from RNA-seq and WES that did not map to the human genome/transcriptome. Generated data were correlated to patients’ clinicopathologic parameters as well as immune profiling using previously validated multiplex IHC panels based on Vectra 3.0™ multispectral microscopy IHC panels and image analysis (InForm™ 2.2.1 software). Our analyses revealed that 18.29% (n=15/82) of tumors contained bacterial signatures. The most frequent bacterial signature was related to Escherichia (n=9/15). Moreover, 6.49% (n= 5/77) of tumors had evidence of human viral signatures, including the Epstein-Barr virus (n=1/5). No tumors contained fungal signatures. Preliminary clinicopathologic analyses suggested that patients whose tumors harbor bacterial signatures had a trend towards decreased overall survival (p=0.12). Tumors from former smokers were also more likely to contain bacterial signatures (p=0.11). Preliminary multiplex immune cell IHC analyses did not highlight statistically significant associations with the presence of intra-tumoral bacteria. Our results suggest that a significant proportion of localized NSCLC tumors may harbor components of the human microbiome. Further studies using larger cohorts and dedicated intra-tumoral microbiome and virome methodologies will be needed to better define these findings and to delineate associations with the local immune infiltrate.

Patient-reported outcomes (PROs) provide information on patient treatment experience. Our aim in this analysis was to assess the longitudinal relationship between body mass index (BMI) with patient-reported symptom severity and interference during treatment. Between May 1, 2017 and December 7, 2018, patients with mNSCLC at a single institution were enrolled in a real-world Advanced Non-Small Cell Lung Holistic Registry (ANCHoR) and completed the MDASI-LC prior to start of therapy and at routine clinic visits. MDASI-LC consists of 16 symptom severity and 6 symptom interference items rated on 0-10 scales (0 = no symptom or interference, 10 = worst imaginable symptom or complete interference). BMI was measured at the same schedule as MDASI-LC. Mixed-effects models were used to examine the longitudinal association between BMI and symptom levels during treatment. 103 patients completed the MDASI-LC prior to start of therapy and at least 2 follow-up assessments. Mean patient age was 64.3 years (standard deviation = 11.5) and 50% were males. 22% of patients received chemotherapy (CTX), 34% immunotherapy (IM), 23% CTX+IM or angiogenesis inhibitor, and 20% targeted therapy. The median pre-treatment BMI was 25.2 (inter quartile range, 5.2). BMI did not change during treatment and no significant difference was found among treatment groups. Compared with the obese group (BMI≥30), the overweight group (25≤BMI<30) experienced lowest levels of fatigue (estimation(est)=-1.23, standard error (SE)=0.49, p=0.016), disturbed sleep (est=-1.66, SE=0.49, p=0.002), distress (est=-0.90, SE=0.40, p=0.030) and less interference on mood (est=-1.03, SE=0.46, p=0.030) and interference with walking (est=-1.50, SE=0.51, p=0.005). The normal group (BMI<25) demonstrated lower levels of fatigue (est=-1.05, standard error (SE)=0.47, p=0.032) and disturbed sleep (est=-1.15, SE=0.47, p=0.018), compared with the obese group. For patients with mNSCLC, obesity was related with higher symptom burden during active treatment. This analysis provides pilot data for future studies on balanced weight control and patients' wellbeing during cancer treatment.

