Deep learning on tertiary lymphoid structures in hematoxylin-eosin predicts cancer prognosis and immunotherapy response
Abstract Tertiary lymphoid structures (TLSs) have been associated with favorable immunotherapy responses and prognosis in various cancers. Despite their significance, their quantification using multiplex immunohistochemistry (mIHC) staining of T and B lymphocytes remains labor-intensive, limiting it...
| Main Authors: | , , , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2024-03-01
|
| Series: | npj Precision Oncology |
| Online Access: | https://doi.org/10.1038/s41698-024-00579-w |
| _version_ | 1797247486728863744 |
|---|---|
| author | Ziqiang Chen Xiaobing Wang Zelin Jin Bosen Li Dongxian Jiang Yanqiu Wang Mengping Jiang Dandan Zhang Pei Yuan Yahui Zhao Feiyue Feng Yicheng Lin Liping Jiang Chenxi Wang Weida Meng Wenjing Ye Jie Wang Wenqing Qiu Houbao Liu Dan Huang Yingyong Hou Xuefei Wang Yuchen Jiao Jianming Ying Zhihua Liu Yun Liu |
| author_facet | Ziqiang Chen Xiaobing Wang Zelin Jin Bosen Li Dongxian Jiang Yanqiu Wang Mengping Jiang Dandan Zhang Pei Yuan Yahui Zhao Feiyue Feng Yicheng Lin Liping Jiang Chenxi Wang Weida Meng Wenjing Ye Jie Wang Wenqing Qiu Houbao Liu Dan Huang Yingyong Hou Xuefei Wang Yuchen Jiao Jianming Ying Zhihua Liu Yun Liu |
| author_sort | Ziqiang Chen |
| collection | DOAJ |
| description | Abstract Tertiary lymphoid structures (TLSs) have been associated with favorable immunotherapy responses and prognosis in various cancers. Despite their significance, their quantification using multiplex immunohistochemistry (mIHC) staining of T and B lymphocytes remains labor-intensive, limiting its clinical utility. To address this challenge, we curated a dataset from matched mIHC and H&E whole-slide images (WSIs) and developed a deep learning model for automated segmentation of TLSs. The model achieved Dice coefficients of 0.91 on the internal test set and 0.866 on the external validation set, along with intersection over union (IoU) scores of 0.819 and 0.787, respectively. The TLS ratio, defined as the segmented TLS area over the total tissue area, correlated with B lymphocyte levels and the expression of CXCL13, a chemokine associated with TLS formation, in 6140 patients spanning 16 tumor types from The Cancer Genome Atlas (TCGA). The prognostic models for overall survival indicated that the inclusion of the TLS ratio with TNM staging significantly enhanced the models’ discriminative ability, outperforming the traditional models that solely incorporated TNM staging, in 10 out of 15 TCGA tumor types. Furthermore, when applied to biopsied treatment-naïve tumor samples, higher TLS ratios predicted a positive immunotherapy response across multiple cohorts, including specific therapies for esophageal squamous cell carcinoma, non-small cell lung cancer, and stomach adenocarcinoma. In conclusion, our deep learning-based approach offers an automated and reproducible method for TLS segmentation and quantification, highlighting its potential in predicting immunotherapy response and informing cancer prognosis. |
| first_indexed | 2024-04-24T19:59:28Z |
| format | Article |
| id | doaj.art-a91d33dd21914b1dba5d286047bfe39c |
| institution | Directory Open Access Journal |
| issn | 2397-768X |
| language | English |
| last_indexed | 2024-04-24T19:59:28Z |
| publishDate | 2024-03-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | npj Precision Oncology |
| spelling | doaj.art-a91d33dd21914b1dba5d286047bfe39c2024-03-24T12:09:15ZengNature Portfolionpj Precision Oncology2397-768X2024-03-018111110.1038/s41698-024-00579-wDeep learning on tertiary lymphoid structures in hematoxylin-eosin predicts cancer prognosis and immunotherapy responseZiqiang Chen0Xiaobing Wang1Zelin Jin2Bosen Li3Dongxian Jiang4Yanqiu Wang5Mengping Jiang6Dandan Zhang7Pei Yuan8Yahui Zhao9Feiyue Feng10Yicheng Lin11Liping Jiang12Chenxi Wang13Weida Meng14Wenjing Ye15Jie Wang16Wenqing Qiu17Houbao Liu18Dan Huang19Yingyong Hou20Xuefei Wang21Yuchen Jiao22Jianming Ying23Zhihua Liu24Yun Liu25MOE Key Laboratory of Metabolism and Molecular Medicine, Department of Biochemistry and Molecular Biology, School of Basic Medical Sciences and Shanghai Xuhui Central Hospital, Fudan UniversityState Key Laboratory of Molecular Oncology, National Cancer Center, National Clinical Research Center for Cancer, Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical CollegeMOE Key Laboratory of Metabolism and Molecular Medicine, Department of Biochemistry and Molecular Biology, School of Basic Medical Sciences and Shanghai Xuhui Central Hospital, Fudan UniversityDepartment