AI May Improve Diagnostic Interpretation of Lung Function Tests

Current testing requires a series of assays including a spirometry test (measures the volume amount and the flow speed of air during breathing), followed by a body plethysmography test (measures static lung volumes and airways resistance), then finally a diffusion test (measures amount of oxygen and other gases that cross the lungs' air sacs). Analysis of the results of these tests is largely based on expert opinion and international guidelines, attempting to detect patterns.

The new study included data from 968 people who were undergoing complete lung function testing for the first time. All participants received a first clinical diagnosis based on lung function tests and additional tests (e.g. CT scans, electrocardiogram, etc.). Final diagnosis was validated by a consensus of the large group of expert clinicians.

The researchers subsequently investigated whether “machine learning” (utilizing algorithms that can learn from data analysis and perform predictions) could be used to effectively analyze a complete set of lung function tests. They developed an algorithm process in addition to the routine lung function parameters and clinical variables of smoking history, body mass index, and age. Based on the pattern of both the clinical and lung function data, the algorithm provided a suggestion for the most likely diagnosis.

Wim Janssens from University of Leuven (Leuven, Belgium), study senior author, said: "We have demonstrated that AI can provide us with a more accurate diagnosis in this new study. The beauty of our development is that the algorithm can simulate the complex reasoning that a clinician uses to give their diagnosis, but in a more standardized and objective way."

Clinicians currently rely on analyzing the results using population-based parameters. With AI, the machine can observe a combination of patterns simultaneously to help produce a more accurate diagnosis. This has been done in other healthcare fields, such as with an automated interpretation of results from an electrocardiogram being routinely used in clinical practice as a decision support system.

Marko Topalovic, University of Leuven, study first author, said: "The benefit of this method is a more accurate and automated interpretation of pulmonary function tests, and thus better disease detection. Not only can this help non-experienced clinicians, but it also has many benefits for healthcare overall as it is time saving in achieving a final diagnosis as it could decrease the number of redundant additional tests clinicians are taking to confirm the diagnosis."

The next step will be to test the algorithm in different populations and increase the system’s decision power using continuous updates on lung function data with validated clinical diagnoses.

Related Links:
University of Leuven