Osimertinib, a third-generation EGFR inhibitor, has become the first-line therapy for patients with metastatic EGFR-mutant NSCLCs since 2018. Osimertinib is well-tolerated, therefore, it opens opportunities to be combined with other therapeutic agents to enhance the treatment outcome. In preclinical models, it has been shown that upregulated VEGF signaling mediates acquired resistance to EGFR therapies. In xenograft models, combination of anti-VEGF medications with EGFR inhibitors were significantly more effective than erlotinib or gefitinib alone. Ramucirumab, a monoclonal antibody targeting VEGF receptor 2, is approved with docetaxel in as second line treatment for NSCLCs. In clinical trial evaluations, the phase 3 RELAY trial (NCT02411448) studying ramucirumab plus erlotinib in patients with metastatic untreated EGFR-mutant NSCLC patients showed a statistically significant improvement in progression-free survival in the combination group compared to erlotinib alone. A phase I study of osimertinib with ramucirumab (NCT02789345) demonstrated safety and feasibility of this combination. With strong preclinical and clinical evidence showing dual inhibition of VEGF/EGFR signaling prolongs progression-free survival for EGFR-mutant lung cancers, and demonstrated safety, we are conducting a phase 2 trial to evaluate the osimertinib ramucirumab combination's efficacy in treatment-naïve EGFR-mutant NSCLC. The OSI+RAM trial is a randomized phase 2 study with the primary endpoint being progression-free survival in osi+ram group as compared to osimertinib monotherapy group. The major inclusion criteria include patients with metastatic NSCLC harboring EGFR mutations (L858R/Exon 19 del). The major exclusion criteria include prior anti-EGFR or anti-VEGF treatments. Patients with stable CNS metastasis are allowed. Based on the results from erlotinib bevacizumab (NEJ026) study, we expect an improvement of PFS from 18.9 months to 29.7 months, corresponding to a hazard ratio of 0.65. The trial plans to enroll total of 150 patients, with 100 allocating to osi+ram arm and 50 to osimertinib monotherapy. Total of 9 study sites in the USA are planned. Hoosier Cancer Research Network will facilitate the execution of the trial. The trial protocol has received IND exemption from US FDA and has been approved by IRB at MD Anderson Cancer Center. The first subject is expected to be enrolled in May 2019. A planned interim analysis will be performed after the first 75 subjects are enrolled. NCT03909334. Section not applicable Section not applicable

Small cell lung cancer (SCLC) is an aggressive cancer. Although sensitive to initial therapy, recurrence is almost inevitable. The molecular mechanisms underlying recurrence are unknown. We have previously demonstrated that complex genomic and T cell receptor (TCR) intratumor heterogeneity (ITH) was associated with increased risks of relapse in non-small cell lung cancers (NSCLC). Genomic ITH and TCR architecture of SCLC and its clinical impact have not been well studied, largely due to lack of tumor specimens as surgery is rarely used to treat SCLC. We performed multiregion whole-exome sequencing and TCR sequencing of 49 tumor samples from 18 resected limited-stage SCLCs to delineate the immunogenomic ITH of SCLC. We compared the results to those in NSCLC and assessed the association of genomic and TCR attributes with patient's survival. On average, 544 mutations/sample were detected. The median proportion of trunk mutations (mutations identified in all regions within the same tumors) was 80.4% versus 70% in NSCLC (TRACERx, Jamal-Hanjani, NEJM, 2017, p=0.08) and all TP53 and RB1 mutations were trunk mutations, suggesting these mutations were early events during carcinogenesis of this cohort of SCLCs. A higher non-synonymous tumor mutational burden (TMB) was associated with a higher T cell density (infiltration) in the tumor (r=0.46, p=0.005). Compared to the TCR repertoire of NSCLC (Reuben, WCLC, 2017), these SCLC tumors demonstrated significantly lower T-cell density (0.05 versus 0.24, p<0.0001), richness (diversity, 1,043 versus 3,666, p<0.0001) and clonality (reactivity, average 0.02 versus 0.15, p<0.0001) despite similar non-synonymous TMB (average 187 in SCLC versus 176 mutations/sample in NSCLC). Only 0.2% to 14.6% of T cells were detectable across all regions from the same tumors, suggesting substantial TCR ITH. Jaccard index (JI), a parameter quantifying TCR ITH was significantly lower in SCLC than in NSCLC (0.06 versus 0.1, p<0.0001) implying higher level of TCR ITH in SCLC than NSCLC. Interestingly, higher T-cell density, richness or clonality appeared to be associated with lower risk of recurrence numerically. Furthermore, higher TCR JI (less degree of ITH) was associated with significantly longer overall survival (HR=0.15, p=0.04). Limited-stage SCLC tumors have distinct TCR repertoire and genomic ITH architecture. Overall, SCLC may have a more pronounced immunosuppressive microenvironment and higher level of TCR repertoire ITH than NSCLC. Nevertheless, higher degree of T cell infiltration and clonal expansion as well as more homogeneous T cell response may be associated with more favorable clinical outcome in patients with limited-stage SCLC.

