RESUMO
BACKGROUND: Human epidermal growth factor receptor 2 (HER2)-low breast cancer has emerged as a new subtype of tumor, for which novel antibody-drug conjugates have shown beneficial effects. Assessment of HER2 requires several immunohistochemistry tests with an additional in situ hybridization test if a case is classified as HER2 2+. Therefore, novel cost-effective methods to speed up the HER2 assessment are highly desirable. METHODS: We used a self-supervised attention-based weakly supervised method to predict HER2-low directly from 1437 histopathological images from 1351 breast cancer patients. We built six distinct models to explore the ability of classifiers to distinguish between the HER2-negative, HER2-low, and HER2-high classes in different scenarios. The attention-based model was used to comprehend the decision-making process aimed at relevant tissue regions. RESULTS: Our results indicate that the effectiveness of classification models hinges on the consistency and dependability of assay-based tests for HER2, as the outcomes from these tests are utilized as the baseline truth for training our models. Through the use of explainable AI, we reveal histologic patterns associated with the HER2 subtypes. CONCLUSION: Our findings offer a demonstration of how deep learning technologies can be applied to identify HER2 subgroup statuses, potentially enriching the toolkit available for clinical decision-making in oncology.
Assuntos
Biomarcadores Tumorais , Neoplasias da Mama , Aprendizado Profundo , Imuno-Histoquímica , Receptor ErbB-2 , Humanos , Receptor ErbB-2/metabolismo , Receptor ErbB-2/genética , Neoplasias da Mama/patologia , Neoplasias da Mama/metabolismo , Neoplasias da Mama/genética , Feminino , Biomarcadores Tumorais/metabolismo , Imuno-Histoquímica/métodos , Aprendizado de Máquina SupervisionadoRESUMO
Breast cancer (BC) is the most common cancer in women, with incidence rates increasing globally in recent years. Therefore, it is important to find new molecules with prognostic and therapeutic value to improve therapeutic response and quality of life. The polyunsaturated fatty acids (PUFAs) metabolic pathway participates in various physiological processes, as well as in the development of malignancies. Although aberrancies in the PUFAs metabolic pathway have been implicated in carcinogenesis, the functional and clinical relevance of this pathway has not been well explored in BC. To evaluate the clinical significance of soluble epoxide hydrolase (EPHX2) expression in Mexican patients with BC using tissue microarrays (TMAs) and digital pathology (DP). Immunohistochemical analyses were performed on 11 TMAs with 267 BC samples to quantify this enzyme. Using DP, EPHX2 protein expression was evaluated solely in tumor areas. The association of EPHX2 with overall survival (OS) was detected through bioinformatic analysis in public databases and confirmed in our cohort via Cox regression analysis. Clear nuclear expression of EPHX2 was identified. Receiver operating characteristics (ROC) curves revealed the optimal cutoff point at 2.847062 × 10-3 pixels, with sensitivity of 69.2% and specificity of 67%. Stratification based on this cutoff value showed elevated EPHX2 expression in multiple clinicopathological features, including older age and nuclear grade, human epidermal growth factor receptor 2 (HER2) and triple negative breast cancer (TNBC) subtypes, and recurrence. Kaplan-Meier curves demonstrated how higher nuclear expression of EPHX2 predicts shorter OS. Consistently, multivariate analysis confirmed EPHX2 as an independent predictor of OS, with a hazard ratio (HR) of 3.483 and a 95% confidence interval of 1.804-6.724 (p < 0.001). Our study demonstrates for the first time that nuclear overexpression of EPHX2 is a predictor of poor prognosis in BC patients. The DP approach was instrumental in identifying this significant association. Our study provides valuable insights into the potential clinical utility of EPHX2 as a prognostic biomarker and therapeutic target in BC.
