Thoracic Cancer Research Articles

Diagnostic trajectories of patients with rare cancer in the Netherlands: results from a nationwide cross-sectional survey.

Diagnosing rare cancers is challenging and often leads to prolonged diagnostic trajectories. This study investigated the diagnostic trajectory of patients with rare cancers in The Netherlands. Data from 1541 patients were recruited via patient advocacy in a national online survey on their diagnostic trajectory, such as first general practitioner (GP) consultation to hospital referral and number of hospital visits before final diagnosis. Differences between solid vs. non-solid tumours and EURACAN domains were explored. Diagnostic timelines varied from less than 3 months to over 12 months. Most patients (76.0%) first consulted their GP before going to a hospital. 76.3% of all patients were referred to a hospital within less than 3 months. 32.1% reported receiving an incorrect diagnosis, and 44.6% of them underwent treatment or medication for the (perceived) incorrect diagnosis. Patients with solid vs. non-solid rare cancers trajectories differed significantly for treatment hospital, route to diagnosis, correctness of initial diagnosis, and number of hospital visits before correct diagnosis (all p < 0.001). Patients with neuroendocrine (NET; 21.7%) and endocrine tumours (17.5%) experienced longer GP-to-hospital visit waiting times. Patients with non-solid cancers often received a correct diagnosis after one hospital visit (75%) when compared with patients with solid cancer (2+ = 57.7%). Those with rare skin cancer and non-cutaneous melanoma, head and neck, and thoracic cancer visited multiple hospitals before an accurate diagnosis (56.7%, 53.8%, and 50.0%). Patients with rare cancers face significant challenges with diagnostic delays and inaccuracies. Researching symptom signatures and investing in regional clinical networks might improve diagnostic timelines.

Emergency department use before cancer diagnosis in Ontario, Canada: a population-based study.

Although suspicions of cancer may be raised in patients who visit the emergency department, little is known about emergency department use before a cancer diagnosis. We sought to describe emergency department use among patients in Ontario within the 90 days before confirmed cancer diagnosis and to evaluate factors associated with this emergency department use. We conducted a retrospective, population-based study of patients aged 18 years or older who had a confirmed cancer diagnosis in Ontario from 2014 to 2021 using linked administrative databases. The primary outcome was any emergency department visit within 90 days before the cancer diagnosis date. We used multivariable logistic regression to evaluate factors associated with emergency department use, such as demographics (e.g., age, sex, rurality, Ontario Health region, indicators of marginalization), comorbidities, previous emergency department visits and hospital admissions, continuity of primary care, type of cancer, and year of cancer diagnosis. We included 651 071 patients with cancer. Of these, 229 683 (35.3%) had an emergency department visit within 90 days before diagnosis, 51.4% of whom were admitted to hospital from the emergency department. Factors associated with increased odds of emergency department use before cancer diagnosis included rurality (odds ratio [OR] 1.15, 95% confidence interval [CI] 1.13-1.17), residence in northern Ontario (North East region OR 1.14, 95% CI 1.10-1.17 and North West region OR 1.27, 95% CI 1.21-1.32, v. Toronto region), and living in the most marginalized areas (material resources OR 1.37, 95% CI 1.35-1.40 and housing OR 1.09, 95% CI 1.06-1.11, v. least marginalized quintile). We observed significant variation in emergency department use by cancer type, with high odds of emergency department use among patients with intracranial, pancreatic, liver or gallbladder, or thoracic cancer. Emergency department use is common before cancer diagnosis, with about one-third of patients with cancer in Ontario using the emergency department before diagnosis. Understanding why patients visit the emergency department before cancer diagnosis is important, particularly for patients who live in rural or marginalized areas, or those who have specific cancer types.

Postoperative lymphopaenia as a risk factor for postoperative infections in cancer surgery: A prospective multicentre cohort study (the EVALYMPH study).

