RESUMEN
AIMS: Urothelial carcinoma (UC) demonstrates significant molecular and histologic heterogeneity. The WHO 2022 classification has hinted at adding molecular signatures to the morphologic diagnosis. As morphology and associated molecular repertoire may potentially translate to choices of and response to therapy and relapse rate, broader acceptability of recognizing these key features among uropathologists is needed. This prompted an international survey to ascertain the practice patterns in classical/subtype UC among uropathologists across the globe. METHODS AND RESULTS: A survey instrument was shared among 98 uropathologists using SurveyMonkey software. Anonymized respondent data were analysed. The response rate was 85%. A majority were in concordance with the profiles of luminal (93%) and basal (82%) types. Opinion on the FGFR3 testing platform was variable. While 95% concurred that TERT promoter mutation is the key driver in UC, 72% had the opinion that APOBEC mutagenesis is the main signature in muscle invasive bladder cancer (MIBC). Uropathologists have divergent opinions on MIBC and ERCC2 mutations. Among the participants, 94% would quantify aggressive micropapillary and sarcomatoid histology, while 88% would reevaluate another transurethral resection of the bladder tumour specimen in nonmuscle invasive tumour with micropapillary, small cell, or sarcomatoid histology. A leading number agreed to specific molecular signatures of micropapillary (93%), plasmacytoid (97%), and small cell (86%) subtypes. Ninety-six percent of participants agreed that a small-cell component portends a more aggressive course and should be treated with neoadjuvant chemotherapy and 63% would perform HER2/neu testing only on oncologist's request in advanced tumours. Ninety percent agreed that microsatellite instability testing, although not a standard protocol, should be considered in young patients with upper tract UC. Eighty-six percent agreed that UC with high tumour mutational burden would be a better candidate for immunotherapy. CONCLUSION: In the era of precision medicine, enhanced understanding of molecular heterogeneity of UC will contribute to better therapeutic options, novel biomarker discovery, innovative management protocols, and outcomes. Our survey provides a broad perspective of pathologists' perceptions and experience regarding incorporation of histomolecular approaches to "personalize" therapy. Due to variable clinical adoption, there is a need for additional data using uniform study criteria. This will drive generation of best practice guidelines in this area for widespread and consistent clinical utility.
RESUMEN
The mechanisms underlying the onset and progression of vasculitis remain poorly understood. This condition is characterized by damage to the vascular wall, recruitment of inflammatory cells, and subsequent structural remodeling, which are hallmarks of vasculitis. The histopathological classification of vasculitis relies on the size of the affected vessel and the predominant type of inflammatory cell involved - neutrophils in acute cases, lymphocytes in chronic conditions, and histiocytes in granulomatous forms. Pathological changes progress in every context, and a single vasculitic pattern can be associated with various systemic conditions. Conversely, a single causative agent may lead to multiple distinct clinical and pathological manifestations of vasculitis. Moreover, many cases of vasculitis have no identifiable cause. A foundational understanding of the normal structure of the cutaneous vascular network is crucial. Similarly, identifying the cellular and molecular participants and their roles in forming the "dermal microvascular unit" is propedeutical.This review aims to elucidate the complex mechanisms involved in the initiation and progression of vasculitis, offering a comprehensive overview of its histopathological classification, underlying causes, and the significant role of the cutaneous vascular network and cellular dynamics. By integrating the latest insights from studies on NETosis and the implications of lymphocytic infiltration in autoimmune diseases, we seek to bridge gaps in current knowledge and highlight areas for future research. Our discussion extends to the clinical implications of vasculitis, emphasizing the importance of identifying etiological agents and understanding the diverse histopathological manifestations to improve diagnostic accuracy and treatment outcomes.
