E-Nose Technology Detects Early-Stage Lung Cancer with High Reliability in MSK-Exclusive Trial

E-Nose Technology Detects Early-Stage Lung Cancer with High Reliability in MSK-Exclusive Trial

Electronic nose (e-nose) technology identified early-stage lung cancer with high reliability in a prospective observational clinical trial conducted at Memorial Sloan Kettering Cancer Center (MSK).

E-nose diagnostic predictions agreed with histopathological results in 86% of 100 patients with clinical stage 1 lung cancer, according to the study published recently in the Journal of Thoracic Oncology(1)

In a previous MSK Clinical Updates and Insights article on e-nose technology for diagnosing lung cancer, we described how the technology works, reported on first-generation evidence in a cancer screening population, and discussed the scope of research underway at MSK to determine the diagnostic accuracy of next-generation devices.

Here, we share results from the first phase 2 trial investigating the next-generation BIONOTE (BIOsensor-based multisensorial system for mimicking NOse, Tongue, and Eyes) e-nose platform (NCT04734145) in a real-world clinical setting among patients diagnosed with early-stage lung cancer.

“These robust results demonstrate e-nose technology could be a reliable complement to existing diagnostic methods for early-stage lung cancer,” said principal investigator MSK thoracic surgeon Gaetano Rocco, MD. “We anticipate this non-invasive, low-cost, portable, and highly reliable diagnostic method will revolutionize the diagnosis and clinical management of early-stage lung cancer, making it accessible to more patients, especially those resistant to CT scans or for whom a biopsy is infeasible. Our immediate next step is applying for funding to support our efforts to advance research and development.”

The Urgent Need for Improved Diagnostic Methods for Early-Stage Lung Cancer

Standard diagnostics for lung cancer include computed tomography (CT) and plasma circulating tumor DNA (ctDNA). CT has high specificity but low sensitivity in patient populations with a low disease prevalence, with an overall predictive value of only 0.25. (2) (3)

Previous lung screening studies have reported varying sensitivity and specificity rates for CT due to differences in disease prevalence, patient population, study design, techniques, and geographic area.  (4)

Currently, plasma biomarkers associated with lung cancer are unreliable for detecting early-stage disease. In fact, detection rates have ranged from 48% for stage 1 to 73% for stage 3 lung cancer, with false-negative rates as high as 78% in patients with stage 1 disease.  (5) (6) (7)

How E-Nose Technology Works

Human breath contains up to 3,000 volatile organic compounds (VOCs) in parts per million, including some in concentrations as low as parts per trillion. (8) (9) Many VOCs are altered in cancer, including those created by cytochrome P450 and oxidative stress, (8) (10) the Warburg effect, (8) (11) and genomic mutation metabolic pathways. (8) (12)

The e-nose device contains electronic chemical sensors connected to a pattern-recognition platform, which mimics the combinatorial selectivity of the human olfactory system. The VOCs in exhaled breath react on the sensors’ surface, changing their conductivity. Transducers convert the input into electrical signals, creating signatures called breathprints. The platform compares the combination of breathprints to a large VOC library specific to a disease like lung cancer. The e-nose provides a yes/no answer on the probability of detecting a lung cancer-associated breathprint, at a low cost of about $10 per test. (1)

Dr. Rocco has been investigating e-nose technology for identifying lung cancer for almost a decade. Before coming to MSK, he directed the Department of Thoracic Surgery at the Istituto Nazionale Tumori, Fondazione Pascale, IRCCS, where he evaluated the first-generation BIONOTE platform created by bioengineers at the University Campus Bio-Medico (UCBM) in Rome, Italy. The multisensorial platform demonstrated an overall sensitivity of 86% and specificity of 95% in a lung cancer screening initiative in a high-risk patient population that included smokers, older adults, and patients with chronic obstructive pulmonary disease. (13)

Study Design

The present MSK-exclusive trial included 100 treatment-naïve adult patients ages 21 to 85 evaluated at MSK between 2020 and 2023. All patients were diagnosed with a single, greater than 50% solid, stage 1 lung nodule visible on a CT or positron emission tomography scan by an experienced thoracic radiologist or confirmed as malignant on biopsy. Exhalates were prospectively collected during presurgical visits before patients received the standard-of-care diagnostic workup, which included transthoracic needle aspiration biopsy, robotic bronchoscopy, or proceeding directly to surgery. (1)  Mediastinal lymph node staging was performed according to current guidelines.  (14)

Patients exhaled into the mouthpiece of the breath-collection device for three minutes. The exhalate was collected in individual cartridges, which were refrigerated before shipment to the UCBM laboratory for analysis. The VOC breathprints from each cartridge were compared to a lung-cancer-specific breathprint library developed by UCBM investigators. The analysis leveraged a training dataset based on 657 breathprints from healthy controls and individuals diagnosed with lung cancer, chronic obstructive pulmonary disease, and obstructive sleep apnea syndrome. The evaluation assigned a probability score, or e-nose score, for detecting lung cancer, ranging from 0 to 1, where a score of 0.2 or greater indicates cancer and less than 0.2 indicates no cancer. (1)

Study Findings

E-nose results agreed with histopathologic results for 86% of 100 cases analyzed. The device earned an F1 score of 92.5% based on 86 true positives, two false negatives, and 12 false positives.  (1)

The investigators also compared the clinical utility of risk stratification into low-, intermediate-, and high-risk probabilities for the e-nose compared to the radiologic-based Swensen  (15) and Brock (16) prediction models. The e-nose would have referred only two patients with malignant nodules to observation compared to 9 with the Swensen model and 11 with the Brock model. The e-nose would also have referred 27 more high-risk patients with malignant nodules to treatment without needing biopsy compared to 19 and 6 patients with these other prediction models, respectively. (1)

Implications and Future Outlook

The high accuracy, high F1 score, high agreement with cytopathologic analysis, and low cost of the e-nose technology support its further research and development.