Stress due to surgical trauma decreases postoperative lymphocyte counts (LCs), potentially favouring the occurrence of postoperative infections (PIs). We aimed to determine whether postoperative lymphopaenia following thoracic or gastrointestinal cancer surgery is an independent risk factor for PIs and to identify modifiable factors related to anaesthesia and surgical procedures that might affect its occurrence. The EVALYMPH study was a prospective, multicentre cohort study with a 30-day patient follow-up. Multivariate analyses were performed to determine the risk factors for PIs and for postoperative lymphopaenia. Patients were included from January 2016 to September 2017 in 25 French centres. Adult patients admitted for thoracic or gastrointestinal cancer surgery were eligible for inclusion. PIs within 30 days after surgery were defined as urinary tract infections, pneumonia, surgical site infections and other infections (bloodstream infections or pleurisy). Of 1207 patients included, 273 (22.6%) developed at least one infection within 30 days after surgery, with a median [IQR] time to onset of 8 [5 to 11] days. An increased risk of PI was significantly associated with an ASA score of IV: hazard ratio (HR) 4.27 (95% confidence interval (CI), 1.87 to 9.72), surgery > 200 min (HR 1.58 (1.15 to 2.17) and lymphopaenia on postoperative day 1 (POD1) (HR 1.56 (1.08 to 2.25). This risk was associated with changes in postoperative LC over time (P = 0.001) but not with preoperative LC (P = 0.536).POD1 lymphopenia was related to patient characteristics and duration of surgery but not to potentially modifiable other surgical or anaesthetics factors. POD1 lymphopaenia was associated with PIs in patients undergoing thoracic or gastrointestinal cancer surgery. To individualise care, patient characteristics and surgery duration should be taken into account. ClinicalTrials.gov Identifier: NCT02799251.

Screening, identification, and diagnosis of malnutrition in hospitalized patients with solid tumors: A retrospective cohort study.

Malnutrition is common in hospitalized patients with cancer and adversely affects clinical outcomes. We evaluated the prevalence of malnutrition risk, dietitian-identified malnutrition (DIMN), and physician-diagnosed malnutrition (PDMN) at admission. This retrospective study included adults diagnosed with a stage I-IV solid tumor malignancy and admitted to Atrium Health Carolinas Medical Center from January 2016 to May 2019. Malnutrition risk was determined by a score ≥2 on the Malnutrition Screening Tool (MST) administered by a registered nurse during the intake process. Registered dietitian nutritionist (RDN) assessments were reviewed for DIMN and grade (mild, moderate, or severe). PDMN included malnutrition International Classification of Diseases, Tenth Revision codes in the discharge summary. Univariate models were estimated; multivariate logistic regression models identified associations between clinicodemographic factors and malnutrition prevalence with stepwise selection. A total of 5143 patients were included. Median age was 63 (range 18-102) years, 48% were female, 70% were White, and 24% were Black. Upper gastrointestinal (21%), thoracic (18%), and genitourinary (18%) cancers were most common. A total of 28% had stage IV disease. MST scores were available for 4085 (79%); 1005 of 4085 (25%) were at malnutrition risk. Eleven percent (n = 557) had malnutrition coded by a physician or documented by an RDN; 4% (n = 223) of these were identified by both clinicians, 4% (n = 197) by RDNs only, and 3% (n = 137) by physicians only. Malnutrition appears to be underdiagnosed by both RDNs and physicians. Underdiagnosis of malnutrition may have significant clinical, operational, and financial implications in cancer care.

Is extended resection for locally advanced thoracic cancer with cardiopulmonary bypass justified?

BackgroundResection of intrathoracic tumor with cardiopulmonary bypass (CPB) remains a relatively under-reported intervention in literature, and its role in managing locally advanced mediastinal and lung cancers is a topic of ongoing debate. Our aim was to review our experience and assess the role of CPB for treating locally advanced mediastinal and lung cancers.MethodsBetween 2015 and 2020, this study initially included 10 patients with primary locally advanced thoracic malignancies with apparent adjacent cardiovascular invasion demonstrated by thoracic imaging scans. Operation was performed based on a multidisciplinary tumor board consensus. Eventually, 8 patients (3 primary lung cancers and 5 mediastinal cancers) received either salvage or elective resection with CPB; two completed surgery without requiring CPB.ResultsRegarding the extent of adjacent structure involvement, 4 patients presented with involvement of the superior vena cava (SVC), 1 involved the right atrium (RA), 2 involved the SVC and RA, and 1 involved the SVC, the origin of main pulmonary artery, and the ascending aorta. Thirty-day mortality occurred in two of three patients receiving salvage surgery due to respiratory insufficiency. With the long-term follow-up, one patient died of recurrence 25 months postoperatively, one survived with recurrence 30 months postoperatively, and four were alive without recurrence for 35, 36, 49, and 107 months after operations.ConclusionIn certain patients, particularly for elective surgical candidates rather than salvage resection, CPB allows for extended resection of locally advanced thoracic cancers with acceptable perioperative safety and survival.