Asunto(s)
Piel , Vasculitis , Humanos , Vasculitis/patología , Vasculitis/etiología , Piel/patología , Piel/irrigación sanguínea , Neutrófilos/patología , Linfocitos/patología , Linfocitos/inmunología , Enfermedades Cutáneas Vasculares/patología , Enfermedades Cutáneas Vasculares/inmunología , Enfermedades Cutáneas Vasculares/etiología , Enfermedades Cutáneas Vasculares/diagnósticoRESUMEN
Introduction. The identification of mitotic figures is essential for the diagnosis, grading, and classification of various different tumors. Despite its importance, there is a paucity of literature reporting the consistency in interpreting mitotic figures among pathologists. This study leverages publicly accessible datasets and social media to recruit an international group of pathologists to score an image database of more than 1000 mitotic figures collectively. Materials and Methods. Pathologists were instructed to randomly select a digital slide from The Cancer Genome Atlas (TCGA) datasets and annotate 10-20 mitotic figures within a 2â mm2 area. The first 1010 submitted mitotic figures were used to create an image dataset, with each figure transformed into an individual tile at 40x magnification. The dataset was redistributed to all pathologists to review and determine whether each tile constituted a mitotic figure. Results. Overall pathologists had a median agreement rate of 80.2% (range 42.0%-95.7%). Individual mitotic figure tiles had a median agreement rate of 87.1% and a fair inter-rater agreement across all tiles (kappa = 0.284). Mitotic figures in prometaphase had lower percentage agreement rates compared to other phases of mitosis. Conclusion. This dataset stands as the largest international consensus study for mitotic figures to date and can be utilized as a training set for future studies. The agreement range reflects a spectrum of criteria that pathologists use to decide what constitutes a mitotic figure, which may have potential implications in tumor diagnostics and clinical management.
RESUMEN
CONTEXT.: Artificial intelligence algorithms hold the potential to fundamentally change many aspects of society. Application of these tools, including the publicly available ChatGPT, has demonstrated impressive domain-specific knowledge in many areas, including medicine. OBJECTIVES.: To understand the level of pathology domain-specific knowledge for ChatGPT using different underlying large language models, GPT-3.5 and the updated GPT-4. DESIGN.: An international group of pathologists (n = 15) was recruited to generate pathology-specific questions at a similar level to those that could be seen on licensing (board) examinations. The questions (n = 15) were answered by GPT-3.5, GPT-4, and a staff pathologist that recently passed their Canadian pathology licensing exams. Participants were instructed to score answers on a 5-point scale and to predict which answer was written by ChatGPT. RESULTS.: GPT-3.5 performed at a similar level to the staff pathologist, while GPT-4 outperformed both. The overall score for both GPT-3.5 and GPT-4 was within the range of meeting expectations for a trainee writing licensing examinations. In all but one question, the reviewers were able to correctly identify the answers generated by GPT-3.5. CONCLUSIONS.: By demonstrating the ability of ChatGPT to answer pathology-specific questions at a level similar to (GPT-3.5) or exceeding (GPT-4) a trained pathologist, this study highlights the potential of large language models to be transformative in this space. In the future, more advanced iterations of these algorithms with increased domain-specific knowledge may have the potential to assist pathologists and enhance pathology resident training.
RESUMEN
AIMS: Pathology education is a core component of medical training, and its literature is critical for refining educational modalities. We performed a cross-sectional bibliometric analysis to explore publications on pathology education, focusing on new medical education technologies. METHODS: The analysis identified 64 pathology journals and 53 keywords. Relevant articles were collected using a web application, PaperScraper, developed to accelerate literature search. Citation data were collected from multiple sources. Descriptive statistics, with time period analysis, were performed using Microsoft Excel and visualised with Flourish Studio. Two article groups were further investigated with a bibliometric software, VOSViewer, to establish co-authorship and keyword relationships. RESULTS: 8946 citations were retrieved from 905 selected articles. Most articles were published in the last decade (447, 49.4%). The top journals were Archives of Pathology & Laboratory Medicine (184), Human Pathology (122) and the American Journal of Clinical Pathology (117). The highest number of citations was found for Human Pathology (2120), followed by Archives of Pathology & Laboratory Medicine (2098) and American Journal of Clinical Pathology (1142). Authors with different backgrounds had the greatest number of articles and citations. 12 co-authorship, 3 keyword and 8 co-citation clusters were found for the social media/online resources group, 8 co-authorship, 4 keyword and 7 co-citation clusters for the digital pathology/virtual microscopy/mobile technologies group. CONCLUSIONS: The analysis revealed a significant increase in publications over time. The emergence of digital teaching and learning resources played a major role in this growth. Overall, these findings underscore the transformative potential of technology in pathology education.
