Search

The aim of this retrospective study was to demonstrate that irAEs, specifically gastrointestinal and pulmonary, examined through International Classification of Disease (ICD) data leads to underrepresentation of true irAEs and overrepresentation of false irAEs, thereby concluding that ICD claims data are a poor approach to electronic health record (EHR) data mining for irAEs in immunotherapy clinical research. This retrospective analysis was conducted in 1,063 cancer patients who received ICIs between 2011 and 2017. We identified irAEs by manual review of medical records to determine the incidence of each of our endpoints, namely colitis, hepatitis, pneumonitis, other irAE, or no irAE. We then performed a secondary analysis utilizing ICD claims data alone using a broad range of symptom and disease-specific ICD codes representative of irAEs. 16% (n = 174/1,063) of the total study population was initially found to have either pneumonitis 3% (n = 37), colitis 7% (n = 81) or hepatitis 5% (n = 56) on manual review. Of these patients, 46% (n = 80/174) did not have ICD code evidence in the EHR reflecting their irAE. Of the total patients not found to have any irAEs during manual review, 61% (n = 459/748) of patients had ICD codes suggestive of possible irAE, yet were not identified as having an irAE during manual review. Examining gastrointestinal and pulmonary irAEs through the International Classification of Disease (ICD) data leads to underrepresentation of true irAEs and overrepresentation of false irAEs.

BACKGROUND: Checkpoint inhibitor pneumonitis (CIP) is an immune related adverse event that may complicate treatment with immune checkpoint inhibitors (ICI) and can cause significant morbidity. We sought to identify predictors for the development of CIP, and whether the use of inhaled corticosteroids (ICS) at time of ICI may be protective. METHODS: Patients with advanced cancer treated with ICI from 2011 and 2018 were included in this study. CIP attribution to ICI was determined by treating physician at time of diagnosis. Predictors were assessed by univariate and multivariable Cox proportional hazard models. RESULTS: We identified 837 pts treated with ICI, of whom 30 (3.6%) developed grade 2 or higher CIP. 82 patients (9.8%) were receiving ICS at time of ICI and had increased risk of developing CIP with hazard ration (HR) of 4.22 (95% CI: 1.93-9.21, p

Background The microbiome has been shown to affect the response to Immune Checkpoint Inhibitors (ICIs) in a small number of cancers and in preclinical models. Here, we sought to broadly survey cancers to identify those in which the microbiome may play a prognostic role using retrospective analyses of patients with advanced cancer treated with ICIs. Methods We conducted a retrospective analysis of 690 patients who received ICI therapy for advanced cancer. We used a literature review to define a causal model for the relationship between medications, the microbiome, and ICI response to guide the abstraction of electronic health records. Medications with precedent for changes to the microbiome included antibiotics, corticosteroids, proton pump inhibitors, histamine receptor blockers, non-steroid anti-inflammatories and statins. We tested the effect of medication timing on overall survival (OS) and evaluated the robustness of medication effects in each cancer. Finally, we compared the size of the effect observed for different classes of antibiotics to taxa that have been correlated to ICI response using a literature review of culture-based antibiotic susceptibilities. Results Of the medications assessed, only antibiotics and corticosteroids significantly associated with shorter OS. The hazard ratios (HRs) for antibiotics and corticosteroids were highest near the start of ICI treatment but remained significant when given prior to ICI. Antibiotics and corticosteroids remained significantly associated with OS even when controlling for multiple factors such as Eastern Cooperative Oncology Group performance status, Charlson Comorbidity Index score, and stage. When grouping antibiotics by class, β-lactams showed the strongest association with OS across all tested cancers. Conclusions The timing and strength of the correlations with antibiotics and corticosteroids after controlling for confounding factors are consistent with the microbiome involvement with the response to ICIs across several cancers.

The original version of this article unfortunately contained a mistake. The second sentence of the section "irAEs and ICI efficacy" should read as.

MICROABSTRACT: Serum albumin in cancer patients is associated with clinical outcomes. In a single-center retrospective study of 210 advanced NSCLC patients receiving immune checkpoint inhibitor (ICI) with or without chemotherapy as first-line therapy, we found a significant association of overall survival with pretreatment albumin and early albumin decrease in treatment with ICI monotherapy but not with chemoimmunotherapy. BACKGROUND: Cancer cachexia exhibits decreased albumin and associates with short overall survival (OS) in patients with non-small cell lung cancer (NSCLC), but whether on-treatment albumin changes associate with OS in NSCLC patients treated with immune checkpoint inhibitors (ICIs) and combination chemoimmunotherapy has not been thoroughly evaluated. PATIENTS AND METHODS: We conducted a single-center retrospective study of patients with advanced NSCLC who received first-line ICI with or without chemotherapy between 2013 and 2020. The association of pretreatment albumin and early albumin changes with OS was evaluated using Kaplan-Meier method and Cox regression models. RESULTS: A total of 210 patients were included: 109 in ICI cohort and 101 in ICI+Chemo cohort. Within a median of 21 days from treatment initiation, patients with ≥10% of albumin decrease had significantly shorter OS compared to patients without albumin decrease in ICI cohort. Pretreatment albumin and albumin decrease within the first or second cycle of treatment were significantly and independently associated with OS in ICI cohort, but not in ICI+Chemo cohort. The lack of association between albumin and OS with the addition of chemotherapy was more pronounced among patients with ≥ 1% PD-L1 expression in subgroup analysis. CONCLUSION: Pretreatment serum albumin and early albumin decrease in ICI monotherapy was significantly associated with OS in advanced NSCLC. Early albumin change, as a routine lab value tested in clinic, may be combined with established biomarkers to improve outcome predictions of ICI monotherapy. The underlying mechanism of the observed association between decreased albumin and ICI resistance warrants further investigation.