Cannabidiol (CBD) Protects Lung Endothelial Cells from Irradiation-Induced Oxidative Stress and Inflammation In Vitro and In Vivo.

Objective: Radiotherapy, which is commonly used for the local control of thoracic cancers, also induces chronic inflammatory responses in the microvasculature of surrounding normal tissues such as the lung and heart that contribute to fatal radiation-induced lung diseases (RILDs) such as pneumonitis and fibrosis. In this study, we investigated the potential of cannabidiol (CBD) to attenuate the irradiation damage to the vasculature. Methods: We investigated the ability of CBD to protect a murine endothelial cell (EC) line (H5V) and primary lung ECs isolated from C57BL/6 mice from irradiation-induced damage in vitro and lung ECs (luECs) in vivo, by measuring the induction of oxidative stress, DNA damage, apoptosis (in vitro), and induction of inflammatory and pro-angiogenic markers (in vivo). Results: We demonstrated that a non-lethal dose of CBD reduces the irradiation-induced oxidative stress and early apoptosis of lung ECs by upregulating the expression of the cytoprotective mediator heme-oxygenase-1 (HO-1). The radiation-induced increased expression of inflammatory (ICAM-2, MCAM) and pro-angiogenic (VE-cadherin, Endoglin) markers was significantly reduced by a continuous daily treatment of C57BL/6 mice with CBD (i.p. 20 mg/kg body weight), 2 weeks before and 2 weeks after a partial irradiation of the lung (less than 20% of the lung volume) with 16 Gy. Conclusions: CBD has the potential to improve the clinical outcome of radiotherapy by reducing toxic side effects on the microvasculature of the lung.

Prognostic Impact of Low Muscle Mass and Inflammatory Markers in Stage III Nonsmall Cell Lung Cancer Turkish Oncology Group and Turkish Society of Radiation Oncology Thoracic Cancer Study Group (08-005).

The aim of this retrospective multicenter study was to evaluate the prognostic significance of low muscle mass, and inflammatory markers in patients with stage III nonsmall cell lung cancer (NSCLC) who received definitive chemoradiotherapy (CRT). Furthermore, the study aimed to determine the threshold value of disease-specific low muscle mass. A total of 461 patients with stage III NSCLC were evaluated. Low muscle mass, prognostic nutritional index (PNI), and biochemical inflammatory markers were assessed. The Kaplan-Meier method and Cox regression analysis were used to analyze overall survival (OS) and progression-free survival (PFS). This study found a disease-specific low muscle mass threshold of LSMI <38.7cm²/m² for women and <45.1cm²/m² for men, with 25.2% of patients having disease-specific low muscle mass. Multivariate cox regression analysis revealed that low PNI was found to be an independent unfavorable prognostic factor for both PFS (HR=0.67; 95% CI: 0.48-0.92, P= 0.015) and OS (HR=0.67; 95% CI: 0.50-0.91, P=0.008). Other factors including ECOG PS 3 (HR=7.76; 95% CI: 1.73-34.76, P=0.007), induction CT (HR=0.66; 95% CI: 0.49-0.88, P= 0.004), and disease-specific low muscle mass (HR=1.40; 95% CI: 1.02-1.92, P= 0.038) also had independent effects on prognosis. The present study provides evidence that the presence of low muscle mass and low PNI significantly impacts the prognosis of patients with stage III NSCLC who undergo definitive CRT. Furthermore, our study is notable for being the first multicenter investigation to identify a disease-specific low muscle mass threshold.

MA14.04 Global Burden and Trends of Asbestos-Related Thoracic Cancers & Their Associations with National Asbestos Ban Status: A Cross-Sectional Study

MA14.04 Global Burden and Trends of Asbestos-Related Thoracic Cancers & Their Associations with National Asbestos Ban Status: A Cross-Sectional Study

Pharmacist-facilitated Patient Reported Outcome Measure (PROM) monitoring: developing an EHR SmartForm© to monitor side effects of oral oncolytics during routine telehealth encounters.