Asunto(s)
Bibliometría , Humanos , Estados Unidos , Estudios TransversalesRESUMEN
One of the goals of pathology is to standardize laboratory practices to increase the precision and effectiveness of diagnostic testing, which will ultimately enhance patient care and results. Standardization is crucial in the domains of tissue processing, analysis, and reporting. To enhance diagnostic testing, innovative technologies are also being created and put into use. Furthermore, although problems like algorithm training and data privacy issues still need to be resolved, digital pathology and artificial intelligence are emerging in a structured manner. Overall, for the field of pathology to advance and for patient care to be improved, standard laboratory practices and innovative technologies must be adopted. In this paper, we describe the state-of-the-art of automation in pathology laboratories in order to lead technological progress and evolution. By anticipating laboratory needs and demands, the aim is to inspire innovation tools and processes as positively transformative support for operators, organizations, and patients.
RESUMEN
Objective: To evaluate intra-observer diagnostic reproducibility using traditional slides (TS) versus whole slide images (WSI). Methods: TS and WSI of 1427 prostatic biopsies (107 consecutive patients) were evaluated by a single pathologist. Agreement between readings was evaluated with Gwet's Agreement coefficient (AC) and Landis and Koch benchmark scale. Results: The positive/negative agreement between the readings was almost perfect (AC1= 0.962; 95% CI[0.949,0.974]), with method independent distribution of discrepancies. Among positive biopsies, 212 had identical Gleason score (GS) on TS and WSI and discordant GS in 69 cases (AC2 = 0.932; 95% CI[0.907, 0.956]). Concordant negative and positive patient classification was observed in 39 and 64 cases, respectively; two cases were assigned to the positive group on TS and 2 on WSI configuring an almost perfect agreement (AC1=0.929; 95% C1[0.860, 0.998]). ISUP Grade group (ISUP GG) agreement was evaluated in the 60 concordantly positive cases: in 45 cases it was identical on TS and WSI; in 10 biopsies the discrepancy implied a modification of the assigned ISUP GG of ≤ 1 class and in 5 the discrepancy implied a modification of 2 classes. Gwet's agreement coefficient was (95% CI [0.834, 0.962]), i.e.: almost perfect agreement. Conclusions: Our data show almost perfect agreement between digital and traditional diagnostic activity in a routine setting, confirming that digital pathology can be safely introduced into routine workflows.
Asunto(s)
Patólogos , Próstata , Masculino , Humanos , Reproducibilidad de los Resultados , Flujo de Trabajo , BiopsiaRESUMEN
AIMS: We investigated the trend in case reports (CRs) publication in a sample of pathology journals. Furthermore, we proposed an alternative publishing route through new digital communication platforms, represented by the 'social media case report'. METHODS: 28 pathology journals were selected from SCImago database and searched in PubMed to identify the number of published CRs. Four reference decades (1981-2020) were selected. The 5-year impact factor (IF) was retrieved from the Academic Accelerator database. RESULTS: CRs increased during the first three decades (6752, 8698 and 11148, respectively; mean values: 355, 27.3%; 334, 26.4%; 398, 28.8%) as the number of CR-publishing journals (19, 26 and 28, respectively). In the last decade, CRs significantly decreased (9341; mean 334, 23.6%) without variation in the number of CR-publishing journals (28). Half of the journals reduced CRs (from -1.1% to -37.9%; mean decreasing percentage -14.7%), especially if active since the first decade (11/14, 79%); the other half increased CRs (from +0.5% to +34.2%; mean increasing percentage +11.8%), with 8/14 (57%) starting publishing in the first decade. The 5-year IF ranged from 0.504 to 5.722. Most of the journals with IF ≥2 (10/14, 71%) reduced the CRs number, while 71% of journals with IF <2 increased CRs publication (especially journals with IF <1, +15.1%). CONCLUSIONS: CRs publication decreased during the last decade, especially for journals which are older or have higher IF. Social media CRs may represent a valid alternative and by using standardised templates to enter all relevant data may be organised in digital databases and/or transformed in traditional CRs.
