1
|
Lee J, Cha S, Kim J, Kim JJ, Kim N, Jae Gal SG, Kim JH, Lee JH, Choi YD, Kang SR, Song GY, Yang DH, Lee JH, Lee KH, Ahn S, Moon KM, Noh MG. Ensemble Deep Learning Model to Predict Lymphovascular Invasion in Gastric Cancer. Cancers (Basel) 2024; 16:430. [PMID: 38275871 PMCID: PMC10814827 DOI: 10.3390/cancers16020430] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/13/2023] [Revised: 01/11/2024] [Accepted: 01/16/2024] [Indexed: 01/27/2024] Open
Abstract
Lymphovascular invasion (LVI) is one of the most important prognostic factors in gastric cancer as it indicates a higher likelihood of lymph node metastasis and poorer overall outcome for the patient. Despite its importance, the detection of LVI(+) in histopathology specimens of gastric cancer can be a challenging task for pathologists as invasion can be subtle and difficult to discern. Herein, we propose a deep learning-based LVI(+) detection method using H&E-stained whole-slide images. The ConViT model showed the best performance in terms of both AUROC and AURPC among the classification models (AUROC: 0.9796; AUPRC: 0.9648). The AUROC and AUPRC of YOLOX computed based on the augmented patch-level confidence score were slightly lower (AUROC: -0.0094; AUPRC: -0.0225) than those of the ConViT classification model. With weighted averaging of the patch-level confidence scores, the ensemble model exhibited the best AUROC, AUPRC, and F1 scores of 0.9880, 0.9769, and 0.9280, respectively. The proposed model is expected to contribute to precision medicine by potentially saving examination-related time and labor and reducing disagreements among pathologists.
Collapse
Affiliation(s)
- Jonghyun Lee
- Department of Medical and Digital Engineering, Hanyang University College of Engineering, Seoul 04763, Republic of Korea;
| | - Seunghyun Cha
- Department of Pre-Medicine, Chonnam National University Medical School, 322 Seoyang-ro, Hwasun-eup, Hwasun-gun, Gwangju 58128, Republic of Korea;
| | - Jiwon Kim
- NetTargets, 495 Sinseong-dong, Yuseong, Daejeon 34109, Republic of Korea
| | - Jung Joo Kim
- AMGINE, Inc., Jeongui-ro 8-gil 13, Seoul 05836, Republic of Korea;
| | - Namkug Kim
- Department of Convergence Medicine, Asan Medical Center, University of Ulsan College of Medicine, Seoul 25440, Republic of Korea; (N.K.); (S.G.J.G.)
| | - Seong Gyu Jae Gal
- Department of Convergence Medicine, Asan Medical Center, University of Ulsan College of Medicine, Seoul 25440, Republic of Korea; (N.K.); (S.G.J.G.)
| | - Ju Han Kim
- Division of Biomedical Informatics, Seoul National University Biomedical Informatics (SNUBI), Seoul National University College of Medicine, Seoul 03080, Republic of Korea;
| | - Jeong Hoon Lee
- Department of Radiology, Stanford University School of Medicine, Stanford, CA 94305-5101, USA;
| | - Yoo-Duk Choi
- Department of Pathology, Chonnam National University Medical School, Gwangju 61469, Republic of Korea;
| | - Sae-Ryung Kang
- Department of Nuclear Medicine, Clinical Medicine Research Center, Chonnam National University Hospital, 671 Jebongno, Gwangju 61469, Republic of Korea;
| | - Ga-Young Song
- Departments of Hematology-Oncology, Chonnam National University Hwasun Hospital, 322 Seoyangro, Hwasun 58128, Republic of Korea; (G.-Y.S.); (D.-H.Y.)
| | - Deok-Hwan Yang
- Departments of Hematology-Oncology, Chonnam National University Hwasun Hospital, 322 Seoyangro, Hwasun 58128, Republic of Korea; (G.-Y.S.); (D.-H.Y.)
