Fast-Track Development and Multi-Institutional Clinical Validation of an Artificial Intelligence Algorithm for Detection of Lymph Node Metastasis in Colorectal Cancer.

Lymph node metastasis (LNM) detection can be automated using artificial intelligence (AI)-based diagnostic tools. Only limited studies have addressed this task for colorectal cancer (CRC). This study aimed to develop of a clinical-grade digital pathology tool for LNM detection in CRC using the original fast-track framework. The training cohort included 432 slides from one department. A segmentation algorithm detecting 8 relevant tissue classes was trained. The test cohorts consisted of materials from 5 pathology departments digitized by 4 different scanning systems. A high-quality, large training data set was generated within 7 days and a minimal amount of annotation work using fast-track principles. The AI tool showed very high accuracy for LNM detection in all cohorts, with sensitivity, negative predictive value, and specificity ranges of 0.980 to 1.000, 0.997 to 1.000, and 0.913 to 0.990, correspondingly. Only 5 of 14,460 analyzed test slides with tumor cells over all cohorts were classified as false negative (3/5 representing clusters of tumor cells in lymphatic vessels). A clinical-grade tool was trained in a short time using fast-track development principles and validated using the largest international, multi-institutional, multiscanner cohort of cases to date, showing very high precision for LNM detection in CRC. We are releasing a part of the test data sets to facilitate academic research.

Standardized Classification of Lung Adenocarcinoma Subtypes and Improvement of Grading Assessment Through Deep Learning.

The histopathologic distinction of lung adenocarcinoma (LADC) subtypes is subject to high interobserver variability, which can compromise the optimal assessment of patient prognosis. Therefore, this study developed convolutional neural networks capable of distinguishing LADC subtypes and predicting disease-specific survival, according to the recently established LADC tumor grades. Consensus LADC histopathologic images were obtained from 17 expert pulmonary pathologists and one pathologist in training. Two deep learning models (AI-1 and AI-2) were trained to predict eight different LADC classes. Furthermore, the trained models were tested on an independent cohort of 133 patients. The models achieved high precision, recall, and F1 scores exceeding 0.90 for most of the LADC classes. Clear stratification of the three LADC grades was reached in predicting the disease-specific survival by the two models, with both Kaplan-Meier curves showing significance (P = 0.0017 and 0.0003). Moreover, both trained models showed high stability in the segmentation of each pair of predicted grades with low variation in the hazard ratio across 200 bootstrapped samples. These findings indicate that the trained convolutional neural networks improve the diagnostic accuracy of the pathologist and refine LADC grade assessment. Thus, the trained models are promising tools that may assist in the routine evaluation of LADC subtypes and grades in clinical practice.

Presence of focal usual interstitial pneumonia is a key prognostic factor in progressive pulmonary fibrosis.

AIMS: Progressive pulmonary fibrosis (PPF) is a newly recognised clinical phenotype of interstitial lung diseases in the 2022 interstitial pulmonary fibrosis (IPF) guidelines. This category is based entirely on clinical and radiological factors, and the background histopathology is unknown. Our objective was to investigate the histopathological characteristics of PPF and to examine the correlation between usual interstitial pneumonia (UIP) and prognosis in this new disease type. We hypothesised that the presence of UIP-like fibrosis predicts patients' survival in PPF cases. METHODS AND RESULTS: We selected 201 cases fulfilling the clinical criteria of PPF from case archives. Cases diagnosed as IPF by a multidisciplinary team were excluded. Whole slide images were evaluated by three pathologists who were blinded to clinical and radiological data. We measured areas of UIP-like fibrosis and calculated what percentage of the total lesion area they occupied. The presence of focal UIP-like fibrosis amounting to 10% or more of the lesion area was seen in 148 (73.6%), 168 (83.6%) and 165 (82.1%) cases for each pathologist, respectively. Agreement of the recognition of UIP-like fibrosis in PPF cases was above κ = 0.6 between all pairs. Survival analysis showed that the presence of focal UIP-like fibrosis correlated with worsened survival under all parameters tested (P < 0.001). CONCLUSIONS: The presence of UIP-like fibrosis is a core pathological feature of clinical PPF, and its presence within diseased areas is associated with poorer prognosis. This study highlights the importance of considering the presence of focal UIP-like fibrosis in the evaluation and management of PPF.