Background The neutrophil to lymphocyte ratio (NLR) is known to be prognostic for patients with advanced cancers treated with immune checkpoint inhibitors (ICI), but has generally been evaluated as a single threshold value at baseline. We evaluated NLR at baseline and within first month during treatment in patients who received ICI for advanced cancer to evaluate the prognostic value of baseline and of changes from baseline to on-treatment NLR. Methods A retrospective review of patients with advanced cancer treated with ICI from 2011 to 2017 at the Ohio State University was performed. NLR was calculated at the initiation of ICI and repeated at median of 21 days. Overall survival (OS) was calculated from the initiation of ICI to date of death or censored at last follow-up. Significance of Cox proportional hazards models were evaluated by log-rank test. Calculations were performed using the survival and survminer packages in R, and SPSS. Results 509 patients were identified and included in the analysis. Patients with baseline and on-treatment NLR

Background A low absolute lymphocyte/monocyte ratio (LMR) in the peripheral blood is associated with poor prognosis in various cancers; however, its role as a predictive biomarker has not been well defined in the era of treatment with immune checkpoint inhibitors (ICI). Methods We queried a database of advanced cancer patients treated with at least one dose of ICI from 2011 to 2017 to study the association of LMR with overall survival (OS). We calculated LMR at baseline and a median of 21 days after the first cycle of ICI (on-treatment LMR), and defined low if

Introduction Immune checkpoint inhibitors (ICI) are associated with unique immune-related adverse events (irAEs). Immune-related thrombocytopenia (irTCP) is an understudied and poorly understood toxicity; little data are available regarding either risk of irTCP or the effect of irTCP on clinical outcomes of patients treated with ICI. Methods We conducted a retrospective review of sequential cancer patients treated with ICI between 2011 and 2017 at our institution. All patients who received ICI alone or in combination with other systemic therapy in any line of treatment were included; those with thrombocytopenia ≥ grade 3 at baseline were excluded. We calculated the incidence of ≥ grade 3 irTCP and overall survival (OS). Patient factors associated with irTCP were assessed. Results We identified 1,038 patients that met eligibility criteria. Overall, 89 (8.6%) patients developed grade ≥ 3 thrombocytopenia; eighteen were attributed to ICI (1.73% overall). Patients who developed grade ≥ 3 irTCP had worse overall survival compared to those whose thrombocytopenia was unrelated to ICI (4.17 vs. 10.8 month; HR. 1.94, 95% CI 1.13, 3.33; log-rank p = 0.0164). Patients with grade ≥ 3 irTCP also had worse survival compared to those without thrombocytopenia (4.17 vs. 13.31 months; HR 2.22, 95% CI 1.36, 3.62; log-rank p = 0.001). The incidence of irTCP appeared lowest among those treated with PD-1/L1 monotherapy (p = 0.059) and was not associated with cancer type, smoking status, age, gender, race, or line of therapy. Conclusions Unlike other irAEs, we found that irTCP was associated with worse overall survival. The incidence of irTCP appeared lowest among those treated with PD-1/L1 monotherapy.

Background Some cancer patients who are diagnosed with thromboembolism may require dual treatment with vascular endothelial growth factor receptor (VEGFR) tyrosine kinase inhibitors (TKIs) and factor Xa inhibitors (low-molecular-weight heparin [LMWH] or direct oral anticoagulants [DOACs]). However, to the authors' knowledge, the safety of such combinations has not been well characterized. Methods Patients with advanced cancer who were treated with concurrent VEGFR TKIs and factor Xa inhibitors between 2010 and 2018 at The Ohio State University Comprehensive Cancer Center were included. Charts were reviewed retrospectively for clinically significant bleeding events occurring during concurrent treatment compared with those occurring during factor Xa inhibitor therapy alone, using each patient as their own control. The Fisher exact test was used to compare distribution of bleeding severities. The Cox proportional hazards model was used to compare bleeding risk between groups. Results Among 86 patients, there were 29 clinically significant bleeding events (including 8 major bleeding events) reported during concurrent treatment and 17 events (including 4 major bleeding events) reported during factor Xa inhibitor therapy alone over a median follow-up of 63 days. Concurrent treatment was associated with significantly higher risks of overall bleeding (hazard ratio, 2.45; 95% confidence interval, 1.28-4.69 [P = .007]) and first-onset bleeding (hazard ratio, 2.23; 95% confidence interval, 1.13-4.42 [P = .02]). Analysis of 6-month bleeding risk and the subgroups of patients treated with concurrent TKIs and LMWH versus LMWH alone demonstrated a similar trend. The sample size was inadequate for comparisons between treatment with concurrent TKIs and DOACs versus DOACs alone. Conclusions Concurrent treatment with VEGFR TKIs and LMWH was found to be associated with a significantly increased risk of bleeding events when compared with LMWH therapy alone.

The original version of this article unfortunately contained a mistake. The second sentence of the section "irAEs and ICI efficacy" should read as.