| | - Jae-Hyuk Lee
- Department of Pathology, Chonnam National University Hwasun Hospital and Medical School, 322 Seoyang-ro, Hwasun-eup, Hwasun-gun, Hwasun 58128, Republic of Korea (K.-H.L.)
| | - Kyung-Hwa Lee
- Department of Pathology, Chonnam National University Hwasun Hospital and Medical School, 322 Seoyang-ro, Hwasun-eup, Hwasun-gun, Hwasun 58128, Republic of Korea (K.-H.L.)
| | - Sangjeong Ahn
- Department of Pathology, Korea University Anam Hospital, Korea University College of Medicine, 73 Goryeodae-ro, Seongbuk-gu, Seoul 02841, Republic of Korea;
| | - Kyoung Min Moon
- Division of Pulmonary and Allergy Medicine, Department of Internal Medicine, Chung-Ang University Hospital, Chung-Ang University College of Medicine, Seoul 06973, Republic of Korea
- Artificial Intelligence, ZIOVISION Co., Ltd., Chuncheon 24341, Republic of Korea
| | - Myung-Giun Noh
- Department of Pathology, Chonnam National University Hwasun Hospital and Medical School, 322 Seoyang-ro, Hwasun-eup, Hwasun-gun, Hwasun 58128, Republic of Korea (K.-H.L.)
| |
Collapse
|
2
|
Yoshida M, Suwa T, Shimada S, Shiotsuki K, Shigeta K, Nakamura H, Takada K, Kishida Y, Ito S, Imai K, Hotta K, Ono H, Sugino T. Desmin immunostaining is effective for improving interobserver variability in the depth assessment of the submucosal invasion of colorectal cancers. Hum Pathol 2023; 141:149-157. [PMID: 37633534 DOI: 10.1016/j.humpath.2023.08.006] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 06/04/2023] [Revised: 08/15/2023] [Accepted: 08/21/2023] [Indexed: 08/28/2023]
Abstract
Tumor depth evaluation is essential for pathological tumor staging because it affects clinical management as an independent risk factor for lymph node metastasis in colorectal cancers. However, poor interobserver variability of invasion depth has been reported. This study aimed to clarify the effectiveness of desmin immunostaining in the histological diagnosis of colorectal cancer. Overall, 63 sets of slides of colorectal cancer stained with hematoxylin and eosin (H&E) and desmin were prepared and independently reviewed by four examiners. After reviewing the desmin-stained slides, the interobserver variability of H&E slides alone was significantly improved for all examiners. For the assessment of Tis vs. T1, the sensitivity and accuracy were significantly improved for all examiners by combining H&E and desmin immunostaining. For the diagnosis of T1b vs. Tis or T1a, specificity and accuracy were significantly improved by adding desmin immunostaining. Ancillary desmin staining to assess submucosal invasion in colorectal cancers significantly improved interobserver agreement, led to efficient screening of T1 cancers, and reduced excessive T1b diagnoses. The combination of desmin immunostaining and H&E staining is highly recommended for diagnosing invasive colorectal cancer.
Collapse
Affiliation(s)
- Masao Yoshida
- Division of Endoscopy, Shizuoka Cancer Center, Shizuoka, 411-8777 Japan.