Real-World Implementation of Digital Pathology: Results From an Intercontinental Survey.

The past 70 years have been characterized by rapid advancements in computer technology, and the health care system has not been immune to this trend. However, anatomical pathology has remained largely an analog discipline. In recent years, this has been changing with the growing adoption of digital pathology, partly driven by the potential of computer-aided diagnosis. As part of an international collaboration, we conducted a comprehensive survey to gain a deeper understanding of the status of digital pathology implementation in Europe and Asia. A total of 127 anatomical pathology laboratories participated in the survey, including 75 from Europe and 52 from Asia, with 72 laboratories having established digital pathology workflow and 55 without digital pathology. Laboratories using digital pathology for diagnostic (n = 29) and nondiagnostic (n = 43) purposes were thoroughly questioned about their implementation strategies and institutional experiences, including details on equipment, storage, integration with laboratory information system, computer-aided diagnosis, and the costs of going digital. The impact of the digital pathology workflow was also evaluated, focusing on turnaround time, specimen traceability, quality control, and overall satisfaction. Laboratories without access to digital pathology were asked to provide insights into their perceptions of the technology, expectations, barriers to adoption, and potential facilitators. Our findings indicate that although digital pathology is still the future for many, it is already the present for some. This decade may be a time when anatomical pathology finally embraces digital revolution on a larger scale.

Artificial Intelligence-Based Tool for Tumor Detection and Quantitative Tissue Analysis in Colorectal Specimens.

Digital pathology adoption allows for applying computational algorithms to routine pathology tasks. Our study aimed to develop a clinical-grade artificial intelligence (AI) tool for precise multiclass tissue segmentation in colorectal specimens (resections and biopsies) and clinically validate the tool for tumor detection in biopsy specimens. The training data set included 241 precisely manually annotated whole-slide images (WSIs) from multiple institutions. The algorithm was trained for semantic segmentation of 11 tissue classes with an additional module for biopsy WSI classification. Six case cohorts from 5 pathology departments (4 countries) were used for formal and clinical validation, digitized by 4 different scanning systems. The developed algorithm showed high precision of segmentation of different tissue classes in colorectal specimens with composite multiclass Dice score of up to 0.895 and pixel-wise tumor detection specificity and sensitivity of up to 0.958 and 0.987, respectively. In the clinical validation study on multiple external cohorts, the AI tool reached sensitivity of 1.0 and specificity of up to 0.969 for tumor detection in biopsy WSI. The AI tool analyzes most biopsy cases in less than 1 minute, allowing effective integration into clinical routine. We developed and extensively validated a highly accurate, clinical-grade tool for assistive diagnostic processing of colorectal specimens. This tool allows for quantitative deciphering of colorectal cancer tissue for development of prognostic and predictive biomarkers and personalization of oncologic care. This study is a foundation for a SemiCOL computational challenge. We open-source multiple manually annotated and weakly labeled test data sets, representing a significant contribution to the colorectal cancer computational pathology field.

MIXTURE of human expertise and deep learning-developing an explainable model for predicting pathological diagnosis and survival in patients with interstitial lung disease.