Asunto(s)
Bases de Datos Factuales , HumanosRESUMEN
We aimed to overcome intratumoral heterogeneity in clear cell renal cell carcinoma (clearRCC). One hundred cases of clearRCC were sampled. First, usual standard sampling was applied (1 block/cm of tumor); second, the whole tumor was sampled, and 0.6 mm cores were taken from each block to construct a tissue microarray; third, the residual tissue, mapped by taking pieces 0.5 × 0.5 cm, reconstructed the entire tumor mass. Precisely, six randomly derived pieces of tissues were placed in each cassette, with the number of cassettes being based on the diameter of the tumor (called multisite 3D fusion). Angiogenic and immune markers were tested. Routine 5231 tissue blocks were obtained. Multisite 3D fusion sections showed pattern A, homogeneous high vascular density (10%), pattern B, homogeneous low vascular density (8%) and pattern C, heterogeneous angiogenic signatures (82%). PD-L1 expression was seen as diffuse (7%), low (33%) and absent (60%). Tumor-infiltrating CD8 scored high in 25% (pattern hot), low in 65% (pattern weak) and zero in 10% of cases (pattern desert). Grading was upgraded in 26% of cases (G3-G4), necrosis and sarcomatoid/rhabdoid characters were observed in, respectively, 11 and 7% of cases after 3D fusion (p = 0.03). CD8 and PD-L1 immune expressions were higher in the undifferentiated G4/rhabdoid/sarcomatoid clearRCC subtypes (p = 0.03). Again, 22% of cases were set to intermediate to high risk of clinical recurrence due to new morphological findings of all aggressive G4, sarcomatoid/rhabdoid features by using 3D fusion compared to standard methods (p = 0.04). In conclusion, we propose an easy-to-apply multisite 3D fusion sampling that negates bias due to tumor heterogeneity.
RESUMEN
The concept of "tigroid" background is used in cytology to describe a peculiar smear background characterized by the presence of a relatively granular, reticulated material that was described as "foamy, lazy, tiger-striped or astrakhan". It was used to describe the background seen in smears obtained from seminoma. In addition to seminoma, we now know that it can be present in different tumours, mostly carcinomas and round cell sarcomas. These share with seminoma a cytoplasm with high glycogen content and many times clear cell morphology. The "tigroid" background is seen when smears are air-dried and Romanowsky-based stains are used (May-Grunwald-Giemsa and Diff-Quik stains). It is only seen in fine needle aspiration or intraoperative squashing or scrapping samples, but not in specimens obtained from effusions or liquid-based cytology. Wet-fixed cytologic samples with alcohol or with formaldehyde tend to dissolve the background so it is not usually present in Papanicolaou stained smears. In this review, we discuss tumours in which the "tigroid" background is observed and its potential diagnostic utility and aetiology. It is interesting to remark that except for parathyroid adenoma and adenomatoid tumour all the neoplasms in which this background has been observed are malignant.
