Search Results (59)

The need to improve medical diagnosis is of utmost importance in medical research, consisting of the optimization of accurate classification models able to assist clinical decisions. To minimize the errors that can be caused by using a single classifier, the voting ensemble technique can be used, combining the classification results of different classifiers to improve the final classification performance. This paper aims to compare the existing voting ensemble techniques with a new game-theory-derived approach based on Shapley values. We extended this method, originally developed for binary tasks, to the multi-class setting in order to capture complementary information provided by different classifiers. In heterogeneous clinical scenarios such as thyroid nodule diagnosis, where distinct models may be better suited to identify specific subtypes (e.g., benign, malignant, or inflammatory lesions), ensemble strategies capable of leveraging these strengths are particularly valuable. The motivating application focuses on the classification of thyroid cancer nodules whose cytopathological clinical diagnosis is typically characterized by a high number of false positive cases that may result in unnecessary thyroidectomy. We apply and compare the performance of seven individual classifiers, along with four ensemble voting techniques (including Shapley values), in a real-world study focused on classifying thyroid cancer nodules using proteomic features obtained through mass spectrometry. Our results indicate a slight improvement in the classification accuracy for ensemble systems compared to the performance of single classifiers. Although the Shapley value-based voting method remains comparable to the other voting methods, we envision this new ensemble approach could be effective in improving the performance of single classifiers in further applications, especially when complementary algorithms are considered in the ensemble. The application of these techniques can lead to the development of new tools to assist clinicians in diagnosing thyroid cancer using proteomic features derived from mass spectrometry. Full article

Mitotic figures in tumor tissues are an important criterion for diagnosing malignant lesions, and physicians often search for the presence of mitosis in whole slide imaging (WSI). However, prolonged visual inspection by doctors may increase the likelihood of human error. With the advancement of deep learning, AI-based automatic cytopathological diagnosis has been increasingly applied in clinical settings. Nevertheless, existing diagnostic models often suffer from high computational costs and suboptimal detection accuracy. More importantly, when assessing cellular abnormalities, doctors frequently compare target cells with their surrounding cells—an aspect that current models fail to capture due to their lack of intercellular information modeling, leading to the loss of critical medical insights. To address these limitations, we conducted an in-depth analysis of existing models and propose an Inter–Intra Hypergraph Neural Network (II-HGNN). Our model introduces a block-based feature extraction mechanism to efficiently capture deep representations. Additionally, we leverage hypergraph convolutional networks to process both intracellular and intercellular information, leading to more precise diagnostic outcomes. We evaluate our model on publicly available datasets under varying imaging conditions, and experimental results demonstrate that our approach consistently outperforms baseline models in terms of accuracy. Full article

Background/Objectives: This study aims to evaluate whether fine-needle aspiration biopsy (FNAB) of melanocytic uveal tumors (MUTs) using 27-gauge (27G) needles yields aspirates like those obtained using 25-gauge (25G) needles for cytology. Methods: A retrospective review was conducted on 32 primary uveal melanomas (PUMs). Tumors were sampled at three adjacent sites, first using a 27G needle for gene expression profile (GEP) testing, second and third with 27G and 25G needles for cytology. The endpoints evaluated were the sufficiency of aspirates for cytopathology and GEP. Results: Among the 32 patients, 17 tumors were choroidal, 6 ciliochoroidal, 7 iridociliochoroidal, and 2 exclusively iridic. Tumor diameter ranged from 3.3 mm to 23 mm (mean 13.2 mm), and thickness ranged from 0.5 mm to 12 mm (mean 6.4 mm). Aspirates from both needle sizes were sufficient for cytopathological diagnosis and GEP in 31 of 32 cases (96.9%). The single insufficient aspirate was insufficient with both the 27G and 25G needles. The cytopathology was identical in all other cases. The tumors were Class 1 in 22 cases (71.0%) and Class 2 in 9 cases (29.0%). Conclusions: FNAB aspirates of MUTs using 27G needles appear sufficient for cytology and GEP in most cases, showing a similar diagnostic yield compared to 25G needles. Full article