Assuntos
Biomarcadores Tumorais , Neoplasias da Mama , Epóxido Hidrolases , Humanos , Epóxido Hidrolases/metabolismo , Epóxido Hidrolases/genética , Feminino , Neoplasias da Mama/patologia , Neoplasias da Mama/metabolismo , Neoplasias da Mama/mortalidade , Neoplasias da Mama/genética , Pessoa de Meia-Idade , Prognóstico , Biomarcadores Tumorais/metabolismo , Idoso , Adulto , Núcleo Celular/metabolismo , Regulação para Cima , Regulação Neoplásica da Expressão Gênica , Curva ROC , Idoso de 80 Anos ou mais , Estimativa de Kaplan-MeierRESUMO
BACKGROUND AND OBJECTIVE: Cell segmentation in bright-field histological slides is a crucial topic in medical image analysis. Having access to accurate segmentation allows researchers to examine the relationship between cellular morphology and clinical observations. Unfortunately, most segmentation methods known today are limited to nuclei and cannot segment the cytoplasm. METHODS: We present a new network architecture Cyto R-CNN that is able to accurately segment whole cells (with both the nucleus and the cytoplasm) in bright-field images. We also present a new dataset CytoNuke, consisting of multiple thousand manual annotations of head and neck squamous cell carcinoma cells. Utilizing this dataset, we compared the performance of Cyto R-CNN to other popular cell segmentation algorithms, including QuPath's built-in algorithm, StarDist, Cellpose and a multi-scale Attention Deeplabv3+. To evaluate segmentation performance, we calculated AP50, AP75 and measured 17 morphological and staining-related features for all detected cells. We compared these measurements to the gold standard of manual segmentation using the Kolmogorov-Smirnov test. RESULTS: Cyto R-CNN achieved an AP50 of 58.65% and an AP75 of 11.56% in whole-cell segmentation, outperforming all other methods (QuPath 19.46/0.91%; StarDist 45.33/2.32%; Cellpose 31.85/5.61%, Deeplabv3+ 3.97/1.01%). Cell features derived from Cyto R-CNN showed the best agreement to the gold standard (D¯=0.15) outperforming QuPath (D¯=0.22), StarDist (D¯=0.25), Cellpose (D¯=0.23) and Deeplabv3+ (D¯=0.33). CONCLUSION: Our newly proposed Cyto R-CNN architecture outperforms current algorithms in whole-cell segmentation while providing more reliable cell measurements than any other model. This could improve digital pathology workflows, potentially leading to improved diagnosis. Moreover, our published dataset can be used to develop further models in the future.
Assuntos
Algoritmos , Processamento de Imagem Assistida por Computador , Redes Neurais de Computação , Humanos , Processamento de Imagem Assistida por Computador/métodos , Núcleo Celular , Neoplasias de Cabeça e Pescoço/diagnóstico por imagem , Neoplasias de Cabeça e Pescoço/patologia , Carcinoma de Células Escamosas de Cabeça e Pescoço/diagnóstico por imagem , Carcinoma de Células Escamosas de Cabeça e Pescoço/patologia , Citoplasma , Reprodutibilidade dos Testes , Carcinoma de Células Escamosas/diagnóstico por imagem , Carcinoma de Células Escamosas/patologiaRESUMO
Identifying marker combinations for robust prognostic validation in primary tumour compartments remains challenging. We aimed to assess the prognostic significance of CSC markers (ALDH1, CD44, p75NTR, BMI-1) and E-cadherin biomarkers in OSCC. We analysed 94 primary OSCC and 67 metastatic lymph node samples, including central and invasive tumour fronts (ITF), along with clinicopathological data. We observed an increase in ALDH1+/CD44+/BMI-1- tumour cells in metastatic lesions compared to primary tumours. Multivariate analysis highlighted that elevated p75NTR levels (at ITF) and reduced E-cadherin expression (at the tumour centre) independently predicted metastasis, whilst ALDH1high exhibited independent predictive lower survival at the ITF, surpassing the efficacy of traditional tumour staging. Then, specifically at the ITF, profiles characterized by CSChighE-cadherinlow (ALDH1highp75NTRhighE-cadherinlow) and CSCintermediateE-cadherinlow (ALDH1 or p75NTRhighE-cadherinlow) were significantly associated with worsened overall survival and increased likelihood of metastasis in OSCC patients. In summary, our study revealed diverse tumour cell profiles in OSCC tissues, with varying CSC and E-cadherin marker patterns across primary tumours and metastatic sites. Given the pivotal role of reduced survival rates as an indicator of unfavourable prognosis, the immunohistochemistry profile identified as CSChighE-cadherinlow at the ITF of primary tumours, emerges as a preferred prognostic marker closely linked to adverse outcomes in OSCC.
