Aim This study aims to evaluate the effectiveness of combined artificial intelligence (AI)-based tools for early patient identification, risk stratification and tracking in increasing the follow-up rate of incidentally detected lung nodules, potentially leading to earlier diagnoses of lung cancer, particularly non-small cell lung cancer (NSCLC).Patients and methods We conducted a retrospective cohort study involving all patients who underwent CT scans at an academic medical centre over an 8-month period. Real-world practice was compared with modelling of a hypothetical intervention with AI tools. This study was complemented by a multi-reader multi-case analysis to enhance the robustness of our findings.Results The implementation of AI tools significantly increased the rates of guideline-concordant follow-up for detected nodules, rising from 34% without the tool to 94% with the AI intervention (p<0.0001, McNemar’s test). Furthermore, the median time to diagnosis of NSCLC was reduced from 129 days to 25 days (p<0.001, Wilcoxon signed-rank test).Conclusion These findings provide compelling evidence that AI tools can enhance the follow-up rates for patients with incidentally detected lung nodules and expedite the diagnosis of lung cancer. The integration of AI in clinical practice may significantly improve patient outcomes in lung cancer detection and management.

RATIONALE Many lung mass and lung nodule management pathways struggle with imperfect risk stratification and aligning complex care coordination, with over 50% of incidentally detected high-risk lung nodule patients not receiving timely follow-up. These diagnostic delays may lead to missed treatment opportunities, with patients missing neo-adjuvant and peri-operative treatment strategies despite evidence of their success. To address this clinical challenge, we evaluated an automated method of identifying patients with pulmonary masses from free-text CT radiology reports using Natural Language Processing (NLP). Employing such a tool in clinical practice could immediately identify and centralize management of high-risk patients upon imaging. This could help ensure that patients with potential stage II or III lung cancer receive timely diagnosis and multi-disciplinary consideration for their treatment plan. METHODS A retrospective dataset of 37,138 chest CT radiology reports was collected from Atrium Health Wake Forest Baptist between July-Nov. 2017. Reports were filtered by chest-related CPT codes (71250, 71260, 71270, and 71275) and for patients aged 60 and above. Patients were excluded if their CT image was acquired as part of the Lung Cancer Screening program. This resulted in 2,725 reports, of which 941 consecutive reports were used for internal validation, and the remaining 1,784 for NLP model optimization. The optimization subset was used to configure model parameters to achieve the highest positive predictive value (PPV) on lung mass identification, where a lung mass was defined as an abnormal lesion in the lung, pleura or mediastinum larger than 3cm. The validation reports were annotated by a clinician with the presence/absence of a reported lung mass. RESULTS The model identified 131 (7.3%) lung mass reports in the optimization subset, achieving a PPV of 98.5%. On the internal validation subset of 941, the model identified 71 (7.6%) reports containing a lung mass, with sensitivity of 89.3%, specificity of 99.5%, PPV of 94.4%, and NPV of 99.1%. According to the indications in the 71 identified reports, 26 (36.7%) were existing patients being managed for cancer, 28 (39.4%) were existing non-cancerous lung mass patients, 13 (18.3%) were new lung mass patients, and 4 (5.6%) were false positives. CONCLUSION The NLP model was able to identify lung mass patients from CT radiology reports with high precision and sensitivity. By deploying this platform across different care settings, we propose a way to increase the number of appropriate and timely lung mass referrals, potentially improving clinical guideline adherence and patient outcomes.

The use of endobronchial ultrasound (EBUS) is standard practice for lung cancer diagnosis and staging. Next generation sequencing (NGS) for detection of genetic alterations is recommended in advanced, non-squamous, non-small-cell lung cancer (NSCLC). Existing protocols for NGS testing are minimal and reported yields vary. This study aimed to determine the yield of EBUS samples obtained for NGS using a sampling protocol at our institution and assess predictive factors to form collection protocols. We reviewed EBUS bronchoscopies from 2016 to 2021 with non-squamous NSCLC diagnoses. For target lesions suspected to be malignant, the sampling protocol was: (a) two slides for on-site evaluation, (b) three to five fine needle aspirations rinsed into saline for immunohistochemical staining and in-house molecular markers, and (c) additional three to five rinses for NGS. Sufficiency for NGS processing was determined by the pathology department. Two hundred and seventy-eight non-squamous NSCLC samples were obtained by EBUS (205 adenocarcinoma; 73 not otherwise specified). EBUS was performed under general anesthesia in 75.5

