The invention discloses a
PARP inhibitor sensitivity prediction method based on a specific
cell cluster and application. The method comprises the following steps: acquiring an
ovarian cancer tumor tissue sample of a subject; detecting whether a specific
cell cluster exists in the sample or not, and determining the abundance of the specific
cell cluster, wherein the specific
cell cluster comprises an IFN epithelial
cell cluster CN12 taking an IFN tumor cell as a center and / or a TAM enrichment cluster CN29 taking an M2 type tumor related macrophage as a center; according to the detection result, predicting according to at least one of the following rules: if the CN12 abundance before treatment is greater than 0.8%, the CN29 abundance before treatment is greater than 32.7%, the CN12 abundance
after treatment is greater than 0.009%, the CN29 abundance
after treatment is greater than 2.9%, or the SPP1
cell density before treatment is greater than 1.7%, predicting that the
drug resistance is not sensitive or easily generated. The prediction model is established from a
cell cluster space
tissue level, and the method has the advantages of high prediction accuracy, flexible sample requirements, high
operability and the like.