Patient-Reported Outcomes (PROs) provide information on patient treatment experience. We have established a real-world Advanced Non-Small Cell Lung Holistic Registry (ANCHoR) to understand how the advent of immunotherapy impacts treatment choice, clinical outcomes, and PROs of metastatic non-small cell lung cancer (mNSCLC). The aim of this analysis is to report early results of baseline symptom status and quality of life among mNSCLC patients using the MD Anderson Symptom Inventory lung cancer module (MDASI-LC) and EuroQol-5D 5-level version (EQ-5D-5L). During 2017, patients with mNSCLC at a single institution were enrolled in ANCHoR and completed the PRO questionnaires at clinic visits. MDASI-LC consists of thirteen core and three lung cancer-specific symptom severity questions, and six interference items rated on 0-10 scales (0 = no symptom or interference, 10 = worst imaginable symptom or complete interference). EQ-5D-5L captures five health state dimensions: mobility, self-care, usual activities, pain/discomfort, and anxiety/depression rated on a five-level scale (1= no problems, 5= extreme problems). A single visual analogue scale (VAS) on EQ-5D-5L records patient self-rated health between "best imaginable" (100) and "worst imaginable" (0) health state. Descriptive statistics for PRO scores at baseline are summarized. Forty-two patients completed baseline PROs before the start of therapy. Mean patient age was 63 years and 45% were males. For MDASI-LC, the mean scores for the core symptom, lung cancer-specific symptom, and interference subscales at baseline were 2.2 (standard deviation [SD] = 2.80), 2.1 (SD = 2.80), and 2.8 (SD = 3.10), respectively. Fatigue was the most severe symptom reported at baseline (mean = 4.1, SD = 3.01), followed by shortness of breath (mean = 3.2, SD = 2.81) and pain (mean = 3.19, SD = 3.00). The highest percentages of patients reporting moderate to severe symptom levels (score of ≥5) were 38% for fatigue, 33% for pain, 31% for drowsiness, 29% for shortness of breath and disturbed sleep, and 26% coughing. For EQ-5D-5L, 91% of patient reported problems with self-care, 81% with mobility, 48% with usual activity and anxiety, and 33% with pain. Mean EQ-5D VAS was 73.9 (SD = 18.2). Prior to the start of treatment, fatigue, pain, drowsiness, disturbed sleep, and coughing were the most common symptoms with fatigue, shortness of breath, and pain being the most severe. Additional follow up will confirm and expand these findings and will also allow us to examine change in PROs after first-line treatment is administered.

An understanding of the patient experience is lacking for newly developed cancer treatments, such as targeted therapies. We profiled the patient-reported outcome (PRO)-measured symptom burden experienced by patients with metastatic non-small cell lung cancer (mNSCLC) during 6 months of conventional chemotherapy or targeted therapy. During 2017, patients with mNSCLC at a single institution were recruited and completed the MD Anderson Symptom Inventory lung cancer module (MDASI-LC) at clinic visits. The MDASI-LC assesses the severity of 13 core and 3 lung-cancer-specific symptoms and 6 interference items on 0‒10 scales (0=no symptom or interference, 10=worst imaginable symptom or complete interference). Descriptive statistics for MDASI-LC scores over 6 months of treatment were summarized. Symptom trajectories for the chemotherapy patients versus the targeted-therapy patients were compared via linear mixed-effects models. Of 65 patients receiving chemotherapy and 27 receiving targeted therapy, the targeted-therapy group had more women (74% vs. 49%, P=0.029) and younger patients (57.6±12.2 vs. 64.2±9.9 years, P=0.012). Before treatment, both groups reported similar symptom burden, although sadness was worse in the targeted-therapy group (2.4±1.6 vs. 0.8±1.5, P=0.021). During the first 60 days of treatment, patients receiving chemotherapy reported significant increase in pain (estimate (est)=0.03, P=0.037) and interference with walking (est=0.04, P=0.025). Compared with those receiving chemotherapy, patients receiving targeted therapy experienced significantly less severe pain (est=‒1.17, P=0.024), fatigue (est=‒1.16, P=0.019), and shortness of breath (est=‒1.23, P=0.028) and less interference with walking (est=‒1.23, P=0.042) (figure 1). More severe dry mouth was reported by patients undergoing targeted therapy (est=1.17, P=0.027). This real-world data demonstrates that, compared with conventional chemotherapy, targeted therapy correlates with less impairment of physiological condition and functioning in patients with mNSCLC. Additional follow up will confirm and expand these findings about the patient experience relative to treatment response.

