Project Details
Projekt
Next-generation morphometry and pathomics for nephropathology
Subject Area
Nephrology
Term
since 2021
Project identifier
Deutsche Forschungsgemeinschaft (DFG) - Project number 445703531
During the previous funding period, we developed a deep-learning-based approach enabling automated large-scale quantification of kidney histology, termed Next Generation Morphometry (NGM). NGM enables -omics analysis of tissue morphology at the microscopic level, i.e. pathomics. We have proven the utility of pathomics in several preclinical studies, including a preclinical randomized multicenter control trial. Tubular atrophy is a major predictor of kidney disease progression but is only assessed in a qualitative or semi-quantitative manner in biopsies. Much less is known about other pathological processes, such as hypertrophy or hyperplasia, also because tools for large-scale analyses are missing. We hypothesize that pathological processes lead to distinct morphological alterations that can be precisely quantified using pathomics, defining kidney disease transition and progression. Here, we will combine our complementary expertise in pathology (Boor) and computer vision, AI, and advanced geometry (Kobbelt) to advance kidney pathomics. In WP1, we will improve the technical pipeline to enable high-throughput segmentation of histological structures and automated quality control. In WP2, we will significantly extend the pathomics feature set and enable pathomics on the level of single cells (together with P9). We will also develop biostatistical analysis tools for the pathomics data (with P2 and P8). In WP3, we will take advantage of our multicenter clinical cohorts of kidney biopsies, including IgA nephropathy, to validate the clinical utility of pathomics. In WP4, we will lay the basis for a kidney pathomics cell atlas and define parameters for disease state transitions, e.g. a threshold for pathological glomerular or tubular hyper- or atrophy. We will also develop approaches for integrating pathomics and transcriptomics/proteomics data (with P1, P2, P3). In WP5, we will continue providing nephropathology and image analysis expertise for the consortium. In summary, we will develop a pathomics pipeline to extract quantitative measures from histological images, laying the basis for a quantitative kidney cell atlas. We will integrate morphological with molecular data and define thresholds for the pathomics features indicative of disease transition, progression and recovery. Pathomics features could become surrogate endpoints in clinical trials or diagnostics, advancing the field of precision computational nephropathology.
DFG Programme
Clinical Research Units