Additional file 1: Figure S1. Causal diagram with references. Figure S2. Number of medications prescribed across all cancers and the frequency of multiple medications. Figure S3. Number of antibiotics, separated by class, prescribed across all cancers and the frequency of multiple antibiotics. Figure S4. Number of corticosteroids, separated by class, prescribed across all cancers and the frequency of multiple antibiotics.

Additional file 2: Table S1. References for the antibiotic susceptibilities of bacterial taxa found to be enriched in responders or non-responders to ICI therapy

Immune-related adverse events (irAEs) comprise a distinct spectrum of auto-inflammatory manifestations triggered due to immune checkpoint inhibitors (ICI). Current data on the association of irAEs with outcomes in NSCLC treated with nivolumab are limited.We pooled data from 531 metastatic NSCLC patients from five centers treated with nivolumab after failing platinum-based chemotherapy. The primary objective was to investigate the relationship between irAEs with clinical benefit to nivolumab as well as to elucidate patterns of irAE-related ICI discontinuations and their impact on survival.33.0% (173/531) of patients treated with nivolumab were noted to have an irAE. Patients with irAEs had a significantly longer median PFS [6.1 vs. 3.1 months, HR 0.68 95% CI (0.55-0.85); p = 0.001] and OS [14.9 vs. 7.4 months, HR 0.66 95% CI (0.52-0.82); p 0.001)] compared to those without irAEs. In multivariate analysis, the presence of irAEs showed a significantly better PFS [HR 0.69, 95% CI (0.55-0.87); p = 0.002] and a trend for better OS [HR 0.62, 95% CI (0.55-1.03); p = 0.057]. Patients with permanent ICI discontinuation secondary to index irAE had a significantly shorter median PFS [2.3 vs. 6.6 months, HR 1.74 95% CI (1.06-2.80); p = 0.02] and median OS [3.6 vs. 17.6 months; HR 2.61 95% CI (1.61-4.21); p 0.001] compared to those that did not have permanent ICI discontinuation.Our pooled exploratory analysis demonstrates improved clinical benefit to nivolumab in NSCLC patients experiencing irAEs. We also observed negative impact of irAE-related treatment discontinuation on survival in this group of patients.

e14549 Background: Immune checkpoint inhibitors (ICI) are a standard of care therapy for patients with many different cancers. ICI are generally well tolerated but are associated with unique immune-related adverse events (irAEs). Immune-related thrombocytopenia (irTCP) is an understudied and poorly understood toxicity and risk factors for irTCP are unknown. Although other irAE have been associated with clinical benefit, little data is available regarding either risk of irTCP or the effect of irTCP on clinical outcomes of patients treated with ICI. Methods: We conducted a retrospective review of sequential patients with any solid or hematologic cancer treated with ICI between 2011 and 2017 at The Ohio State University Comprehensive Cancer Center. All patients who received ICI alone or in combination with other systemic therapy in any line of treatment were included; those with thrombocytopenia ≥ grade 3 at baseline were excluded. Thrombocytopenia grading was performed utilizing the Common Terminology Criteria for Adverse Events version 5.0. Attribution to ICI was defined as timing after ICI initiation, lack of alternative causes, and improvement with either holding treatment or use of immune suppressive therapy. Overall survival (OS) was calculated from the date of initiation of ICI to death from any cause or date of the last follow-up examination. Cox proportional hazard models were used to examine the associations between platelet categories with OS. Median OS was estimated using Kaplan-Meier method with 95% CI. Results: We identified 1,038 patients treated with ICI therapy after excluding patients with baseline thrombocytopenia. The median age was 61.4; 613 (59.1%) were male and 613 (59.1%) had history of smoking. The most common cancer types were melanoma (n = 337, 32.5%), lung (n = 213, 20.5%), and renal cell carcinoma (n = 114, 11%). The most common ICI were PD1/L1 (n = 729, 70.2%) and CTLA4 (n = 191, 18.4%). Overall, 89 (8.6%) patients developed grade ≥3 thrombocytopenia; twenty were attributed to ICI (1.93% overall). Patients who developed ≥3 irTCP had worse overall survival compared to those whose thrombocytopenia was unrelated to ICI (5.45 vs 11.3 months; HR. 2.077, 95% CI 1.231, 3.503; log-rank p = 0.005). Patients with ≥3 irTCP also had shorter survival compared to those without thrombocytopenia of any etiology (5.45 vs 13.3 months; HR 2.247, 95% CI: 1.414, 3.571; p < 0.001). irTCP was more common among those treated with PD1/L1 (p = 0.03) but was not associated with cancer type, smoking status, age, gender, race, or line of therapy. Conclusions: Type of immunotherapy is related to irTCP, whereas cancer type and line of therapy are not. Unlike other irAEs, we found that irTCP was associated with shorter overall survival. These findings provide important insight into a poorly understood and rare toxicity. irTCP should be evaluated in more extensive studies to better inform clinical decision making.