PURPOSE:Lung nodules, whether discovered through lung cancer screening or incidentally on CT (computed tomography), present a huge cost burden on health care.Given the prevalence of pulmonary nodules, a variety of diagnostic decision tools are used to risk stratify lung nodules.These tools include the Mayo Clinic calculator, Brock calculator, artificial intelligence software (AI), blood-based biomarkers, bronchial epithelial biomarkers, and nasal epithelial cell biomarkers.We have created a lung nodule registry to examine the effect of these varying modalities on clinical outcomes.This abstract focuses on time to diagnosis between the various modalities. METHODS:We reviewed a nodule registry of over 600 nodules and isolated those with calculated Mayo scores with and without ancillary lung nodule classifiers.We analyzed this group with regards to lung cancer prediction and time to final diagnosis.Malignant diagnoses were based on positive tissue sampling (biopsy or surgery) or those instances when a lesion was presumed positive based on risk factors but could not safely undergo definitive surgical diagnosis.Benign diagnoses were made if a lesion was resolved on imaging, culture positive, determined inflammatory by clinical presentation, or was stable after 24 months of imaging follow-up. RESULTS:Our preliminary analysis includes 136 patients.With regards to time to final diagnosis (in months) for benign nodules we observed the following: Mayo 9mo (n¼8), AI 5mo (n¼9), Bronchial biomarker 11mo (n¼47), Blood-based 7mo (n¼2).With regards to time to final diagnosis (in months) for malignant nodules we observed the following: Mayo 6mo (n¼17), AI 4mo (n¼7), Bronchial biomarker 8mo (n¼29), .With regards to time to final diagnosis (in months) for biopsy negative but presumed malignant nodules we observed the following: Mayo 6mo (n¼3), AI 14mo (n¼1), Bronchial biomarker 12mo (n¼10), Blood-based 8mo (n¼2).Due to our initial sample size being low, statistical analysis could not be ascertained, but will be performed on our final analysis as further data collection is underway. CONCLUSIONS:Comparing the pulmonary nodule risk prediction modalities there appears to be a longer time to diagnosis with bronchial biomarker classifier compared to other modalities, but final statistical analysis with the existing larger sample size needs to be completed.CLINICAL IMPLICATIONS: Given the various pulmonary nodule risk prediction modalities, choosing an accurate, timely, and cost-effective modality is important to clinical outcomes.Other clinical outcomes of interest include the number of imaging studies and procedures patients undergo and what effect these different modalities have on these numbers.

Artificial intelligence (AI) radiomics-based tools demonstrate promise for indeterminate pulmonary nodule (PN) malignancy risk stratification.1 We performed a secondary analysis of a previous multi-reader, multi-case study2 to evaluate the effect of an AI tool on clinicians' PN management decisions. The details of this study have been previously described.2 Briefly, 12 readers (6 radiologists, 6 pulmonologists) independently evaluated 300 indeterminate PN cases using solely axial CT chest scan imaging data. PNs were 5–30 mm in maximal diameter, and 50% were malignant. The AI tool assessed was the Lung Cancer Prediction Convolutional Neural Network (Virtual Nodule Clinic, version 2.0.0; Optellum Ltd, Oxford, UK).3, 4 This tool calculates a Lung Cancer Prediction (LCP) score describing PN malignancy risk on a decile scale from 1 to 10 assuming a malignancy prevalence of 30%. For each case, each reader independently provided estimates of malignancy risk (0%–100%) and management decision (no follow-up, ≥6-month CT follow-up, 6-week to 6-month CT follow-up, immediate imaging follow-up, non-surgical biopsy, or surgical resection) before and after being shown the LCP score. We defined appropriate management of malignant PNs as non-surgical biopsy and surgical resection. For benign PNs, no follow-up or imaging follow-up were deemed appropriate. We classified immediate imaging as appropriate management for all PNs. The median LCP score for malignant PNs was 9 (IQR, 8–10) and 5 (IQR, 2–7) for benign PNs (p < 0.001). Among malignant PNs, the average reader malignancy risk estimate was 60.2% (SD, 31.7%) without the AI tool compared to 69.0% (SD, 28.6%) with it (p < 0.001). Among benign PNs, the average reader malignancy risk estimate was 23.4% (SD, 28.1%) without the AI tool compared to 21.0% (SD, 26.9%) with it (p = 0.01). The distributions of management decisions are displayed in Figure 1. Overall, the proportion of cases with appropriate management decisions increased from 79.5% (SD, 5.7%) to 84.1% (SD, 6.6%) with AI (p = 0.008). Among malignant PNs, on average readers selected immediate imaging, biopsy, or surgical resection in 71.9% (SD, 14.0%) of cases without use of AI compared to 81.4% (SD, 13.7%) with the AI tool (p < 0.001). Among benign PNs, on average readers selected no action, short-term, long-term, or immediate follow-up imaging in 87.2% (SD, 10.4%) of cases without and 88.7% (SD, 11.1%) with the AI tool, respectively (p = 0.19). We found that use of an AI tool was associated with an increase of the average proportion of cases with appropriate management decisions from 79.5% to 84.1%. This was largely driven by a 10 percentage point increase in malignant PNs appropriately managed with immediate imaging or tissue sampling. On the other hand, we did not observe a statistically significant difference in the management of benign PNs with use of the AI tool. Taken together, these results suggest that the previously demonstrated improvement in diagnostic accuracy with use of an AI tool may translate into meaningful changes in clinical management decisions and promote earlier diagnostic evaluation of malignant PNs, which may ultimately lead to increased timeliness of appropriate clinical treatment for thoracic malignancies. Roger Kim: Conceptualization (lead); formal analysis (lead); methodology (equal); visualization (lead); writing – original draft (lead); writing – review and editing (lead). Jason L. Oke: Formal analysis (supporting); methodology (supporting); writing – review and editing (equal). Travis L. Dotson: Data curation (equal); writing – review and editing (equal). Christina Bellinger: Data curation (equal); writing – review and editing (equal). Anil Vachani: Conceptualization (equal); data curation (equal); formal analysis (supporting); methodology (supporting); supervision (lead); writing – original draft (supporting); writing – review and editing (equal). Research funding: Roger Y. Kim was supported by the National Cancer Institute of the National Institutes of Health (Award Number 5UM1CA221939). Jason L. Oke was part-funded by the NIHR Oxford Biomedical Research Centre, Oxford University Hospitals NHS Foundation Trust. This study was funded by Optellum Ltd. Roger Y. Kim reports research funding from Siemens outside of the submitted work. Anil Vachani reports research funding from MagArray, Inc., Broncus Medical, and PreCyte, Inc. and a consulting role with Novocure and Johnson & Johnson, outside of the submitted work. The remaining authors reported no relevant conflicts of interest. The use of deidentified imaging studies complied with Health Insurance Portability and Accountability Act guidelines, and the need for informed consent was waived by local institutional review boards.