Patient reported outcome measures (PROMs) are increasingly used in oncology care, but pharmacists providing direct patient care have been overlooked. We engaged pharmacists and adults receiving oral oncolytics (chemotherapy medication taken by mouth) to develop a SmartForm© in the electronic health record (EHR) for PROM monitoring. Pharmacists verbally ask the patient side effect questions during routine telehealth encounters and enter responses in real time. Our development process was guided by the Knowledge to Action Framework. In phase 1 (Knowledge Inquiry), we prioritized side effects to assess in the EHR SmartForm© via interviews with patients and a Delphi panel with pharmacists. Adults receiving oral oncolytics for breast (n = 12), thoracic (n = 12), or hematological (n = 12) cancer were interviewed, with purposeful sampling for adults who were aged 65 + years or Black. Interviews were coded with content analysis. We conducted three Delphi rounds, with 11/19, 13/19, and 19/19 pharmacists, respectively. In phase 2 (Knowledge Synthesis), PROM items were selected and the EHR SmartForm© programmed. In phase 3 (Knowledge Tailoring), we conducted usability testing with pharmacists. Pharmacists and patients were consistent in prioritizing side effects of oral oncolytics and 10 were retained. Patients advocated asking whether they can do their usual activities, while pharmacists added medication adherence. Usability testing yielded suggestions to simplify the SmartForm©. By presenting screenshots of our SmartForm©, our findings are useful to other healthcare systems looking for a PROM solution integrated in the EHR, with a reasonable pharmacist/clinician workload, and no requirement for patients to have internet access/comfort.

Epidemiological Study of Adenoid Cystic Carcinoma and Its Outcomes: Insights from the Surveillance, Epidemiology, and End Results (SEER) Database

Objective: Adenoid cystic carcinoma (ACC) is a rare malignant tumor that mainly arises in the head and neck area. We aimed to compare the long-term survival of patients with ACC based on their geographic regions within the United States using the Surveillance, Epidemiology, and End Results (SEER) registry data. Methods: We queried the SEER database to evaluate the geographic distribution of ACC patients based on inpatient admissions. The states included in the study were divided into four geographical regions (Midwest, Northeast, South, and West) based on the U.S. Census Bureau-designated regions and divisions. Demographic and clinical variables were compared between the groups. Kaplan–Meier curves and Cox regression were used to assess late mortality. Results: A total of 5150 patients were included (4.2% from the Midwest, 17.2% from the Northeast, 22.5% from the South, and 56.1% from the West regions). The median follow-up was 12.3 (95% CI: 11.6–13.1 years). Median overall survival was 11.0 (95% CI: 9.2-NR years), 14.3 (95% CI: 12.4–16.4 years), 11.3 (95% CI: 9.7–14.8 years), and 12.0 (95% CI: 11.3–13.0 years) for Midwest, Northeast, South, and West regions, respectively. In multivariable analysis, older age, male sex, thoracic cancer, the presence of regional and distal disease, receiving chemotherapy, not undergoing surgical resection, and being treated in the West vs. Northeast region were found to be independent predictors of poor survival. We identified a significant survival difference between the different regions, with the West exhibiting the worst survival compared to the Northeast region. Conclusions: In addition to the well-known predictors of late mortality in ACC (tumor location, stage, and treatment modalities), our study identified a lack of social support (being unmarried) and geographic location (West region) as independent predictors of late mortality in multivariable analysis. Further research is needed to explore the causal relationships.

Longitudinal outcomes after financial toxicity and health-related social risks screening and referral.