Previous studies indicate that neoadjuvant chemotherapy improves survival in patients with loco-regionally advanced non-small cell lung cancer (NSCLC). The amount of residual viable tumor has been associated with long-term overall survival. This histopathologic measure has potential to become a standard method for evaluation of the effectiveness of neoadjuvant therapy regimens. However, adequate comparison of chemotherapy-treated and untreated lung cancers is lacking. We analyzed histopathologic characteristics of resected NSCLC with and without prior neoadjuvant chemotherapy. Histopathologic assessment was performed of specimens obtained from patients enrolled on the immunogenomic lung cancer study (ICON), which integrates clinical, pathologic, immune, genomic and outcome data from surgically resected NSCLC. Cases included material from 10 patients who underwent neoadjuvant chemotherapy and 10 patients treated with primary surgery (adenocarcinoma, n=5; squamous cell carcinoma, n=5; for each cohort). Hematoxylin and eosin-stained tumor sections (mean, 6; range, 3-10) were evaluated and semiquantitatively scored for parameters commonly attributed to treatment response. The percentage of viable tumor was estimated by comparison to the proportion of fibrosis and necrosis on each slide. Additional parameters analyzed included the presence of inflammation, tertiary lymphoid structures (TLS), macrophages, lymphovascular invasion (LVI), cholesterol clefts, giant cells and neovascularization (score 0-3). For each patient, the results for all slides were averaged to determine a mean value. P values were calculated using the Mann-Whitney test. All histopathologic parameters typically associated with treatment response could also be identified in untreated specimens, albeit in different proportions. Compared to the untreated cohort, samples after chemotherapy were characterized by lower proportion of viable tumor (42.4% vs 67.7%, p=0.04) and higher degrees of fibrosis (46.6% vs 26.6%, p=0.08), and necrosis (11.0 % vs 5.6%, p=0.35). Among the additional parameters, similar scores were seen for inflammation (1.54 vs 1.46, p=0.60), TLS (1.00 vs 0.80, p=0.47), LVI (0.16 vs 0.23, p=0.62), and neovascularization (both 0) while macrophages (0.94 vs 0.12, p=0.20), cholesterol clefts (0.92 vs 0.13, p= 0.03) and giant cells (0.80 vs 0.40, p=0.17) were more common among the neoadjuvant cohort. Histopathologic variables commonly associated with chemotherapy treatment response can also be identified in treatment naïve lung cancers. However, the amount of viable tumor, fibrosis and cholesterol clefts are parameters strongly associated with neoadjuvant therapy. These results highlight the importance of assessing the type and extent of treatment response. Analysis of larger patient cohorts will reveal potential prognostic value in primary tumors, chemotherapy-treated, and eventually immunotherapy-treated tumors.

ctDNA is a blood-based biomarker with promising potential in lung cancer for minimal residual disease (MRD) assessment and early detection of recurrence. However, data regarding feasibility are limited, especially for stage I-II disease. We performed longitudinal plasma ctDNA profiling of early-stage lung cancer patients (pts) that underwent resection at MD Anderson Cancer Center from Apr 2016 to Jan 2017. Plasma ctDNA was analyzed from pre-operative and multiple post-operative time points until disease recurrence. ctDNA profiling was performed using a 30kb Digital Sequencing panel (Guardant Health) covering SNVs in 21 genes and indels in 9 genes that are commonly present in lung cancer. ctDNA profiles from ∼30,000 lung cancer pts were used to train a classifier to exclude non-tumor related mutations. A total of 40 pts were included in this analysis, comprised of the first 17 pts with recurrence in the longitudinal study and 23 consecutive pts without recurrence. This cohort was primarily stage I and II (15 [38%], 16 [40%]). Histology included adenocarcinoma (29 [73%]), SCC (6 [15%]), and SCLC (2 [5%]). 58% had adjuvant therapy. Median follow-up was 17.7 (3.4 – 24.5) months and median time to recurrence was 7.1 (3.4 – 16.5) months in this selected cohort. At least one ctDNA alteration was detected in 55% (21/38) of pts with evaluable pre-op samples and in 22% (8/37) of pts at 4 weeks post-op. Presence of ctDNA at 4 weeks post-op heralded eventual recurrence with 43% sensitivity and 91% specificity (75% PPV, 73% NPV) and was significantly associated with worse recurrence free survival (p=0.022, HR 6.52; 95% CI 1.3 – 32.6), while also accounting for stage. In the absence of the variant classifier, an additional 7/37 pts had non-tumor alterations detected at 4 weeks post-op with a recurrence sensitivity and specificity of 57.1% and 69.6%. ctDNA was identified in 76% (13/17) of pts prior to or at the time of recurrence. The median interval between ctDNA detection and radiographic recurrence was 91 days. Detection of post-op ctDNA, as early as 4 weeks after resection of early-stage lung cancer, is associated with significantly increased risk of recurrence. Accurate detection of ctDNA in this MRD setting is enabled by a highly sensitive sequencing platform that incorporates a novel variant classifier to enhance clinical specificity.