Assuntos
Família Aldeído Desidrogenase 1 , Biomarcadores Tumorais , Caderinas , Carcinoma de Células Escamosas , Neoplasias Bucais , Adulto , Idoso , Feminino , Humanos , Masculino , Pessoa de Meia-Idade , Família Aldeído Desidrogenase 1/metabolismo , Biomarcadores Tumorais/metabolismo , Caderinas/metabolismo , Carcinoma de Células Escamosas/metabolismo , Carcinoma de Células Escamosas/patologia , Carcinoma de Células Escamosas/mortalidade , Receptores de Hialuronatos/metabolismo , Imuno-Histoquímica , Metástase Linfática , Neoplasias Bucais/patologia , Neoplasias Bucais/metabolismo , Neoplasias Bucais/mortalidade , Neoplasias Bucais/diagnóstico , Células-Tronco Neoplásicas/metabolismo , Células-Tronco Neoplásicas/patologia , Proteínas do Tecido Nervoso/metabolismo , Complexo Repressor Polycomb 1/metabolismo , Complexo Repressor Polycomb 1/genética , Prognóstico , Receptores de Fator de Crescimento Neural/metabolismo , Retinal Desidrogenase/metabolismoRESUMO
Morphometry of striated muscle fibres is critical for monitoring muscle health and function. Here, we evaluated functional parameters of skeletal and cardiac striated muscle in two experimental models using the Morphometric Analysis of Muscle Fibre tool (MusMA). The collagen-induced arthritis model was used to evaluate the function of skeletal striated muscle and the non-alcoholic fatty liver disease model was used for cardiac striated muscle analysis. After euthanasia, we used haeamatoxylin and eosin stained sections of skeletal and cardiac muscle to perform muscle fibre segmentation and morphometric analysis. Morphometric analysis classified muscle fibres into six subpopulations: normal, regular hypertrophic, irregular hypertrophic, irregular, irregular atrophic and regular atrophic. The percentage of atrophic fibres was associated with lower walking speed (p = 0.009) and lower body weight (p = 0.026), respectively. Fibres categorized as normal were associated with maximum grip strength (p < 0.001) and higher march speed (p < 0.001). In the evaluation of cardiac striated muscle fibres, the percentage of normal cardiomyocytes negatively correlated with cardiovascular risk markers such as the presence of abdominal adipose tissue (p = .003), miR-33a expression (p = .001) and the expression of miR-126 (p = .042) Furthermore, the percentage of atrophic cardiomyocytes correlated significantly with the Castelli risk index II (p = .014). MusMA is a simple and objective tool that allows the screening of striated muscle fibre morphometry, which can complement the diagnosis of muscle diseases while providing functional and prognostic information in basic and clinical research.
Assuntos
Fibras Musculares Esqueléticas , Animais , Masculino , Prognóstico , Fibras Musculares Esqueléticas/patologia , Doenças Cardiovasculares/patologia , Doenças Cardiovasculares/fisiopatologia , Miócitos Cardíacos/patologia , Fatores de Risco de Doenças CardíacasRESUMO
OBJECTIVES: ⢠Gather a panel of Latin American experts in testing and treating BRAF-melanoma. ⢠Describe the current landscape of BRAF-mutated melanoma in Latin America. ⢠Outline the current gaps in testing and recommend improvements for testing and treating BRAF-mutated melanoma in the region. INTRODUCTION: Melanoma prevalence in Latin America is lower than in high- and middle-income countries. However, recent data indicate that the region's incidence and mortality are rising, with more stage IV patients being diagnosed. According to international clinical practice guidelines, conducting BRAF-mutation testing in patients with stage III or stage IV melanoma and high-risk resected disease is imperative. Still, BRAF-mutation testing and targeted therapies are inconsistently available in the region. METHODS: Americas Health Foundation convened a meeting of Latin American experts on BRAF-mutated melanoma to develop guidelines and recommendations for diagnosis through treatment. RESULTS AND CONCLUSIONS: Some recommendations for improving diagnostics through improving access and reducing the cost of BRAF-mutation testing, enhancing efficiency in pathology laboratories, and creating country-specific local guidelines. The panel also gave treatment recommendations for neo-adjuvant therapy, adjuvant therapy, and therapy for patients with metastatic disease in Latin America.
Assuntos
Melanoma , Mutação , Proteínas Proto-Oncogênicas B-raf , Humanos , Melanoma/genética , Melanoma/terapia , Melanoma/diagnóstico , Proteínas Proto-Oncogênicas B-raf/genética , América Latina/epidemiologia , Neoplasias Cutâneas/genética , Neoplasias Cutâneas/terapia , Neoplasias Cutâneas/diagnóstico , Guias de Prática Clínica como AssuntoRESUMO
The most widely accepted and used type of digital pathology (DP) is whole-slide imaging (WSI). The USFDA granted two WSI system approvals for primary diagnosis, the first in 2017. In Latin America, DP has the potential to reshape healthcare by enhancing diagnostic capabilities through artificial intelligence (AI) and standardizing pathology reports. Yet, we must tackle regulatory hurdles, training, resource availability, and unique challenges to the region. Collectively addressing these hurdles can enable the region to harness DP's advantages-enhancing disease diagnosis, medical research, and healthcare accessibility for its population. Americas Health Foundation assembled a panel of Latin American pathologists who are experts in DP to assess the hurdles to implementing it into pathologists' workflows in the region and provide recommendations for overcoming them. Some key steps recommended include creating a Latin American Society of Digital Pathology to provide continuing education, developing AI models trained on the Latin American population, establishing national regulatory frameworks for protecting the data, and standardizing formats for DP images to ensure that pathologists can collaborate and validate specimens across the various DP platforms.