17 Background: Financial toxicity (FT) is associated with delayed/omitted care, reduced quality of life, and worse mortality. Routine screening for FT and health-related social risks (HRSR) is recommended during clinical care, but there is limited longitudinal evidence on outcomes after patients screen positive and are referred for assistance. Methods: FT and HRSR screening for patients with breast, gynecologic, gastrointestinal, or thoracic cancers was implemented at an urban comprehensive cancer center. Screening included the Comprehensive Score for Financial Toxicity (COST) tool (0-44, lower scores=worse FT) and a HRSR checklist to identify food, housing, and transportation insecurity. COST score ≤16-20 or reported HRSR was a “positive” screen and counseling referral was offered: patients could accept the referral by identifying their current concerns. Analysis sample was limited to receiving chemotherapy, radiotherapy, or surgery 120 days prior to baseline survey and had a follow-up survey 4-12 months later. Data were collected 10/2022-12/2023. Results: 2196 patients met inclusion criteria, median time between surveys was 5.5 months (IQR 4.4, 6.9). 20% (n=438) screened positive and accepted referral; 11% (n=246) screened positive and declined referral; the remaining 69% (n=1512) screened negative. Sociodemographic factors of age (median 58 vs 62, p&lt;.001), sex (82% vs 74% female, p =.015), and race (31% vs 17% racial/ethnic minority, p&lt;.001) were associated with accepting vs declining referral. Those who accepted referral had lower COST scores (median 10 vs 19) and more HRSR (39% vs 20% housing; 32% vs 18% food; 37% vs 17% transportation) at baseline (all p≤.001). At follow-up, both cohorts had COST decrease, indicating worsened FT [mean difference (95% CI): accepted referral -2.0 (-2.8, -1.2) vs declined referral -2.8 (-4.0, 1.5)]. Across all collected HRSR, those who accepted referral had small benefits in previously reported HRSR, however were more likely to have newly designated HRSR (Table shows changes). Conclusions: Screening for FT and HRSR identified one-third of patients on recent cancer treatment with financial or social needs. Those who accepted financial assistance referral had significantly higher FT and more reported HRSR. Although referral did appear to have some benefit on initial needs, new HRSR were identified at follow-up. Future work will focus on optimizing available assistance to meet ongoing patient needs during and after cancer treatment. Accepted Referraln = 438 (%) Declined Referraln = 246 (%) Housing Resolved need 67 (15) 31 (13) New need 54 (12) 15 (6.1) Stable unmet need 102 (23) 18 (7.3) Stable no need 215 (49) 182 (74) Food Resolved need 64 (15) 31 (13) New need 69 (16) 16 (6.5) Stable unmet need 77 (18) 14 (5.7) Stable no need 228 (52) 185 (75) Transportation Resolved need 73 (17) 34 (14) New need 56 (13) 20 (8.1) Stable unmet need 87 (20) 8 (3.3) Stable no need 222 (51) 184 (75)

Clinical utility and impact of a diagnostic oncology clinic.

74 Background: Oncologists have expertise in the work-up and management of patients with a suspected diagnosis of cancer. Patients diagnosed with cancer during acute care hospitalization often have poorer outcomes than those diagnosed as outpatients. For these patients, time to appropriate cancer diagnosis, management of cancer-related symptoms, and timely initiation of treatment are key to improved outcomes. Outpatient new diagnosis clinics may improve patient outcomes by improving access to subspecialty care. Methods: Patients seen by the inpatient oncology consult service over a two-month period (8/1/23-8/31/23 and 10/1/23-10/31/23) were included in this analysis. Inpatient management of disease was defined as an invasive procedure to treat the disease or a complication of the disease, initiation of systemic anti-neoplastic treatment, or radiation therapy. Surveys regarding perceived barriers to timely outpatient oncology care were distributed to oncology fellows and attendings. A Diagnostic Oncology Clinic (DOC) was started 2/1/2024 with the goal to decrease time to first outpatient oncology by a median of 12 days to less than 10 days, decrease hospital length of stay (LOS), improve outpatient management of cancer related symptoms, and improve time to treatment. Results: A total of 95 patients were evaluated. 74 (78%) of patients had a new diagnosis of cancer. Most patients had gastrointestinal cancer (n = 40, 42%) or thoracic cancer (n = 24, 25%). 51 patients (53.7%) had metastatic disease. 48 patients (50.5%) had an ECOG performance status of 0 or 1 at the time of consultation. The median hospital LOS was 11 days (IQR, 5-19). 13 patients (13.7%) had an inpatient PET scan and 43 patients (45.3%) required inpatient management of their disease or disease-related complication. 57 patients (60%) were discharged home, 21 (22%) to an extended care facility, and 17 (18%) expired or transitioned to hospice. 35 patients (37%) were seen as new outpatients with a median time to visit of 12 days (IQR 7-20). The median time to initiation of anti-neoplastic therapy as an outpatient was 24 days (IQR 18-33). There were 29 survey respondents corresponding to a response rate of 69%. 16 (55%) felt that about half the time or less patients will have timely follow-up with an oncologist after discharge, and 19 (65%) felt half the time or less patients will have appropriate cancer-related symptom management after discharge. Conclusions: Most patients seen by the inpatient oncology consult service with a new diagnosis of cancer have advanced and symptomatic disease. Historically, there has been significant time to outpatient care and cancer-specific management led by an oncologist. Creation of a diagnostic oncology clinic may shift some diagnostic testing and symptom management from the inpatient to outpatient setting, decrease time to first outpatient visit, and improve provider workflow for a vulnerable patient population. Analysis of this intervention is ongoing.

Validating a simplified financial toxicity screening tool for patients with cancer.