PD-L1 tumor expression is a leading biomarker in metastatic non-small lung cancer (NSCLC). Its role and expression in surgically resectable lung cancers is not yet defined. The association between PD-L1 expression on tumor and CD68+ tumor-associated macrophages (TAMs) and the inflammatory cells within the tumor microenvironment continues to be studied. We analyzed 97 surgically resected lung cancers utilizing immunofluorescence profiling and flow cytometry (n=47) with the aim of defining PD-L1 expression and its association with tumor inflammatory cells. Multiplex immunofluorescence profiling of lung cancers was performed with the focus on malignant cells (MC), MC PD-L1%, CD3+, CD8+, PD-1+ cells, CD68+, CD68+ PD-L1%, and CD20+ cells. Data on cell populations were expressed as the number of cells per mm2, PD-L1 expression as percentage. Flow cytometry was performed on freshly disaggregated tumor samples. The associations of cell populations with clinical and pathologic characteristics were assessed using Spearman's rank correlation coefficient and Wilcoxon rank-sum test. 97 patients, 55 (57%) female and 42 (43%) male, with median tumor size 4.0 cm underwent surgical resection for pathologic stage I (N=39), stage II (N=34), and stage III (N=24) NSCLC. 85 (88%) were former smokers, 12 (12%) never smokers. 62 (65%) had adenocarcinoma, 25 (25%) squamous cell carcinoma, 10 (10%) other histology. Neoadjuvant chemotherapy was administered in 16 (16%) patients. R0 resection was achieved in 89 (92%) patients. At the median follow-up duration of 16 months, 18 patients experienced recurrence. CD68+ cells were less abundant than MC within tumor environment (median 120 cell/mm2 vs 4699, p<0.0001). However, PD-L1% expression was significantly higher on CD68+ vs MC within the tumor (median 33% vs 0.02%, p<0.0001); this was true for all stages. CD68+ PD-L1% in SCC was higher compared to adenocarcinoma (median 55% vs 30%, p=0.26). Induction chemotherapy increased CD68+ PD-L1% (median 31% no chemo vs 58%, p=0.05) without affecting the proportion of effector CD8+ TIL expressing its receptor, PD-1 (p=0.757). Tumors with > median CD68+ PD-L1% expression were associated with higher CD3+ (p=0.006), CD8+ (p=0.06), and CD68+ (p=0.004) cell numbers within the tumor. In early NSCLC PD-L1% expression appears to be predominant in CD68+ TAMs rather than in malignant cells. Higher than median PD-L1% expression on CD68+ is associated with increased in CD3+ and CD8+ T cells. Further studies are required to understand the role of CD68+PD-L1 cells within tumor microenvironment, the influence of neoadjuvant chemotherapy or immunotherapy regimens on these cells, and their effect on outcomes.

