We are pleased to share a comprehensive validation of the latest version of PerfLab, conducted using the standardized datasets from the Open Source Initiative for Perfusion Imaging (OSIPI) DCE challenge (see the benchmark paper in Magnetic Resonance in Medicine).

In this rigorous assessment, we systematically evaluated the accuracy and repeatability of the key steps in our DCE-MRI analysis pipeline—AIF estimation, and pharmacokinetic modeling.

The results show that the current production version of PerfLab yields higher accuracy and lower parameter variability than our original submission to the OSIPI challenge. These optimizations translate directly into more robust and reproducible perfusion metrics for your imaging studies.

👉 [Read the detailed validation report here] to explore the benchmark results and see how PerfLab can enhance your research workflow.

Two collaborative projects supported within the canSERV framework, in partnership with RCSI and LIH, are currently in progress with active contributions from our team. As part of these projects, we are providing data processing services through PerfLab, enabling efficient and standardized analysis workflows for advanced imaging data.

In parallel, we are systematically collecting feedback from project users. Their input is being directly incorporated into ongoing development, helping us refine the software, improve usability, and address practical needs identified during deployment.

We look forward to continuing collaboration with our partners and further enhancing PerfLab based on user experience and emerging research requirements. For more information on the canSERV initiative, please visit the official website: canSERV - Cancer Survivorship Research

Users can now benefit from a new version of PerfLab computational core. The new core adds multiple new features, like automatic processing and report creation. In the backend the core runs on updated Python version and follows stricter coding rules to ensure better reproducibility and maintainability. With the release of V4 a V5_dev is now out for testing and open for feature additons. The V3 is now considered legacy and will no longer be maintained.

Users can now run all or part of the processing chain automatically by providing a link to a reference processing. In steps where user-defined regions of interest are preferred, the processing pauses and waits for input.

In cooperation with Medoro s.r.o., we have implemented a secure connection between PerfLab and the hospital's PACS. This allows DCE-MRI data to be sent to PerfLab with a single click from the MR system's console. Once the perfusion map calculation is complete, the results are automatically sent back to the PACS.

Reference installation: Masaryk Memorial Cancer Institute