318 Background: Medical-related financial hardship (called financial toxicity) can negatively impact patient clinical outcomes and overall well-being. Routine financial toxicity screening is recommended to facilitate referrals to resources for patients experiencing, or at risk for, financial toxicity. The validated Comprehensive Score for Financial Toxicity (COST) tool is widely used in research to measure financial toxicity; however, its 11-question length and lack of standardized screening threshold make it challenging to implement in clinical practice. This analysis sought to develop and validate an abbreviated version of the tool, suitable for routine screening in clinical practice. Methods: The full COST was administered as part of clinical care from June 2022-August 2023. Patients receiving treatment for breast, gynecologic, thoracic, or gastrointestinal cancer at an urban comprehensive cancer center were queried via patient portal or in clinic. Using 38,330 observations, we calculated the correlation of abbreviated tools of varying length with the full COST score (0-44, lower scores=higher FT). We then used a validation dataset of 19,196 observations to assess the positive and negative predictive values (PPV, NPV) of an abbreviated tool against the full COST score, when compared to thresholds for risk previously used in the literature (score ≤15, ≤20). Results: Including two questions (Q3: “I worry about the financial problems I will have in the future as a result of my illness or treatment”; Q6: “I am satisfied with my current financial situation”) yielded r=0.922 correlation with the full COST. For this 2-item tool, PPV ranged from 74-91%, and NPV ranged from 91-98% when compared to the full COST tool. Table 1 shows the PPV and NPV of varying score thresholds for the abbreviated COST against full scores ≤15 and ≤20. Conclusions: Using a large and clinically diverse sample of patients with cancer,we demonstrate the predictive value of an abbreviated COST tool and show its approximate accuracy when compared to the full tool. Simplified, clinically integratedscreening for financial hardship can efficiently ensure appropriate intervention for patients experiencing medical-related financial hardship and can also be a first step toward mitigating hardship among those at risk for developing it. Predictive value of 2-item COST vs. full COST score in the validation set. Full COST score ≤ 15 Yes No 2-question, COST score ≤ 1 Yes 2337 (14%) 253 (1.5%) PPV = 90% No 822 (4.8%) 13844 (80%) 1-NPV = 5.6% Full COST score ≤ 15 Yes No 2-question, COST score ≤ 2 Yes 2924 (17%) 1045 (6.1%) PPV = 74% No 235 (1.4%) 13052 (76%) 1-NPV = 1.8% Full COST score ≤ 20 Yes No 2-question, COST score ≤ 2 Yes 3626 (21%) 343 (2.0%) PPV = 91% No 1199 (6.9%) 12088 (70%) 1-NPV = 9.0% Full COST score ≤ 20 Yes No 2-question, COST score ≤ 3 Yes 4359 (25%) 1272 (7.4%) PPV = 77% No 466 (2.7%) 11159 (65%) 1-NPV = 4.0%

The cancer diagnostic clinic: A rapid access clinic to improve outcomes for patients with a suspected cancer diagnosis.

77 Background: Patients with suspected cancer often face challenges navigating complex health systems, resulting in delays in diagnosis and time to treatment, which are associated with adverse outcomes and patient dissatisfaction. To improve patient access and bridge this healthcare gap, the Cancer Diagnostic (CDx) clinic was established to streamline the evaluation for patients with a new suspected cancer. The CDx clinic prioritizes expedited cancer evaluation, timely and accurate diagnostic testing, and transition of care to subspecialty Oncologists, to improve the patient experience and cancer-related outcomes. Methods: Leaders in the Abramson Cancer Center (ACC) initiated a pilot project to facilitate timely appointments to Oncologists for patients in the Emergency Department with a new suspected cancer. Patient access lead time and time to diagnosis and treatment were tracked. This pilot was formally expanded across the health system to become the CDx clinic for eligible adult patients with suspicious symptoms, lab abnormalities, or imaging findings concerning for cancer. Patient data was extracted using an EMR-based form and supplemented by chart review. Study period was March 2023 to April 2024 and included socio-demographics, cancer subtype, referral type, time from referral to new patient appointment (access lead time), time from first appointment to diagnosis and treatment (surgery, systemic or radiation therapy), patient retention, complications, and mortality. Results: Over 12 months, 108 patients were evaluated in the CDx clinic. Most patients were male (56%), White (58%), and non-residents of Philadelphia County (65%). There were 8 referrals (range: 6-13) per month with the majority of patients from external referrals (60.2%). Compared to baseline, CDx resulted (median pre/post in days) in similar time to first appointment (7 vs. 7 days), improved time to diagnosis (21 vs. 8 days), and improved time to treatment (42 vs. 29.5 days). 71 patients were diagnosed with cancer with Gastrointestinal (47%) and Thoracic (24%) cancers being most common. 98.6% (70/71) of patients were retained at ACC and transitioned to subspecialty care. 80.3% (57/71) of patients received cancer-directed treatment. After the initial consult date, for the patients diagnosed with cancer, there were 4 ED visits and 20 hospitalizations within 30 days, 19 hospice enrollments, and 16 deaths. Conclusions: The CDx clinic has been a valuable resource for patients with suspected cancer by optimizing time to diagnosis and time to treatment. Additional navigation pathways are underway to expand the CDx patient population to target a more underserved population in an effort to promote health equity in cancer care.