Interstitial pneumonia is a heterogeneous disease with a progressive course and poor prognosis, at times even worse than those in the main cancer types. Histopathological examination is crucial for its diagnosis and estimation of prognosis. However, the evaluation strongly depends on the experience of pathologists, and the reproducibility of diagnosis is low. Herein, we propose MIXTURE (huMan-In-the-loop eXplainable artificial intelligence Through the Use of REcurrent training), an original method to develop deep learning models for extracting pathologically significant findings based on an expert pathologist's perspective with a small annotation effort. The procedure of MIXTURE consists of three steps as follows. First, we created feature extractors for tiles from whole slide images using self-supervised learning. The similar looking tiles were clustered based on the output features and then pathologists integrated the pathologically synonymous clusters. Using the integrated clusters as labeled data, deep learning models to classify the tiles into pathological findings were created by transfer-learning the feature extractors. We developed three models for different magnifications. Using these extracted findings, our model was able to predict the diagnosis of usual interstitial pneumonia, a finding suggestive of progressive disease, with high accuracy (AUC 0.90 in validation set and AUC 0.86 in test set). This high accuracy could not be achieved without the integration of findings by pathologists. The patients predicted as UIP had poorer prognosis (5-year overall survival [OS]: 55.4%) than those predicted as non-UIP (OS: 95.2%). The Cox proportional hazards model for each microscopic finding and prognosis pointed out dense fibrosis, fibroblastic foci, elastosis, and lymphocyte aggregation as independent risk factors. We suggest that MIXTURE may serve as a model approach to different diseases evaluated by medical imaging, including pathology and radiology, and be the prototype for explainable artificial intelligence that can collaborate with humans.

The future of artificial intelligence in digital pathology - results of a survey across stakeholder groups.

AIMS: Artificial intelligence (AI) provides a powerful tool to extract information from digitised histopathology whole slide images. During the last 5 years, academic and commercial actors have developed new technical solutions for a diverse set of tasks, including tissue segmentation, cell detection, mutation prediction, prognostication and prediction of treatment response. In the light of limited overall resources, it is presently unclear for researchers, practitioners and policymakers which of these topics are stable enough for clinical use in the near future and which topics are still experimental, but worth investing time and effort into. METHODS AND RESULTS: To identify potentially promising applications of AI in pathology, we performed an anonymous online survey of 75 computational pathology domain experts from academia and industry. Participants enrolled in 2021 were queried about their subjective opinion on promising and appealing subfields of computational pathology with a focus upon solid tumours. The results of this survey indicate that the prediction of treatment response directly from routine pathology slides is regarded as the most promising future application. This item was ranked highest in the overall analysis and in subgroups by age and professional background. Furthermore, prediction of genetic alterations, gene expression and survival directly from routine pathology images scored consistently high throughout subgroups. CONCLUSIONS: Together, these data demonstrate a possible direction for the development of computational pathology systems in clinical, academic and industrial research in the near future.

A collaborative workflow between pathologists and deep learning for the evaluation of tumour cellularity in lung adenocarcinoma.

AIMS: The reporting of tumour cellularity in cancer samples has become a mandatory task for pathologists. However, the estimation of tumour cellularity is often inaccurate. Therefore, we propose a collaborative workflow between pathologists and artificial intelligence (AI) models to evaluate tumour cellularity in lung cancer samples and propose a protocol to apply it to routine practice. METHODS AND RESULTS: We developed a quantitative model of lung adenocarcinoma that was validated and tested on 50 cases, and a collaborative workflow where pathologists could access the AI results and adjust their original tumour cellularity scores (adjusted-score) that we tested on 151 cases. The adjusted-score was validated by comparing them with a ground truth established by manual annotation of haematoxylin and eosin slides with reference to immunostains with thyroid transcription factor-1 and napsin A. For training, validation, testing the AI and testing the collaborative workflow, we used 40, 10, 50 and 151 whole slide images of lung adenocarcinoma, respectively. The sensitivity and specificity of tumour segmentation were 97 and 87%, respectively, and the accuracy of nuclei recognition was 99%. One pathologist's visually estimated scores were compared to the adjusted-score, and the pathologist's scores were altered in 87% of cases. Comparison with the ground truth revealed that the adjusted-score was more precise than the pathologists' scores (P < 0.05). CONCLUSION: We proposed a collaborative workflow between AI and pathologists as a model to improve daily practice and enhance the prediction of tumour cellularity for genetic tests.

Cicatricial organising pneumonia associated with fibrosing interstitial pneumonia - a clinicopathological study.