Asunto(s)
Seminoma , Neoplasias Testiculares , Biopsia con Aguja Fina , Citodiagnóstico , Humanos , MasculinoRESUMEN
AIMS: According to The Cancer Genome Atlas (TCGA), around 9% of bladder carcinomas usually show abnormalities of the murine double minute 2 (MDM2) gene, but a few studies have been investigated them. We profiled MDM2 gene amplification in a series of urothelial carcinomas (UC) considering the molecular subtypes and expression of programmed death ligand 1 (PD-L1). METHODS: 117 patients with muscle-invasive UC (pT2-3) without (N0) or with (N+) lymph-node metastases were revised. Only cases with availability of in toto specimens and follow-up were studied. Tissue microarray was built. p53, ER, RB1, GATA-3, CK20, CK5/6, CD44 and PD-L1 (clone sp263) immunoexpression was evaluated. Fluorescent in situ hybridisation was assessed by using the HER-2/neu, FGFR-3, CDKN2A and MDM2 probes. True (ratio 12q/CEP12 >2) MDM2 gene amplification was distinguished from polyploidy/gains (ratio <2, absolute copy number of MDM-2 >2). MDM2 and PD-L1 values were correlated to the TCGA molecular phenotypes. Statistical analysis was performed. RESULTS: 6/50 (12%) cases (5 N0 and 1 N+) were amplified for MDM2 without matching to molecular phenotypes. Of 50, 14 (37%) cases expressed PD-L1 at 1% cut-off; 3/50 (9%) at >50% cut-off; of these, 2 cases on side of neoplasia among inflammatory cells. Only one out of six (17%) cases amplified for MDM2 showed expression (>50% cut-off) of PD-L1. MDM2 amplification was independent to all documented profiles (k test=0.3) and was prevalent in recurrent UC. CONCLUSION: MDM2 amplification has been seen in both PD-L1 positive and negative muscle-invasive bladder UC independently from the TCGA molecular phenotypes. MDM2 and PD-L1 might be assessed in order to predict a better response to combo/single targeted therapies.
Asunto(s)
Antígeno B7-H1/metabolismo , Proteínas Proto-Oncogénicas c-mdm2/genética , Neoplasias de la Vejiga Urinaria/genética , Anciano , Anciano de 80 o más Años , Antígeno B7-H1/genética , Biomarcadores/metabolismo , Femenino , Amplificación de Genes , Humanos , Inmunohistoquímica , Hibridación Fluorescente in Situ , Metástasis Linfática , Masculino , Persona de Mediana Edad , Músculos/patología , Fenotipo , Proteínas Proto-Oncogénicas c-mdm2/metabolismo , Análisis de Matrices Tisulares , Vejiga Urinaria/patología , Neoplasias de la Vejiga Urinaria/patología , Urotelio/patologíaRESUMEN
The term 'infographics' is a blend of the two words "information" and "graphics". Infographics can be described as 'information visualizations', conceived as visual translation of data including text, numbers, graphs, charts, drawings and so on. Visual representations are a fundamental part of scientific communication. They match the need to organize different pieces of information in a coherent and synthetic structure and constitute one of the most effective methods scientists rely on to divulge their findings. In particular, infographics provide an overview of key points regarding specific topics in a form that promotes quick learning and knowledge retention. They can be presented in printed or digital formats, being the latter particularly suitable for a global-scale diffusion via social media or websites.In recent years, many pathologists have started developing digital infographics as a strategy for providing free educational contents on Facebook, Twitter or websites. In the present review, we focus on the value of digital infographics to summarize various aspects of Surgical and General Pathology. They shed light on diagnostic criteria, differentials and predictive/prognostic markers for many diseases, being a useful learning tool both for residents and practicing pathologists. In this paper, the model of infographics ideation, processing and sharing to an online audience is described and the impact of infographics on knowledge processes in Pathology is investigated.
