Integrating cytological images and spatial transcriptomics for cell segmentation with DISSECT.
New imaging approach reveals how immune cells rearrange around tumors after immunotherapy, deepening understanding of treatment response.
DISSECT addresses a key bottleneck in spatial transcriptomics—accurate cell segmentation—by combining cytological image features with transcriptomic profiles in a deep generative model that outperforms existing algorithms across heterogeneous tissues. Applied to gastric cancer samples before and after anti-PD-1 immunotherapy, it demonstrated utility for characterising spatial immune microenvironment changes relevant to treatment response.
What the study was
- Study design
- computational_validation_applied_study
- Category
- ai_ml_diagnostics
- Maturity
- Validated
- Journal
- Nat Comput Sci
Why it surfaced
Nature Computational Science; technically novel integration of image and transcriptome modalities for cell segmentation; gastric cancer application with anti-PD-1 context links to precision oncology and immunotherapy monitoring; population/unmet_need and clinical_applicability capped as a research methodology paper.
A plain-language summary of published research — not medical advice. Talk to a clinician about your own care.