RESUMO
Cervical cancer is the second most common form of cancer and a leading cause of premature death among women aged 15 to 44 worldwide. In Brazil, there is a high prevalence of infection by the human papillomavirus - HPV. Digital pathology optimizes time and space for reading cervicovaginal cytology slides. We evaluated the feasibility of using whole slide images (WSI) for the routine interpretation of cytology exams. A total of 99 cases of vaginal cytology were selected from a reference laboratory in Northeastern Brazil. Three cytotechnicians participated in the study. Cellular atypia was the one that most presented concordance values. Two observers almost perfectly agreed (k = 0.86 and k = 0.84, respectively) on the negative diagnoses. The performance of the evaluators for NILM (negative for intraepithelial lesion and malignancy) showed high reproducibility and sensitivity in the digital slides, mainly between evaluators A and C. In contrast, the microbiology group showed disagreement between the diagnoses by digital slides and the standard- gold. The concordance between the digital diagnoses and the gold standard for ASCUS was 89%. In the inflammatory category, Spearman's test showed similar results between raters A, B, and C (rs = 0.47, rs = 0.41, and rs = 0.47, respectively). This study reports the diagnostic validation using digital slides in view of the need to optimize the cytology visualization process. Our experience shows good diagnostic agreement between digital and optical microscopy in several analyzed categories, but mainly in relation to cellular atypia and inflammatory processes.
Assuntos
Processamento de Imagem Assistida por Computador , Neoplasias do Colo do Útero , Feminino , Humanos , Reprodutibilidade dos Testes , Processamento de Imagem Assistida por Computador/métodos , Colo do Útero/patologia , Citodiagnóstico/métodos , Neoplasias do Colo do Útero/diagnóstico , Neoplasias do Colo do Útero/patologia , Papillomavirus Humano , Esfregaço Vaginal/métodosRESUMO
INTRODUCTION: Digital cytopathology is being progressively implemented in centres worldwide, but impediments such as the three-dimensionality of specimens and the size of scanned images have prevented its use from becoming widespread. This study aimed to validate the use of digital whole slide image scanning of cytopathology samples for routine sign-out. METHODS: Specimens were scanned using the Leica Aperio GT 450 System. The following sample types were used: liquid-based cytology, direct conventional smears from fine needle aspirates and cytospins. Cases were validated by the same pathologist who originally rendered the conventional diagnosis, with a washout of at least 3 months. Final digital diagnoses were compared to the original analogical diagnoses, and cases were considered concordant up to a one-degree difference between the original and digital diagnoses. Reasons for the unsuccessful scanning of slides were also noted. The technical procedures followed the College of American Pathologists' guidelines for digital pathology validation. RESULTS: A total of 730 slides from 383 cases (337 female, 51 male; median age 42) were successfully scanned. These cases consisted of the following sample types: 81 (21.1%) conventional smears, 240 (62.7%) liquid-based cytology samples and 62 (16.2%) cytospins. There were only five discordant cases, with a 98.7% agreement between original and digital diagnoses using the difference rate of up to one degree. Seventy-seven slides (10.5%) had to be rescanned due to technical problems. The main reasons for unsuccessful scanning were paucicellular samples (44; 57.1%), the thickness of the smears (18; 23.4%) and issues with the coverslip (15; 19.5%). CONCLUSION: Cytological specimens can be successfully scanned and used for digital pathology, with excellent agreement with the original diagnoses.
