1
|
Hezi H, Gelber M, Balabanov A, Maruvka YE, Freiman M. CIMIL-CRC: A clinically-informed multiple instance learning framework for patient-level colorectal cancer molecular subtypes classification from H&E stained images. COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE 2025; 259:108513. [PMID: 39581068 DOI: 10.1016/j.cmpb.2024.108513] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Received: 06/21/2024] [Revised: 11/02/2024] [Accepted: 11/09/2024] [Indexed: 11/26/2024]
Abstract
BACKGROUND AND OBJECTIVE Treatment approaches for colorectal cancer (CRC) are highly dependent on the molecular subtype, as immunotherapy has shown efficacy in cases with microsatellite instability (MSI) but is ineffective for the microsatellite stable (MSS) subtype. There is promising potential in utilizing deep neural networks (DNNs) to automate the differentiation of CRC subtypes by analyzing hematoxylin and eosin (H&E) stained whole-slide images (WSIs). Due to the extensive size of WSIs, multiple instance learning (MIL) techniques are typically explored. However, existing MIL methods focus on identifying the most representative image patches for classification, which may result in the loss of critical information. Additionally, these methods often overlook clinically relevant information, like the tendency for MSI class tumors to predominantly occur on the proximal (right side) colon. METHODS We introduce 'CIMIL-CRC', a DNN framework that: (1) solves the MSI/MSS MIL problem by efficiently combining a pre-trained feature extraction model with principal component analysis (PCA) to aggregate information from all patches, and (2) integrates clinical priors, particularly the tumor location within the colon, into the model to enhance patient-level classification accuracy. We assessed our CIMIL-CRC method using the average area under the receiver operating characteristic curve (AUROC) from a 5-fold cross-validation experimental setup for model development on the TCGA-CRC-DX cohort, contrasting it with a baseline patch-level classification, a MIL-only approach, and a clinically-informed patch-level classification approach. RESULTS Our CIMIL-CRC outperformed all methods (AUROC: 0.92±0.002 (95% CI 0.91-0.92), vs. 0.79±0.02 (95% CI 0.76-0.82), 0.86±0.01 (95% CI 0.85-0.88), and 0.87±0.01 (95% CI 0.86-0.88), respectively). The improvement was statistically significant. To the best of our knowledge, this is the best result achieved for MSI/MSS classification on this dataset. CONCLUSION Our CIMIL-CRC method holds promise for offering insights into the key representations of histopathological images and suggests a straightforward implementation.
Collapse
Affiliation(s)
- Hadar Hezi
- Faculty of Biomedical Engineering, Technion - Israel Institute of Technology, Haifa, Israel
| | - Matan Gelber
- Faculty of Electrical and Computer Engineering, Technion - Israel Institute of Technology, Haifa, Israel
| | - Alexander Balabanov
- Faculty of Electrical and Computer Engineering, Technion - Israel Institute of Technology, Haifa, Israel
| | - Yosef E Maruvka
- Faculty of Food Engineering and Biotechnology, Technion - Israel Institute of Technology, Haifa, Israel
| | - Moti Freiman
- Faculty of Biomedical Engineering, Technion - Israel Institute of Technology, Haifa, Israel.