Asunto(s)
Patología Quirúrgica , Medios de Comunicación Sociales , Visualización de Datos , HumanosRESUMEN
The magnitude of the diagnostic benefit conferred by performing histopathological examinations after medico-legal/forensic autopsies remains debatable. We have tried to address this issue by reviewing a series of histopathology referrals concerning medico-legal autopsies in real-world routine practice. We present an audit of the consultations provided to forensics by clinical pathologists at our institute between 2015 and 2018. Over this period, 493 post-mortem examinations were performed by forensic pathologists. Of these cases, 52 (11%) were referred for histopathology. Gross assessment was requested in 22/52 (42%) cases. Histopathology examination was performed on single organs in 15/52 (29%) cases, primarily on the lung and heart, whereas parenchymatous multi-organ analysis was carried out in 14/52 (27%) cases. Bone-marrow sampling was studied in 4/52 (8%) cases. Immunohistochemistry was needed in 16/52 (31%) cases, special stains in 9/52 (21%) cases and molecular analysis in 4/52 (8%) cases. Focusing on technical processes, standard methodology on pre-analytical procedures was changed in 10/52 (19%) cases in order to answer specific diagnostic questions. We showed that although most of the time the diagnosis is clear by the end of dissection on the basis of the macroscopic findings, histopathology can provide, modify or confirm the cause of death in many medico-legal/forensic cases. Therefore, it is desirable that forensic pathologists and clinical pathologists establish robust working relationships in a cooperative environment. We conclude that it is important to implement guidelines based on real-world routine practice in order to identify cases where histopathology can provide useful contributions, which in our experience applied to 11% of forensic cases.
Asunto(s)
Autopsia , Patologia Forense/métodos , Patología Clínica/métodos , Derivación y Consulta , Guías como Asunto , Humanos , Patólogos/clasificación , Patólogos/normasRESUMEN
Pure invasive papillary carcinoma (IPC) is a rare subtype of breast carcinoma with good prognosis compared with classical invasive breast carcinoma (IBC) of no special type. The majority of IPC are estrogen receptor and progesterone receptor (ER/PR) positive and HER2 negative (luminal A-like). We report the case of a 72-year-old women who was referred to the Senology Clinic for a routine workup following surgery for an intraductal papilloma. The core needle biopsy (CNB) showed a lesion mainly composed of irregular papillae and micropapillae with apocrine epithelial cells of low-to-intermediate nuclear grade, without a myoepithelial cell layer within the papillae and at the periphery, as demonstrated with multiple immunostains. The diagnosis of apocrine papillary lesion of uncertain malignant potential was made. The subsequent lumpectomy showed an IBC with the same cyto-architectural features as the CNB. In addition, lymphovascular invasion and papillary/micropapillary apocrine in situ lesion were noted. Notably, the tumor was ER/PR and HER2 negative and strongly positive for androgen receptor. A final diagnosis of mixed apocrine papillary/micropapillary carcinoma with triple-negative status was made. To the best of our knowledge, this is the first report of an IBC with these features. Breast pathologists should be aware of this entity when dealing with CNB samples characterized by a complex papillary lesion with apocrine atypia that lacks a myoepithelial cell layer on multiple immunostains. These lesions should be classified at least as of uncertain malignant potential based on the cyto-architectural features prompting a surgery for removal.
Asunto(s)
Glándulas Apocrinas/patología , Carcinoma Papilar/diagnóstico , Glándulas Mamarias Humanas/patología , Neoplasias Complejas y Mixtas/diagnóstico , Neoplasias de la Mama Triple Negativas/diagnóstico , Anciano , Glándulas Apocrinas/cirugía , Biopsia con Aguja Gruesa , Carcinoma Papilar/patología , Carcinoma Papilar/cirugía , Femenino , Humanos , Glándulas Mamarias Humanas/cirugía , Mastectomía Segmentaria , Neoplasias Complejas y Mixtas/patología , Neoplasias Complejas y Mixtas/cirugía , Resultado del Tratamiento , Neoplasias de la Mama Triple Negativas/patología , Neoplasias de la Mama Triple Negativas/cirugíaRESUMEN
In recent years, digital and communication technology has been changing the way to spread knowledge in Medicine. In the field of Pathology, several remote learning resources have been made available through multiple social media platforms (Facebook, Twitter and others), YouTube channels and dedicated Websites, with a growing number of freely available lectures or tutorials, broadcasted live and/or archived for on-demand viewing. All these internet-based resources enable the pursuit of a flexible, independent, self-motivated and self-directed way of learning that fits perfectly with the increasing limitations of time, space and speed of modern day learners.These resources have played a significant role in filling the void of conventional education during the ongoing Covid-19 pandemic. Moreover, with their widespread diffusion throughout communities of Pathologists from all over the world they help to reduce the educational gap between resource-rich and resource-poor countries, having the potential to become standardized knowledge-sharing platforms and to be incorporated into curricula at any level.pathCast is one of the most robust and reliable open-access online remote learning platforms for pathologists, which live-streams lectures across the world. In the present paper we describe its structure, its acceptance by the global community of pathologists, what innovation elements has introduced regarding methodologies for education and its powerful and positive impact for residency training and continuing life-long education of practicing pathologists. A comprehensive list of the pathCast lectures with the respective links is also provided along with a brief discussion on other freely accessible online educational resources for pathologists.
