Search papers, labs, and topics across Lattice.
This study introduces Loki-OT, a novel method that enhances cell classification by integrating region-level tissue reasoning through Unbalanced Optimal Transport, addressing the challenge of cell mimicry in pathology. By leveraging MLLM-derived density priors, Loki-OT effectively reassigns ambiguous cells while utilizing the discriminative features of pretrained cell foundation models. The method outperforms existing classifiers, achieving lower patient-level mean absolute error and improved F1 scores in complex tissue environments, demonstrating its potential for more accurate pathological assessments.
Loki-OT surpasses traditional classifiers by achieving superior accuracy in distinguishing morphologically similar lymphocytes through innovative tissue context reasoning.
Cell mimicry arises when different cell types appear morphologically similar. Human pathologists resolve this ambiguity using surrounding tissue context, whereas current vision models either lack contextual reasoning (cell foundation models) or cannot operate at the cell level (pathology MLLMs). We present Loki-OT, which propagates region-level tissue reasoning to individual cell predictions via Unbalanced Optimal Transport, using MLLM-derived density priors as soft guidance for ambiguous cell reassignment. Loki-OT is motivated by the observation that pretrained cell foundation model features already encode discriminative information, including tissue context, but standard cell-level supervision fails to use tissue context effectively. The resulting transport plan is distilled into a lightweight student MLP classifier that learns context-aware decision boundaries within the pretrained feature space. On the independent TCGA-BRCA cohort, Loki-OT achieved lower patient-level MAE than the fully supervised in-domain PanopTILs classifier and improved F1 in epithelium-rich mimicry tissues, using 278 weak region-level MLLM estimates built on a general-domain cell foundation model. Code: https://github.com/xiangli980/Lymphocyte_Mimicry_Correction_via_Loki_OT
