Computational pathology^1,2 has witnessed considerable progress in the development of both task-specific predictive models and task-agnostic self-supervised vision encoders^3,4. However, despite the explosive growth of generative artificial intelligence (AI), there have been few studies on building general-purpose multimodal AI assistants and copilots⁵ tailored to pathology. Here we present PathChat, a vision-language generalist AI assistant for human pathology. We built PathChat by adapting a foundational vision encoder for pathology, combining it with a pretrained large language model and fine-tuning the whole system on over 456,000 diverse visual-language instructions consisting of 999,202 question and answer turns. We compare PathChat with several multimodal vision-language AI assistants and GPT-4V, which powers the commercially available multimodal general-purpose AI assistant ChatGPT-4 (ref. ⁶). PathChat achieved state-of-the-art performance on multiple-choice diagnostic questions from cases with diverse tissue origins and disease models. Furthermore, using open-ended questions and human expert evaluation, we found that overall PathChat produced more accurate and pathologist-preferable responses to diverse queries related to pathology. As an interactive vision-language AI copilot that can flexibly handle both visual and natural language inputs, PathChat may potentially find impactful applications in pathology education, research and human-in-the-loop clinical decision-making.

Fine-needle aspiration cytology (FNAC) has long been recognized as a minimally invasive, cost-effective, and reliable diagnostic tool for breast lesions. However, with the advent of core-needle biopsy (CNB), the role of FNAC has diminished in some clinical settings. This review aims to re-evaluate the diagnostic value of FNAC in the current era, focusing on its complementary use alongside CNB, the adoption of new approaches such as the International Academy of Cytology Yokohama System, and the implementation of rapid on-site evaluation to reduce inadequate sample rates. Advances in liquid-based cytology, receptor expression testing, molecular diagnostics, and artificial intelligence are discussed, highlighting their potential to enhance the diagnostic accuracy of FNAC. Despite challenges, FNAC remains a valuable diagnostic method, particularly in low-resource settings and specific clinical scenarios, and its role continues to evolve with technology.

Ahn S, et al. J Breast Cancer. 2021 Jun;24(3):266-279. https://doi.org/10.4048/jbc.2021.24.e29

Background: High-grade serous carcinoma (HGSC) of the ovary is the most common subtype of epithelial ovarian cancer (EOC) and has an overall poor prognosis. There is increasing awareness of the importance of immune cell populations and tumor-infiltrating lymphocytes (TILs) in various immune pathways in the tumor microenvironment. The present study evaluated immune-related gene expressions and TIL levels, as well as associated chemotherapeutic responses, to elucidate the correlation between gene expression and TIL levels in HGSC. Materials and methods: Fresh tissue samples from 12 HGSC patients were included in this study. Depending on their response to adjuvant chemotherapy, the patients were divided into two groups: chemosensitive (CS) or chemoresistant (CR). The expression levels of 770 genes were analyzed using the nCounter® PanCancer Immune Profiling Panel of the NanoString nCounter® Analysis System. Quantitative real-time polymerase chain reaction (qPCR) was performed to validate the NanoString data obtained. The TIL levels in representative sections were examined via hematoxylin and eosin staining. Gene and TIL levels were subsequently correlated with the chemotherapeutic response. Results: Several genes were differentially expressed in the two study groups. Eleven representative genes were selected for further evaluation. Of those, 9 genes (IRF1, CXCL9, LTB, CCL5, IL-8, GZMA, PSMB9, CD38, and VCAM1) were significantly overexpressed in the CS group; whereas expressions of 2 genes (CD24 and CD164) were increased in the CR group. Results of qPCR were consistent with those of the NanoString nCounter® analysis. Stromal TIL levels were significantly associated with adjuvant chemotherapeutic response (p = 0.001). Conclusions: Significant differences between the CS and CR groups were observed in the expression levels of immune-related genes. Immune-related gene expressions were significantly higher in the CS group, which also had higher levels of TILs. We, therefore, suggest that, in patients with HGSC, immune-related gene expressions and TIL levels may be associated with chemotherapeutic sensitivity.

The prognostic significance of tumor-infiltrating lymphocytes (TILs) has been determined in breast cancers. Interferons can affect T-cell activity through direct and indirect mechanisms. Myxovirus resistance A (MxA) is an excellent marker of interferon activity. Here,we evaluated TILs and MxA expression in human epidermal growth factor receptor 2 (HER2)–positive breast cancers. Ninety cases of hormone receptor (HR)+/HER2+ tumors and 78 cases of HR–/HER2+ tumors were included. The TILs level was assessed using hematoxylin and eosin–stained full face sections, and MxA expressionwas evaluated by immunohistochemistrywith a tissue microarray. MxA protein expression was significantly higher in tumors with high histologic grade (p=0.023) and high levels of TILs (p=0.002). High levels of TILs were correlated with high histological grade (p=0.001), negative lymphovascular invasion (p=0.007), negative lymph node metastasis (p=0.007), absence of HR expression (p < 0.001), abundant tertiary lymphoid structures (TLSs) around ductal carcinoma in situ (p=0.018), and abundant TLSs around the invasive component (p < 0.001). High levels of TILs were also associated with improved disease-free survival, particularly in HR–/HER2+ breast cancers. However, MxA was not a prognostic factor. High expression of MxA in tumor cells was associated with high levels of TILs in HER2-positive breast cancers. Additionally, a high level of TILs was a prognostic factor for breast cancer, whereas the level of MxA expression had no prognostic value.

Soft tissue sarcomas (STS) are rare. We evaluated the WT1 protein expression level in various types of STS and elucidated the value of WT1 as a prognostic factor and a possible therapeutic target.