Assuntos
Citologia , Microscopia , Masculino , Humanos , Feminino , Adulto , Microscopia/métodos , CitodiagnósticoRESUMO
BACKGROUND: The usage of whole-slide images has recently been gaining a foothold in medical education, training, and diagnosis. OBJECTIVES: The first objective of the current study was to compare academic performance on virtual microscopy (VM) and light microscopy (LM) for learning pathology, anatomy, and histology in medical and dental students during the COVID-19 period. The second objective was to gather insight into various applications and usage of such technology for medical education. MATERIALS AND METHODS: Using the keywords "virtual microscopy" or "light microscopy" or "digital microscopy" and "medical" and "dental" students, databases (PubMed, Embase, Scopus, Cochrane, CINAHL, and Google Scholar) were searched. Hand searching and snowballing were also employed for article searching. After extracting the relevant data based on inclusion and execution criteria, the qualitative data were used for the systematic review and quantitative data were used for meta-analysis. The Newcastle Ottawa Scale (NOS) scale was used to assess the quality of the included studies. Additionally, we registered our systematic review protocol in the prospective register of systematic reviews (PROSPERO) with registration number CRD42020205583. RESULTS: A total of 39 studies met the criteria to be included in the systematic review. Overall, results indicated a preference for this technology and better academic scores. Qualitative analyses reported improved academic scores, ease of use, and enhanced collaboration amongst students as the top advantages, whereas technical issues were a disadvantage. The performance comparison of virtual versus light microscopy meta-analysis included 19 studies. Most (10/39) studies were from medical universities in the USA. VM was mainly used for teaching pathology courses (25/39) at medical schools (30/39). Dental schools (10/39) have also reported using VM for teaching microscopy. The COVID-19 pandemic was responsible for the transition to VM use in 17/39 studies. The pooled effect size of 19 studies significantly demonstrated higher exam performance (SMD: 1.36 [95% CI: 0.75, 1.96], p < 0.001) among the students who used VM for their learning. Students in the VM group demonstrated significantly higher exam performance than LM in pathology (SMD: 0.85 [95% CI: 0.26, 1.44], p < 0.01) and histopathology (SMD: 1.25 [95% CI: 0.71, 1.78], p < 0.001). For histology (SMD: 1.67 [95% CI: -0.05, 3.40], p = 0.06), the result was insignificant. The overall analysis of 15 studies assessing exam performance showed significantly higher performance for both medical (SMD: 1.42 [95% CI: 0.59, 2.25], p < 0.001) and dental students (SMD: 0.58 [95% CI: 0.58, 0.79], p < 0.001). CONCLUSIONS: The results of qualitative and quantitative analyses show that VM technology and digitization of glass slides enhance the teaching and learning of microscopic aspects of disease. Additionally, the COVID-19 global health crisis has produced many challenges to overcome from a macroscopic to microscopic scale, for which modern virtual technology is the solution. Therefore, medical educators worldwide should incorporate newer teaching technologies in the curriculum for the success of the coming generation of health-care professionals.
RESUMO
Introducción: la citología permite examinar células de un tejido de manera mínimamente invasiva, sin embargo, la capacidad de realizar técnicas complementarias como la inmunocitoquímica (ICQ) no está exenta de dificultades. Es el objetivo de nuestro trabajo presentar una metodología que permita la utilización de ICQ automatizada asociada a un análisis automatizado mediante técnica de patología digital. Métodos: se incluyeron 5 sujetos sanos y se obtuvieron muestras de superficie ocular utilizando un citocepillo. La muestra fue procesada de manera automatizada mediante citología en fase líquida. Posteriormente se realizó ICQ automatizada para detectar la positividad nuclear del receptor de vitamina D. Para la evaluación, se utilizaron dos métodos: cuantificación directa bajo microscopio de luz y análisis automatizado usando analizador de imágenes en las diapositivas digitales obtenidas con un Scanner. El porcentaje de positividad encontrado con ambos métodos fueron comparados utilizando la prueba de Kappa. Resultados: todas las muestras presentaron una celularidad adecuada. En todos los casos fue posible realizar ICQ automatizada, más aún, todas las muestras presentaron una calidad óptima. Al comparar ambos métodos (manual versus automatizado) se observó un nivel de acuerdo sustancial (Kappa=0,69). Conclusiones: la metodología presentada en este manuscrito permite la evaluación automatizada de marcadores inmunohistoquímicos de la superficie ocular de manera mínimamente invasiva, siendo similar al conteo manual, pero más objetivo y reproducible. Esta técnica podría ser útil para el estudio proteómico en patologías como la enfermedad por ojo seco.
Introduction: Cytology tests use small amounts of tissue samples for diagnosis as a minimally invasive technique; however, the ability to perform complementary methods such as immunocytochemistry (ICC) is not without difficulties. The aim of our work is to present a method that allows the use of automated ICC associated with an automated image analysis using digital pathology. Methods: Five healthy subjects were included, and ocular surface samples were obtained using a cytobrush. The sample was processed as liquid-based cytology. Automated ICC was subsequently performed to detect vitamin D receptor nuclear positivity. Two methods were used for evaluation: manual counting under a light microscope and automated analysis using an image analyzer on digitized slides. The percentage of positivity found in both methods was compared using the Kappa test. Results: All samples presented adequate cellularity. In all cases, it was possible to perform automated ICC; moreover, all samples presented optimal quality. When comparing both methods (manual versus automated), a substantial level of agreement was seen (Kappa=0.69). Conclusions. The method presented in this manuscript allows the minimally invasive automated evaluation of ocular surface ICC markers, being like manual counting but more objective and reproducible. This technique could be useful for proteomic study in pathologies such as dry eye disease.