Impact of applying machine learning to the electronic medical record on prediction of cancer-associated thrombosis.

409 Background: Cancer-associated thrombosis (CAT) is preventable among high-risk individuals with prophylactic anticoagulation. Risk assessment tools focus only on the initial period after cancer diagnosis; however, some patients may face different risks during different periods of their cancer journey. We developed a longitudinal machine learning system to predict the risk of CAT throughout cancer treatment. Methods: Using electronic health record data at Princess Margaret Cancer Centre, we assembled a cohort of people with thoracic and gastrointestinal cancer receiving systemic treatments between August 1, 2017 and December 31, 2019. We manually labelled 5,000 CT scans and 500 Doppler ultrasounds for the occurrence of CAT. We fine-tuned open-source large language models (LLMs) on those labelled reports, which we then used to automatically detect CAT among all 37,000 CT scans and Dopplers. Finally, we trained longitudinal machine learning systems to predict CAT within 90 days after each cancer treatment. LLMs and the CAT risk prediction system were evaluated among the held-out test cohort of people whose first treatment occurred in 2019. Results: The overall cohort included 2,031 patients and 21,375 treatment sessions. When classifying radiology reports, the fine-tuned LLM achieved an area under the receiver operating characteristic curve (AUROC) of 0.758 for DVT and 0.988 for PE in the test cohort. CAT occurred within 90 days after 6.23% of treatment sessions. The best ML system predicted the risk of CAT within 90 days with an AUROC of 0.651 (95% CI, 0.620-0.678). Considering only the first treatment per patient, treatments within the first 90 days after the first treatment, and treatments after 90 days, the system achieved AUROC of 0.639 (95% CI, 0.568-0.704), 0.599 (95% CI, 0.559-0.642), and 0.741 (95% CI, 0.699-0.783), respectively. Conclusions: Machine learning can longitudinally predict CAT among patients receiving systemic therapy for aerodigestive cancers. These systems could personalize prophylactic anticoagulation by identifying specific periods when patients face sufficient risk to warrant prevention.

Time toxicity faced by patients with advanced cancer across primary cancer sites.

268 Background: Time toxicity, defined as the burden of time for patients interacting with the health care systems, is particularly important for patients with limited life expectancy. Little is known about the relative impact of time toxicity among different cancer types at the end of life. This study examines median contact days for inpatient, outpatient, and home health services among different primary cancer types for patients with advanced cancer with limited prognosis. Methods: A retrospective cohort analysis was performed on patients with advanced (stage 4) cancer who died within one year of diagnosis at Kaiser Permanente Northern California between January 1, 2015, and June 30, 2023. Time toxicity was defined as comprehensive contact days with healthcare system over the course of patients’ survival period between cancer diagnosis and death (i.e., contact days divided by survival or contact %). Types of contact included institutional stay (emergency department visit, hospitalization, skilled nursing facility, rehabilitation); outpatient visits (ambulatory care, labs, imaging); and in-home care (home-health, home hospice, telephone and virtual visit). The study quantified overall median contact % and contact % by specific care settings, stratified by cancer primary site. Results: The study included 10102 patients. The median age at diagnosis was 74 years with 47% female. Patients were predominantly white (65%), followed by Asian (12%), Hispanic (12%), and black (8%). The median Elixhauser comorbidity score was 4. With the exception of head and neck cancers, most cancers had higher contact % near diagnosis and at end of life (U-shaped curve). Head and neck cancers had the highest percent of patients undergoing surgery (36%) and radiation (52%), which may explain its sinuosity-like curve with a rapid increase of weekly contact % after diagnosis (a small peak), followed by rapid decrease and another increase by the end of their life. Survival days were comparatively long for head and neck cancers (158 days, IQR: 84-265) and short for thoracic (80 days, IQR: 35-166) and upper gastrointestinal cancers (80 days, IQR: 38-173). Malignant hematology patients had notably higher comprehensive contact % (74%, IQR: 47-100) which was driven primarily by emergency room department visits and hospital stays (p&lt;0.001 for all contact types). Conclusions: The study highlights the high burden of time toxicity in patients with advanced cancer with limited prognosis. The significant variations in time toxicity among different cancer types, especially in malignant hematology and head and neck patients, emphasize the need to tailor goals of care discussions for patients with cancer at the end of life. By focusing on reducing unwanted time toxicity, healthcare providers can enhance the quality of life for patients with cancer, ensuring more efficient and patient-centered care delivery.

