| Creators: |
Dinser, Moritz and Hieber, Daniel and Pryss, Rüdiger and Monoranu, Camelia-Maria and Liesche-Starnecker, Friederike and Schobel, Johannes and Nickl, Vera |
| Title: |
Automated Locoregional Antigen Determination as a Prerequisite for CAR T-Cell Therapy Target Selection |
| Item Type: |
Conference or Workshop Item |
| Event Title: |
(Abstractband der) 107. Jahrestagung der Deutschen Gesellschaft für Pathologie (DGPatho) "Next Generation Pathology" |
| Event Location: |
München, Germany |
| Event Dates: |
23.-25. Mai 2024 |
| Projects: |
DigiHealth, NAP |
| Page Range: |
p. 159 |
| Date: |
2024 |
| Divisions: |
Gesundheitsmanagement |
| Abstract (ENG): |
Background:
Clinical studies show that the genetic modification of T-cells to express chimeric antigen receptors (CARs) provide
promising results in tumor therapy [1]. CAR T-cells are retargeted to recognize tumor associated antigens (TAAs).
However, the selection of suitable TAAs for CAR T-cell therapy poses a challenge. Striking a balance is crucial, aiming for minimal overlap of target antigens to mitigate the risk of off-tumor toxicity, while simultaneously ensuring comprehensive coverage across the entire tumor to avoid antigen escape and recurrence.
This work presents an automated Artificial Intelligence pipeline that determines TAAs with maximum coverage and
minimal overlap based on immunohistochemical (IHC) stained pathological slides.
Methods:
The pipeline utilizes one Hematoxylin and Eosin (HE) stained slide and six IHC slides, stained for different TAAs. Initial alignment of all slides was performed using the VALIS framework [2], accounting for varying tissue distribution. Subsequently, a Machine Learning (ML) model identified the tumor area in the HE-slide, and the resulting tumor mask 160 was applied to all IHC-slides. Tumor regions from all IHC-slides were then extracted, segmented into smaller tiles, and subjected to a color deconvolution algorithm [3]. A quantification algorithm analyzed the antigen coverage [4]. Finally, combining antigen expressions from all IHC-slides, antigen pairs or triplets with minimal overlap and maximum coverage were determined.
Results:
The proposed pipeline is currently evaluated in a medical study determining optimal TAAs for a dual or triple CAR T-cell therapy strategy in Glioblastoma (GBM) patients. The study evaluates six TAAs. A GBM-specific ML model is used for tumor segmentation [5]. First results align with medical analyses from an accompanying study.
Conclusion:
This work presents an automated and feasible pipeline for selecting TAAs with a maximum coverage and minimal
overlap for optimizing CAR T-cell therapy strategies. The initial results are promising and consistent with accompanying medical analyses. An upcoming large-scale evaluation of GBM slides will confirm its suitability for clinical use. The pipeline uses a minimum of ML, limited to tumor detection, ensuring easy-to-understand results. |
| Forthcoming: |
No |
| Language: |
German |
| Link eMedia: |
Download |
| Citation: |
Dinser, Moritz and Hieber, Daniel and Pryss, Rüdiger and Monoranu, Camelia-Maria and Liesche-Starnecker, Friederike and Schobel, Johannes and Nickl, Vera
(2024)
Automated Locoregional Antigen Determination as a Prerequisite for CAR T-Cell Therapy Target Selection.
In: (Abstractband der) 107. Jahrestagung der Deutschen Gesellschaft für Pathologie (DGPatho) "Next Generation Pathology", 23.-25. Mai 2024, München, Germany, p. 159.
|