The Percepta Genomic Sequencing Classifier (GSC) was developed to up-classify as well as down-classify the risk of malignancy for lung lesions when bronchoscopy is non-diagnostic. We evaluated the performance of Percepta GSC in risk re-classification of indeterminate lung lesions. This multicenter study included individuals who currently or formerly smoked undergoing bronchoscopy for suspected lung cancer from the AEGIS I/ II cohorts and the Percepta Registry. The classifier was measured in normal-appearing bronchial epithelium from bronchial brushings. The sensitivity, specificity, and predictive values were calculated using predefined thresholds. The ability of the classifier to decrease unnecessary invasive procedures was estimated. A set of 412 patients were included in the validation (prevalence of malignancy was 39.6%). Overall, 29% of intermediate-risk lung lesions were down-classified to low-risk with a 91.0% negative predictive value (NPV) and 12.2% of intermediate-risk lesions were up-classified to high-risk with a 65.4% positive predictive value (PPV). In addition, 54.5% of low-risk lesions were down-classified to very low risk with >99% NPV and 27.3% of high-risk lesions were up-classified to very high risk with a 91.5% PPV. If the classifier results were used in nodule management, 50% of patients with benign lesions and 29% of patients with malignant lesions undergoing additional invasive procedures could have avoided these procedures. The Percepta GSC is highly accurate as both a rule-out and rule-in test. This high accuracy of risk re-classification may lead to improved management of lung lesions.

Background Limited data are available regarding whether computer-aided diagnosis (CAD) improves assessment of malignancy risk in indeterminate pulmonary nodules (IPNs). Purpose To evaluate the effect of an artificial intelligence-based CAD tool on clinician IPN diagnostic performance and agreement for both malignancy risk categories and management recommendations. Materials and Methods This was a retrospective multireader multicase study performed in June and July 2020 on chest CT studies of IPNs. Readers used only CT imaging data and provided an estimate of malignancy risk and a management recommendation for each case without and with CAD. The effect of CAD on average reader diagnostic performance was assessed using the Obuchowski-Rockette and Dorfman-Berbaum-Metz method to calculate estimates of area under the receiver operating characteristic curve (AUC), sensitivity, and specificity. Multirater Fleiss κ statistics were used to measure interobserver agreement for malignancy risk and management recommendations. Results A total of 300 chest CT scans of IPNs with maximal diameters of 5-30 mm (50.0% malignant) were reviewed by 12 readers (six radiologists, six pulmonologists) (patient median age, 65 years; IQR, 59-71 years; 164 [55%] men). Readers' average AUC improved from 0.82 to 0.89 with CAD (P < .001). At malignancy risk thresholds of 5% and 65%, use of CAD improved average sensitivity from 94.1% to 97.9% (P = .01) and from 52.6% to 63.1% (P < .001), respectively. Average reader specificity improved from 37.4% to 42.3% (P = .03) and from 87.3% to 89.9% (P = .05), respectively. Reader interobserver agreement improved with CAD for both the less than 5% (Fleiss κ, 0.50 vs 0.71; P < .001) and more than 65% (Fleiss κ, 0.54 vs 0.71; P < .001) malignancy risk categories. Overall reader interobserver agreement for management recommendation categories (no action, CT surveillance, diagnostic procedure) also improved with CAD (Fleiss κ, 0.44 vs 0.52; P = .001). Conclusion Use of computer-aided diagnosis improved estimation of indeterminate pulmonary nodule malignancy risk on chest CT scans and improved interobserver agreement for both risk stratification and management recommendations. © RSNA, 2022 Online supplemental material is available for this article. See also the editorial by Yanagawa in this issue.