Immune checkpoint blockade (IO) has demonstrated durable clinical benefit in metastatic non-small cell lung cancer (NSCLC). Tumors with driver mutations such as EGFR exon 19 and 21 mutations and ALK translocation tend to have low response rates to IO. However, IO response in NSCLC patients with rare driver mutations, such as EGFR exon 20 (∼2%), HER-2 (∼2%) and BRAF (∼3%), representing approximately 7% of lung adenocarcinomas, has been poorly addressed. We queried GEMINI (MD Anderson Lung Cancer Moon Shot funded database for prospective collection of clinical information on NSCLC) for patients with mutations in EGFR exon 19, 20, 21, HER-2 and BRAF treated with PD-1/PD-L1 checkpoint inhibitors. We assessed progression-free survival (PFS), overall response rate (ORR) and overall survival (OS) in each molecular group. Between 2014-2018, 108 patients with classic EGFR mutations (exon 19 del + exon 21 L858R, n=37), EGFR exon 20 mutations (n=36; no T790M included), HER-2 mutations (n=22) and BRAF mutations (n=13; V600E: 3pts; non-V600E: 10pts) had been treated with PD-1/PD-L1 inhibitors. EGFR exon 20 mutants and BRAF mutants demonstrated significantly higher PFS (EGFR exon 20: HR 0.4, p<0.001; BRAF: HR 0.2, p<0.001), higher disease control rate at 6 and 12 months as well as higher ORR when compared to classic EGFR mutants (Table). These differences remained significant in multivariate analysis after adjusting for age, smoking, PD-L1 status, radiation prior to treatment initiation, treatment with concurrent agents and prior treatment with TKIs. HER-2 mutants had similar PFS compared to EGFR classic mutants (HR 0.8, p=0.35) (Table). EGFR exon 20 and BRAF mutations are associated with superior outcome from PD-1/PD-L1 checkpoint inhibitors compared to classic EGFR and HER-2 mutations. Further studies on co-mutational status and tumor mutation burden in these molecularly-defined groups are ongoing to address potential underlying mechanisms associated with these findings.

Our previous work has demonstrated that higher level of genomic complexity is associated with more heterogeneous neoantigen repertoire, suppressed T cell repertoire and postsurgical relapse in localized non-small cell lung cancers (NSCLC) highlighting the complex interaction of tumor molecular and immune landscape and their impact on cancer biology and patient survival. We launched the ICON Project (Immune Genomic Profiling of NSCLC) to prospectively delineate the molecular and immune landscape of early stage NSCLC and their impact on patient survival through a multidisciplinary approach. Here we report the updated genomic and immune analyses. Surgical specimens from stage I-III NSCLC were subjected to whole-exome and RNA sequencing for mutational analysis, in silico neoantigen prediction and gene expression analysis as well as T cell receptor sequencing, cytometry by time-of-flight and multiplex immunofluorescence staining. From 2016-2018, 127 patients were accrued and 50 surgical samples have undergone WES, RNAseq, TCR sequencing and immune phenotyping. Median age is 66 yrs (range: 39-86), 52% (26/50) were female and 76% (38/50) former smokers. 76% (38/50) are non-squamous carcinomas and 24% (12/50) squamous cell carcinomas. 34% have stage I disease (17/50), 30% stage II (15/50), 34% stage III (17/50) and 2% stage IV (1/50). The majority of patients had upfront surgery (45/50; 90%). With median follow-up of 19 months, 15 patients have relapsed. Median tumor mutational burden is 7.8mut/Mb and predicted neoantigen burden was 10/sample (range: 0-250). Predicted neoantigen burden is significantly correlated with tumor mutational burden (r=0.41, p=0.002). The most commonly mutated genes are TP53, KRAS, CDKN2A, PIK3CA, EGFR, BRAF, GRIN2A and ATM. C->A transversions and C->T transitions were the most common mutational subtypes. PD-1 expression and regulatory T-cell (CD4+/FoxP3+) infiltration are significantly increased in tumor tissue compared to normal tissue (p=0.003 and p=0.02 respectively), while CD3, CD8, granzyme B and CD45RO are decreased in tumor tissue compared to normal lung. NSCLC tumors have an immunosuppressive microenvironment compared to tumor adjacent normal lung tissues. Clinical data will be adequate to conduct genomic and immune profiling comparisons across different clinical subgroups. Mutational and neoantigen profiling are consistent with previously reported studies and correlations between molecular and immune landscapes and its impact on patient survival are ongoing.

