April 2026, Pharmacia, an open-access, peer-reviewed journal published by Pensoft, has published the article: “Real-world effectiveness of Ruxolitinib: analysis and comparison with COMFORT-II.”

This Real-World Evidence (RWE) study, co-authored by researchers from the National Council on Prices and Reimbursement of Medicinal Products (NCPRMP), Medical University – Sofia, Sofia University “St. Kliment Ohridski”, and the National Specialized Hospital for Active Therapy of Hematological Diseases, evaluates the long-term effectiveness of ruxolitinib in Bulgarian patients with myelofibrosis.

Sqilline’s Danny Platform powered the study by extracting, structuring, and harmonizing structured and unstructured hospital electronic health records using AI and Large Language Models (LLMs). The platform transformed routine clinical data into research-ready datasets with ≥95% extraction accuracy, enabling robust real-world analyses and comparison with the pivotal COMFORT-II clinical trial. The retrospective study included 466 patients treated between 2018 and 2024, with a trial-aligned subgroup of 68 patients for direct comparison.

Introduction

Myelofibrosis (MF) is a rare and progressive myeloproliferative neoplasm characterized by bone marrow fibrosis, splenomegaly, constitutional symptoms, and reduced survival. Although ruxolitinib has become the global standard first-line therapy following the COMFORT studies, additional real-world evidence is essential to evaluate its effectiveness in routine clinical practice across broader and more heterogeneous patient populations.

Key Summary Points

Objectives

  • Evaluate long-term real-world effectiveness of ruxolitinib in Bulgarian patients with myelofibrosis.
  • Assess progression-free survival (PFS), overall survival (OS), clinical benefit rate (CBR), and objective response rate (ORR).
  • Compare Bulgarian real-world outcomes with the pivotal COMFORT-II clinical trial.

Challenges

  • Clinical trial populations differ substantially from routine clinical practice, limiting direct comparison.
  • Hospital data originate from heterogeneous structured and free-text electronic medical records.
  • Many clinical variables required for trial matching are incomplete in routine documentation.
  • Advanced cohort alignment was required to generate statistically meaningful comparisons with COMFORT-II.

Solution

Danny Platform enabled:

  • AI-powered extraction of structured and unstructured EHR data using embedded Machine Learning and Large Language Models.
  • Harmonization of nationwide hospital data into standardized, research-ready datasets.
  • Validation of extracted clinical data with ≥95% accuracy.
  • Application of Iterative Proportional Fitting (IPF) and trial eligibility criteria to generate a comparable real-world cohort aligned with COMFORT-II.

Results

Patient Cohort

  • 466 patients with myelofibrosis receiving ruxolitinib
  • Study period: 2018–2024
  • Median age: 71 years
  • Female: 47%
  • Male: 53%
  • Trial-matched comparison cohort: 68 patients

Real-World Outcomes

Progression-Free Survival (Entire Cohort)

  • Median PFS: 73.5 months
  • 12-month PFS: 88%
  • 20-month PFS: 83%
  • 30-month PFS: 75%
  • 40-month PFS: 65%
  • 50-month PFS: 61%

Overall Survival (Entire Cohort)

  • Median OS: 48.2 months
  • 12-month OS: 87%
  • 20-month OS: 78%
  • 30-month OS: 70%
  • 40-month OS: 56%

Clinical Effectiveness

  • Clinical Benefit Rate (CBR): 96%
  • Objective Response Rate (ORR): 79.2% (53 evaluable patients)

Comparison with COMFORT-I

Following application of inclusion/exclusion criteria and IPF adjustment:

  • Eligible comparison cohort: 68 patients
  • Trial cohort: 146 patients

Trial-Aligned Outcomes

Progression-Free Survival

  • 6-month PFS: 93%
  • 12-month PFS: 84%
  • Median PFS: 35.3 months

Overall Survival

  • 6-month OS: 90% (vs 97% in COMFORT-II)
  • 12-month OS: 74% (vs 95%)
  • 18-month OS: 59% (vs 88%)

The study demonstrated that real-world outcomes were generally consistent with published international evidence while reflecting the broader and more clinically complex patient population encountered in routine practice.

Conclusions

  • Confirms the effectiveness of ruxolitinib as the standard first-line treatment for myelofibrosis in routine clinical practice.
  • Real-world outcomes are consistent with COMFORT-II and other international real-world registries.
  • Demonstrates the value of AI-powered data harmonization for generating robust evidence from heterogeneous hospital records.
  • Shows how Danny Platform enables high-quality real-world evidence studies capable of supporting clinical decision-making and health technology assessment.

Authors & Affiliations

  • Antoan Rangelov — Sofia University “St. Kliment Ohridski”; National Council on Prices and Reimbursement of Medicinal Products (NCPRMP), Sofia
  • Boryana Ivanova — NCPRMP; Medical University – Sofia
  • Manoela Manova — NCPRMP; Medical University – Sofia
  • Galya Kondeva — National Specialized Hospital for Active Therapy of Hematological Diseases, Sofia
  • Alexandra Savova — NCPRMP; Medical University – Sofia