Rationale: Electromagnetic navigational bronchoscopy (ENB) is an important, minimally invasive diagnostic tool for malignant and benign peripheral lung lesions, offering lower complication risks than transthoracic needle aspirations. As a relatively new technology, the best sampling modality and lesion characteristics for ENB has yet to be determined. We evaluated the sensitivity and diagnostic yield of different sampling modalities (needle aspiration, brush biopsy, transbronchial forceps biopsies) and radiographical lesion characteristics by Tsuboi classification. We also evaluated the difference in yield and sensitivity with the addition of radial probe EBUS to augment ENB. Methods: We completed a retrospective chart review of all patients that had ENB performed at our institution since its implementation in 2011. We reviewed the lesion size, location, Tsuboi classification, cytology, pathology results and analyzed biopsy specimen tool types. Results: We included a total of 248 patients who had ENB performed between 2011 and 2018. Average age was 67 years and 50% female. A total of 270 lesions were targeted with a mean size of 24 +/- 12 mm. Sensitivity for malignancy was 59.2% with a diagnostic yield of 72.3%. Sensitivity and diagnostic accuracy trended higher with combined sampling modalities (brush and transbronchial needle aspiration and forcep biopsy). Lesions with type I and type II Tsuboi classification of bronchus sign had higher sensitivity compared to type III classification (67.9% [n = 101 type I], 64.6% [n = 65 type II], 37.9% [n = 36 type III]), p = 0.01 and p = 0.04. Conclusion: For navigation bronchoscopy, sensitivity is higher in bronchus sign lesions that end directly into lesion (Tsuboi type I) and travel through malignant lesions (Tsuboi type II) compared to tangentially circumventing the lesion (Tsuboi type III).

Improving stratification of patients with indeterminate pulmonary nodules (IPNs) can lead both to earlier diagnosis of lung cancer and to reduced scanning and reduced intervention in cases of benign disease. AI-based decision support software has been shown to outperform conventional risk models at classifying IPNs as low or high risk, but its performance in addition to clinician assessment has yet to be investigated. We report the results of a Multiple-Reader Multiple-Case reader evaluation comparing reader performance for both radiologists and pulmonologists on an IPN risk stratification task with and without AI assistance from the previously-published Lung Cancer Prediction Convolutional Neural Network (LCP-CNN).

The Comprehensive, Computable NanoString Diagnostic gene panel (C2Dx) is a promising solution to address the need for a molecular pathological research and diagnostic tool for precision oncology utilizing small volume tumor specimens. We translate subtyping-related gene expression patterns of Non-Small Cell Lung Cancer (NSCLC) derived from public transcriptomic data which establish a highly robust and accurate subtyping system. The C2Dx demonstrates supreme performance on the NanoString platform using microgram-level FNA samples and has excellent portability to frozen tissues and RNA-Seq transcriptomic data. This workflow shows great potential for research and the clinical practice of cancer molecular diagnosis.

Chemo-immunotherapy is central to the treatment of small cell lung cancer (SCLC). Despite modest progress made with the addition of immunotherapy, current cytotoxic regimens display minimal survival benefit and new treatments are needed. Thymidylate synthase (TS) is a well-validated anti-cancer drug target, but conventional TS inhibitors display limited clinical efficacy in refractory or recurrent SCLC. We performed RNA-Seq analysis to identify gene expression changes in SCLC biopsy samples to provide mechanistic insight into the potential utility of targeting pyrimidine biosynthesis to treat SCLC. We identified systematic dysregulation of pyrimidine biosynthesis, including elevated TYMS expression that likely contributes to the lack of efficacy for current TS inhibitors in SCLC. We also identified E2F1-3 upregulation in SCLC as a potential driver of TYMS expression that may contribute to tumor aggressiveness. To test if TS inhibition could be a viable strategy for SCLC treatment, we developed patient-derived organoids (PDOs) from human SCLC biopsy samples and used these to evaluate both conventional fluoropyrimidine drugs (e.g., 5-fluorouracil), platinum-based drugs, and CF10, a novel fluoropyrimidine polymer with enhanced TS inhibition activity. PDOs were relatively resistant to 5-FU and while moderately sensitive to the front-line agent cisplatin, were relatively more sensitive to CF10. Our studies demonstrate dysregulated pyrimidine biosynthesis contributes to drug resistance in SCLC and indicate that a novel approach to target these pathways may improve outcomes.

Purpose/Objectives: We aimed to assess the predictive value of a lung cancer gene panel for the development of brain metastases. Materials/Methods: Between 2011 and 2015, 102 patients with lung cancer were prospectively enrolled in a clinical trial in which a diagnostic fine-needle aspirate was obtained. Gene expression was conducted on all samples that rendered a diagnosis of non-small cell lung cancer (NSCLC). Subsequent retrospective analysis of brain metastases-related outcomes was performed by reviewing patient electronic medical records. A competing risk multivariable regression was performed to estimate the adjusted hazard ratio for the development of brain metastases and non-brain metastases from NSCLC. Results: A total of 49 of 102 patients had died by the last follow-up. Median time of follow-up was 13 months (range 0.23–67 months). A total of 17 patients developed brain metastases. Median survival time after diagnosis of brain metastases was 3.58 months (95% confidence interval (CI) 2.17, not available). A total of 30 patients developed metastases without any evidence of brain metastases until the time of death or last follow-up. Competing risk analysis identified three genes that were downregulated differentially in the patients with brain metastases versus non-brain metastatic disease: CD37 (0.017), cystatin A (0.022), and IL-23A (0.027). Other factors associated with brain metastases include: stage T ( P ⩽ 8.3e-6) and stage N ( P= 6.8e-4). Conclusions: We have identified three genes, CD37, cystatin A, and IL-23A, for which downregulation of gene expression was associated with a greater propensity for developing brain metastases. Validation of these biomarkers could have implications on surveillance patterns in patients with brain metastases from NSCLC.