e14525 Background: Immune checkpoint inhibitors (ICIs) have improved the survival of patients with multiple cancer types, however ICI treatment is associated with a unique set of immune-related adverse events (irAEs) that can affect any organ. Few studies have evaluated the risk factors and outcomes of ICI induced hepatitis (ICIH). Methods: We utilized an institutional database of patients with advanced cancers treated with ICI between 2011 and 2017 at The OSU Comprehensive Cancer Center to identify patients with ICIH. Any patient who received at least one dose of ICI alone or in combination with other systemic therapies either as part of clinical trial or standard of care were included. Clinical data were extracted through chart abstraction. irAEs were graded using the Common Terminology Criteria for Adverse Events v5. Overall survival (OS) was calculated from the date of ICI initiation to death from any cause or the date of the last follow-up. OS with 95% confidence intervals were estimated using the Kaplan–Meier method. OS was also evaluated by occurrence of ICIH using the log-rank test. Results: We identified 1,096 patients treated with at least one dose of ICI. Most common cancers included lung (n=224, 20%) and melanoma (n=342, 31%). The most common ICIs were PD1/L1 (n=774, 71%) and CTLA-4 inhibitors (n=195, 18%). ICIH of any grade occurred in 64 (6%) patients. Overall, 46 (71%) were male and median age was 60 years. Severity of hepatitis was grade 1-2 in 30 patients (47%) (Table 1). The incidence of ≥grade 3 ICIH in the entire cohort was 3.1%. Median time to ICIH diagnosis was 63 days. ICIH occurred alone in 24 patients, and co-occurred with other irAEs in 40 patients. The most common co-occurring irAEs were pneumonitis (n=7); colitis (n=15), thyroid abnormality (n=14); and dermatitis (n=15). ICIH was more common in women (p=0.038), in patients treated with combination ICIs (p

e18593 Background: Bone metastases and skeletal related events (SREs) are a common cause of morbidity among patients with metastatic non-small cell lung cancer (mNSCLC). SREs include pathologic bone fractures, spinal cord compression, orthopedic surgery/stabilization, and palliative radiation to bone. Bone modifying agents (BMA) such as bisphosphonates and RANK ligand inhibitors are often used for mNSCLC patients with bone metastases to reduce the incidence of SRE. Data on the associations between BMA use and overall survival (OS) and the development of SRE among patients with mNSCLC treated with immune checkpoint inhibitors (ICI) is limited. Methods: We conducted a retrospective study of patients with mNSCLC treated with first line pembrolizumab or pembrolizumab plus chemotherapy at The Ohio State University from 2017-2020. The associations between SRE and categorical variables were studied using Fisher’s exact test and with continuous variables using the Kruskal-Wallis test. The association between BMA and risk of osseous progression was studied using Fisher’s exact test. The association between BMA and OS was examined using a Cox proportional hazard model. Median OS was estimated using the Kaplan-Meier method with 95% CI. OS was calculated from date of ICI initiation to death from any cause or last follow up. Results: We identified a cohort of 228 patients: 107 (47%) treated with pembrolizumab and 121 (53%) treated with pembrolizumab plus chemotherapy; 48 (21%) had squamous histology and 179 (79%) had nonsquamous histology (1 had NSCLC histology not specified); median age 62.6 years; median OS was 26.8 months (95% CI: 17.2 – 34.5 months). A total of 95/228 (42%) patients had bone metastases at time of ICI initiation; 45/95 (42%) had treatment with BMA and 50/95 (58%) were not treated with BMA. Among patients with bone metastases at time of ICI, 45 (42%) developed SRE after ICI initiation and BMA use was not associated with risk of SRE (p = 1). BMA use was not significantly associated with risk of osseous progression (p = 0.21). For patients with baseline bone metastases prior to ICI, the use of BMA was not associated with OS (p = 0.24); the type of BMA used (bisphosphonate vs RANK ligand inhibitor) was also not associated with OS (p = 0.13). Conclusions: In patients with metastatic NSCLC treated with first line pembrolizumab alone or with chemotherapy, bone modifying agents were not associated with OS, decreased risk of osseous progression, or development of SRE in patients with baseline bone metastases prior to ICI initiation. The type of BMA used was also not associated with survival. Further studies evaluating the utility and impact of BMA use in this patient population are needed.

IMPORTANCE: The spectrum of individual immune-related adverse events (irAEs) from anti–programmed cell death 1 (PD-1) and programmed cell death ligand 1 (PD-L1) immune checkpoint inhibitors (ICIs) has been reported widely, and their development is associated with improved patient survival across tumor types. The spectrum and impact on survival for patients with non–small cell lung cancer (NSCLC) who develop multisystem irAEs from ICIs, has not been described. OBJECTIVE: To characterize multisystem irAEs, their association with survival, and risk factors for multisystem irAE development. DESIGN, SETTING, AND PARTICIPANTS: This retrospective cohort study carried out in 5 academic institutions worldwide included 623 patients with stage III/IV NSCLC, treated with anti–PD-(L)1 ICIs alone or in combination with another anticancer agent between January 2007 and January 2019. EXPOSURES: Anti–PD-(L)1 monotherapy or combinations. MAIN OUTCOMES AND MEASURES: Multisystem irAEs were characterized by combinations of individual irAEs or organ system involved, separated by ICI-monotherapy or combinations. Median progression-free (PFS) and overall survival (OS) were estimated using the Kaplan-Meier method. Differences in PFS and OS between irAE groups were assessed by multivariable models. Risk for multisystem irAE was estimated as odds ratios by multivariable logistic regression. RESULTS: The 623 patients included in the study were mostly men (60%, n = 375) and White (77%, n = 480). The median (range) age was 66 (58-73) years, and 148 patients (24%) developed a single irAE, whereas 58 (9.3%) developed multisystem irAEs. The most common multisystem irAE patterns in patients receiving anti–PD-(L)1 monotherapy were pneumonitis thyroiditis (n = 7, 14%), hepatitis thyroiditis (n = 5, 10%), dermatitis pneumonitis (n = 5, 10%), and dermatitis thyroiditis (n = 4, 8%). Favorable Eastern Cooperative Oncology Group (ECOG) performance status (PS) (ECOG PS = 0/1 vs 2; adjusted odds ratio [aOR], 0.27; 95% CI, 0.08-0.94; P = .04) and longer ICI duration (aOR, 1.02; 95% CI, 1.01-1.03; P