AIMS: The recent recognition of cicatricial organising pneumonia (ciOP) indicates that the ciOP may resemble or simulate fibrotic interstitial pneumonia; however, there has been great uncertainty regarding the affected populations, pathogenesis, clinical relevance and characteristics. In this study, we compared the characteristics of fibrotic interstitial pneumonia with and without ciOP. METHODS AND RESULTS: We enrolled 121 patients from the consultation archive whose pathological findings were fibrotic interstitial pneumonia and for whom follow-up clinical data were available. We reviewed these cases histopathologically and classified them according to whether or not they showed ciOP. We compared the clinicopathological features between the two groups. CiOP, histopathologically characterised by deposition of dense collagenous fibres within the alveolar space without destruction of the lung structure, was found in 48 patients (39.7%). None of the cases with ciOP experienced acute exacerbation during 12 months' follow-up. The group with ciOP had more severe diffusion impairment but this, together with restrictive ventilatory impairment, improved significantly compared to the group without ciOP. CONCLUSION: CiOP is a histopathological finding commonly found in fibrotic interstitial pneumonia. It does not relate to acute exacerbation or decrease in pulmonary function.

Quality control stress test for deep learning-based diagnostic model in digital pathology.

Digital pathology provides a possibility for computational analysis of histological slides and automatization of routine pathological tasks. Histological slides are very heterogeneous concerning staining, sections' thickness, and artifacts arising during tissue processing, cutting, staining, and digitization. In this study, we digitally reproduce major types of artifacts. Using six datasets from four different institutions digitized by different scanner systems, we systematically explore artifacts' influence on the accuracy of the pre-trained, validated, deep learning-based model for prostate cancer detection in histological slides. We provide evidence that any histological artifact dependent on severity can lead to a substantial loss in model performance. Strategies for the prevention of diagnostic model accuracy losses in the context of artifacts are warranted. Stress-testing of diagnostic models using synthetically generated artifacts might be an essential step during clinical validation of deep learning-based algorithms.

Detection of Lung Cancer Lymph Node Metastases from Whole-Slide Histopathologic Images Using a Two-Step Deep Learning Approach.

The application of deep learning for the detection of lymph node metastases on histologic slides has attracted worldwide attention due to its potentially important role in patient treatment and prognosis. Despite this attention, false-positive predictions remain problematic, particularly in the case of reactive lymphoid follicles. In this study, a novel two-step deep learning algorithm was developed to address the issue of false-positive prediction while maintaining accurate cancer detection. Three-hundred and forty-nine whole-slide lung cancer lymph node images, including 233 slides for algorithm training, 10 slides for validation, and 106 slides for evaluation, were collected. In the first step, a deep learning algorithm was used to eliminate frequently misclassified noncancerous regions (lymphoid follicles). In the second step, a deep learning classifier was developed to detect cancer cells. Using this two-step approach, errors were reduced by 36.4% on average and up to 89% in slides with reactive lymphoid follicles. Furthermore, 100% sensitivity was reached in cases of macrometastases, micrometastases, and isolated tumor cells. To reduce the small number of remaining false positives, a receiver-operating characteristic curve was created using foci size thresholds of 0.6 mm and 0.7 mm, achieving sensitivity and specificity of 79.6% and 96.5%, and 75.5% and 98.2%, respectively. A two-step approach can be used to detect lung cancer metastases in lymph node tissue effectively and with few false positives.

Low prevalence of p16-positive HPV-related head-neck cancers in Thailand: tertiary referral center experience.

