Predictive model with standard clinical parameters differentiates PSC and SSC

The researchers hope that the model has “the potential to streamline patient referral, guide personalized treatment strategies, and enhance clinical trial eligibility.”

Study presenter Miki Scaravaglio (University of Milano-Bicocca, Italy) told medwireNews that there are several reasons why it is important to differentiate between primary sclerosing cholangitis (PSC) and secondary sclerosing cholangitis (SSC), the first one being that “they are completely different diseases, and their clinical management is completely different.”

She explained: “For PSC, we don’t have a specific treatment, but we have to surveil for cancer development, because it’s a disease with higher risk of cancer, and the risk is higher in the first years after diagnosis. So it’s crucial to make a diagnosis as soon as possible.”

On the other hand, treatments are available for SSC and so not making the correct diagnosis “means that we delay a treatment that might change the natural history of the disease,” Scaravaglio said, as well as contribute to patient anxiety.

She continued that, from a research perspective, clinical trials for PSC seek to exclude patients with SSC, and vice versa, “but we don’t really have a standardized way” to do that. “So this, of course, is a problem, because we end up having a heterogeneous cohort in both types of clinical trials,” Scaravaglio said. This is “even more important for PSC,” due the lack of approved treatments.

The researchers therefore aimed to identify a set of readily available clinical parameters to support the early and accurate diagnosis of PSC that generalists can easily collect as a form of “initial screening in order to speed up referral to the specialists,” Scaravaglio explained.

They studied adult cases of sclerosing cholangitis from three tertiary liver centers in Italy, with the patients classified as PSC or SSC based on current expert consensus. In all, 155 PSC patients and 79 with SSC were included. There were numerous baseline differences between the two groups.

The team found that, on multivariate logistic regression analysis, six baseline variables were significantly associated with an increased likelihood of PSC rather than SSC: younger age at diagnosis; the presence of IBD; no history of abdominal surgery; autoimmune comorbidities; a family history of autoimmune diseases; and the absence of pancreatic abnormalities.

The resulting predictive model for PSC diagnosis, termed the PSC-DM, was tested in a validation cohort from the UK comprising 98 PSC patients and 44 with SSC. The model was able to predict the presence of PSC at an area under the receiver operating characteristic curve of 0.945.

Scaravaglio noted that, despite both diseases being rare, the study is limited by the small sample size, and so they plan to test the model further in a larger, prospective cohort.