| |
Collapse
|
2
|
Xiao G, Li J, Deng L, Gao S, Tan C, He G, Du R. Microsatellite instability evaluation by a novel PCR-based 8-loci test kit in colorectal cancer. Biotechnol Appl Biochem 2024; 71:860-867. [PMID: 38556769 DOI: 10.1002/bab.2582] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/08/2023] [Accepted: 03/18/2024] [Indexed: 04/02/2024]
Abstract
Microsatellite instability (MSI) assessment is strongly recommended for colorectal cancer patients, as MSI status is crucial in determining optimal treatment and predicting prognosis. This study evaluated the reliability and accuracy of a novel polymerase chain reaction (PCR)-based 8-loci MSI test kit, a rapid test kit designed to detect MSI, by comparing its performance with immunohistochemistry (IHC) and the National Cancer Institute (NCI) 2B3D Panel. MSI status was determined in 186 formalin-fixed paraffin-embedded (FFPE) colorectal cancer tissue samples with known mismatch repair (MMR) status by IHC using the novel PCR-based 8-loci MSI test kit. Additionally, the consistency between the NCI 2B3D Panel and the novel PCR-based 8-loci panel was compared using 69 FFPE tumor tissues paired with adjacent non-cancerous tissue. The novel PCR-based 8-loci MSI test kit and IHC demonstrated high concordance (overall agreement: 97.8%). However, four samples displayed discordant results, exhibiting MMR deficiency using IHC and microsatellite stability using the novel PCR-based 8-loci MSI test kit. Of the 69 samples reanalyzed using the NCI 2B3D Panel, high concordance with the novel PCR-based 8-loci MSI test kit was observed in 67 of 69 cases (overall agreement: 97.1%). The novel PCR-based 8-loci MSI test kit is a rapid and reliable tool for accurately detecting MSI status in colorectal cancer.
Collapse
Affiliation(s)
- Gaofang Xiao
- Department of Pathology, Shantou University Affiliated Yuebei People's Hospital, Shaoguan, Guangdong, China
| | - Jing Li
- Department of Pathology, Shantou University Affiliated Yuebei People's Hospital, Shaoguan, Guangdong, China
| | - Lijun Deng
- Department of Medical Engineering, Shantou University Affiliated Yuebei People's Hospital, Shaoguan, Guangdong, China
| | - Shuangquan Gao
- Department of Pathology, Shantou University Affiliated Yuebei People's Hospital, Shaoguan, Guangdong, China
| | - Caiyun Tan
- Department of Pathology, Shantou University Affiliated Yuebei People's Hospital, Shaoguan, Guangdong, China
| | - Guiqing He
- Department of Pathology, Shantou University Affiliated Yuebei People's Hospital, Shaoguan, Guangdong, China
| | - Richang Du
- Department of Pathology, Shantou University Affiliated Yuebei People's Hospital, Shaoguan, Guangdong, China
| |
Collapse
|
3
|
Li X, Zhang S, Zeng J, Song SS, Liu X, Kang W, Liang M, Yang R, Li H, Liang L. Heterogeneous expression of mismatch repair proteins and interpretation of immunohistochemical results in colorectal cancer and endometrial cancer. Pathol Res Pract 2023; 248:154647. [PMID: 37437501 DOI: 10.1016/j.prp.2023.154647] [Citation(s) in RCA: 2] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 02/04/2023] [Revised: 06/18/2023] [Accepted: 06/25/2023] [Indexed: 07/14/2023]
Abstract
To investigate the heterogeneous expression patterns of four mismatch repair (MMR) proteins in colorectal cancer (CRC) and endometrial cancer (EC), and their effects on the interpretation of immunohistochemical (IHC) results. A total of 1636 CRC and EC specimens were collected from two hospitals. Seventy-eight cases had heterogeneous expression of MMR proteins, including 49 CRC and 29 EC cases. Polymerase chain reaction-capillary electrophoresis (PCR-CE) was then performed to detect the microsatellite instability (MSI) status, and 44 cases were further verified by targeted next-generation sequencing (NGS). Heterogeneous expression of MMR proteins was observed in 66 cases (66/78, 84.6%) of proficient MMR (pMMR) and 12 cases (12/78, 15.4%) of deficient MMR (dMMR). The proportion of heterogeneous MMR protein expression in EC (12.0%) was higher than that in CRC (3.5%). The heterogeneous expression patterns were divided into focal clonal heterogeneity (6/78, 7.7%) and glandular mosaic heterogeneity (72/78, 92.3%). Surprisingly, three pMMR CRC cases showed isolated small focal clonal heterogeneity of mutS homologue 6 (MSH6), with < 15% positive tumour cells, which was validated as high MSI (MSI-H) with PCR-CE and NGS. However, the other three EC pMMR cases with > 50% focal clonal heterogeneity of MMR proteins were verified as microsatellite stable (MSS) or low MSI (MSI-L). Fifteen EC cases with glandular mosaic heterogeneous expression of MMR proteins included two MSI-H cases, which were validated using PCR-CE and NGS. Among the dMMR cases, only two EC cases with mutL homologue 1 (MLH1)/PMS1 homologue 2, mismatch repair system component (PMS2) loss and MSH2/MSH6 mosaic heterogeneous expression were confirmed as MSS using PCR-CE and NGS, which may be related to the mechanism of MLH1 promoter methylation. Thus, in CRC, only cases with small focal clonal heterogeneous expression of MSH6 have a high likelihood of MSI-H, and further PCR-CE or NGS testing is recommended. The possibility of MSI-H cannot be ruled out in EC cases with glandular mosaic heterogeneous expression of MMR proteins; PCR-CE or NGS detection is therefore necessary.