BACKGROUND: There has been a sharp rise in the incidence of human papillomavirus (HPV) associated oropharyngeal squamous cell carcinoma (OPSCC) in many countries. Patients with HPV-positive OPSCC have a more favorable prognosis compared with HPV-negative OPSCC, leading to investigation and adoption of de-escalation treatment protocols. The baseline rate of HPV prevalence in certain populations is of epidemiologic significance. We aimed to evaluate the rate of high-risk HPV in a large cohort of Thai patients, including OPSCC, oral SCC (OSCC) and laryngeal SCC (LSCC). METHODS: In total, 504 patients with HN cancer (110 OPSCC, 260 OSCC and 134 LSCC) who had been treated in Chulalongkorn University between 2010 and 2016 formed the sample set. All histological slides were reviewed to validate the diagnosis and render the histological type as keratinizing (K), non-keratinizing (NK) or non-keratinizing with maturation (NK-M). Immunohistochemistry with p16 was performed in all cases and scored semiquantatively. Positive and equivocal cases were tested by the high-risk HPV DNA in situ hybridization (ISH). Validation with quantitative polymerase-chain reaction (qPCR) was performed in p16-positive OPSCC. RESULTS: The OPSCC were represented by NK (7.3%), NK-M (16.4%) and K (76.4%) types, with an HPV incidence of 100, 22.2 and 4.7%, respectively. The average HPV prevalence in OPSCC was 14.5%. The concordance with p16/ISH was 51.6%, while concordance of the NK morphology with positive HPV ISH was 100%. ISH-qPCR concordance in p16-positive OPSCC was 72.7%. Patients with HPV-positive OPSCC had significantly more tumors with a NK histologic type, tonsillar location, earlier clinical stage, less association with smoking, and, finally, better outcome and longer survival time. In non-OPSCC, p16-positive HPV-associated cancers were found in only 1.5% of OSCC (4/260) and LSCC (2/134). CONCLUSION: A low rate of HPV-related OPSCC was found in Thai patients. The NK morphology was an excellent predictor of high-risk HPV infection in OPSCC. For OPSCC patients, HPV-positive ones had a significantly longer survival time than HPV-negative ones. There was a lack of p16-positive HPV-related OSCC and LSCC. Morphology and p16 status had a poor predictive value for detecting HPV in OSCC and LSCC.

Interobserver and intraobserver variation in the morphological evaluation of noninvasive follicular thyroid neoplasm with papillary-like nuclear features in Asian practice.

To evaluate the current diagnostic criteria in reporting nuclear features of noninvasive follicular thyroid neoplasm with papillary-like nuclear features (NIFTP), nine Asian pathologists with expertise in thyroid reviewed virtual slides of 30 noninvasive follicular patterned thyroid lesions according to the nuclear scoring system originally proposed by an international expert and a more detailed scoring system proposed by the Asian Working Group. The interobserver agreement for nuclear grading score was generally moderate (kappa value = 0.452). The best consistency fell on the chromatin features (kappa value = 0.658-1.000). A fair to moderate interobserver agreement was demonstrated in the evaluation of nuclear elongation, nuclear overlapping, membrane irregularities and distribution of papillary thyroid carcinoma (PTC) type nuclear features. A slight agreement was rendered for the evaluation of the nuclear enlargement. Intraobserver agreement was substantial to perfect when comparing results of both scoring systems. However, both scoring systems were not able to reliably separate NIFTP from an encapsulated follicular variant PTC with minimal lymph node metastasis or BRAFV600E mutation. Although the three-point nuclear scoring system for the diagnosis of NIFTP is widely used in Asian practice, interobserver variation was considerable. Ancillary immunohistochemical or molecular testing might be helpful in differentiating NIFTP from true PTC.

CLINICAL IMPACT OF NON-INVASIVE FOLLICULAR THYROID NEOPLASM WITH PAPILLARY-LIKE NUCLEAR FEATURES ON THE RISK OF MALIGNANCY IN THE BETHESDA SYSTEM FOR REPORTING THYROID CYTOPATHOLOGY: A META-ANALYSIS OF 14,153 RESECTED THYROID NODULES.