RESUMO
Background: Deep learning methods have demonstrated remarkable performance in pathology image analysis, but they are computationally very demanding. The aim of our study is to reduce their computational cost to enable their use with large tissue image datasets. Methods: We propose a method called Network Auto-Reduction (NAR) that simplifies a Convolutional Neural Network (CNN) by reducing the network to minimize the computational cost of doing a prediction. NAR performs a compound scaling in which the width, depth, and resolution dimensions of the network are reduced together to maintain a balance among them in the resulting simplified network. We compare our method with a state-of-the-art solution called ResRep. The evaluation is carried out with popular CNN architectures and a real-world application that identifies distributions of tumor-infiltrating lymphocytes in tissue images. Results: The experimental results show that both ResRep and NAR are able to generate simplified, more efficient versions of ResNet50 V2. The simplified versions by ResRep and NAR require 1.32× and 3.26× fewer floating-point operations (FLOPs), respectively, than the original network without a loss in classification power as measured by the Area under the Curve (AUC) metric. When applied to a deeper and more computationally expensive network, Inception V4, NAR is able to generate a version that requires 4× lower than the original version with the same AUC performance. Conclusions: NAR is able to achieve substantial reductions in the execution cost of two popular CNN architectures, while resulting in small or no loss in model accuracy. Such cost savings can significantly improve the use of deep learning methods in digital pathology. They can enable studies with larger tissue image datasets and facilitate the use of less expensive and more accessible graphics processing units (GPUs), thus reducing the computing costs of a study.
RESUMO
INTRODUCTION: The subjective evaluation of nuclear features in follicular-patterned lesions of the thyroid is a reason for diagnosis discordance. The assessment of nuclear features also varies whether the observation is performed optically or digitally. Our objective was to study the concordance among pathologists regarding the nuclear score (NS) evaluation in a series of follicular-patterned lesions, using optical versus three digital scanning protocols. METHODS: Three pathologists evaluated the NS in a 3mm2 area randomly selected from 20 hematoxylin-eosin slides representative of the respective 20 follicular-patterned thyroid lesions. The NS evaluation was performed using optical and three different scanning protocols in two scanners: P1000_20x, P1000_40x and DP200_20x. Kappa statistic (κ) and intraclass correlation coefficient (ICC) were obtained for intra- and interpathologist concordance. RESULTS: We recorded a good agreement among pathologists in the optical evaluation of the NS (ICC of 0.73). The concordance between optical versus digital observation had an almost perfect agreement for P1000_20x [κ = 0.85 (0.67-1.02); p < 0.0001] and a substantial agreement for both P1000_40x [κ = 0.69 (0.43-0.95) p = 0.002] and DP200_20x [κ = 0.77 (0.57-0.97); p = 0.001]. The P1000_20x protocol had the best intrapathologist concordance with the optical method, classified as almost perfect agreement for pathologists A (80%) and B (85%), and substantial agreement for pathologist C (70%). CONCLUSION: Digital observation of the WSI is valid for the NS evaluation in follicular-patterned thyroid lesions, with good agreement among pathologists and between optical and scanning protocols. Performance studies and validation procedures cannot be avoided in this setting to prevent diagnostic discordance due to the scanning process.
Assuntos
Núcleo Celular , Glândula Tireoide , Núcleo Celular/patologia , Humanos , Variações Dependentes do Observador , Glândula Tireoide/diagnóstico por imagem , Glândula Tireoide/patologiaRESUMO
BACKGROUND AND OBJECTIVE: Deep learning methods have demonstrated remarkable performance in pathology image analysis, but they require a large amount of annotated training data from expert pathologists. The aim of this study is to minimize the data annotation need in these analyses. METHODS: Active learning (AL) is an iterative approach to training deep learning models. It was used in our context with a Tumor Infiltrating Lymphocytes (TIL) classification task to minimize annotation. State-of-the-art AL methods were evaluated with the TIL application and we have proposed and evaluated a more efficient and effective AL acquisition method. The proposed method uses data grouping based on imaging features and model prediction uncertainty to select meaningful training samples (image patches). RESULTS: An experimental evaluation with a collection of cancer tissue images shows that: (i) Our approach reduces the number of patches required to attain a given AUC as compared to other approaches, and (ii) our optimization (subpooling) leads to AL execution time improvement of about 2.12×. CONCLUSIONS: This strategy enabled TIL based deep learning analyses using smaller annotation demand. We expect this approach may be used to build other analyses in digital pathology with fewer training samples.