SESSION TITLE: Late Breaking Abstracts SESSION TYPE: Original Investigations PRESENTED ON: 10/23/2019 10:45 AM - 11:45 AM PURPOSE: Guidelines recommend patients with intermediate-risk nodules undergo bronchoscopic workup, yet the diagnostic yield is only around 50%, underscoring the need for additional tools to inform patient management. We previously established the clinical accuracy of Percepta Bronchial Genomic Classifier, a molecular test that utilizes a brushing of bronchial epithelium to improve nodule management in smokers. We have since developed the second-generation Percepta Genomic Sequencing Classifier (GSC), which utilizes 1232 gene transcripts from whole-transcriptome RNA sequencing, as well as clinical factors, to achieve improved performance. Here we report the clinical validation and demonstrate improvement for nodule diagnosis. METHODS: We utilized over 1600 patient samples to train the new Percepta GSC, which is an ensemble of four machine-learning models that incorporate genomic and clinical features as well as their interactions. We established diagnostic accuracy on an independent validation set consisting of 412 patient samples from three cohorts: AEGIS I and II (246 patients) and the Percepta Registry (166 patients)—all with nondiagnostic lung nodules from bronchoscopy. We measured positive predictive value (PPV) and negative predictive value (NPV) of Percepta GSC results in physician-reported pre-test risk categories of low, intermediate or high risk. In AEGIS I/II cohorts, where patients with diagnostic bronchoscopy results are available, we calculated the sensitivities of bronchoscopy and Percepta GSC alone, as well as in combination. RESULTS: In the validation set, 80 patients (19%) had a low pre-test risk (cancer prevalence 5.0%), 188 (35%) had an intermediate pre-test risk (cancer prevalence 28.2%), and 144 (46%) had a high pre-test risk (cancer prevalence 73.6%). Percepta GSC down-classified patients with low pre-test risk with >99% NPV and intermediate pre-test risk with a 91.0% NPV. In addition, the classifier up-classified patients with intermediate pre-test risk with a 65.4% PPV and high pre-test risk with a 91.5% PPV. In total, 41.6% of intermediate pre-test risk patients were either down- or up-classified by the test. In the AEGIS cohorts, the sensitivity among low and intermediate pre-test risk patients was: 40.9% (bronchoscopy alone), 92.3% (Percepta GSC alone), and 95.5% (combined). CONCLUSIONS: Percepta GSC is clinically validated to accurately up-classify and down-classify the probability of malignancy for a substantial portion of lung nodule patients with nondiagnostic bronchoscopy. Percepta GSC significantly improves the sensitivity of bronchoscopy overall and demonstrates a combined 95%+ sensitivity for low/intermediate pre-test risk groups. CLINICAL IMPLICATIONS: Percepta GSC compliments bronchoscopy and when used together can enable improved management of pulmonary nodules. DISCLOSURES: No relevant relationships by Christina Bellinger, source=Web Response No relevant relationships by Michael Bernstein, source=Web Response Employee relationship with Veracyte, Inc Please note: >$100000 Added 06/22/2019 by Sangeeta Bhorade, source=Web Response, value=Salary Employee relationship with Veracyte Inc Please note: >$100000 Added 03/21/2019 by Yoonha Choi, source=Web Response, value=Salary Removed 03/21/2019 by Yoonha Choi, source=Web Response Employee relationship with Veracyte Inc Please note: >$100000 Added 03/21/2019 by Yoonha Choi, source=Web Response, value=Salary No relevant relationships by Travis Dotson, source=Web Response Consultant relationship with AstraZeneca Please note: $5001 - $20000 Added 11/29/2018 by David Feller-Kopman, source=Web Response, value=Consulting fee Consultant relationship with Veracyte Please note: $5001 - $20000 Added 11/29/2018 by David Feller-Kopman, source=Web Response, value=Consulting fee Consultant relationship with Veran Medical Please note: $5001 - $20000 Added 11/29/2018 by David Feller-Kopman, source=Web Response, value=Consulting fee Employee relationship with Veracyte Inc Please note: >$100000 Added 03/15/2019 by Jing Huang, source=Web Response, value=Salary no disclosure on file for Giulia Kennedy; Consultant relationship with Veracyte Please note: $5001 - $20000 Added 03/18/2019 by Hans Lee, source=Web Response, value=Consulting fee Consultant relationship with Veran medical Please note: $5001 - $20000 Added 03/18/2019 by Hans Lee, source=Web Response, value=Consulting fee Employee relationship with Veracyte Please note: >$100000 Added 03/15/2019 by Lori Lofaro, source=Web Response, value=Salary Advisory Committee Member relationship with Grail Please note: $1-$1000 Added 06/22/2019 by Peter Mazzone, source=Web Response, value=Consulting fee Consultant relationship with SEER Please note: $1-$1000 Added 06/22/2019 by Peter Mazzone, source=Web Response, value=Consulting fee Research support to my institution relationship with Veracyte Please note: $5001 - $20000 Added 06/22/2019 by Peter Mazzone, source=Web Response, value=Grant/Research Support Research support to my institution relationship with Oncocyte Please note: $5001 - $20000 Added 06/22/2019 by Peter Mazzone, source=Web Response, value=Grant/Research Support Research support to my institution relationship with Exact Sciences Please note: $1001 - $5000 Added 06/22/2019 by Peter Mazzone, source=Web Response, value=Grant/Research Support Employee relationship with Veracyte Please note: >$100000 Added 03/18/2019 by Daniel Pankratz, source=Web Response, value=Salary No relevant relationships by Momen Wahidi, source=Web Response Employee relationship with Veracyte Please note: >$100000 Added 03/18/2019 by Patric Walsh, source=Web Response, value=Salary

