2017 FLUIDDA’s year in review: IPF, Machine Learning and Share Prices

With the end of the year approaching it is always good to take a moment and reflect on the year that has passed.I would describe 2017 for FLUIDDA in two words: significant progress.

On many fronts, 2017 has been a transformative year characterized by strengthening of the corporate structure and by several clinical study results that support the hypothesis around the value of our Functional Respiratory Imaging (FRI) approach. In this blog I’ll highlight three topics that stood out for me: Idiopathic Pulmonary Fibrosis (IPF), Machine learning and Share Prices.

Idiopathic Pulmonary Fibrosis (IPF)

Idiopathic Pulmonary Fibrosis or IPF is a terrible lung disease that received significant attention in 2017. Several promising drug candidates advanced to the next stage of development. With FLUIDDA, we were able to confirm our initial hypothesis, which led to an FDA letter of support, that FRI parameters provide clinically relevant information about the disease and drug efficacy which cannot be obtained by conventional measures such as the Forced Vital Capacity (FVC). We showed that significant disease is already present in the lung at a time when FVC is still normal. This knowledge will lead to better screening of patients and earlier detection of the disease. The earlier a disease like IPF can be diagnosed, the higher the probability of effectively halting disease progression and, eventually, reversing the damage. In terms of monitoring drug efficacy, it turned out that especially changes in FRI-based airway volumes and resistances appeared to be quite sensitive measures to monitor drug efficacy in an early stage requiring only small numbers of patients (~20 to 30 per arm). With several studies lined up for 2018 we look forward to helping the IPF field in any way we can and to advance the science for the benefit of the patient.

Machine learning

In a previous blog I described how I believe machine learning can and will revolutionize healthcare and in particular how it adds value to FRI. In September 2017 we had a very successful ERS conference with no less than 9 presentations on FRI.

In one of the presentations we presented the results of our work in lung transplantation where we combined FRI with machine learning. In the field of lung transplantation, chronic rejection of the transplanted lung is the main cause of mortality. We studied 41 patients who received a lung transplantation, 15 patients developed Bronchiolitis Obliterans  Syndrome (aka chronic rejection) while 26 did not. Based on an inspiratory/expiratory scan taken a few weeks after the transplant surgery, the combination of FRI and machine learning resulted in an accurate (85%) prediction of the fate of individual patients (BOS vs non-BOS). It seems that signs of rejection are already present in an early stage and that if the diagnostic tests are sensitive enough these signs can be detected.

In another poster presented at ERS, we assessed the power of FRI and machine learning to predict response to a novel drug in IPF after 48 weeks of treatment. In the study population of 66 IPF patients, again high accuracies of up to 86% were found using an inspiratory/expiratory scan taken at baseline. In order words, in more than 8 out of 10 cases FRI using machine learning algorithms, correctly predicted whether a patient would respond to a drug or not. Even though the patient numbers are small and more studies are needed to confirm the outcome, these initial results are very encouraging for patients suffering from this dreadful condition for which limited therapeutic options are available.

Finally, in a very recent study FRI and machine learning was used to assess the probability of COPD patients developing an exacerbation in the weeks following an FRI assessment. Patients with frequent exacerbations are at a higher risk of dying compared to patients with fewer exacerbations. Furthermore, exacerbations are the biggest cost driver in COPD. In a cohort of 62 patients the FRI, machine learning combo was able to accurately (80%) predict whether an exacerbation in an individual patient was imminent (n = 23) or not (n = 39). If additional studies confirm these findings then this could lead to better management of COPD patients resulting in fewer hospital (re)admissions.

Share Prices

Having a sensitive tool that accurately describes disease stage, progression and treatment efficacy not only helps doctors and drug developers to develop better drugs and treat patients more efficiently, it also helps investors to assess the market potential of a new therapy. In 2017 two of our clients, Galapagos and Bellerophon, released the results of clinical studies they performed with FRI. In both cases the FRI results supported the mode of action of the drugs and underscored the potential of the treatments to better manage IPF and Pulmonary Hypertensions associated with IPF and COPD. In the weeks and months following the press release the share prices of both companies went up with 40% (GLPG) and even 250% (BLPH). In several meetings we had with analysts and investors following these press releases, good endpoints was always mentioned as a critical tool for any valuation exercise. In that regard we believe the FRI has an important role to play across the spectrum and that many stakeholders can use the tool to their benefit.

Looking forward to 2018 we are excited to continue our mission to bring imaging to the forefront of respiratory medicine, to provide drug developers with better tools, to enable regulatory agencies to approve safe drugs in a shorter timeframe and to assist doctors in treating patients more efficiently.


I wish you and all your loved ones a Wonderful New Year!


Dr. Jan De Backer,
Chief Executive Officer

Sorry, comments are closed for this post.