Robin Peretzke

Robin Peretzke

Auto-generated profile, based on publicly available information

Robin Peretzke

German Cancer Research Center (DKFZ) Heidelberg

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
2 December 2024 in Molecular Psychiatry
Verified
Journal Article
Open access
PDF file
RPRobin Peretzke
Robin Peretzke
PNPeter F. Neher
Peter F. Neher
GBGeva A. Brandt
Geva A. Brandt
SFStefan Fritze
Stefan Fritze
SVSebastian Volkmer
Sebastian Volkmer
DHDusan Hirjak
Dusan Hirjak
Catatonia is a severe psychomotor disorder characterized by motor, affective and cognitive-behavioral abnormalities. Although previous magnetic resonance imaging (MRI) studies suggested white matter (WM) dysconnectivity in the pathogenesis of catatonia, it is unclear whether microstructural alterations of WM tracts connecting psychomotor regions might contribute to a better classification of catatonia patients. Here, diffusion-weighted MRI data were collected from two independent cohorts (whiteCAT/replication cohort) of patients with (n = 45/n = 13) and without (n = 56/n = 26) catatonia according to ICD-11 criteria. Catatonia severity was examined using the Northoff (NCRS) and Bush-Francis (BFCRS) Catatonia Rating Scales. We used tract-based spatial statistics (TBSS), tractometry (TractSeg) and machine-learning (ML) to classify catatonia patients from tractometry values as well as tractomics features generated by the newly developed tool RadTract. Catatonia patients showed fractional anisotropy (FA) alterations measured via TractSeg in different corpus callosum segments (CC_1, CC_3, CC_4, CC_5 and CC_6) compared to non-catatonia patients across both cohorts. Our classification results indicated a higher level of performance when trained on tractomics as opposed to traditional tractometry values. Moreover, in the CC_6, we successfully trained two classifiers using the tractomics features identified in the whiteCAT data. These classifiers were applied separately to the whiteCAT and replication cohorts, demonstrating comparable performance with Area Under the Receiver Operating Characteristics (AUROC) values of 0.79 for the whiteCAT cohort and 0.76 for the replication cohort. In contrast, training on FA tractometry resulted in lower AUROC values of 0.66 for the whiteCAT cohort and 0.51 for the replication cohort. In conclusion, these findings underscore the significance of CC WM microstructural alterations in the pathophysiology of catatonia. The successful use of an ML based classification model to identify catatonia patients has the potential to improve diagnostic precision.

4 November 2025 in The Journal of Pathology
Verified
Journal Article
Open access
MFMaximilian Fischer
Maximilian Fischer
AMAlexander Muckenhuber
Alexander Muckenhuber
RPRobin Peretzke
Robin Peretzke
LFLuay Farah
Luay Farah
CUConstantin Ulrich
Constantin Ulrich
LMLena Maier-Hein
Lena Maier-Hein
Pancreatic ductal adenocarcinoma (PDAC) subtyping typically relies on immunohistochemistry (IHC) staining for critical markers like HNF1A and KRT81, a labor‐intensive manual staining process that introduces variability. Virtual staining methods offer promising alternatives by generating synthetic IHC images from routine hematoxylin and eosin (H&E) slides. However, most current approaches evaluate success by image quality measures rather than assessing diagnostically relevant features. Here, we introduce a novel cycleGAN framework utilizing a contrastive‐inspired approach trained on semipaired datasets derived from consecutive tissue sections. Our method significantly enhances PDAC subtyping accuracy based on synthetic IHC images generated from standard H&E inputs, improving the classification F1‐score from 0.66 to 0.77 for KRT81 and from 0.61 to 0.73 for HNF1A, compared with classification directly on H&E images. This approach also substantially outperforms baseline CycleGAN models. These results underscore the clinical potential of contrastive virtual staining to streamline PDAC diagnostics and improve their robustness. © 2025 The Author(s). The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

13 April 2025
Verified
Book Chapter
Closed access
PDF file
NDNico Albert Disch
Nico Albert Disch
RPRobin Peretzke
Robin Peretzke
SRSaikat Roy
Saikat Roy
CUConstantin Ulrich
Constantin Ulrich
DZDavid Zimmerer
David Zimmerer
LMLena Maier-Hein
Lena Maier-Hein

18 December 2025 in arXiv
Verified
Preprint
Open access
PDF file
NDNico Albert Disch
Nico Albert Disch
SRSaikat Roy
Saikat Roy
CUConstantin Ulrich
Constantin Ulrich
YKYannick Kirchhoff
Yannick Kirchhoff
MRMaximilian Rokuss
Maximilian Rokuss
LMLena Maier-Hein
Lena Maier-Hein
Forecasting how 3D medical scans evolve over time is important for disease progression, treatment planning, and developmental assessment. Yet existing models either rely on a single prior scan, fixed grid times, or target global labels, which limits voxel-level forecasting under irregular sampling. We present CRONOS, a unified framework for many-to-one prediction from multiple past scans that supports both discrete (grid-based) and continuous (real-valued) timestamps in one model, to the best of our knowledge the first to achieve continuous sequence-to-image forecasting for 3D medical data. CRONOS learns a spatio-temporal velocity field that transports context volumes toward a target volume at an arbitrary time, while operating directly in 3D voxel space. Across three public datasets spanning Cine-MRI, perfusion CT, and longitudinal MRI, CRONOS outperforms other baselines, while remaining computationally competitive. We will release code and evaluation protocols to enable reproducible, multi-dataset benchmarking of multi-context, continuous-time forecasting.