SESSION TITLE: Advances in the Diagnosis of Lung Cancer SESSION TYPE: Original Investigations PRESENTED ON: 10/09/2018 02:30 PM - 03:30 PM PURPOSE: Bronchoscopy is frequently used for evaluation of pulmonary lesions, but its sensitivity for detecting lung cancer can be limited. A bronchial genomic classifier (Percepta) has been validated as a complement to lung cancer diagnostic bronchoscopy to improve its sensitivity and negative predictive value. When bronchoscopy is inconclusive, Percepta can identify patients who can be considered for CT surveillance instead of undergoing another invasive diagnostic procedure. We report here on the clinical utility of Percepta among patients enrolled in the Percepta Registry at up to 12 months post bronchoscopy. METHODS: Patients were prospectively enrolled at 40 medical centers when Percepta was ordered due to an inconclusive bronchoscopy. The classifier sample was obtained by brushing the right mainstem bronchus during bronchoscopy, regardless of nodule size or location. Pre- and post-classifier clinical management recommendations were recorded and follow-up clinical, procedure, and imaging data were collected. RESULTS: 399 patients had an inconclusive bronchoscopy and were within indication (no prior cancer and current or former smoker). The majority of lesions were <30mm (77%), peripherally located (72%), solid (73%), and upper lobe (55%). Advanced bronchoscopic technologies were used in 68% of cases and PET was used prior to bronchoscopy in 37% of patients. This interim analysis focuses on the 289 patients (72%) with intermediate (245) or low (44) pre-test risk of malignancy. 32% of intermediate pre-test risk patients were down classified by Percepta to low risk, and 52% of low pre-test risk patients were down classified by Percepta to very low risk. These results are consistent with the results from the AEGIS 1 and 2 studies (Silvestri et al, NEJM 2015): 38% and 54% down classification, p = 0.85 and p =0.30 respectively. Among patients where risk of malignancy was down-classified by Percepta, physicians significantly reduced invasive procedure recommendations from 41% to 18% in the intermediate pre-test risk and 9% to 0% in the low pre-test risk group. This results in an overall procedure reduction of 34% to 14% (relative reduction of 59%, p=0.0005). 83% of those who were down-classified remained procedure free at 12 months follow up. CONCLUSIONS: We observed a significant reduction in additional invasive procedures compared to the pre-test management plan for patients who were down classified by Percepta after an inconclusive bronchoscopy. This reduction in procedures has been durable over 12 months. Additional data will help further determine the ultimate clinical utility of the test. CLINICAL IMPLICATIONS: A bronchial genomic classifier can reduce the number of unnecessary invasive procedures that are performed following an inconclusive bronchoscopy for suspect lung cancer. DISCLOSURES: No relevant relationships by Sadia Benzaquen, source=Web Response No relevant relationships by Michael Bernstein, source=Web Response Consultant relationship with Medtronic ILS Please note: $5001 - $20000 Added 02/25/2018 by Krish Bhadra, source=Web Response, value=Consulting fee Advisory Committee Member relationship with Biodesix Please note: $5001 - $20000 Added 02/25/2018 by Krish Bhadra, source=Web Response, value=Consulting fee Consultant relationship with Boston Scientific Please note: $1001 - $5000 Added 02/25/2018 by Krish Bhadra, source=Web Response, value=Consulting fee Consultant relationship with Merit Endotek Please note: $1001 - $5000 Added 02/25/2018 by Krish Bhadra, source=Web Response, value=Consulting fee Consultant relationship with BodyVision Please note: $1001 - $5000 Added 02/25/2018 by Krish Bhadra, source=Web Response, value=Consulting fee Consultant Consultant relationship with Auris Surgical Robotics Please note: $1001 - $5000 Added 03/03/2018 by Krish Bhadra, source=Web Response, value=Consulting fee No relevant relationships by Travis Dotson, source=Web Response No relevant relationships by Mark Esterle, source=Web Response researcher relationship with veracyte Please note: $1-$1000 Added 03/03/2018 by Joshiah Gordon, source=Web Response, value=Grant/Research Support Employee relationship with Veracyte, Inc. Please note: $20001 - $100000 Added 03/05/2018 by Bailey Griscom, source=Web Response, value=Salary Speaker/Speaker's Bureau relationship with Boston Scientific Please note: $20001 - $100000 Added 03/01/2018 by D Hogarth, source=Web Response, value=Honoraria Speaker/Speaker's Bureau relationship with Shire Please note: $5001 - $20000 Added 03/01/2018 by D Hogarth, source=Web Response, value=Honoraria Consultant relationship with Auris Please note: $1001 - $5000 Added 03/01/2018 by D Hogarth, source=Web Response, value=Consulting fee Consultant relationship with Auris Please note: $20001 - $100000 Added 03/01/2018 by D Hogarth, source=Web Response, value=Ownership interest Unrestricted Education Grant relationship with Boston Scientific Please note: $20001 - $100000 Added 03/01/2018 by D Hogarth, source=Web Response, value=Unrestricted Education Grant Consultant relationship with BronchiSense Please note: $20001 - $100000 Added 03/01/2018 by D Hogarth, source=Web Response, value=Ownership interest Consultant relationship with LX Medical Please note: $5001 - $20000 Added 03/01/2018 by D Hogarth, source=Web Response, value=Ownership interest Consultant relationship with Biodesix Please note: $5001 - $20000 Added 03/02/2018 by D Hogarth, source=Web Response, value=Consulting fee Consultant relationship with Body Vision Please note: $20001 - $100000 Added 03/03/2018 by D Hogarth, source=Web Response, value=Ownership interest Consultant relationship with Medtronic Please note: $5001 - $20000 Added 03/03/2018 by D Hogarth, source=Web Response, value=Consulting fee Consultant relationship with Auris Please note: $5001 - $20000 Added 03/03/2018 by D Hogarth, source=Web Response, value=Consulting fee Consultant relationship with Auris Please note: >$100000 Added 03/03/2018 by D Hogarth, source=Web Response, value=Ownership interest Consultant relationship with Preora Please note: $20001 - $100000 Added 03/03/2018 by D Hogarth, source=Web Response, value=Ownership interest Speaker/Speaker's Bureau relationship with Grifols Please note: $20001 - $100000 Added 03/03/2018 by D Hogarth, source=Web Response, value=Honoraria Consultant relationship with Heritage Biologics Please note: $20001 - $100000 Added 03/03/2018 by D Hogarth, source=Web Response, value=Consulting fee Consultant relationship with Boston Scientific Please note: $20001 - $100000 Added 03/03/2018 by D Hogarth, source=Web Response, value=Consulting fee Consultant relationship with Gala Therapeutics Please note: $5001 - $20000 Added 03/03/2018 by D Hogarth, source=Web Response, value=Consulting fee Consultant relationship with Matrix Analytics Please note: $20001 - $100000 Added 03/03/2018 by D Hogarth, source=Web Response, value=Ownership interest Consultant relationship with OncoCyte Please note: $5001 - $20000 Added 03/03/2018 by D Hogarth, source=Web Response, value=Consulting fee Owner/Founder relationship with Medical Opinion Systems Please note: $20001 - $100000 Added 03/03/2018 by D Hogarth, source=Web Response, value=Ownership interest Consultant relationship with Neurotronic Please note: $1001 - $5000 Added 03/03/2018 by D Hogarth, source=Web Response, value=Consulting fee Speaker/Speaker's Bureau relationship with Veracyte Please note: $5001 - $20000 Added 03/03/2018 by D Hogarth, source=Web Response, value=Honoraria Employee relationship with Veracyte Inc Please note: >$100000 Added 03/04/2018 by Jing Huang, source=Web Response, value=Salary Employee relationship with Veracyte Please note: >$100000 Added 03/09/2018 by Marla Johnson, source=Web Response, value=Salary Removed 03/09/2018 by Marla Johnson, source=Web Response Employee relationship with Veracyte Please note: $20001 - $100000 Added 03/09/2018 by Marla Johnson, source=Web Response, value=Salary Employee relationship with Veracyte Please note: >$100000 Added 03/27/2018 by Giulia Kennedy, source=Admin input, value=Salary Consultant relationship with Veracyte Please note: $5001 - $20000 Added 03/09/2018 by Hans Lee, source=Web Response, value=Consulting fee Consultant relationship with Veran Medical Please note: $20001 - $100000 Added 03/09/2018 by Hans Lee, source=Web Response, value=Grant/Research Support Employee relationship with Veracyte Please note: >$100000 Added 03/03/2018 by Lori Lofaro, source=Web Response, value=Salary Advisory Committee Member relationship with Exact Sciences Please note: $1001 - $5000 Added 03/05/2018 by Peter Mazzone, source=Web Response, value=Consulting fee Research support relationship with Veracyte Please note: $5001 - $20000 Added 03/05/2018 by Peter Mazzone, source=Web Response, value=Grant/Research Support Research support relationship with Oncocyte Please note: $5001 - $20000 Added 03/05/2018 by Peter Mazzone, source=Web Response, value=Grant/Research Support Consultant relationship with Veracyte Please note: $20001 - $100000 Added 04/30/2018 by Avrum Spira, source=Admin input, value=Salary Consultant relationship with Janssen Pharmaceuticals Please note: $20001 - $100000 Added 04/30/2018 by Avrum Spira, source=Admin input, value=Consulting fee No relevant relationships by Patrick Whitten, source=Web Response

