Computer Model Seen to Accurately Distinguish PAH From PVOD

Gene expression patterns in the lungs can help distinguish pulmonary arterial hypertension (PAH) from pulmonary veno-occlusive disease (PVOD) — a subtype in which the narrowing of veins, instead of arteries, causes pulmonary hypertension — with high accuracy, a study reported.

Findings also showed that lung diseases share similar alterations in gene expression, suggesting a common disease-causing mechanism, and that PVOD is actually much more similar to idiopathic pulmonary fibrosis (IPF) than to PAH. (Of note, gene expression is the process by which information in a gene is converted into a product, like a protein.)

The study, “Molecular Profiling of Vascular Remodeling in Chronic Pulmonary Disease,” was published in the American Journal of Pathology

While the management and prognosis of pulmonary hypertension depends on whether the fibrotic narrowing of blood vessels happens in arteries or veins, it’s challenging to distinguish between PAH and PVOD, particularly at early stages.

Despite affecting distinct blood vessel compartments, these diseases share similar clinical signs in their early stages. But the disease’s likely course, its prognosis, is markedly worse for those with PVOD than for PAH patients.

People with PVOD are also at risk of life-threatening pulmonary edema — when fluid collects in the air sacs of the lungs — particularly when treated with PAH medications.

“The pathogenesis of PVOD and PAH is poorly understood and the clinical differentiation between both diseases remains challenging because of similar clinical presentation. Our study is the first to put forward a molecular model with the ability to differentiate between the PH subtypes PAH and PVOD,” Lavinia Neubert, MD, lead study investigator, said in a press release.

Aiming to create an algorithm able to differentiate these two conditions, a team led by researchers at the University of Hannover, in Germany, examined lung samples from 20 PAH and 19 PVOD patients undergoing a lung transplant — the only curative treatment for PVOD and advanced lung diseases with severe remodeling of pulmonary vessels.

Samples from 16 lung donors collected before transplant, and from 13 IPF and 15 chronic obstructive pulmonary disease (COPD) patients, were also examined. These latter two lung diseases are also frequently associated with vascular remodeling and pulmonary hypertension.

Using lung donors as controls, the team started by measuring differences in a panel of genes involved in inflammation and in those coding for regulatory proteins called kinases.

Two genes — SGK196 and MAST2 — were consistently overactive in all patient groups compared to control samples, suggesting that the two kinases (enzymes) coded by these genes play an essential role in the formation of pulmonary vascular lesions.

A total of 10 genes were abnormal in samples from PVOD patients, but decreased MMP9 activity was the only alteration unique to these patients. The MMP9 protein is involved in modifying and degrading the extracellular matrix, and its absence may contribute to vascular remodeling by facilitating matrix deposition in veins.

PAH samples had 93 genes with atypical activity, and 50 of them were unique to the disease. IPF and COPD showed 46 and 60 genes, respectively, with significantly different expression compared to control samples.

To determine whether these genetic signatures might distinguish PAH from PVOD, researchers split these samples into a training and a test set. First, eight PAH and seven PVOD samples were used to create a computer model based on a set of six selected genes.

This model was then tested on 12 samples from each group, and it correctly identified samples with a 96% accuracy. In fact, the model correctly identified all PVOD samples and 11 out of 12 PAH samples, meaning it had a 100% sensitivity and 92% specificity for differentiating PVOD from PAH samples.

“We are confident that the classification accuracy of our molecular approach is close to the gold standard of histopathological diagnosis,” said Neubert, who is also a member of the German Center for Lung Research, Biomedical Research in Endstage and Obstructive Lung Disease Hannover.

“Our findings promise to help develop novel target-specific interventions and innovative approaches to facilitate clinical diagnostics in an elusive group of diseases,” Neubert added.

An analysis of biological functions active and inactive in a given disease found that PVOD was 11 times more similar to IPF than to PAH, despite being considered a PAH subtype.

Based on the proteins produced in each disease, PAH revealed an environment that causes cell death, with cancer-like characteristics and increased cell stress. The environment in IPF and PVOD, in contrast, prevented cell death and inflammation, and suggested lower migration of immune cells.

Through this work, the team was “able to generate a neuronal network that differentiated PVOD from PAH samples,” and to find “close functional similarities … between PVOD and IPF as well as between PAH and COPD,” the researchers wrote.

If these genetic and protein alterations are also observed in blood or urine samples, the researchers believes this approach could provide an alternative, non-invasive way to distinguish PAH from PVOD with high accuracy. If proven feasible, it could replace the invasive and risky lung biopsies used to diagnose patients.

“Since lung biopsies remain as high-risk interventions for PH patients and non-invasive approaches currently do not allow for a definite diagnostic accuracy, these findings may facilitate diagnosis of PVOD by molecular analysis, provided that our findings are reproducible in blood or urine samples,” Neubert concluded.