| | - Tetsuya Suwa
- Division of Endoscopy, Shizuoka Cancer Center, Shizuoka, 411-8777 Japan
| | - Seitaro Shimada
- 3rd Department of Internal Medicine, University of Toyama, Toyama, 930-0194 Japan
| | - Kazuo Shiotsuki
- Department of Gastroenterology Kitakyushu Municipal Medical Center, Fukuoka, 802-8561 Japan
| | - Kohei Shigeta
- Division of Endoscopy, Shizuoka Cancer Center, Shizuoka, 411-8777 Japan; 3rd Department of Internal Medicine, University of Toyama, Toyama, 930-0194 Japan
| | - Haruka Nakamura
- Division of Endoscopy, Shizuoka Cancer Center, Shizuoka, 411-8777 Japan
| | - Kazunori Takada
- Division of Endoscopy, Shizuoka Cancer Center, Shizuoka, 411-8777 Japan
| | - Yoshihiro Kishida
- Division of Endoscopy, Shizuoka Cancer Center, Shizuoka, 411-8777 Japan
| | - Sayo Ito
- Division of Endoscopy, Shizuoka Cancer Center, Shizuoka, 411-8777 Japan
| | - Kenichiro Imai
- Division of Endoscopy, Shizuoka Cancer Center, Shizuoka, 411-8777 Japan
| | - Kinichi Hotta
- Division of Endoscopy, Shizuoka Cancer Center, Shizuoka, 411-8777 Japan
| | - Hiroyuki Ono
- Division of Endoscopy, Shizuoka Cancer Center, Shizuoka, 411-8777 Japan
| | - Takashi Sugino
- Department of Pathology, Shizuoka Cancer Center, Shizuoka, 411-8777 Japan
| |
Collapse
|
3
|
Gonzalez J, Bahmad HF, Ocejo S, Abreu A, Popp M, Gogola S, Fernandez V, Recine M, Poppiti R. The Usefulness of Elastin Staining to Detect Vascular Invasion in Cancer. Int J Mol Sci 2023; 24:15264. [PMID: 37894944 PMCID: PMC10607141 DOI: 10.3390/ijms242015264] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/21/2023] [Revised: 10/15/2023] [Accepted: 10/16/2023] [Indexed: 10/29/2023] Open
Abstract
Tumor prognosis hinges on accurate cancer staging, a pivotal process influenced by the identification of lymphovascular invasion (LVI), i.e., blood vessel and lymphatic vessel invasion. Protocols by the College of American Pathologists (CAP) and the World Health Organization (WHO) have been established to assess LVI in various tumor types, including, but not limited to, breast cancer, colorectal cancer (CRC), pancreatic exocrine tumors, and thyroid carcinomas. The CAP refers to blood vessel invasion as "angioinvasion" (vascular invasion) to differentiate it from lymphatic vessel invasion (lymphatic invasion). For clarity, the latter terms will be used throughout this review. The presence of lymphatic and/or vascular invasion has emerged as a pivotal prognostic factor; therefore, its accurate identification is crucial not only for staging but also for providing the patient with an honest understanding of his/her prognosis. Given the prognostic importance of the correct identification of LVI, specific staining techniques are employed to distinguish lymphatic vessel invasion from angioinvasion and to differentiate true LVI from artifact. These encompass hematoxylin and eosin (H&E) staining, elastic staining, Factor VIII staining, Ulex europaeus I agglutinin staining, CD31, CD34, D2-40, ERG, and D2-40 (podoplanin) immunohistochemical (IHC) stains among others. Based on a review of numerous publications regarding the efficacy of various methods for LVI detection, elastin staining demonstrated superior accuracy and prognostic value, allowing for more targeted treatment strategies. The clinical significance of accurately detecting LVI cannot be overstated, as it is strongly linked to higher cancer-related mortality and an increased risk of tumor recurrence. This review aims to examine the existing literature on the use of elastin stains in the detection of vascular invasion among different types of tumors and its prognostic value.
Collapse
Affiliation(s)
- Jeffrey Gonzalez
- Herbert Wertheim College of Medicine, Florida International University, Miami, FL 33199, USA; (J.G.); (S.O.); (A.A.); (M.P.); (S.G.)
| | - Hisham F. Bahmad
- Arkadi M. Rywlin M.D. Department of Pathology and Laboratory Medicine, Mount Sinai Medical Center, Miami Beach, FL 33140, USA; (V.F.); (M.R.); (R.P.)
| | - Stephanie Ocejo
- Herbert Wertheim College of Medicine, Florida International University, Miami, FL 33199, USA; (J.G.); (S.O.); (A.A.); (M.P.); (S.G.)