Objectives Targeted therapies for non-small-cell lung cancers (NSCLCs) are based on the presence of driver mutations such as epidermal growth factor receptor (EGFR) and the echinoderm microtubule-associated protein-like 4-anaplastic lymphoma kinase (EML4-ALK) translocation. Endobronchial ultrasound-guided-transbronchial needle aspiration (EBUS-TBNA) is a first-line modality for diagnosing and staging NSCLC. A quality improvement protocol maximizing tissue acquisition for molecular analysis has not been previously described. Methods We instituted a standardized protocol designed from a multidisciplinary meeting of the pulmonology, oncology, and pathology departments for the acquisition and on-site processing of samples obtained through EBUS-TBNA to improve the yield for genetic analysis of EGFR and ALK testing. Results Preprotocol there were 50 NSCLCs (29 adenocarcinomas) and postprotocol there were 109 NSCLCs (52 adenocarcinomas). A statistically significant increase in yield for molecular analysis was seen in both EGFR (36% preprotocol and 80% postprotocol, P < 0.01) and ALK (41% preprotocol and 80% postprotocol, P < 0.01). There was no difference in complications preprotocol and postprotocol. Conclusions Implementation of a standardized protocol with EBUS-TBNA was associated with an increase in adequacy for molecular genetic analysis in NSCLC.

