Colorectal cancer recurrence after surgery gets a spatial host-microbiome readout: AUC 0.81 versus 0.45 for standard pathology
Synopsis
Whether colorectal cancer recurs after resection has been judged mainly from routine pathology such as lymphovascular invasion, perineural invasion, tumor budding and Ki-67. AlphaFISH pairs a sequencing-free nanoprobe biochip with a transformer model to read 41 host transcripts and 51 bacterial genera across 149 colorectal biopsies, and in 28 patients followed for one year after surgery it predicted recurrence or metastasis with an AUC of 0.81, against 0.45 for the combined routine clinical markers. The result comes from a preprint that has not been peer reviewed, on a limited sample.
Interpretation
Recurrence prediction after colorectal cancer surgery gains a signal from tissue spatial structure: where host gene expression and bacteria sit relative to each other in the same biopsy is used to judge whether a patient will recur or metastasize within a year. That judgment previously rested on a combination of routine clinical markers taken at surgery, namely lymphovascular invasion, perineural invasion, tumor budding and the Ki-67 index. In 28 colorectal cancer patients followed for one year after resection, AlphaFISH predicted recurrence or metastasis with an AUC of 0.81, while the combined routine markers reached an AUC of 0.45.
The prediction rests on a workflow that presses tissue directly onto a chip to read RNA, capturing host mRNA and bacterial 16S rRNA at the same location. Sequencing-based spatial platforms typically use poly(T) probes to capture eukaryotic mRNA, while bacterial RNA is hard to detect reliably because of cell walls, low abundance and confusion with dominant host RNA. A 1×1 mm² chip carries about 60,000 nanoprobes of 0.8 µm diameter at 4 µm pitch, each functionalized with polyT and 16S bait sequences; captured targets are decoded by 7-channel confocal imaging, and the platform produced spatial transcriptomic and spatial microbiome data across 149 colorectal biopsies (51 tumor, 51 adjacent non-tumor, 47 non-cancer controls).
Perspective
The demonstrated scope is 149 colorectal biopsies from 68 subjects, including paired tumor and adjacent non-tumor tissue from 51 colorectal cancer patients plus Crohn's disease and healthy controls; recurrence prediction was evaluated only in 28 colorectal cancer patients followed for one year after surgery. The intended readers are researchers and engineers working on spatial omics and tumor prognosis modeling.
The recurrence evaluation covers 28 patients over one year of follow-up, which the authors describe as a prospective longitudinal analysis; the preprint states explicitly that the results have not been peer reviewed and should not guide clinical practice. The authors also note that the current panel covers only 41 host transcripts and 51 bacterial genera, that labels rely mainly on TNM stage, and that the model is constrained by the information carried by the preselected molecular anchors.