of General Surgery/Gastric Cancer Center, Zhongshan Hospital, Fudan UniversityDepartment of Pathology, Zhongshan Hospital, Fudan UniversityDepartments of Pathology, International Peace Maternity and Child Health Hospital Affiliated to Shanghai Jiao Tong University School of MedicineMOE Key Laboratory of Metabolism and Molecular Medicine, Department of Biochemistry and Molecular Biology, School of Basic Medical Sciences and Shanghai Xuhui Central Hospital, Fudan UniversityMOE Key Laboratory of Metabolism and Molecular Medicine, Department of Biochemistry and Molecular Biology, School of Basic Medical Sciences and Shanghai Xuhui Central Hospital, Fudan UniversityDepartment of Pathology, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical CollegeState Key Laboratory of Molecular Oncology, National Cancer Center, National Clinical Research Center for Cancer, Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical CollegeThoracic Surgery Department, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical CollegeMOE Key Laboratory of Metabolism and Molecular Medicine, Department of Biochemistry and Molecular Biology, School of Basic Medical Sciences and Shanghai Xuhui Central Hospital, Fudan UniversityState Key Laboratory of Molecular Oncology, National Cancer Center, National Clinical Research Center for Cancer, Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical CollegeState Key Laboratory of Molecular Oncology, National Cancer Center, National Clinical Research Center for Cancer, Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical CollegeMOE Key Laboratory of Metabolism and Molecular Medicine, Department of Biochemistry and Molecular Biology, School of Basic Medical Sciences and Shanghai Xuhui Central Hospital, Fudan UniversityDivision of Rheumatology and Immunology, Huashan Hospital, Fudan UniversityDepartments of Thoracic Surgery, Fudan University Shanghai Cancer CenterShanghai Xuhui Central HospitalShanghai Xuhui Central HospitalDepartment of Pathology, Fudan University Shanghai Cancer Center, Shanghai, China; Department of Oncology, Shanghai Medical College, Fudan University, Shanghai, China; Institute of Pathology, Fudan UniversityDepartment of Pathology, Zhongshan Hospital, Fudan UniversityDepartment of General Surgery/Gastric Cancer Center, Zhongshan Hospital, Fudan UniversityState Key Laboratory of Molecular Oncology, National Cancer Center, National Clinical Research Center for Cancer, Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical CollegeDepartment of Pathology, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical CollegeState Key Laboratory of Molecular Oncology, National Cancer Center, National Clinical Research Center for Cancer, Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical CollegeMOE Key Laboratory of Metabolism and Molecular Medicine, Department of Biochemistry and Molecular Biology, School of Basic Medical Sciences and Shanghai Xuhui Central Hospital, Fudan UniversityAbstract Tertiary lymphoid structures (TLSs) have been associated with favorable immunotherapy responses and prognosis in various cancers. Despite their significance, their quantification using multiplex immunohistochemistry (mIHC) staining of T and B lymphocytes remains labor-intensive, limiting its clinical utility. To address this challenge, we curated a dataset from matched mIHC and H&E whole-slide images (WSIs) and developed a deep learning model for automated segmentation of TLSs. The model achieved Dice coefficients of 0.91 on the internal test set and 0.866 on the external validation set, along with intersection over union (IoU) scores of 0.819 and 0.787, respectively. The TLS ratio, defined as the segmented TLS area over the total tissue area, correlated with B lymphocyte levels and the expression of CXCL13, a chemokine associated with TLS formation, in 6140 patients spanning 16 tumor types from The Cancer Genome Atlas (TCGA). The prognostic models for overall survival indicated that the inclusion of the TLS ratio with TNM staging significantly enhanced the models’ discriminative ability, outperforming the traditional models that solely incorporated TNM staging, in 10 out of 15 TCGA tumor types. Furthermore, when applied to biopsied treatment-naïve tumor samples, higher TLS ratios predicted a positive immunotherapy response across multiple cohorts, including specific therapies for esophageal squamous cell carcinoma, non-small cell lung cancer, and stomach adenocarcinoma. In conclusion, our deep learning-based approach offers an automated and reproducible method for TLS segmentation and quantification, highlighting its potential in predicting immunotherapy response and informing cancer prognosis.https://doi.org/10.1038/s41698-024-00579-w |
