Modern Pathology最新文献

This manuscript serves as an introduction to a comprehensive seven-part review article series on artificial intelligence (AI) and machine learning (ML) and their current and future influence within pathology and medicine. This introductory review provides a comprehensive grasp of this fast-expanding realm and its potential to transform medical diagnosis, workflow, research, and education. Fundamental terminology employed in AI-ML is covered using an extensive dictionary. The article also provides a broad overview of the main domains in the AI-ML field, encompassing both generative and non-generative (traditional) AI. Thereby serving as a primer to the other six review articles in this series that describe the details about statistics, regulations, bias, ethical dilemmas, and ML-Ops in AI-ML. The intent of these review articles is to better equip individuals who are or will be working in an AI-enabled healthcare system.

Embryonic-type neuroectodermal tumors (ENTs) arising from testicular germ cell tumors (GCTs) are a relatively common type of somatic transformation in GCTs with poor prognosis and limited therapeutic options, particularly when patients develop disease recurrence or metastasis. Knowledge of key events driving this transformation is limited to the paucity of comprehensive genomic data. We performed a retrospective database search in a Clinical Laboratory Improvement Amendments- and College of American Pathologists-certified laboratory for testicular GCT-derived ENTs that had previously undergone next-generation sequencing-based comprehensive genomic profiling during the course of clinical care. Clinicopathological and genomic data were centrally rereviewed. Here, we report the molecular features of 10 ENTs of testicular GCT origin. All tumors harbored gain of chromosome 12p, often with KRAS, CCND2, and KMD5A coamplification, supporting a germ cell origin. The tumors were microsatellite-stable and exhibited a low tumor mutational burden. Three tumors (30%) exhibited MYCN or MYC amplification with co-occurring inactivation of the p53 pathway via either TP53 mutations or MDM2 amplification in 2 tumors. Three additional tumors (30%) had activation of the PI3K pathway via PIK3CA and PIK3CG mutations or PIK3C2B amplification; 1 tumor with co-occurring CDK4 amplification. Gene rearrangements were detected in 3 tumors (30%), with novel BRD4::MAU2 and BCOR::CLIP2 fusions as well as an internal truncating ATRX rearrangement, respectively. In summary, ENTs arising from GCTs are molecularly heterogeneous; however, a large fraction of testicular ENTs could be stratified by 2 distinct sets of genetic alterations, including MYCN/MYC amplification with concurrent suppression of the p53 pathway, and activation of the PI3K pathway with co-occurring CDK4 amplification. Moreover, the novel gene fusions identified in a subset of testicular GCT-derived ENTs overlap with molecularly defined tumors of embryonic-type neuroectodermal features in the central nervous system, indicating the potential common driving events for tumorigenesis from different anatomical sites.

For 2 decades, the American Society of Clinial Oncology-College of American Pathologists human epidermal growth factor receptor 2 (HER2) testing criteria have included 0 and 1+ scores, but this distinction was inconsequential. Now, based on the DESTINY Breast-04 Trial (DB-04) results, for patients with metastatic breast cancer it underpins eligibility for trastuzumab-deruxtecan treatment. Discerning 0 from 1+ immunohistochemistry (IHC) staining is challenging, as HER2 low is not a biologically distinct cancer subset, there are no reference standards or controls, and second-tier tests (eg, in situ hybridization) do not apply. Prior reports cast doubt on the reliability of pathologists' IHC scoring, with resulting treatment misalignments. With institutional review board approval, our group of 9 breast pathologists from 8 Australian laboratories had previously established HER2-low-focused scoring conventions, based on the American Society of Clinial Oncology-College of American Pathologists 2018 HER2 guidelines, and specifying common staining pitfalls. We reported the results of the first set of 60 breast cancers evaluated with these methods. After a 5-month washout, for the present validation study, we have compiled a second set of 64 HER2-negative invasive breast cancer core biopsies, all assessed with the Ventana 4B5 HER2 assay. We have each scored digitized images of HER2 IHC slides of the cases. Using the majority opinion as the target score, we have calculated our performance metrics. We have compared the results of our performance in set 1 and set 2 to assess the effectiveness of our approach and learning retention. The cases in this validation set included 40 (62.5%) HER2 low, 10 (17.2%) ultralow (UL), and 13 (18.8%) null cancers. Concordance was not achieved in 1 case. For distinguishing HER2 low from other cancers (UL and null combined) the mean values of our performance metrics were accuracy 89.58%, sensitivity 90.83%, specificity 87.50%, positive predictive value 95.63%, negative predictive value 83.59%, and Cohen kappa score 0.81. Comparing these results with our initial study, we have maintained our high level of performance across these parameters. Our mean kappa score is now in the excellent range for concordance. Maintaining high performance across a range of measures in 2 separate data sets validates the effectiveness of our HER2-low-focused scoring conventions. Having validated our approach, we will use these reference case sets with expert-level consensus scores for peer training and updating our national HER2 IHC external quality assurance program. In our ongoing studies, we are also assessing the performance of software algorithms to determine their suitability for the prescreening of HER2 IHC slides.