Background: Malignant pleural effusions (MPEs) are common in advanced lung cancer patients. Cytological examination of pleural fluid is essential for identifying cell types but presents diagnostic challenges, particularly when reactive mesothelial cells mimic neoplastic cells. AI-powered diagnostic systems have emerged as valuable tools in digital cytopathology. This study explores the applicability of machine-learning (ML) models and highlights the importance of accessible tools for clinicians, enabling them to develop AI solutions and make advanced diagnostic tools available even in resource-limited settings. The focus is on differentiating normal/reactive cells from neoplastic cells in pleural effusions linked to lung adenocarcinoma. Methods: A dataset from the Cytopathology Unit at the Sant’Andrea University Hospital comprising 969 raw images, annotated with 3130 single mesothelial cells and 3260 adenocarcinoma cells, was categorized into two classes based on morphological features. Object-detection models were developed using YOLOv8 and the latest YOLOv11 instance segmentation models. Results: The models achieved an Intersection over Union (IoU) score of 0.72, demonstrating robust performance in class prediction for both categories, with YOLOv11 showing performance improvements over YOLOv8 in different metrics. Conclusions: The application of machine learning in cytopathology offers clinicians valuable support in differential diagnosis while also expanding their ability to engage with AI tools and methodologies. The diagnosis of MPEs is marked by substantial morphological and technical variability, underscoring the need for high-quality datasets and advanced deep-learning models. These technologies have the potential to enhance data interpretation and support more effective clinical treatment strategies in the era of precision medicine. Full article

Background/Objectives: Mucoepidermoid carcinoma (MEC) is the most common malignant tumor of the minor salivary glands, often affecting the hard palate. Preoperative diagnosis and surgical planning are challenging due to anatomical complexity and limitations in sampling, generally obtained by fine-needle aspiration (FNA). This study retrospectively evaluated the diagnostic and therapeutic performance of a high-definition ultrasound (HDUS)-guided fine-needle aspiration cytology/biopsy (FNAC/FNAB) protocol in diagnosing intraoral MEC, based on the Milan System for Reporting Salivary Gland Cytopathology (MSRSGC), with the relative clinical outcomes. Methods: A cohort of 64 patients with histologically confirmed MEC of the minor salivary glands, treated between 2000 and 2022, was retrospectively analyzed. All patients underwent HDUS-guided FNAC/FNAB, imaging (CT, MRI, and panoramic X-ray), and subsequent surgical treatment. The cytological specimens were classified using the MSRSGC. Surgical margins, histopathological findings, lymph node status, and follow-up outcomes were recorded. Results: Of 64 MECs, 42 cases were finally diagnosed as low-grade (LG)/intermediate grade (IG) and 22 as high-grade (HG) carcinomas, using a two-tier histological classification system. HDUS accurately delineated the lesion size, infiltration depth, and bone proximity, with excellent correlation with surgical specimens (difference ≤ 0.6 mm). MSRSGC classification distributed the cases across all categories, with 28 classified as malignant (category VI). Repeat FNAC improved the diagnostic yield in non-diagnostic and atypical cases. FNAB confirmed the cytological findings in all cases, with immunohistochemistry investigation with Ki-67 supporting tumor grading. Surgical margins were clear in all resections. Lymph node metastases were identified in all patients who underwent neck dissection (n = 18), all with HG-MEC. No recurrences occurred among the LG/IG-MEC patients during a median 2-year follow-up. Conclusions: The combined use of HDUS and FNAC/FNAB, interpreted through the MSRSGC framework, offers a highly accurate, minimally invasive approach for preoperative diagnosis and surgical planning in intraoral MEC. HDUS-guided cytology significantly improves diagnostic reliability, particularly in LG/IG and cystic variants, facilitating tailored surgical management. Also, the clinical outcomes may support the possibility of using a simplified grading classification for two histopathological types. Full article