Assuntos
Linfócitos do Interstício Tumoral , Neoplasias , Humanos , Processamento de Imagem Assistida por Computador , Linfócitos do Interstício Tumoral/patologia , Neoplasias/diagnóstico por imagem , Neoplasias/patologia , Aprendizagem Baseada em ProblemasRESUMO
Gastrointestinal ischemia may be presented as a complication associated with late shock detection in patients in critical condition. Prolonged ischemia can cause mucosal integrity to lose its barrier function, triggering alterations that can induce organ dysfunction and lead to death. Electrical impedance spectroscopy has been proposed to identify early alteration in ischemia-induced gastric mucosa in this type of patients. This work analyzed changes in impedance parameters, and tissue and molecular alterations that allow us to identify the time of ischemia in which the gastric mucosa still maintains its barrier function. The animals were randomly distributed in four groups: Control, Ischemia 60, 90, and 120 min. Impedance parameters were measured and predictive values were determined to categorize the degree of injury using a receiver operating characteristic curve. Markers of inflammatory process and apoptosis (iNOS, TNFα, COX-2, and Caspase-3) were analyzed. The largest increase in impedance parameters occurred in the ischemia 90 and 120 min groups, with resistance at low frequencies (RL) and reactance at high frequencies (XH) being the most related to damage, allowing prediction of the occurrence of reversible and irreversible tissue damage. Histological analysis and apoptosis assay showed progressive mucosal deterioration with irreversible damage (p < 0.001) starting from 90 min of ischemia. Furthermore, a significant increase in the expression of iNOS, TNFα, and COX-2 was identified in addition to apoptosis in the gastric mucosa starting from 90 min of ischemia. Tissue damage generated by an ischemia time greater than 60 min induces loss of barrier function in the gastric mucosa.
Assuntos
Mucosa Gástrica/patologia , Isquemia/patologia , Traumatismo por Reperfusão/patologia , Animais , Ciclo-Oxigenase 2/metabolismo , Impedância Elétrica , Mucosa Gástrica/metabolismo , Isquemia/metabolismo , Masculino , Ratos Wistar , Traumatismo por Reperfusão/metabolismo , Fatores de TempoRESUMO
BACKGROUND: With the emergence of the new coronavirus pandemic (COVID-19), distance learning, especially that mediated by information and digital communication technologies, has been adopted in all areas of knowledge and at all levels, including medical education. Imminently practical areas, such as pathology, have made traditional teaching based on conventional microscopy more flexible through the synergies of computational tools and image digitization, not only to improve teaching-learning but also to offer alternatives to repetitive and exhaustive histopathological analyzes. In this context, machine learning algorithms capable of recognizing histological patterns in kidney biopsy slides have been developed and validated with a view to building computational models capable of accurately identifying renal pathologies. In practice, the use of such algorithms can contribute to the universalization of teaching, allowing quality training even in regions where there is a lack of good nephropathologists. The purpose of this work is to describe and test the functionality of SmartPathk, a tool to support teaching of glomerulopathies using machine learning. The training for knowledge acquisition was performed automatically by machine learning methods using the J48 algorithm to create a computational model of an appropriate decision tree. RESULTS: An intelligent system, SmartPathk, was developed as a complementary remote tool in the teaching-learning process for pathology teachers and their students (undergraduate and graduate students), showing 89,47% accuracy using machine learning algorithms based on decision trees. CONCLUSION: This artificial intelligence system can assist in teaching renal pathology to increase the training capacity of new medical professionals in this area.
Assuntos
COVID-19 , Educação a Distância , Inteligência Artificial , Humanos , Aprendizado de Máquina , SARS-CoV-2 , EnsinoRESUMO
The role of digital pathology in remote reporting has seen an increase during the COVID-19 pandemic. Recently, recommendations had been made regarding the urgent need of reorganizing head and neck cancer diagnostic services to provide a safe work environment for the staff. A total of 162 glass slides from 109 patients over a period of 5 weeks were included in this validation and were assessed by all pathologists in both analyses (digital and conventional) to allow intraobserver comparison. The intraobserver agreement between the digital method (DM) and conventional method (CM) was considered almost perfect (κ ranged from 0.85 to 0.98, with 95% CI, ranging from 0.81 to 1). The most significant and frequent disagreements within trainees encompassed epithelial dysplasia grading and differentiation among severe dysplasia (carcinoma in situ) and oral squamous cell carcinoma. The most frequent pitfall from DM was lag in screen mirroring. The lack of details of inflammatory cells and the need for a higher magnification to assess dysplasia were pointed in one case each. The COVID-19 crisis has accelerated and consolidated the use of online meeting tools, which would be a valuable resource even in the post-pandemic scenario. Adaptation in laboratory workflow, the advent of digital pathology and remote reporting can mitigate the impact of similar future disruptions to the oral and maxillofacial pathology laboratory workflow avoiding delays in diagnosis and report, to facilitate timely management of head and neck cancer patients. Graphical abstract.