Anal squamous cell carcinoma (SCC) incidence has increased, and treatment has shifted from surgery to chemoradiotherapy (CRT), with salvage abdominoperineal resection (APR) being reserved for persistent/recurrent cases. This study evaluates the utility of different Tumor Regression Scoring Systems (TRSS) in predicting survival in anal SCC patients, using pathologists' observations and digital pathology. Cases managed surgically from 2005 to 2019 were collected. Residual tumor was assessed by multiple methods (gross tumor size, largest focus of tumor on H&E slide, average of residual tumor in all submitted H&E slides, JES, Chirieac, Schneider, Hermann, CAP scoring system). Three expert pathologists individually estimated ("eyeballed") the residual tumor % based on residual tumor/tumor bed (single representative H&E slide). The QuPath software was used to measure tumor volume on the same slide. American Joint Committee on Cancer (AJCC) 8th staging, and outcome data were retrieved from electronic medical records. The study involved 48 participants, predominantly female (56%), with a median age of 57. Most were Caucasian. HPV-positive was present in 77% of those assessed (17/22). Initial treatment included chemoradiation (CRT), followed by APR (79%) or pelvic exenteration (21%). Complications (13%), persistent disease (33%), and recurrence (54%) led to surgical interventions. 51% had moderately differentiated SCC, whereas 42% were poorly differentiated. Lymphovascular invasion (44%), perineural invasion (38%), and lymph node metastasis (13%) were present. Distant metastasis was rare (2%). Median overall survival was 3.2 years. Positive margins (HR 4.12, 95% CI 1.83, 9.28) and larger tumor size (HR 1.02 95% CI 1.01, 1.03) were associated with an increased hazard of death. Most residual tumor measurement methods were not significantly associated with overall survival. Interobserver agreement (based on "eyeballing") was moderate (kappa 0.4). Computational pathology-based residual tumor percentage was the only method significantly associated with outcome, with each 10% increase in the residual tumor percentage corresponding to a 1.23-fold higher hazard death (95% CI 1.03, 1.46; p=0.024). This study highlights computational pathology's important role in predicting outcomes in anal SCC treated with CRT and surgery. Specifically, the computational assessment of the residual tumor percentage proves to be a strong predictor of overall survival, outperforming other established TRSS methods.

Molecular testing of tumor samples for targetable biomarkers is restricted by a lack of standardization, turnaround time, cost, and tissue availability across cancer types. Additionally, targetable alterations of low prevalence may not be tested in routine workflows. Algorithms that predict DNA alterations from routinely generated hematoxylin and eosin-stained images could prioritize samples for confirmatory molecular testing. Costs and the necessity of a large number of samples containing mutations limit approaches that train individual algorithms for each alteration. In this work, models were trained for simultaneous prediction of multiple DNA alterations from hematoxylin and eosin images using a multitask approach. Compared with biomarker-specific models, this approach performed better on average, with pronounced gains for rare mutations. The models reasonably generalized to independent temporal holdout, externally stained, and multisite The Cancer Genome Atlas test sets. Additionally, whole slide image embeddings derived using multitask models demonstrated strong performance in downstream tasks that were not a part of training. Overall, this is a promising approach to develop clinically useful algorithms that provide multiple actionable predictions from a single slide.

Penile squamous cell carcinoma (pSCC) represents an uncommon malignancy characterized by stagnant mortality, psychosexual distress, and a highly variable prognosis. Currently, the World Health Organization distinguishes between human papillomavirus (HPV)-related and HPV-independent pSCC. Recently, there has been an evolving line of research documenting the enrichment of HPV-independent pSCC with a high tumor mutational burden (TMB) and programmed death ligand-1 expression, as well as clusters of genes associated with HPV status. In this study, we conducted comprehensive next-generation sequencing DNA profiling of 146 pSCC samples using a panel consisting of 355 genes associated with tumors. This profiling was correlated with immunohistochemical markers and prognostic clinical data. A survival analysis of recurrent genomic events (found in ≥10 cases) was performed. TP53, CDKN2A, ATM, EPHA7, POT1, CHEK1, GRIN2A, and EGFR alterations were associated with significantly shortened overall survival in univariate and multivariate analysis. HPV positivity, diagnosed through both p16 immunohistochemistry and HPV DNA analysis, displayed no impact on survival but was associated with high-grade, lymphatic invasion, programmed death ligand-1 negativity/weak expression, and low TMB. FAT1, TP53, CDKN2A, CASP8, and HRAS were more often mutated in HPV-independent pSCC. In contrast, HPV-associated pSCCs were enriched by EPHA7, ATM, GRIN2A, and CHEK1 mutations. PIK3CA, FAT1, FBXW7, and KMT2D mutations were associated with high TMB. NOTCH1, TP53, CDKN2A, POT1, KMT2D, ATM, CHEK1, EPHA3, and EGFR alterations were related to adverse clinicopathologic signs, such as advanced stage, high tumor budding, and lymphovascular invasion. We detected 160 alterations with potential treatment implications, with 21.2% of samples showing alterations in the homologous recombination repair pathway. To the best of our knowledge, this study describes the largest cohort of pSCC with complex molecular pathologic, clinical, and prognostic analysis correlating with prognosis.