e19305 Background: Tyrosine kinase inhibitors (TKIs) of vascular endothelial growth factor receptor (VEGFR) may interfere with the coagulation system making patients susceptible to both hemorrhage and thromboembolism. Some cancer patients require dual treatment with VEGFR TKIs and factor Xa inhibitors. However, the safety of such combination treatment (e.g. bleeding risk) has not been well characterized in the literature. Methods: This retrospective study identified metastatic cancer patients who received VEGFR TKIs (pazopanib, sunitinib, sorafenib, axitinib, regorafenib, vandetanib, lenvatinib, cabozantinib, tivozanib) and Xa inhibitors (low-molecular weight heparin [LMWH] or direct oral anticoagulants [DOACs]- rivaroxaban or apixaban) from 2001 to 2018 at the Ohio State University. We assessed bleeding risks of dual therapy vs factor Xa inhibitors alone, using the same patients as self-controls. We defined bleeding based on criteria from the International Society on Thrombosis and Haemostasis. Fisher’s exact test was used to compare distribution of bleeding severities and the Cox model was used to analyze bleeding risk. Results: A total of 86 patients were included for analysis, with a median age of 56 (range, 34-88). The patient population was predominantly of male gender (69.7%), Caucasian race (89.3%) and renal cell cancer type (48.9%). As some patients had been on both LMWH and DOAC at different time periods, we found that 81 patients received concurrent TKI and LMWH; 19 patients had concurrent TKI and DOAC. Overall, 25 patients developed 29 clinically significant bleeding events (including 8 major bleeding) during concurrent treatment, and 15 patients developed 17 clinically significant bleeding events (4 major bleeding) during factor Xa inhibitor alone over a median follow up of 63 days. We found a statistically higher bleeding risk with concurrent treatment (HR = 2.23, 95% CI, 1.13–4.42, p = 0.02) compared with Xa inhibitor alone. In a stratified analysis, concurrent VEGFR TKI plus LMWH or DOAC both showed a strong trend for elevated bleeding risk. Although there were increased bleeding events with concurrent treatment, there was no significant difference in the proportion of major bleeding (27.6% vs 23.5%, p = 0.76). Median time to first bleeding event was 56 days for concurrent therapy and 72 days for Xa inhibitor alone. Conclusions: Concurrent treatment of VEGFR TKI and factor Xa inhibitors is associated with significantly increased bleeding risk when compared with factor Xa inhibitors alone in metastatic cancer patients.

3047 Background: Immunosuppressive factors within the tumor microenvironment (TME) pose a barrier to response to treatment with immune checkpoint inhibitors (ICI). Monocytes alter the TME to promote cancer progression through local immune suppression and angiogenesis. Peripheral blood lymphocyte-to-monocyte ratio (LMR) may reflect the interaction between host immunity, represented by lymphocytes, and the tumor microenvironment, represented by monocytes. A low LMR in the peripheral blood may serve as a surrogate biomarker and has been associated with poor prognosis in various cancers; however, its role has not been well defined in the era of treatment with ICI. Methods: We retrospectively evaluated 1034 patients with advanced cancer treated with ICI from 2011 to 2017. We calculated LMR as ratio of absolute lymphocyte/monocyte counts at baseline and median of 21 days after first cycle of ICI (on-treatment LMR) and considered low if < 2. Overall survival (OS) was calculated from the initiation of ICI to date of death or censored at last follow-up. Median OS with 95% confidence intervals (CI) was estimated using the Kaplan-Meier method. Log rank test was used for group comparison. Results: 536 pts (52%) with LMR < 2 at baseline had shorted median OS compared to 498 (48%) with LMR≥2 (median OS 8.4 months vs 17.8 months, p < 0.001). Of 1034 pts with baseline LMR, 837 had follow up LMR evaluable. In patients with baseline and on-treatment LMR, those with baseline LMR < 2, who had on treatment LMR ≥2, had OS of 16.8 months (95% CI 10.3-23.5) compared to median OS 8.0 months (95% CI 6-9.4) for patients with on treatment LMR < 2 after first cycle of ICI, p < 0.001. Patients with baseline LMR≥2, who had on treatment LMR ≥ 2, had median OS of 23 months (95% CI 19.7-28.9), but median OS was 9.4 months (95% CI 7.1-11.1) for patients with on-treatment LMR < 2 after first cycle of ICI, p < 0.001. Conclusions: We observed a statistically significant association between not only baseline LMR but also change in LMR from baseline after first cycle of ICI and overall survival in cancer patients treated with ICI. The role of LMR at baseline and on-treatment LMR should be evaluated in further studies incorporating known additional prognostic factors for ICI therapy. [Table: see text]

