Research AI

We develop and run applications to improve the clinical routine in the diagnostics workflow and to empower research. From comprehensive anatomical segmentation to AI-powered reporting, our tools bridge the gap between innovation and daily practice.

Report Comparison

Resident report review feedback system with track changes visualization and peer group analytics.

PacsCrawler

RIS and PACS data search engine with extensive filtering and bulk export capabilities.

TotalSegmentator

Comprehensive anatomical structure and organ segmentation for CT and MRI imaging.

CuraForge Engine

Modular integration platform for AI solutions with automated reporting and LLM-powered QC.

Report Comparison

Similar to track changes, residents can review changes (additions and deletions) made to preliminary reports by attendings. Dashboards summarize individual findings with reference to peer group performance.

In Production Since 2017-09-04

Reports Processed > 1.1 million

1'108'973

Total reports processed

Daily active users

>200

Average accesses per month

Associated Publications

Quantifying Radiology Residents' Learning Curves in Report Writing Performance Vosshenrich J, Guntli C, Cyriac J, et al. Radiology, 2024 Quantifying the Effects of Structured Reporting on Turnaround Times Vosshenrich J, Brantner P, Cyriac J, et al. Acad Radiol, 2023 Quantifying Radiology Resident Fatigue: Analysis of Preliminary Reports Vosshenrich J, Brantner P, Cyriac J, et al. Radiology, 2021 Revealing the Most Common Reporting Errors Through Data Mining Vosshenrich J, Nesic I, Cyriac J, et al. Eur Radiol, 2021

PacsCrawler

A powerful search tool operating on RIS and PACS in conjunction to identify studies matching selected criteria. Extensive filtering is possible based on RIS, DICOM, and report elements. Bulk export or DICOM transfer to research PACS is supported.

In Production Since 2017

Searches Performed -

Repository GitHub

>3M

Studies indexed

>100

Searches per month

>1,000

Research exports completed

Associated Publications

Publication references to be added

TotalSegmentator

Capable of segmenting over 100 anatomical structures and organs for CT and MRI imaging, down to the segment level. Provides robust segmentation even when pathology like fractures is present.

Public since 2022-01-19

Structures Supported 400+

Repository GitHub

Try TotalSegmentator Online

>80 Mio.

Number of processed images

>800k

Installations worldwide

400+

Anatomical structures supported

Associated Publications

Liver Segment and Lesion Segmentation on CT and MRI: An Open-Source Contribution to TotalSegmentator Nicoli AP, Bach M, Wasserthal J, et al. J Imaging Inform Med, 2025 TotalSegmentator MRI: Robust Sequence-independent Segmentation of Multiple Anatomic Structures Akinci D'Antonoli T, Berger LK, Indrakanti AK, et al. Radiology, 2025 Automatic Segmentation of Cardiovascular Structures on Chest CT: An Update of the TotalSegmentator Hinck D, Segeroth M, Miazza J, et al. Eur J Radiol, 2025 Intra-Individual Reproducibility of Automated Abdominal Organ Segmentation Abel L, Wasserthal J, Meyer MT, et al. J Imaging Inform Med, 2025 TotalSegmentator: Robust Segmentation of 104 Anatomic Structures in CT Images Wasserthal J, Breit HC, Meyer MT, et al. Radiol Artif Intell, 2023

CuraForge Engine

A modular integration platform unifying external AI solutions and data services (RIS, DICOM, HIS). A scripting engine allows users to write Python scripts to transform data and generate text snippets automatically inserted into reporting templates.

In Production Since January 2025

Coverage All CT/MRI Body and Chest

Key Components

Automated reporting of AI results and other data sources

LLM-powered Report QC module for consistency checks

ICD-10 coding assistant

See CuraForge Demo

Impact and Savings

666.5 hrs

Reporting time saved per year

Report writing time freed for each of 31 residents

$76,742

Freed-up labor cost savings per year

Dose and contrast info errors (down from 2.8% and 0.8%)

Gender mismatches (prev. 961 incidents last decade)

>10

Reports with auto-reporting elements

Associated Publications

Publication references to be added

Additional Tools

Clinical Applications

Workload Monitor Real-time departmental workload tracking and analytics

More Coming Soon Additional clinical applications in development

Who We Are

Our Team

A multidisciplinary group of radiologists, engineers, and data scientists working at the intersection of medicine and technology.

Joshy Cyriac

Technology / AI Officer

Full-stack development and system architecture for clinical AI applications.

Michael Bach

Group Leader

Data science and machine learning for medical imaging research.

Jakob Wasserthal

Researcher

Deep learning research and development of segmentation models including TotalSegmentator.

Martin Segeroth

Researcher

Research and development of AI solutions for clinical radiology workflows.

Tobias Heye

Clinical Lead

Clinical oversight and strategic direction for radiology AI integration.

Maurice Pradella

Radiologist

Clinical radiology expertise and validation of AI-driven diagnostic tools.

Ashraya Indrakanti

Researcher

Research contributions to AI-powered medical imaging and diagnostics.

Gian-Luca Leiggener

Intern

Supporting development and research projects within the technology lab.

Get In Touch

Contact Us

Interested in our solutions or want to collaborate? We're always open to discussing new projects, research partnerships, or answering questions about our tools.

Location University Hospital Basel
Clinic for Radiology and Nuclear Medicine
Petersgraben 4, 4031 Basel, Switzerland

Email Joshy.Cyriac@usb.ch

Web www.unispital-basel.ch