The Advanced Non-Small Lung Holistic Registry (ANCHoR) is established to examine the real-world impact of immunotherapy on choice of treatment, clinical outcomes, and patient reported outcomes of patients with Stage IV NSCLC. Stage IV NSCLC patients diagnosed or initiating treatment at MD Anderson from January 1, 2017 are enrolled in the ongoing ANCHoR study. Their demographic, clinicopathological, molecular, and treatment data were populated in a prospective database. Treatment patterns by line and PD-L1 status were summarized in this interim analysis. At the time of data cut off (Dec 31, 2017) 182 patients were enrolled in the registry, of which 150 were tested for PD-L1. Number of patients initiating first-, second-, and third-line treatment were 163, 42 and 7, respectively. Of the 30 patients not tested for PD-L1, 10 did not have enough tissue and 8 had actionable mutations. The emergence of immunotherapy has had a dramatic impact on the first-line treatment of patient with advanced NSCLC. As of December, 2017 up to 41% of patient received immunotherapy either singly (23%) or in combination with chemotherapy. Only 40% of the patients now receive chemotherapy alone. There has been dramatic decrease in the use of chemotherapy with an anti-angiogenesis agent (1.23%). In our dataset 16% of the patients were eligible for targeted therapy as initial treatment.

Non-small cell lung cancer (NSCLC) is characterized by a high mutational load. Accordingly, it is also among the tumor types responding to immune checkpoint blockade, likely through harnessing of the anti-tumor T cell response. However, the lung is continuously exposed to the outside environment, which may result in a continuous state of inflammation against outside pathogens unrelated to the tumor microenvironment. Therefore, further investigation into the T cell repertoire and T cell phenotypes across normal lung and tumor is warranted. We performed T cell receptor (TCR) sequencing on peripheral blood mononuclear cells (PBMC), normal lung, and tumor from 225 NSCLC patients, among which, 96 patients were also subjected to whole exome sequencing (WES) of PBMC, tumor and normal lung tissues. We further performed Cytometry by Time-of-Flight (CyTOF) on 10 NSCLC tumors and paired normal lung tissues to phenotype immune and T cell subsets. Comparison of the T cell repertoire showed 9% (from 4% to 15%) of T cell clones were shared between normal lung and paired tumor. Furthermore, among the top 100 clones identified in the tumor, on average 57 (from 0 to 95) were shared with paired normal lung tissue. Interestingly, T cell clonality was higher in the normal lung in 89% of patients suggesting potential differences in the immune response and immunogenicity. A substantial number of somatic mutations were also identified not only in NSCLC tumors (average 566; from 147 to 2819), but also in morphologically normal lung tissues (average 156; from 50 to 2481). CyTOF demonstrated striking differences in the immune infiltrate between normal lung and tumor, namely a lower frequency of PD-1+CD28+ T cells (both CD4+ and CD8+) in the normal lung (2.7% versus 3.0% in tumor). In addition, a unique GITR+ T cell subset (0.96%) was entirely restricted to the normal lung. Conversely, increases in regulatory T cell frequency (CD4+FoxP3+) were observed in the tumor (10.4% vs 1.7% in normal lung), further highlighting the differences in T cell phenotype and response across normal lung and tumor. These results suggest that a substantial proportion of infiltrating T cells in NSCLC tumors may be residential T cells associated with response to environmental factors. However, normal lung and NSCLC tumors carry T cells of distinct phenotypes including increases in immunosuppressive T cells within the tumor which may further highlight the differences in the anti-tumor immune response.