3140 Background: The identification of risk factors for immune-related adverse events (irAEs) is an important area of research. Among irAEs, pneumonitis carries one of the highest morbidities. There is a lack of strong predictors for pneumonitis in patients (pts) treated with ICI. We sought to identify predictors for the development of pneumonitis, and whether the use of inhaled corticosteroids (ICS) at time of ICI could be protective. Methods: Pts with advanced cancer treated with ICI from 2011 and 2018 were included in this retrospective study. Pneumonitis attribution to ICI was determined by treating physician at time of diagnosis. Time to pneumonitis was defined as days from the start of ICI to pneumonitis diagnosis. Pts who never had pneumonitis were censored at the time of last follow up or death. Predictors of pneumonitis were assessed by univariate Cox proportional hazard models at a significance threshold of alpha = 0.05. Results: A total of 837 pts were identified, and 30 (3.6%) pts developed any grade pneumonitis (12 grade 2, 14 grade 3, 1 grade 4, 3 grade 5) after receiving ICI (Table). Pts with age ≥65 years (y) had increased risk of developing pneumonitis over pts with age < 65y (HR 2.1, 95 CI: 1.02-4.4, p=0.041). 82 (9.7%) of the total cohort were on inhaled corticosteroid (ICS) at time of ICI, and 9 (11%) developed pneumonitis. Rather than being protective, pts on ICS had higher risk of pneumonitis (HR 4.2, 95 CI: 1.9-9.2, p=0.001). Pts with lung cancer had an increased risk for pneumonitis compared to pts with other cancers (HR 3.2, 95 CI: 1.5-6.4, p =0.003). Other risk factors included performance status, smoking history, line of therapy, or prior treatment including radiation were not statistically significant. Conclusions: Rather than a protective effect of ICS, our analysis found a higher risk of pneumonitis in pts treated with ICS. We confirmed an increased risk of pneumonitis for lung cancer pts compared to pts with other cancers, and higher risk of pneumonitis in pts age >65y. We hypothesize that the increased inflammatory status in chronic lung inflammation may predispose pts to pneumonitis that was not ameliorated by ICS. Future study is needed in prospective cohorts to further clarify the underlying inflammatory mechanism. [Table: see text]

3048 Background: Indications for immune checkpoint inhibitor (ICI) in cancer care are expanding rapidly. There is increasing need for accurate decision tool to better guide treatment. We have constructed a new prognostic scoring system, neutrophil-lymphocyte score (NRS), based on the nonlinear dynamic change of neutrophil to lymphocyte ratio (NLR) in relation to survival over the first cycle of ICI treatment. We compared this novel system to existing indices such as NLR, lymphocyte to monocyte ratio (LMR), platelet to lymphocyte ratio (PLR), Advanced Lung Cancer Inflammation Index (ALI), and Systemic Immune-inflammation Index (SII). Methods: This is a retrospective analysis of 837 patients at Ohio State University from 2011-18. Neutrophil (ANC), lymphocyte (ALC), platelet (plt), monocyte (AMC), albumin (alb), and body mass index (BMI) were collected at baseline. Repeat labs were collected at cycle 2. NLR = ANC/ALC, ALI = BMI x alb / NLR, LMR = ALC/AMC, SII = platelet x NLR, PLR = plt/ALC. NLR Ratio = baseline NLR / repeat NLR. Based on the association between NLR and the overall survival, we assigned 1 point (p) for baseline NLR < 0.7, 6p for 0.7 to < 2, 5p for 2 to < 3, 4 p for 3 to < 4, 3 for 4 to 5, 2p for 5 to < 9, and 1p for ≥9. We also assigned 1p for NLR ratio < 0.6, 2p for 0.6 to < 0.8, 3p for 0.8 to < 1.2, 5p for 1.25 to < 1.4, 3p for 1.4 to < 1.6, and 2p for ≥1.6. NLS = sum of these 2 scores . NLS_A = NLS*alb. Time-dependent receiver operator characteristic (ROC) curves with integrated time-dependent area under the curve (TD AUC) values were used to evaluate the predictive accuracy of each index for survival. Results: For baseline and repeat values, all indices were statistically significant (P < 0.001) in predicting survival. Baseline integrated TD AUC were: ALI 0.704, NLR 0.692, SII 0.663, LMR 0.645, and PLR 0.612. All of the repeat indices at cycle 2 had higher prognostic value than their baseline counterparts. Integrated TD AUC for indices at cycle 2 were: ALI 0.740 (with baseline BMI), NLR 0.729, SII 0.694, LMR 0.671, and PLR 0.652. NLS_A was a composite score based on the dynamic change of NLR from cycle 1 to 2 and the treatment alb with integrated TD-AUC at 0.754. Conclusions: Indices constructed from ANC, ALC, AMC, Plt, alb, and BMI can be obtained inexpensively and provide great prognostic value for pts on ICI. We have constructed a novel scoring system (NLS_A) and demonstrated its improvement over the current prognostic indices. Studies with a larger cohort are needed to further improve and validate this system.