The progress of artificial intelligence (AI), particularly its core algorithms—machine learning (ML) and deep learning (DL)—has been significant in the medical field, impacting both scientific research and clinical practice. These algorithms are now capable of analyzing ultrasound images, processing them, and providing outcomes, such as determining the benignity or malignancy of thyroid nodules. This integration into ultrasound machines is referred to as computer-aided diagnosis (CAD). The use of such software extends beyond ultrasound to include cytopathological and molecular assessments, enhancing the estimation of malignancy risk. AI’s considerable potential in cancer diagnosis and prevention is evident. This article provides an overview of AI models based on ML and DL algorithms used in thyroid diagnostics. Recent studies demonstrate their effectiveness and diagnostic role in ultrasound, pathology, and molecular fields. Notable advancements include content-based image retrieval (CBIR), enhanced saliency CBIR (SE-CBIR), Restore-Generative Adversarial Networks (GANs), and Vision Transformers (ViTs). These new algorithms show remarkable results, indicating their potential as diagnostic and prognostic tools for thyroid pathology. The future trend points to these AI systems becoming the preferred choice for thyroid diagnostics. Full article

Background: We address the application of artificial intelligence (AI) techniques in thyroid cytopathology, specifically for diagnosing papillary thyroid carcinoma (PTC), the most common type of thyroid cancer. Methods: Our research introduces deep learning frameworks that analyze cytological images from fine-needle aspiration cytology (FNAC), a key preoperative diagnostic method for PTC. The first framework is a patch-level classifier referred as “TCS-CNN”, based on a convolutional neural network (CNN) architecture, to predict thyroid cancer based on the Bethesda System (TBS) category. The second framework is an attention-based deep multiple instance learning (AD-MIL) model, which employs a feature extractor using TCS-CNN and an attention mechanism to aggregate features from smaller-patch-level regions into predictions for larger-patch-level regions, referred to as bag-level predictions in this context. Results: The proposed TCS-CNN framework achieves an accuracy of 97% and a recall of 96% for small-patch-level classification, accurately capturing local malignancy information. Additionally, the AD-MIL framework also achieves approximately 96% accuracy and recall, demonstrating that this framework can maintain comparable performance while expanding the diagnostic coverage to larger regions through patch aggregation. Conclusions: This study provides a feasibility analysis for thyroid cytopathology classification and visual interpretability for AI diagnosis, suggesting potential improvements in patient outcomes and reductions in healthcare costs. Full article

Persistent oncogenic HPV infection is strongly associated with cervical cancer. Studies have suggested a higher prevalence of HPV in women living with HTLV-1. This study aimed to determine whether HTLV-1 infection is associated with cervicovaginal HPV infection and to characterize HPV types according to oncogenic risk. Vaginal fluid samples were subjected to HPV diagnosis via PCR, and positive samples were subjected to Sanger sequencing and massive sequencing. Papanicolaou smears were examined using light microscopy to identify cell abnormalities. Among the 155 women screened, 79 were HTLV-1-infected and 76 were uninfected. HPV PCR identified 23 positive samples (15/79 vs. 8/76; p = 0.13). Twenty-three HPV types were identified, of which only types 31, 54, and 58 were present in both groups. When the number of HPV58 infections in each group was compared, women with HTLV-1 had a higher prevalence (8/79 versus 1/76; p = 0.03). In total, 61.9% of HTLV-1-infected women had at least one high-risk or probable high-risk HPV type (p = 0.12). Cytopathological findings were not significantly different between the groups. Further research is needed to determine whether HTLV-1 infection affects HPV progression and cervical cancer development and to assess the potential benefits of vaccination for women living with HTLV-1. Full article