Objective: It is still controversial as to how the reclassification of noninvasive follicular thyroid neoplasm with papillary-like nuclear features (NIFTP) affects the risk of malignancy (ROM) in The Bethesda System for Reporting Thyroid Cytopathology (TBSRTC). This meta-analysis was aimed to investigate the impact of NIFTP on the ROM in each TBSRTC category. Methods: We accessed three electronic databases including PubMed, Web of Science, and Scopus to search for relevant data from January, 2016 to July, 2018. Relative risk and meta-analysis of proportions using the DerSimonian-Laird method, and each corresponding 95% confidence interval (CI) was pooled using a random-effect model. Results: A total of 14 studies consisting of 14,153 resected nodules were included for meta-analyses. Overall, there was a significant reduction in ROM in all TBSRTC categories following the NIFTP reclassification, except TBSRTC category I. The largest absolute and relative decrease in ROM was observed in TBSRTC category V (16%; 95% CI = 8 to 24) and category III (32%; 95% CI = 24 to 39), respectively. There was a positive correlation between the rate of NIFTP and resection rate (r = 0.83; P = .02). The decreases in ROM were more prominent in Western than in Asian cohorts. Conclusion: We confirmed the decrease in ROM due to the NIFTP reclassification for most of TBSRTC categories, which was more significant in Western than in Asian practice. The incidence of NIFTP was higher in institutions where surgical resection rates were high in patients with indeterminate cytology nodules. Abbreviations: AUS/FLUS = atypia of undetermined significance/follicular lesion of undetermined significance; CI = confidence interval; FNA = fine-needle aspiration; FN/SFN = follicular neoplasm/suspicious for follicular neoplasm; NIFTP = noninvasive follicular thyroid neoplasm with papillary-like nuclear features; NI-FVPTC = noninvasive follicular variant of papillary thyroid carcinoma; ROM = risk of malignancy; RR = relative risk; SM = suspicious for malignancy; TBSRTC = The Bethesda System for Reporting Thyroid Cytopathology.

Remarkable Alteration of PD-L1 Expression after Immune Checkpoint Therapy in Patients with Non-Small-Cell Lung Cancer: Two Autopsy Case Reports.

Pembrolizumab is an immune checkpoint inhibitor (ICI), currently recommended as the first-line treatment for patients with advanced non-small-cell lung cancer (NSCLC) showing ≥50% expression of programmed death-ligand 1 (PD-L1). Previously it was reported that platinum-based chemotherapy may change PD-L1 expression in solid cancers. However, no reports addressing alteration of PD-L1 expression after ICI therapy in NSCLC are available so far. The patients were Japanese males 83 and 87 years old, who were diagnosed with NSCLC based on the transbronchial lung biopsies showing sarcomatoid feature with high PD-L1 expression. They received Pembrolizumab, however, passed away with disease progression on day 27 and day 9, respectively. PD-L1, PD1, and CD8 antibodies were applied to pretreatment tumor biopsies and autopsy specimens. Immunoexpression of all the markers was evaluated using Aperio ImageScope. We found that PD-L1 expression decreased significantly from 75.6% to 13.2% and from 100% to 58.8%, in patients 1 and 2, respectively. This alteration was less prominent in the perinecrotic tumor area. A considerable decrease of PD-L1 score was linked with a little effect of Pembrolizumab in our patients. This association might be one of the contributing mechanisms of resistance to ICI and needs further investigation in large-scale studies.

Promoter hypermethylation of SOX11 correlates with adverse clinicopathological features of human prostate cancer.

Currently available tools for early diagnosis and prognosis of prostate cancer lack sufficient accuracy. There is a need to identify novel biomarkers for this common malignancy. SOX family genes play an important role in embryogenesis and are also implicated in various cancers. SOX11 has been recently recognized as a potential tumour suppressor that is downregulated in prostate cancer. We hypothesized that hypermethylation may be responsible for SOX11 silencing in human prostate cancer. The aim of the study was to investigate SOX11 promoter methylation in prostate adenocarcinoma by comparing it with benign prostatic hyperplasia (BPH). A total of 143 human prostate tissue samples, 62 from patients with prostate cancer and 81 from patients with BPH were examined by methylation-specific PCR. Associations between SOX11 promoter methylation and clinicopathological parameters were assessed by univariate statistics. Detection rates of SOX11 promoter methylation were 80.6% and 35.8% in prostate cancer and BPH respectively (P < 0.001). SOX11 hypermethylation was associated with adverse clinicopathological characteristics of prostate cancer, including higher PSA level (P < 0.01), Gleason score ≥ 7 (P = 0.03) and perineural invasion (P = 0.03). SOX11 methylation was positively correlated with the PSA level (P = 0.001). Our data indicated that SOX11 can be a promising methylation marker candidate for differential diagnosis and risk stratification for prostate cancer.