While previous reports have established MET and HER2 amplification as two mechanisms of non-T790M driven EGFR TKI resistance in EGFR mutant NSCLC, resistance occurs in the absence of these modifications in a significant number of patients. Therefore, there exists an unmet need to define additional mechanisms of resistance to EGFR TKIs. We hypothesized that targeted next-generation sequencing could detect additional targetable activating mutations in paired tumor samples from patients with acquired resistance to first or second generation EGFR TKIs. We conducted an analysis of clinical and molecular data prospectively collected from 285 EGFR-mutant NSCLC patients enrolled into the MD Anderson Lung Cancer GEMINI database. Of 157 patients treated with first-line therapy (erlotinib, gefitinib, or afatinib), we identified 75 patients with TKI-acquired resistance with matched pre/post-TKI tumor samples. Matched tumor samples were analyzed with targeted gene sequencing. Recurrent alterations were defined as an alteration occurring more than 2 times. Recurrent acquired mutations were expressed in Ba/F3 and EGFR mutant (T790M+/-) NSCLC cells. Mutation expressing Ba/F3 cell lines were assayed for IL-3 independence, and mutation expressing NSCLC cell were screened against combination targeted TKIs. EGFR mutant NSCLC patients treated with first-line therapy had a median PFS of 14 months; and, of the patients with pre/post-TKI tumor molecular data, 47% of patients were T790M negative. There were 30 recurrent acquired alterations identified in 13 different genes. Genes included ARAF, BRAF, EGFR, FGFR, GNAS, JAK2, MCL1, PDGFRα, PIK3CA, RAF1, RB1, SMAD4, and TP53. Of the alterations identified, most occurred in 1 of 4 targetable genes: BRAF (N=3), FGFR (N=5), PDGFRα (N=3), or PIK3CA (N=2). Both previously reported and novel mutations were identified, and preliminary screening of mutant expressing Ba/F3 cell lines found that of the mutations tested (BRAF WT & G469H, FGFR2 A371G, PDGFRα WT & L682F, and PIK3Ca E545K) all grew independent of IL-3. HCC827 and H1975 cell lines expressing acquired mutations in BRAF, FGFR, PDGFRα, or PIK3CA were more sensitive to combination targeted therapy compared to EGFR TKIs or mutation specific TKIs alone unlike control cell lines, supporting the possibility that targeting these mutations would be of therapeutic benefit. Analysis of patient data identified 30 recurrent genomic alterations in 13 different genes including novel alterations in BRAF, EGFR, FGFR, PDGFRα, RB1, and SMAD4, many of which were found to be activating mutations. Our analysis identified potentially targetable mutations of BRAF, FGFR, PDGFRα, and PIK3CA which merits further pre-clinical and clinical investigation.

Anti-PD-1 and anti-PD-L1 antibodies including pembrolizumab, nivolumab and atezolizumab have entered clinical practice in the management of metastatic NSCLC as monotherapy and immunotherapy-based combinations. We have established a real-world Advanced Non-Small Cell Lung Cancer Holistic Registry (ANCHoR) to understand how the emergence of immunotherapy impacts choice of treatment, clinical outcomes, and patient reported outcomes (PROs) in the different histo-molecular subtypes of metastatic NSCLC.

LKB1 is a protein kinase that is mutated and down-regulated in 20-30% of non-small cell lung cancer (NSCLC). LKB1 mutations co-occur with KRAS alterations in 7%-10% of NSCLC, resulting in an aggressive phenotype with short survival. Because LKB1 activates AMPK, the master sensor of cellular energy, many of the best known functions of LKB1 are attributed to its ability to control metabolic alterations in the cells. However LKB1 also plays an important role in regulating angiogenesis, likely as a strategy to overcome energetic depletion of tumor microenvironment. Bevacizumab, the human anti-VEGF antibody, improves the PFS and OS of NSCLC patients combined with chemotherapy, but often the benefit is transient and therapeutic resistance occurs. Our laboratory has previously identified alterations in cell metabolism and in vasculature of LKB1-deficient tumors when compared to LKB1 wild type in NSCLC. LKB1 KO murine NSCLC cell lines were generated using CRISPR/Cas9 system in a KRASG12D mutant background (LKR10 & LKR13). Syngeneic NSCLC models were established via s.c. injection of LKB1 intact and KO murine cells in immunocompetent mice. After tumors reached 150 mm3 mice were randomly assigned to treatment groups consisting of vehicle, mouse anti-VEGF and nintedanib. Tumor volumes were measured and compared using student's t test and samples were collected for vasculature analysis. Survival curves will be calculated using log rank test. Hypoxia experiments were preformed and apoptosis was measured using annexin V and 7ADD staining. Treatment with anti-VEGF or nintedanib significantly inhibited tumor progression in LKB1 wt KRASG12D mutant mouse model (p<0.001) but did not show any therapeutic effect in the LKB1 KO KRASG12D group. Furthermore in the LKB1 wt group, the median survival of anti-VEGF and nintedanib treated mice was 111 days and 84 days respectively and 37 days in the vehicle group. No improvement in survival was detected in the LKB1 KO group after treatment with anti-VEGF. In vitro studies showed that LKB1 loss is associated with a decrease in oxygen consumption and enhanced glycolysis. Furthermore LKB1 KO NSCLC cells showed a decrease in apoptosis under hypoxic and low nutrient conditions compared to LKR13 LKB1 wt cells. NSCLC LKB1-deficient tumors showed resistance to anti-angiogenic therapy and this effect is driven by the regulation of metabolic adaptations that allow cells to survive under hypoxic and low nutrient conditions.