Asunto(s)
Educación a Distancia , Patología/educación , Educación Médica Continua , Internado y ResidenciaRESUMEN
Lung cancer is still one of the leading cause of death worldwide. The clinical variability of lung cancer is high and drives treatment decision. In this context, correct discrimination of pulmonary neuroendocrine tumors is still of critical relevance. The spectrum of neuroendocrine tumors is various, and each type has molecular and phenotypical differences. In order to advance in the discrimination of neuroendocrine from non-neuroendocrine lung tumors, we tested a series of 95 surgically resected and formalin-fixed paraffin embedded lung cancer tissues, and we analyzed the expression of miR205-5p and miR375-3p via TaqMan RT-qPCR. Via a robust mathematical approach, we excluded technical outliers increasing the data reproducibility. We found that miR375-3p levels are higher in low-grade neuroendocrine lung tumor samples compared to non-neuroendocrine lung tumors. However, miR375-3p is not able to distinguish among different types of neuroendocrine lung tumors. In this work, we provide a new molecular marker for distinguishing non-neuroendocrine from low-grade neuroendocrine lung tumors samples establishing an easy miRNA score to be used in clinical settings, enabling the pathologist to classify more accurately lung tumors biopsies, which may be ambiguously cataloged in routine examination.
RESUMEN
Artificial Intelligence is exponentially increasing its impact on healthcare. As deep learning is mastering computer vision tasks, its application to digital pathology is natural, with the promise of aiding in routine reporting and standardizing results across trials. Deep learning features inferred from digital pathology scans can improve validity and robustness of current clinico-pathological features, up to identifying novel histological patterns, e.g., from tumor infiltrating lymphocytes. In this study, we examine the issue of evaluating accuracy of predictive models from deep learning features in digital pathology, as an hallmark of reproducibility. We introduce the DAPPER framework for validation based on a rigorous Data Analysis Plan derived from the FDA's MAQC project, designed to analyze causes of variability in predictive biomarkers. We apply the framework on models that identify tissue of origin on 787 Whole Slide Images from the Genotype-Tissue Expression (GTEx) project. We test three different deep learning architectures (VGG, ResNet, Inception) as feature extractors and three classifiers (a fully connected multilayer, Support Vector Machine and Random Forests) and work with four datasets (5, 10, 20 or 30 classes), for a total of 53, 000 tiles at 512 × 512 resolution. We analyze accuracy and feature stability of the machine learning classifiers, also demonstrating the need for diagnostic tests (e.g., random labels) to identify selection bias and risks for reproducibility. Further, we use the deep features from the VGG model from GTEx on the KIMIA24 dataset for identification of slide of origin (24 classes) to train a classifier on 1, 060 annotated tiles and validated on 265 unseen ones. The DAPPER software, including its deep learning pipeline and the Histological Imaging-Newsy Tiles (HINT) benchmark dataset derived from GTEx, is released as a basis for standardization and validation initiatives in AI for digital pathology.
