The new research was presented at the European Respiratory Society International Congress.
‘Diagnosis of lung disease by artificial intelligence (AI) is more accurate compared to pulmonologists, proving that AI can serve as a second opinion for pulmonologists when they are assessing and diagnosing their patients.’
Dr Marko Topalovic (PhD), a postdoctoral researcher at the Laboratory for Respiratory Diseases, Catholic University of Leuven (KU Leuven), Belgium, told the meeting that "Pulmonary function tests provide an extensive series of numerical outputs and their patterns can be hard for the human eye to perceive and recognise; however, it is easy for computers to manage large quantities of data like these and so we thought AI could be useful for pulmonologists."
Pulmonary function tests (PFT) include:
- Spirometry where the amount of air inhaled and exhaled is measured
- Plethysmography test or a body box which enables doctors to assess lung volume
- Diffusion capacity test which tests the efficiency of the alveoli (small air sacks in the lungs) to see how well a patient's lungs transfer oxygen and carbon dioxide to and from the bloodstream
After these tests are done, doctors have to interpret the results to reach a diagnosis telling what is wrong with the patient.
Obtaining the AI algorithm using good quality data
First, the research team trained the AI algorithm
- to do this they used historical data from 1430 patients from 33 Belgian hospitals.
An expert panel then assessed the data using patients' medical histories, results of all PFTs and any additional tests and came to a conclusion on the correct interpretation and diagnosis for each patient.
The team then measured the interpretations against gold standard guidelines
from the European Respiratory Society and the American Thoracic Society.
The results were used to develop an algorithm to train the AI
, before validating it by incorporating it into real clinical practice. The algorithm was made to recognize patterns up to nine different diseases.
The study involved 120 pulmonologists from 16 hospitals situated in European countries like Belgium, France, The Netherlands, Germany, and Luxembourg).
The doctors made 6000 interpretations of PFT data from 50 randomly selected patients; the same data was examined using AI too.
The study measured the results from both interpretations against the gold standard guidelines in the same way as during development of the algorithm.
- The interpretations of the PFTs by the pulmonologists matched the guidelines in 74 percent of cases while the AI-based software interpretations matched the guidelines with a perfect score of 100 percent
- The doctors diagnosed the primary disease correctly in 45 percent of cases while the AI gave a correct diagnosis in 82 percent of cases
When left to the pulmonologists alone, the task of interpretation of PFTs and the diagnosis of respiratory disease requires more information and further tests to reach a satisfactory level of accuracy. The AI-based software can be used as a powerful decision support tool to improve current clinical practice especially in identifying difficult patterns of rare diseases
It is certain that AI will not replace doctors
- doctors have the ability to see beyond just the PFTs alone. They see a broader perspective which enables them to make decisions based on a combination of many different factors. However, AI will help doctors accomplish more and decrease chances for errors and redundant work.
The researchers plan to get more hospitals to use AI in the future and investigate transferring the technology to primary care, where the data would be captured by general practitioners to help them make correct diagnoses and referrals. Reference :
- AI improves doctors' ability to correctly interpret tests and diagnose lung disease - (https://aitopics.org/doc/news:23E604F8/)