| | - Alvaro Abreu
- Herbert Wertheim College of Medicine, Florida International University, Miami, FL 33199, USA; (J.G.); (S.O.); (A.A.); (M.P.); (S.G.)
| | - Meagan Popp
- Herbert Wertheim College of Medicine, Florida International University, Miami, FL 33199, USA; (J.G.); (S.O.); (A.A.); (M.P.); (S.G.)
| | - Samantha Gogola
- Herbert Wertheim College of Medicine, Florida International University, Miami, FL 33199, USA; (J.G.); (S.O.); (A.A.); (M.P.); (S.G.)
| | - Vielka Fernandez
- Arkadi M. Rywlin M.D. Department of Pathology and Laboratory Medicine, Mount Sinai Medical Center, Miami Beach, FL 33140, USA; (V.F.); (M.R.); (R.P.)
| | - Monica Recine
- Arkadi M. Rywlin M.D. Department of Pathology and Laboratory Medicine, Mount Sinai Medical Center, Miami Beach, FL 33140, USA; (V.F.); (M.R.); (R.P.)
- Department of Pathology, Herbert Wertheim College of Medicine, Florida International University, Miami, FL 33199, USA
| | - Robert Poppiti
- Arkadi M. Rywlin M.D. Department of Pathology and Laboratory Medicine, Mount Sinai Medical Center, Miami Beach, FL 33140, USA; (V.F.); (M.R.); (R.P.)
- Department of Pathology, Herbert Wertheim College of Medicine, Florida International University, Miami, FL 33199, USA
| |
Collapse
|
4
|
Spoerl S, Gerken M, Fischer R, Mamilos A, Spoerl S, Wolf S, Pohl F, Klingelhöffer C, Ettl T, Reichert TE, Spanier G. Lymphatic and vascular invasion in oral squamous cell carcinoma: Implications for recurrence and survival in a population-based cohort study. Oral Oncol 2020; 111:105009. [PMID: 33032181 DOI: 10.1016/j.oraloncology.2020.105009] [Citation(s) in RCA: 8] [Impact Index Per Article: 2.0] [Reference Citation Analysis] [What about the content of this article? (0)] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/20/2020] [Revised: 08/04/2020] [Accepted: 09/09/2020] [Indexed: 12/18/2022]
Abstract
OBJECTIVE Numerous studies analyzed lymphovascular invasion (LVI) in various malignant diseases, however, little is known about the role of lymphatic invasion (LI) as well as vascular invasion (VI) in oral squamous cell carcinoma (OSCC). The aim of this study is to illuminate the role of LI and VI in a population-based cohort study. METHODS We retrospectively analyzed 745 primarily resected OSCC patients in Eastern Bavaria for histopathologically verified LI and VI. Overall survival (OS) and recurrence-free survival (RFS) were calculated, whereas analysis was performed by uni- and multivariate statistics. Mean follow-up time was 7.4 years. RESULTS LI was found in 115 patients (15.4%), VI was diagnosed in 23 cases (3.1%). LI correlated significantly with distinct anatomical sites (p = 0.004), increasing pT-classification (p < 0.001), lymph node involvement (p < 0.001), higher grading (p < 0.001), advanced UICC-stages (p < 0.001) and adjuvant therapies (p < 0.001). Similar results were found for VI. Survival analysis resulted in a significantly decreased five-year OS and RFS in patients with diagnosed LI (OS: 41.1%, RFS: 38.3%) in contrast to LI-negative cases (OS: 66.8%, RFS: 59.7.7%, p < 0.001). Analogous outcomes were seen for patients with VI. Additionally, LI was identified as a predictive parameter, indicating individual patients' response to adjuvant therapies. CONCLUSION This population-based cohort study underlines the unfavorable aspect of LI and VI on outcome in OSCC. Including LI and VI in existing staging systems could help to stratify patients' risk for adverse outcome and consecutively determine adjuvant treatment in malignant disease.
Collapse
|