46 Background: Neutrophil to lymphocyte ratio (NLR) is known to be prognostic for patients with various types of cancer, including those who are treated with immune checkpoint inhibitors (ICI). We evaluated NLR at baseline and during early phase of treatment for patients with advanced non-small cell lung cancer (NSCLC) who received ICI to evaluate its prognostic value in first line ICI therapy and its relationship to programmed death-ligand 1 (PD-L1) expression. Methods: We retrospectively evaluated patients with advanced NSCLC who received ICI as first line therapy from 2016 to 2018. NLR was calculated as ratio of absolute neutrophil/lymphocyte counts and was consider elevated if ≥5. PD-L1 expression by immunohistochemistry was performed as standard of care with 22C3 antibody. Kaplan Meier analysis was used for survival. Wilcoxon-Mann-Whitney test was used to evaluate PD-L1 expression between groups. All Calculation was performed using SAS V9.4. Results: A total of 78 patients were included in this study; 28 received pembrolizumab monotherapy, 45 received pembrolizumab, carboplatin, and pemetrexed, and 5 received other pembrolizumab combinations. Patients with baseline NLR < 5 before initiating ICI had median estimated OS of 46.1 months (79.4% patient still alive and 95% CI not reached) compare to median OS of 10.7 months (95% CI lower limit 6.4 and upper limit not reached) for patients with NLR ≥5, P < 0.001. Similar association between NLR and OS can be seen when NLR were repeated immediate before the cycle 2 with P = 0.001. We observed an association between baseline NLR and PD-L1 expression. The median PD-L1 tumor proportion score (TPS) was 60% (Interquartile Range 1-80) for patients with baseline NLR < 5 vs 20% (IQR 0-60) for patients with NLR ≥5, P = 0.037. Conclusions: We confirmed the prognostic value of NLR for patients with advanced NSCLC at baseline and during early treatment in patients with NSCLC receiving first line ICI treatment. We also found an association between elevated NLR and lower PD-L1 expression. Future study with larger cohort is still needed to further delineate these relationships.

Background: Some cancer patients with thromboembolism require dual treatment of VEGFR TKIs and factor Xa inhibitors (direct or indirect), which may contribute to increased bleeding risks. However, the safety of such combination treatment has not been well characterized in the literature or national guidelines. Methods: This is a single center retrospective study, where we identified metastatic cancer patients (renal cancer, colorectal cancer, sarcoma, etc) who received concurrent VEGFR TKIs (pazopanib, sunitinib, sorafenib, axitinib, regorafenib, vandetanib, lenvatinib or cabozantinib) and Xa inhibitors (low-molecular weight heparin [LMWH] or direct oral anticoagulants [DOACs] including rivaroxaban or apixaban) at the Ohio State University Medical Center. We assessed bleeding risks of dual therapies vs. factor Xa inhibitors alone, using the same patients as self controls. We reviewed medical charts of all identified patients for clinically significant bleeding events (defined as combined major bleeding and clinically relevant non-major bleeding by the International Society of Thrombosis and Haemostasis criteria). The Cox proportional hazard model was used to compare the differences of time to first clinically significant bleeding event between the groups of concurrent use and anticoagulant use only. Results: A total of 86 patients (26 females and 60 males) were included: 78 Caucasians, 6 African Americans, and 2 others. 81 patients had concurrent TKI and LMWH treatment; 20 patients had concurrent TKI and DOACs; and 85 patients have had been on factor Xa inhibitor alone (LMWH or DOACs) at some point. Some patients had been on both LMWH and DOACs at different time periods. Overall, there were 29 clinically significant bleeding events (including 8 major bleeding) during concurrent treatment and 17 events (4 major bleeding) during factor Xa inhibitor alone over a median follow up of 63 days (52 days for concurrent treatment and 99 days for Xa inhibitor alone). In this self-control study, concurrent treatment was associated with a statistically higher risk of clinically significant bleeding events (HR, 2.84; 95% CI, 1.43-5.64, P < 0.01), which reached 37% patient population in the first 3 months, while the bleeding associated with factor Xa inhibitor alone seemed spaced out at the entire length of anticoagulation (8% by 6 months). Similar trend was found in the analysis of patient group of concurrent TKI and LMWH vs. LMWH alone (HR, 1.96; 95% CI, 0.95-4.02, P = 0.067), although significance was not reached likely due to insufficient power. Sample size was inadequate for meaningful comparison between concurrent VEGFR TKI and DOAC vs. DOAC alone. Conclusions: Concurrent treatment of VEGFR TKI and factor Xa inhibitors is associated with a significantly increased bleeding risks when compared with factor Xa inhibitors alone in patients with metastatic cancer. Disclosures No relevant conflicts of interest to declare.

e18201 Background: Neutrophil to Lymphocyte Ratio (NLR) is prognostic for cancer patients treated with immune checkpoint inhibitors (ICI). We showed the change in NLR early during treatment to be a stronger, curvilinear predictor, i.e. patients with an intermediate change in NLR performed better than those with large decreases or increases. This led us to re-examine whether NLR is an optimal predictor of overall survival (OS). Methods: A retrospective review of 467 patients with advanced cancer who received ICIs from 2011 to 2017 at the Ohio State University was performed with IRB approval. NLR was collected at the initiation of ICI and on-treatment (median 21, IQR 8 days) and calculated as ratio of absolute neutrophil to lymphocyte counts. Variable selection machine-learning algorithms included fast and frugal decision trees and random forest, performed in R. Results: The machine-learning algorithm fast and frugal decision trees identified the ratio of NLR on treatment to baseline NLR, the NLR on treatment, the change in NLR and the cubic change in NLR to be the most informative predictors of survival at 12 months. A random forest algorithm identified the same four variables as the most important for prediction accuracy. Age, sex and cancer type were the least informative predictors in the model, suggesting the on-treatment NLR variables are of value across wide range of demographics. Conclusions: NLR measured during treatment, and its derivative values of the ratio to baseline, change from baseline, and the cubic change from baseline, hold more predictive value than NLR measured at baseline. Common control variables such as age, and sex showed little effect on the model, suggesting on-treatment NLR is useful across wide demographic space. [Table: see text]