Asunto(s)
Algoritmos , Inteligencia Artificial , Técnicas Histológicas/métodos , Interpretación de Imagen Asistida por Computador/métodos , Programas Informáticos , Humanos , Reproducibilidad de los ResultadosRESUMEN
BACKGROUND: Acute kidney injury is a treatable entity although difficult to recognize without diagnostic biopsy. We investigated the potential association between clinically defined deceased donors and acute kidney injury with preimplantation histological findings and recipient outcomes. METHODS: Kidney biopsies from donors were classified using the Acute Kidney Injury Network criteria and assessed for percentage glomerulosclerosis, tubular atrophy, interstitial fibrosis, and vascular narrowing with the Remuzzi score and for acute tubular necrosis. Differences in incidence rates of delayed graft function (DGF) and cumulative rejection episodes were compared between recipients transplanted with normal and 3 levels of acute kidney injury using the analysis of variance with Bonferroni correction ( P = .0012). RESULTS: Sixteen out of 335 donors showed a severe acute kidney injury level 3 with a median serum creatinine of 458 µmol/L. Fourteen (88%) had 0-3 Remuzzi score and were used for single kidney transplantation and 2 (12%) were used for dual kidney transplantation (score: 4-6). Recipients who received a kidney from a donor with level 3 acute kidney injury had a higher percentage of DGF (47%) without statistical significance ( P = .008). The rate of cumulative rejection (45%) at 2 years was not significantly increased ( P = .09). CONCLUSIONS: Recipients receiving level 3 acute kidney injury kidneys, selected with Remuzzi histopathological score and acute tubular necrosis assessment, had a greater incidence of DGF but a similar long-term cumulative rejection compared to no injury and level 1 and level 2 acute kidney injury donors. The application of the histopathological examination allowed expansion of the kidney donor pool.
Asunto(s)
Lesión Renal Aguda/patología , Funcionamiento Retardado del Injerto/epidemiología , Rechazo de Injerto/epidemiología , Fallo Renal Crónico/cirugía , Trasplante de Riñón , Riñón/patología , Complicaciones Posoperatorias/epidemiología , Trasplantes/patología , Lesión Renal Aguda/sangre , Adolescente , Adulto , Anciano , Anciano de 80 o más Años , Atrofia , Creatinina/sangre , Femenino , Fibrosis , Humanos , Necrosis de la Corteza Renal/patología , Glomérulos Renales/patología , Masculino , Persona de Mediana Edad , Estudios Retrospectivos , Esclerosis , Índice de Severidad de la Enfermedad , Donantes de Tejidos , Resultado del Tratamiento , Adulto JovenRESUMEN
INTRODUCTION: Autophagy plays a role in various central nervous system diseases. Little is known about its molecular activation in drug addiction. Our aim was to investigate the signalling pathways of autophagy in brain tissues from drug abusers. METHODS: Twenty-five drug abusers with acute lethal intoxication and 10 controls were medico-legally autopsied. Brain-tissue samples from the parietal cortex and cerebellum were obtained. Expression of LC3B, phospho-mTOR (ph-mTOR) and phospho70S6 Kinase (p70S6K) was identified in tissue microarrays, with three tissue spots per case. Blood, urine or vitreous humour were tested in all cases to identify the acute intoxication. Hair analysis was performed in 14 cases to confirm chronic intoxication; the remaining cases had a documented medical history of chronic abuse. RESULTS: The autophagy marker LC3B was always positive on both the cortex and the cerebellum, stratified as strongly in 18 (72%) cases and weakly positive in seven (28%) cases. ph-mTOR was negative in all cases. The p70S6K molecule showed positivity in 14 (56%) cases on cortex tissue. The cerebellum was always negative, except for Purkinje cells. Drug abusers had statistically more double positive cases (LC3B-p70S6K) than controls ( p=0.0094). CONCLUSION: Autophagy pathways were activated in our series, and 56% of drug abusers showed simultaneous LC3B-p70S6K immunoexpression on tissue from the parietal cortex and cerebellum. This may be of value in autopsy practice as an indicator of brain damage due to drug abuse and could serve as alternative or additional double sensitive diagnostic method to detect drug-related deaths using a tissue-based rationale.