An electronic nose (eNose) that measures volatile organic compounds (VOCs) emitted from the lungs successfully distinguished sarcoidosis from interstitial lung disease (ILD) and healthy controls, according to a report in the journal CHEST.
Dr. Iris van der Sar
The approach has the potential to generate clinical data that can’t be achieved through other noninvasive means, such as the serum biomarker soluble interleukin-2 receptor (sIL-2R). sIL-2R is often used to track disease activity, but it isn’t specific for diagnosing sarcoidosis, and it isn’t available worldwide.
Sarcoidosis is a granulomatous inflammatory disease with no known cause and can affect most organs, but an estimated 89%-99% of cases affect the lungs. There is no simple noninvasive diagnostic test, leaving physicians to rely on clinical features, biopsies to obtain tissue pathology, and the ruling out of other granulomatous diagnoses.
The challenge is more difficult because sarcoidosis is a heterogeneous disease, with great variation in the number of organs affected, severity, rate of progression, and response to therapy.
Previous researchers have used VOCs in an attempt to diagnose diseases, since the compounds reflect pathophysiological processes. Gas chromatography/mass spectrometry (GCMS) is one method to identify the individual VOCs, but the process is time consuming and complex. Some nevertheless showed potential in sarcoidosis, but failed to reproduce their performance in validation cohorts.
In the new study, a cross-sectional analysis showed that exhaled breath analysis using an eNose had excellent sensitivity and specificity for distinguishing sarcoidosis from ILD and healthy controls, and identified sarcoidosis regardless of pulmonary involvement, pulmonary fibrosis, multiple organ involvement, immunosuppressive treatment, or whether or not pathology supported the diagnosis.
Dr. Marlies Wijsenbeek
The eNose technology produces a “breath-print” after combining information from a broad range of VOCs. The information originates from an array of metal-oxide semiconductor sensors with partial specificity that artificial intelligence processes to discern patterns. Overall, the system functions similarly to the mammalian olfactory system. The artificial intelligence instead views it as a “breath-print” that it can compare against previously learned patterns.
“It is a quite easy, simple, and quick procedure, which is noninvasive. We can collect a lot of data from the VOCs in the exhaled breath because there are several sensors that cross-react. We can create breath profiles and group patients to see if profiles differ. Ultimately, we can use the profiles to diagnose or detect disease in the earlier stage and more accurately,” said Iris van der Sar, MD. Dr. van der Sar is the lead author on the study and a PhD candidate at Erasmus Medical Center in Rotterdam.
The study requires further prospective validation, but the technology could have important clinical benefits, said senior author and principal investigator Marlies Wijsenbeek, MD, PhD, pulmonologist and head of the Interstitial Lung Disease Center at Erasmus Medical Center. “If we in future can avoid a biopsy, that would be most attractive,” said Dr. Wijsenbeek.
“We hope to come to a point-of-care device that can be used to facilitate early diagnosis at low burden for the patient and health care system,” said Karen Moor, MD, PhD, and post-doc on this project. The researchers also hope to determine if the eNose can help evaluate a patient’s response to therapy.
Dr. Karen Moor
Studies of eNose technology in other chronic diseases have shown promising results, but not all results have been validated yet in independent or external cohorts.
The current study included 569 outpatients, 252 with sarcoidosis and 317 with ILD, along with 48 healthy controls. The researchers constructed a training set using 168 patients with sarcoidosis and 32 healthy controls, and a validation set using 84 patients with sarcoidosis and 16 healthy controls. The eNose differentiated between patients and controls in both groups, with an area under the curve of 1.00 for each regardless of pulmonary involvement or treatment.
It also distinguished those with sarcoidosis and pulmonary involvement from those with ILD, with an AUC of 0.90 (95% confidence interval, 0.87-0.94) in the training set, and an AUC of 0.87 (95% CI, 0.82-0.93) in the validation set.
It differentiated between pulmonary sarcoidosis and hypersensitivity pneumonitis in the training set (AUC 0.95; 95% CI, 0.90-0.99) and the validation set (AUC, 0.88; 95% CI, 0.75-1.00).
The study received no funding. Dr. Wijsenbeek, Dr. van der Sar, and Dr. Moor have no relevant financial disclosures.