Collapse
Affiliation(s)
- Xiangzhao Li
- Department of Pathology, Nanfang Hospital/Basic Medical College, Southern Medical University, Guangzhou 510515, Guangdong Province, People's Republic of China; Guangdong Province Key Laboratory of Molecular Tumor Pathology, Guangzhou 510515, Guangdong Province, People's Republic of China
| | - Shifen Zhang
- Department of Pathology, Nanfang Hospital/Basic Medical College, Southern Medical University, Guangzhou 510515, Guangdong Province, People's Republic of China; Department of Pathology, Shenzhen People's Hospital (The Second Clinical Medical College, Jinan University; The First Affiliated Hospital, Southern University of Science and Technology), Shenzhen 518020, Guangdong, People's Republic of China
| | - Jiamin Zeng
- Department of Pathology, Nanfang Hospital/Basic Medical College, Southern Medical University, Guangzhou 510515, Guangdong Province, People's Republic of China; Department of Pathology, The Second Affiliated Hospital of Guangzhou Medical University, Guangzhou 510260, People's Republic of China
| | - Sha-Sha Song
- Department of Pathology, Yantai Fushan People's Hospital, Yantai 265500, Shandong, People's Republic of China
| | - Xiaoqing Liu
- Shanghai Promega Biological Products Limited, People's Republic of China
| | - Wei Kang
- Department of Anatomical and Cellular Pathology, State Key Laboratory of Translational Oncology, Institute of Digestive Disease, State Key Laboratory of Digestive Disease, Prince of Wales Hospital, The Chinese University of Hong Kong, Hong Kong Special Administrative Region of China
| | - Minyi Liang
- Department of Pathology, Nanfang Hospital/Basic Medical College, Southern Medical University, Guangzhou 510515, Guangdong Province, People's Republic of China; Guangdong Province Key Laboratory of Molecular Tumor Pathology, Guangzhou 510515, Guangdong Province, People's Republic of China
| | - Rui Yang
- Department of Pathology, Nanfang Hospital/Basic Medical College, Southern Medical University, Guangzhou 510515, Guangdong Province, People's Republic of China; Guangdong Province Key Laboratory of Molecular Tumor Pathology, Guangzhou 510515, Guangdong Province, People's Republic of China
| | - Hong Li
- Department of Pathology, Nanfang Hospital/Basic Medical College, Southern Medical University, Guangzhou 510515, Guangdong Province, People's Republic of China; Guangdong Province Key Laboratory of Molecular Tumor Pathology, Guangzhou 510515, Guangdong Province, People's Republic of China
| | - Li Liang
- Department of Pathology, Nanfang Hospital/Basic Medical College, Southern Medical University, Guangzhou 510515, Guangdong Province, People's Republic of China; Guangdong Province Key Laboratory of Molecular Tumor Pathology, Guangzhou 510515, Guangdong Province, People's Republic of China; Jinfeng Laboratory, Chongqing 401329, People's Republic of China.
| |
Collapse
|