Dr. Rocco said the next steps will include refining the device from its current size of about three feet long to a smaller scale, perhaps as small as a USB flash drive fitting easily in a clinician’s jacket pocket. Dr. Rocco anticipates miniaturizing the device will only take a few months and looks forward to forging collaborations to achieve that goal.

Dr. Rocco also noted that importing the multisensorial analytic platform would reduce the turnaround time for results from two or three weeks to a couple of hours. Results from the present and future studies will inform a clearance application to the U.S. Food and Drug Administration. The Pneumopipe collection tube is a crucial part of the system and is patented in the European Union.

“We anticipate the e-nose will be highly useful as a screening tool for determining which patients should proceed to imaging, identifying subtypes and genetic mutations, predicting and evaluating responses to treatments, and guiding the development of precision therapies,” said Dr. Rocco. “The e-nose platform also has promising potential for diagnosing other cancers and lung conditions.”

The paper was supported in part by a National Institutes of Health/National Cancer Institute Cancer Center Support Grant P30 CA008748 and MSK’s Fiona and Stanley Druckenmiller Research Center for Lung Cancer ResearchDr. Rocco reported a financial interest with Scanlan International, Merck, and Medtronic. Refer to the paper for disclosures from other authors.

Refer a Patient
Call our dedicated clinician access number at 646-677-7440 or click the link below, and one of our care advisors will assist you with your referral needs.
  1. Rocco G, Pennazza G, Tan KS, et al. A real-world assessment of stage I lung cancer through electronic nose technology. J Thorac Oncol. Published online May 16, 2024.
  2. National Lung Screening Trial Research Team, Aberle DR, Adams AM, et al. Reduced lung-cancer mortality with low-dose computed tomographic screening. N Engl J Med. 2011;365(5):395-409.
  3. de Koning HJ, van der Aalst CM, de Jong PA, et al. Reduced Lung-Cancer Mortality with Volume CT Screening in a Randomized Trial. N Engl J Med. 2020;382(6):503-513.
  4. Mao Y, Cai J, Heuvelmans MA, et al. Performance of Lung-RADS in different target populations: a systematic review and meta-analysis. Eur Radiol. 2024;34(3):1877-1892.
  5. Klein EA, Richards D, Cohn A, et al. Clinical validation of a targeted methylation-based multi-cancer early detection test using an independent validation set. Ann Oncol. 2021;32(9):1167-1177.
  6. Liu MC, Oxnard GR, Klein EA, et al. Sensitive and specific multi-cancer detection and localization using methylation signatures in cell-free DNA. Ann Oncol. 2020;31(6):745-759. 
  7. Gale D, Heider K, Ruiz-Valdepenas A, et al. Residual ctDNA after treatment predicts early relapse in patients with early-stage non-small cell lung cancer. Ann Oncol. 2022;33(5):500-510.
  8. Haick H, Broza YY, Mochalski P, Ruzsanyi V, Amann A. Assessment, origin, and implementation of breath volatile cancer markers. Chem Soc Rev. 2014;43(5):1423-1449.
  9. Pauling L, Robinson AB, Teranishi R, Cary P. Quantitative analysis of urine vapor and breath by gas-liquid partition chromatography. Proc Natl Acad Sci U S A. 1971;68(10):2374-2376.
  10. Horváth I, Lázár Z, Gyulai N, Kollai M, Losonczy G. Exhaled biomarkers in lung cancer. Eur Respir J. 2009;34(1):261-275.
  11. Robey IF, Lien AD, Welsh SJ, Baggett BK, Gillies RJ. Hypoxia-inducible factor-1alpha and the glycolytic phenotype in tumors. Neoplasia. 2005;7(4):324-330.
  12. Kneepkens CM, Lepage G, Roy CC. The potential of the hydrocarbon breath test as a measure of lipid peroxidation. Free Radic Biol Med. 1994;17(2):127-160.
  13. Rocco R, Incalzi RA, Pennazza G, et al. BIONOTE e-nose technology may reduce false positives in lung cancer screening programmes. Eur J Cardiothorac Surg. 2016;49(4):1112-1117.
  14. Postmus PE, Kerr KM, Oudkerk M, et al. Early and locally advanced non-small-cell lung cancer (NSCLC): ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2017;28(suppl_4):iv1-iv21.
  15. Swensen SJ, Silverstein MD, Ilstrup DM, et al. The probability of malignancy in solitary pulmonary nodules. Application to small radiologically indeterminate nodules. Arch Intern Med. 1997;157(8):849-855.
  16. McWilliams A, Tammemagi MC, Mayo JR, et al. Probability of Cancer in Pulmonary Nodules Detected on First Screening CT. N Engl J Med. 2013;369(10):910-919.