INTRODUCTION:RNA isolation from tumor tissue is used for biomarker analyses and validation. Limited diagnostic material from small volume biopsies combined with an increasing demand for standard histologic, molecular characterization, and next generation sequencing applications often leads to limited material for research. We sought to evaluate small volume sampling of lung cancer tissue collected from a single needle pass during a diagnostic procedure and determine if it can provide RNA of acceptable quantity and quality.METHODS:We enrolled 140 patients with probable primary bronchogenic carcinoma and collected RNA from a dedicated FNA aspiration. Total RNA (ηg), RNA integrity number (RIN), and %Mass in base pairs were evaluated from each patient sample. A customized nanoString nCounter® 95-gene panel was used to profile the expression patterns of feature NSCLC genes. We compared gene expression patterns that distinguish lung adenocarcinoma (LUAD) and squamous cell carcinoma (LUSC) in our cohort with a corresponding Cancer Genome Atlas (TCGA) NSCLC datasets.RESULTS:Of the 149 patients consented. RNA-extraction was performed in 101 eligible patients. A satisfactory total RNA mass and RIN was quantified for all samples with a similar distribution among cellular subtypes. Mean %-Mass over 300 base pairs was noted for all specimens and 96% of samples met criteria to perform genetic evaluation with our commercialized gene expression assay. The FNA-derived transcriptomic results showed excellent consistency with the TCGA counterparts, and the differential expression pattern of LUAD vs LUSC subtypes were highly similar.DISCUSSION:In this study, RNA retrieval from a single-pass FNA regardless of procedural approach showed equivalence and suitability for gene expression assessments. RNA extraction from small volume samples has the potential to provide valuable material for genetic profiling.

The current complexity of non–small cell lung lies in multiple histologies, targetable mutations, and gene expression, all of which are necessary information for structuring treatment and formulating prognosis. Obtaining this information from small-volume biopsies can present unique challenges, and researchers are exploring RNA analysis to elicit this. RNA analysis is already being used in other malignancies such as those of the breast and thyroid. We performed a review of the literature, demonstrating that sufficient material can be obtained through RNA specimens collected during endobronchial ultrasound-guided transbronchial needle aspiration, and that this has potential utility in diagnostic and prognostic roles. Myth: Transbronchial needle aspiration will obtain insufficient sample for RNA analysis.

Background: Endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) presents a minimally invasive way to evaluate abnormal mediastinal and hilar adenopathy. Although EBUS has been established as an effective modality to diagnose lung cancer, its sensitivity for the diagnosis of lymphoma has been demonstrated to be lower. Because of these lower yields uncertainty persists about the ability of EBUS-TBNA to reliably diagnose lymphoma and questions remain regarding the utility of EBUS-TBNA as a first-line biopsy modality for patients suspected of having lymphoma. Methods: We conducted a review of our database (n=806 EBUS-TBNAs) for patients undergoing EBUS-TBNA for mediastinal and/or hilar lymphadenopathy over an 8-year span to identify patients diagnosed with lymphoma. Results: Twenty patients (2.3%) who underwent EBUS-TBNA were ultimately diagnosed with lymphoma. In total, 17 of the 20 patients with lymphoma obtained a diagnosis using EBUS-TBNA. The overall sensitivity of EBUS-TBNA for lymphoma was 85%. The sensitivity for de novo diagnosis was 78% (7/9), and sensitivity for recurrence was 91% (10/11). All patients who achieved a diagnosis by EBUS-TBNA could be adequately subtyped, allowing treatment recommendations. Conclusion: Although the sensitivity of EBUS-TBNA for the diagnosis of lymphoma did not reach values of published data for non–small cell lung cancer, EBUS-TBNA can be considered as a first-line diagnostic tool for patients with mediastinal and/or hilar lymphadenopathy suspected to be lymphoma. Because of the inherent limitations in small volume needle biopsies it is essential that negative samples obtained in the setting of high clinical suspicion warrant further evaluation.