What's new in thyroid pathology 2024: updates from the new WHO classification and Bethesda system.

In line with the release of the 5th edition WHO Classification of Tumors of Endocrine Organs (2022) and the 3rd edition of the Bethesda System for Reporting Thyroid Cytopathology (2023), the field of thyroid pathology and cytopathology has witnessed key transformations. This digest brings to the fore the refined terminologies, newly introduced categories, and contentious methodological considerations pivotal to the updated classification.

Squamous metaplasia is an indicator of acute exacerbation in patients with usual interstitial pneumonia / idiopathic pulmonary fibrosis.

BACKGROUND: Acute exacerbation (AE) is a potentially lethal event in patients with usual interstitial pneumonia/idiopathic pulmonary fibrosis (UIP/IPF). However, to date, no pathological predictors of AE have been identified. This retrospective study aimed to elucidate the pathological features that could predict AE in patients with UIP. METHODS: We reviewed the pathological findings of 91 patients with UIP/IPF and correlated these findings with AE events. Thirteen histological variables related to acute lung injury were evaluated by three independent observers and classified as positive or negative. The patients' clinical data during follow-up were collected and reviewed for AE. A recursive partition using the Gini index for the prediction of AE was performed, with each pathological finding as a candidate for branching. RESULTS: Twenty patients (22%) developed AE during the median follow-up duration of 40 months. Thirty-eight patients died (15 due to AE and 23 for other reasons). The median time interval from surgical lung biopsy to AE onset was 497 (interquartile range: 901-1657) days. Histologically, squamous metaplasia was positively associated with AE (odds ratio: 4.7, P = 0.015) and worse event-free survival in patients with UIP (P = 0.04). Leaf scoring based on the Gini index for recursive partition, including five positive findings (squamous metaplasia, neutrophilic infiltration, septal widening, Kuhn's hyaline, and fibrin), showed a sensitivity of 90% with a specificity of 74.7% (area under curve: 0.89). CONCLUSIONS: We found that squamous metaplasia is an important histopathological finding that predicts AE events and tends to unfavorable outcome in patients with UIP/IPF.

Machine-Learning-Based Classification Model to Address Diagnostic Challenges in Transbronchial Lung Biopsy.

BACKGROUND: When obtaining specimens from pulmonary nodules in TBLB, distinguishing between benign samples and mis-sampling from a tumor presents a challenge. Our objective is to develop a machine-learning-based classifier for TBLB specimens. METHODS: Three pathologists assessed six pathological findings, including interface bronchitis/bronchiolitis (IB/B), plasma cell infiltration (PLC), eosinophil infiltration (Eo), lymphoid aggregation (Ly), fibroelastosis (FE), and organizing pneumonia (OP), as potential histologic markers to distinguish between benign and malignant conditions. A total of 251 TBLB cases with defined benign and malignant outcomes based on clinical follow-up were collected and a gradient-boosted decision-tree-based machine learning model (XGBoost) was trained and tested on randomly split training and test sets. RESULTS: Five pathological changes showed independent, mild-to-moderate associations (AUC ranging from 0.58 to 0.75) with benign conditions, with IB/B being the strongest predictor. On the other hand, FE emerged to be the sole indicator of malignant conditions with a mild association (AUC = 0.66). Our model was trained on 200 cases and tested on 51 cases, achieving an AUC of 0.78 for the binary classification of benign vs. malignant on the test set. CONCLUSION: The machine-learning model developed has the potential to distinguish between benign and malignant conditions in TBLB samples excluding the presence or absence of tumor cells, thereby improving diagnostic accuracy and reducing the burden of repeated sampling procedures for patients.