Artificial Intelligence Meets Whole Slide Images: Deep Learning Model Shapes an Immune-Hot Tumor and Guides Precision Therapy in Bladder Cancer

Abstract

Background. To construct and validate a deep learning cluster from whole slide images (WSI) for depicting the immunophenotypes and functional heterogeneity of the tumor microenvironment (TME) in patients with bladder cancer (BLCA) and to explore an artificial intelligence (AI) score to explore the underlying biological pathways in the developed WSI cluster. Methods. In this study, the WSI cluster was constructed based on a deep learning procedure. Further rerecognition of TME features in pathological images was applied based on a neural network. Then, we integrated the TCGA cohort and several external testing cohorts to explore and validate this novel WSI cluster and a corresponding quantitative indicator, the AI score. Finally, correlations between the AI cluster (AI score) and classical BLCA molecular subtypes, immunophenotypes, functional heterogeneity, and potential therapeutic method in BLCA were assessed. Results. The WSI cluster was identified associated with clinical survival ( P < 0.001 ) and was proved as an independent predictor ( P = 0.031 ), which could also predict the immunology and the clinical significance of BLCA. Rerecognition of pathological images established a robust 3-year survival prediction model (with an average classification accuracy of 86%, AUC of 0.95) for BLCA patients combining TME features and clinical features. In addition, an AI score was constructed to quantify the underlying logic of the WSI cluster ( AUC = 0.838 ). Finally, we hypothesized that high AI score shapes an immune-hot TME in BLCA. Thus, treatment options including immune checkpoint blockade (ICB), chemotherapy, and ERBB therapy can be used for the treatment of BLCA patients in WSI cluster1 (high AI score subtype). Conclusions. In general, we showed that deep learning can predict prognosis and may aid in the precision medicine for BLCA directly from H&E histology, which is more economical and efficient.

Recent Advancements in Deep Learning Using Whole Slide Imaging for Cancer Prognosis

“…A considerable number of investigations have also employed variations of CNN, often categorized under the umbrella term of general deep learning. This group includes diverse studies by Liang et al [ 55 ], Zheng et al [ 49 ], and Jiang et al [ 48 ]. Residual networks, a form of CNN, have been frequently implemented, as seen in works by Liu et al [ 69 ], Wu et al [ 72 ], and Knuutila et al [ 76 ], who utilized the ResNet model due to its ability to effectively train very deep neural networks.…”

“…A primary challenge is the variance in data quality, which is an issue pervasive across all medical imaging analyses [ 48 , 70 ]. A broad range of factors including staining inconsistencies, scanning variations, and differences in slide preparation methodologies can induce a significant level of noise in the data, posing difficulties for the robustness of deep learning models.…”

“…Attention to the tumor microenvironment (TME) is a common theme among several studies. Jiang et al [ 48 ] and Jiao et al [ 31 ] utilized CNNs to assess the TME in bladder cancer and colon adenocarcinoma, respectively. Liang et al [ 55 ] introduced PathFinder, a deep learning framework that underscored the prognostic significance of the necrotic spatial distribution in liver cancer.…”

Machine learning to comprehensively reveal signature genes and regulatory mechanisms in pituitary tumors

“…This report is inconsistent with our results, which needs to be con rmed by subsequent studies. This may be the reason for the small sample size in GEO database.It has been reported that the degree of in ltration of suppressive immune cells in the tumor microenvironment, lifting the immunosuppressive tumor microenvironment, turning "cold tumor" into "hot tumor" to enhance anti-tumor immunity [13,24]. In addition, we investigated the relationship between the expression of key genes and immune cells in pituitary tumors, which may provide new clinical guidance for cancer diagnosis.…”

Molecular classification of urothelial bladder carcinoma

“…The critical role of precise identification of MIBC patients who are most likely to achieve optimal outcomes with neoadjuvant checkpoint inhibitor therapy, either as monotherapy or in conjunction with NAC, cannot be overstated in guiding future treatment approaches. Given the enormous amount of information in recent years, using “multi-omics” analysis, differences in intratumor environment, and heterogeneity, artificial intelligence may appear to be a “game changer” to model and tailor treatment in the near future [ 64 , 65 ].The molecular classification of UC has not yet been officially approved by any oncology society and remains unstandardized, as several groups have created nomenclatures that may be confusing and unsuitable for clinical utilization. There is a need for consensus that takes into account molecular, biological, clinical, and prognostic features of tumors, with close cooperation between expert societies (e.g., EAU, AUA) based on well-designed clinical trials and evidence-based medicine.…”