EP.05B.03 Evaluating the Risk of Pneumonitis from G-CSF in Thoracic Cancer Patients with Pre-Existing ILD

EP.05B.03 Evaluating the Risk of Pneumonitis from G-CSF in Thoracic Cancer Patients with Pre-Existing ILD

P1.09A.01 The Effectiveness of Split-course Stereotactic Body Radiation Therapy in Treating Large Thoracic Cancer

P1.09A.01 The Effectiveness of Split-course Stereotactic Body Radiation Therapy in Treating Large Thoracic Cancer

Oncolytic viruses alter the biogenesis of tumor extracellular vesicles and influence their immunogenicity

Extracellular vesicles (EVs) are mediators of intercellular communication in the tumor microenvironment. Tumor EVs are commonly associated with metastasis, immunosuppression or drug resistance. Viral infections usually increase EV secretion, but little is known about the effect of oncolytic viruses (OVs) on tumor EVs. Here, we investigated the impact of oncolytic vesicular stomatitis virus (VSV) and vaccinia virus (VACV) on EVs secreted by human melanoma and thoracic cancer cells. We found that OV infection increases the production of EVs by tumor cells. These EVs contain proteins of viral origin, such as VSV-G, thus creating a continuum of particles sharing markers of both canonical EVs and viruses. As such, the presence of VSV-G on EVs improves the transfer of their protein content to cell types commonly found in the tumor microenvironment. A proteomic analysis also revealed that EVs-OV secreted during VSV infection are enriched in immunity-related proteins. Finally, CD8+ T cells incubated with EVs-OV from infected cells display slightly enhanced cytotoxic functions. Taken together, these data suggest that OVs enhance the communication mediated by tumor EVs, which could participate to the therapeutic efficacy of OVs. These results also provide rationale for engineering OVs to exploit EVs and disseminate therapeutic proteins within the tumor microenvironment.

Real-World Analysis of Survival and Treatment Efficacy in Stage IIIA-N2 Non-Small Cell Lung Cancer.

Stage IIIA-N2 non-small cell lung cancer (NSCLC) poses a significant clinical challenge, with low survival rates despite advances in therapy. The lack of a standardised treatment approach complicates patient management. This study utilises real-world data from Guy's Thoracic Cancer Database to analyse patient outcomes, identify key predictors of overall survival (OS) and disease-free survival (DFS), and address the limitations of randomised controlled trials. This observational, single-centre, non-randomised study analysed 142 patients diagnosed with clinical and pathological T1/2 N2 NSCLC who received curative treatment from 2015 to 2021. Patients were categorised into three groups: Group A (30 patients) underwent surgery for clinical N2 disease, Group B (54 patients) had unsuspected N2 disease discovered during surgery, and Group C (58 patients) received radical chemoradiation or radiotherapy alone (CRT/RT) for clinical N2 disease. Data on demographics, treatment types, recurrence, and survival rates were analysed. The median OS for the cohort was 31 months, with 2-year and 5-year OS rates of 60% and 30%, respectively. Group A had a median OS of 32 months, Group B 36 months, and Group C 25 months. The median DFS was 18 months overall, with Group A at 16 months, Group B at 22 months, and Group C at 17 months. Significant predictors of OS included ECOG performance status, lymphovascular invasion, and histology. No significant differences in OS were found between treatment groups (p = 0.99). This study highlights the complexity and diversity of Stage IIIA-N2 NSCLC, with no single superior treatment strategy identified. The findings underscore the necessity for personalised treatment approaches and multidisciplinary decision-making. Future research should focus on integrating newer therapeutic modalities and conducting multi-centre trials to refine treatment strategies. Collaboration and ongoing data collection are crucial for improving personalised treatment plans and survival outcomes for Stage IIIA-N2 NSCLC patients.