2577 Background: Immune checkpoint inhibitor (ICI) therapy is often suspended because of immune-mediated diarrhea and colitis (IMDC). We examined the recurrence rate and risk factors for IMDC after ICI resumption. Methods: This retrospective multicenter study examined patients who resumed ICI therapy after improvement of IMDC between 1/2010 and 11/2018. Univariate and multivariate logistic regression analyses assessed the association of clinical covariates and IMDC recurrence. Results: Of the 167 patients in our analysis, 32 resumed an anti-CTLA-4 agent and 135 an anti-PD-1/L1 agent. The median duration from IMDC to restart of ICI treatment was 49 days (IQR, 23-136). IMDC recurred in 57 (34%) patients overall (44% of those resuming an anti–CTLA-4 and 32% resuming an anti–PD-1/L1 agent); 47 of these patients (82%) required immunosuppressive therapy for recurrent IMDC (Table). The median duration from ICI resumption to IMDC recurrence was 53 days (IQR 22-138). On multivariate logistic regression, patients who received anti-PD-1/L1 therapy at initial IMDC had a higher risk of IMDC recurrence (odds ratio [OR], 3.45, 95%CI, 1.59-7.69; P

94 Background: The response to Immunotherapy (IO) is known to be affected by concomitant medications including corticosteroids and antibiotics. We evaluated the medication history of patients undergoing IO to explore other medications that affect overall survival and to estimate the relative impact of each medication when given in combination. Methods: A retrospective review of patients with advanced cancer who received IO from 2011 to 2017 at the Ohio State University was performed with IRB approval. Data were extracted from the medical record, including medication history 180 days around the start of IO therapy. Data were collected in a REDCap database. Overall Survival (OS) was calculated from the initiation of IO. Cox Proportional-Hazards models were used and evaluated by log-rank test at alpha = 0.05. All calculations were performed using the survival and survminer packages in R. Results: Patients who received antibiotics or corticosteroids had decreased OS (p = 0.019 and p = 0.043, respectively) across several cancer types. Medications that were not significantly associated with OS included statins (p = 0.38), proton pump inhibitors (p = 0.94), H2 blockers (p = 0.27) and NSAIDS (p = 0.46). A total of 159 patients had complete data for all medications suitable for modeling relative effects. 149 (94%) of patients received antibiotics within 180 days of IO and 19 (12%) received both corticosteroids and antibiotics. The combination of corticosteroids and antibiotics had lower median OS than antibiotics alone or neither medication (p < 0.0018). A Cox Proportional Hazards model of antibiotics and corticosteroids, controlling for age, BMI and ECOG performance status, showed antibiotics, age and BMI to be significant predictors. Conclusions: Antibiotics and corticosteroids near the start of IO reduced overall survival, and the combination reduced the median overall survival further. However, a combined model that controlled for age, BMI and ECOG showed antibiotics, age and BMI to have a significant effect on OS. Though preliminary, these results suggest that antibiotics and corticosteroids may be affecting OS in the context of IO through overlapping pathways.

118 Background: Immune-related adverse events (irAEs) due to immune-checkpoint inhibitors (ICI) can lead to significant morbidity. Immune pneumonitis (IP) constitutes one of the common irAEs in non-small cell lung cancer (NSCLC) with a reported incidence of 3-5%. However outside of clinical trials there is a paucity of data regarding incidence/outcomes of IP in nivolumab (N) treated metastatic-NSCLC (m-NSCLC). Methods: We retrospectively pooled data of 381 m-NSCLC patients (pts) from four cohorts treated with N within different treatment time periods: East Carolina University, USA: 4/2015- 2/2018; Ohio State University, USA: 6/2013-6/2016; University of Perugia, Italy: 10/2013-9/2017, and Sendai Kousei Hospital, Japan: 1/2016-1/2018. IP identification was based on clinical diagnosis of treating team. Grade (G) was based on Common Terminology Criteria for Adverse Events (CTCAE v 4.03). Time to immune pneumonitis (TTIP) was the time from starting N to diagnosis of IP. Results: The rate of IP was 13.1% (n=50) with 68% of IP presenting as the index irAE. Median TTIP was 9.75 weeks (range: 0.25-73.7) with G1 (24.0%), G2 (36.0%), G3 (36.0%), G4 (4.0%). 57.9% of IP ≥G2 had TTIP ≤ 3 months (m). For all-G (IP), 22.0% pts had stopped N before IP either due to non-IP irAEs or progression. Excluding these, treatment discontinuations (d/c) attributed to IP ≥ G2 were 66.6%. IP ≥G2 had higher treatment d/c when TTIP was ≤ 3m vs. > 3m (51.8% vs. 14.8%; P=0.09). Median overall survival (OS) by log rank test for pts with any grade IP vs. no IP was not statistically significant (13.7 vs. 7.5m; P =0.14). IP ≥G2 occurring at > 3m had longer median OS vs. IP ≥G2 occurring in ≤3m (16.7 vs. 3.6m, P=0.008), which was not significant at a 3m landmark analysis when accounting for time on N (P= 0.31). Conclusions: Our real-world multi-institutional data shows a higher incidence of IP in m-NSCLC. We also see a trend for frequent ICI d/c in IP ≥ G2 occurring within 3 months that may ultimately influence OS per our landmark analysis. Thus aggressive and early symptom management specifically directed at allowing reinitiation of ICI after IP needs to be considered.