| spellingShingle | Ziqiang Chen Xiaobing Wang Zelin Jin Bosen Li Dongxian Jiang Yanqiu Wang Mengping Jiang Dandan Zhang Pei Yuan Yahui Zhao Feiyue Feng Yicheng Lin Liping Jiang Chenxi Wang Weida Meng Wenjing Ye Jie Wang Wenqing Qiu Houbao Liu Dan Huang Yingyong Hou Xuefei Wang Yuchen Jiao Jianming Ying Zhihua Liu Yun Liu Deep learning on tertiary lymphoid structures in hematoxylin-eosin predicts cancer prognosis and immunotherapy response npj Precision Oncology |
| title | Deep learning on tertiary lymphoid structures in hematoxylin-eosin predicts cancer prognosis and immunotherapy response |
| title_full | Deep learning on tertiary lymphoid structures in hematoxylin-eosin predicts cancer prognosis and immunotherapy response |
| title_fullStr | Deep learning on tertiary lymphoid structures in hematoxylin-eosin predicts cancer prognosis and immunotherapy response |
| title_full_unstemmed | Deep learning on tertiary lymphoid structures in hematoxylin-eosin predicts cancer prognosis and immunotherapy response |
| title_short | Deep learning on tertiary lymphoid structures in hematoxylin-eosin predicts cancer prognosis and immunotherapy response |
| title_sort | deep learning on tertiary lymphoid structures in hematoxylin eosin predicts cancer prognosis and immunotherapy response |
| url | https://doi.org/10.1038/s41698-024-00579-w |
| work_keys_str_mv | AT ziqiangchen deeplearningontertiarylymphoidstructuresinhematoxylineosinpredictscancerprognosisandimmunotherapyresponse AT xiaobingwang deeplearningontertiarylymphoidstructuresinhematoxylineosinpredictscancerprognosisandimmunotherapyresponse AT zelinjin deeplearningontertiarylymphoidstructuresinhematoxylineosinpredictscancerprognosisandimmunotherapyresponse AT bosenli deeplearningontertiarylymphoidstructuresinhematoxylineosinpredictscancerprognosisandimmunotherapyresponse AT dongxianjiang deeplearningontertiarylymphoidstructuresinhematoxylineosinpredictscancerprognosisandimmunotherapyresponse AT yanqiuwang deeplearningontertiarylymphoidstructuresinhematoxylineosinpredictscancerprognosisandimmunotherapyresponse AT mengpingjiang deeplearningontertiarylymphoidstructuresinhematoxylineosinpredictscancerprognosisandimmunotherapyresponse AT dandanzhang deeplearningontertiarylymphoidstructuresinhematoxylineosinpredictscancerprognosisandimmunotherapyresponse AT peiyuan deeplearningontertiarylymphoidstructuresinhematoxylineosinpredictscancerprognosisandimmunotherapyresponse AT yahuizhao deeplearningontertiarylymphoidstructuresinhematoxylineosinpredictscancerprognosisandimmunotherapyresponse AT feiyuefeng deeplearningontertiarylymphoidstructuresinhematoxylineosinpredictscancerprognosisandimmunotherapyresponse AT yichenglin deeplearningontertiarylymphoidstructuresinhematoxylineosinpredictscancerprognosisandimmunotherapyresponse AT lipingjiang deeplearningontertiarylymphoidstructuresinhematoxylineosinpredictscancerprognosisandimmunotherapyresponse AT chenxiwang deeplearningontertiarylymphoidstructuresinhematoxylineosinpredictscancerprognosisandimmunotherapyresponse AT weidameng deeplearningontertiarylymphoidstructuresinhematoxylineosinpredictscancerprognosisandimmunotherapyresponse AT wenjingye deeplearningontertiarylymphoidstructuresinhematoxylineosinpredictscancerprognosisandimmunotherapyresponse AT jiewang deeplearningontertiarylymphoidstructuresinhematoxylineosinpredictscancerprognosisandimmunotherapyresponse AT wenqingqiu deeplearningontertiarylymphoidstructuresinhematoxylineosinpredictscancerprognosisandimmunotherapyresponse AT houbaoliu deeplearningontertiarylymphoidstructuresinhematoxylineosinpredictscancerprognosisandimmunotherapyresponse AT danhuang deeplearningontertiarylymphoidstructuresinhematoxylineosinpredictscancerprognosisandimmunotherapyresponse AT yingyonghou deeplearningontertiarylymphoidstructuresinhematoxylineosinpredictscancerprognosisandimmunotherapyresponse AT xuefeiwang deeplearningontertiarylymphoidstructuresinhematoxylineosinpredictscancerprognosisandimmunotherapyresponse AT yuchenjiao deeplearningontertiarylymphoidstructuresinhematoxylineosinpredictscancerprognosisandimmunotherapyresponse AT jianmingying deeplearningontertiarylymphoidstructuresinhematoxylineosinpredictscancerprognosisandimmunotherapyresponse AT zhihualiu deeplearningontertiarylymphoidstructuresinhematoxylineosinpredictscancerprognosisandimmunotherapyresponse AT yunliu deeplearningontertiarylymphoidstructuresinhematoxylineosinpredictscancerprognosisandimmunotherapyresponse |