As artificial intelligence (AI) gains prominence in pathology and medicine, the ethical implications and potential biases within such integrated AI models will require careful scrutiny. Ethics and bias are important considerations in our practice settings, especially as increased number of machine learning (ML) systems are being integrated within our various medical domains. Such machine learning based systems, have demonstrated remarkable capabilities in specified tasks such as but not limited to image recognition, natural language processing, and predictive analytics. However, the potential bias that may exist within such AI-ML models can also inadvertently lead to unfair and potentially detrimental outcomes. The source of bias within such machine learning models can be due to numerous factors but can be typically put in three main buckets (data bias, development bias and interaction bias). These could be due to the training data, algorithmic bias, feature engineering and selection issues, clinical and institutional bias (i.e. practice variability), reporting bias, and temporal bias (i.e. changes in technology, clinical practice or disease patterns). Therefore despite the potential of these AI-ML applications, their deployment in our day to day practice also raises noteworthy ethical concerns. To address ethics and bias in medicine, a comprehensive evaluation process is required which will encompass all aspects such systems, from model development through clinical deployment. Addressing these biases is crucial to ensure that AI-ML systems remain fair, transparent, and beneficial to all. This review will discuss the relevant ethical and bias considerations in AI-ML specifically within the pathology and medical domain.

The syndromes of thrombotic microangiopathy (TMA) are associated with acute kidney injury and end-stage kidney disease. TMAs typically present with thrombocytopenia and microangiopathic hemolytic anemia (ie, systemic TMA). Kidney-limited TMA can occur, although often overlooked and undertreated. In this study, we studied the etiology and outcome of kidney-limited TMA. Patients with TMA on kidney biopsy, either systemic or kidney-limited, were recruited and classified as definite complement-mediated (C-)TMA (ie, ≥1 pathogenic complement gene variant), probable C-TMA (ie, massive ex vivo C5b9 formation without a pathogenic complement gene variant), and non (n)C-TMA (ie, normal ex vivo C5b9 formation). Morphologic features of TMA on kidney biopsy and their clinical correlates were studied. Patients were classified as definite C-TMA (N = 14; 18%), probable C-TMA (N = 21; 27%), or nC-TMA (N = 42; 55%), including 51 (66%) out of 77 patients with kidney-limited TMA. Patients with definite and probable C-TMA often presented with hemolysis (79% and 62% vs 34%; P = .007), glomerular thrombosis (79% and 76% vs 43%), a higher creatinine level (974 and 502 vs 280 μmol/L; P = .001), and a younger age (33 and 33 vs 40 years; P = .029) as compared with nC-TMA. Morphologic features neither defined etiology nor differed between systemic and kidney-limited TMA. Eculizumab improved kidney outcomes in patients with kidney-limited C-TMA but not in those with nC-TMA akin to patients with systemic C-TMA. Kidney outcomes were not affected by chronicity grading on kidney biopsy. Kidney-limited TMA is common in diverse TMAs, including C-TMA. A kidney biopsy is needed to detect TMA at the earliest possible stage of the disease. Morphology does not allow for the identification of etiology, and patients with kidney-limited TMA should therefore be screened for complement dysregulation, having a major impact on treatment and prognosis.

This review article builds upon the introductory piece in our 7-part series, delving deeper into the transformative potential of generative artificial intelligence (Gen AI) in pathology and medicine. The article explores the applications of Gen AI models in pathology and medicine, including the use of custom chatbots for diagnostic report generation, synthetic image synthesis for training new models, data set augmentation, hypothetical scenario generation for educational purposes, and the use of multimodal along with multiagent models. This article also provides an overview of the common categories within Gen AI models, discussing open-source and closed-source models, as well as specific examples of popular models such as GPT-4, Llama, Mistral, DALL-E, Stable Diffusion, and their associated frameworks (eg, transformers, generative adversarial networks, diffusion-based neural networks), along with their limitations and challenges, especially within the medical domain. We also review common libraries and tools that are currently deemed necessary to build and integrate such models. Finally, we look to the future, discussing the potential impact of Gen AI on health care, including benefits, challenges, and concerns related to privacy, bias, ethics, application programming interface costs, and security measures.