Assuntos
COVID-19 , Carcinoma in Situ/patologia , Tecnologia Digital , Interpretação de Imagem Assistida por Computador , Neoplasias Maxilares/patologia , Microscopia , Neoplasias Bucais/patologia , Carcinoma de Células Escamosas de Cabeça e Pescoço/patologia , Telepatologia , Biópsia , Diagnóstico Diferencial , Humanos , Variações Dependentes do Observador , Valor Preditivo dos Testes , Reprodutibilidade dos Testes , Fluxo de TrabalhoRESUMO
BACKGROUND: Time, travel and financial constraints have meant that traditional visiting teaching engagements are more difficult to accomplish. This has been exacerbated with the advent of the COVID-19 pandemic. The use of digital pathology and whole slide imaging (WSI) as an educational tool for distance teaching is underutilised and not fully exploited. This paper highlights the utility and feedback on the use of WSI for distance education/teaching. MATERIALS AND METHODS: Building on an existing relationship with the University of the West Indies (UWI), pathologists at University Health Network, Toronto, provided distance education using WSI, a digitised slide image hosting repository and videoconferencing facilities to provide case-based teaching to 15 UWI pathology trainees. Feedback was obtained from residents via a questionnaire and from teachers via a discussion. RESULTS: There was uniform support from teachers who felt that teaching was not hampered by the 'virtual' engagement. Comfort levels grew with each engagement and technical issues with sound diminished with the use of a portable speaker. The residents were very supportive and enthusiastic in embracing this mode of teaching. While technical glitches marred initial sessions, the process evened out especially when the slide hosting facility, teleconferencing and sound issues were changed. CONCLUSIONS: There was unanimous endorsement that use of WSI was the future, especially for distance teaching. However, it was not meant to supplant the use of glass slides in their current routine, daily practice.
Assuntos
Educação a Distância/métodos , Processamento de Imagem Assistida por Computador/métodos , Patologia Clínica/educação , COVID-19 , Canadá , Tecnologia Digital/métodos , Humanos , SARS-CoV-2 , Índias OcidentaisRESUMO
DNA repair deficiency (DRD) is an important driver of carcinogenesis and an efficient target for anti-tumor therapies to improve patient survival. Thus, detection of DRD in tumors is paramount. Currently, determination of DRD in tumors is dependent on wet-lab assays. Here we describe an efficient machine learning algorithm which can predict DRD from histopathological images. The utility of this algorithm is demonstrated with data obtained from 1445 cancer patients. Our method performs rather well when trained on breast cancer specimens with homologous recombination deficiency (HRD), AUC (area under curve) = 0.80. Results for an independent breast cancer cohort achieved an AUC = 0.70. The utility of our method was further shown by considering the detection of mismatch repair deficiency (MMRD) in gastric cancer, yielding an AUC = 0.81. Our results demonstrate the capacity of our learning-base system as a low-cost tool for DRD detection.
RESUMO
(1) Background: Despite the complementarity between radiology and histopathology, both from a diagnostic and a prognostic perspective, quantitative analyses of these modalities are usually performed in disconnected silos. This work presents initial results for differentiating two major non-small cell lung cancer (NSCLC) subtypes by exploring cross-scale associations between Computed Tomography (CT) images and corresponding digitized pathology images. (2) Methods: The analysis comprised three phases, (i) a multi-resolution cell density quantification to identify discriminant pathomic patterns for differentiating adenocarcinoma (ADC) and squamous cell carcinoma (SCC), (ii) radiomic characterization of CT images by using Haralick descriptors to quantify tumor textural heterogeneity as represented by gray-level co-occurrences to discriminate the two pathological subtypes, and (iii) quantitative correlation analysis between the multi-modal features to identify potential associations between them. This analysis was carried out using two publicly available digitized pathology databases (117 cases from TCGA and 54 cases from CPTAC) and a public radiological collection of CT images (101 cases from NSCLC-R). (3) Results: The top-ranked cell density pathomic features from the histopathology analysis were correlation, contrast, homogeneity, sum of entropy and difference of variance; which yielded a cross-validated AUC of 0.72 ± 0.02 on the training set (CPTAC) and hold-out validation AUC of 0.77 on the testing set (TCGA). Top-ranked co-occurrence radiomic features within NSCLC-R were contrast, correlation and sum of entropy which yielded a cross-validated AUC of 0.72 ± 0.01. Preliminary but significant cross-scale associations were identified between cell density statistics and CT intensity values using matched specimens available in the TCGA cohort, which were used to significantly improve the overall discriminatory performance of radiomic features in differentiating NSCLC subtypes (AUC = 0.78 ± 0.01). (4) Conclusions: Initial results suggest that cross-scale associations may exist between digital pathology and CT imaging which can be used to identify relevant radiomic and histopathology features to accurately distinguish lung adenocarcinomas from squamous cell carcinomas.