Background: Uveal melanoma (UM) is the most common primary intraocular malignancy in adults. The median overall survival time for patients who develop metastasis is approximately one year. In this study, we aim to leverage deep learning (DL) techniques to analyze digital cytopathology images and directly predict the 48 month survival status on a patient level. Methods: Fine-needle aspiration biopsy (FNAB) of the tumor was performed in each patient diagnosed with UM. The cell aspirate was smeared on a glass slide and stained with H&E. Each slide then underwent whole-slide scanning. Within each whole-slide image, regions of interest (ROIs) with UM cells were automatically extracted. Each ROI was converted into super pixels, and the super pixels were automatically detected, segmented and annotated as “tumor cell” or “background” using DL. Cell-level features were extracted from the segmented tumor cells. The cell-level features were aggregated into slide-level features which were learned by a fully connected layer in an artificial neural network, and the patient survival status was predicted directly from the slide-level features. The data were partitioned at the patient level (78% training and 22% testing). Our DL model was trained to perform the binary prediction of yes-versus-no survival by Month 48. The ground truth for patient survival was established via a retrospective chart review. Results: A total of 74 patients were included in this study (43% female; mean age at the time of diagnosis: 61.8 ± 11.6 years), and 207,260 unique ROIs were generated for model training and testing. By Month 48 after diagnosis, 18 patients (24%) died from UM metastasis. Our hold-out test set contained 16 patients, where 6 patients had passed away and 10 patients were alive at Month 48. When using a sensitivity threshold of 80% in predicting UM-specific death by Month 48, our model achieved an overall accuracy of 75%. Within the subgroup of patients who died by Month 48, our model achieved a prediction accuracy of 83%. Of note, one patient in our test set was a clinical surprise, namely death by Month 48 despite having a GEP class 1A tumor, which typically portends a good prognosis. Our model correctly predicted this clinical surprise as well. Conclusions: Our DL model was able to predict the Month 48 survival status directly from digital cytopathology images obtained from FNABs of UM tumors with reasonably robust performance. This approach, if validated prospectively, could serve as an alternative survival prediction tool for patients with UM to whom GEP is not available. Full article

Background: Thyroid cancer is the most common endocrine malignancy, and accurate diagnosis is crucial for effective management. Fine needle aspiration cytology, guided by the Bethesda System for Reporting Thyroid Cytopathology, categorizes thyroid nodules into six categories, with Bethesda III and IV representing indeterminate diagnoses that pose significant challenges for clinical decision-making. Understanding the molecular profiles of these categories may enhance diagnostic accuracy and guide treatment strategies. Methods: This study retrospectively analyzed data from 217 patients with Bethesda III and IV thyroid nodules who underwent ThyroSeq v3 molecular testing followed by thyroid surgery at McGill University teaching hospitals. The analysis focused on the presence of specific molecular mutations, copy number alterations (CNAs), and gene expression profiles (GEPs) within these nodules. The relationship between these molecular findings and the clinico-pathological features of the patients was also examined. Results: This study identified notable differences in the molecular landscape of Bethesda III and IV thyroid nodules. Bethesda IV nodules exhibited a higher prevalence of CNAs and distinct GEPs compared to Bethesda III nodules. Interestingly, the BRAFV600E mutation was found exclusively in Bethesda III nodules, which correlated with more aggressive malignant behavior. These findings underscore the potential of molecular profiling to differentiate between the clinical behaviors of these indeterminate nodule categories. Conclusions: Molecular profiling, including the assessment of CNAs, GEPs, and specific mutations like BRAFV600E, provides valuable insights into the nature of Bethesda III and IV thyroid nodules. The distinct molecular characteristics observed between these categories suggest that such profiling could be instrumental in improving diagnostic accuracy and tailoring treatment approaches, ultimately enhancing patient outcomes in thyroid cancer management. Full article

Lymphadenopathy is a common presentation of both reactive and malignant diseases, and lymph node fine-needle aspiration cytology (LN-FNAC) is an effective and inexpensive screening method. It can prevent unnecessary invasive surgery and excisional biopsy, especially in benign cases. Unfortunately, the lack of universally accepted terminology for reporting results has hindered its widespread support. The Sydney system proposal for lymph node cytopathology categorization and reporting introduced five diagnostic categories to address the lack of universally accepted terminology for reporting results in lymphadenopathy. Our study analyzed 188 lymph node fine-needle cytology (FNC) samples from King Abdulaziz University Hospital, Saudi Arabia, examining clinical follow-up data, pathology records, patient information, and final diagnosis from January 2019 to December 2022. Most specimens were from axillary lymph nodes, with 99.5% tissue correlation. The Sydney system category classification identified 56.9% of cases as malignant, while 26.1% were benign. The final surgical specimen diagnosis revealed a higher percentage of malignant diagnoses, with the highest risk of malignancy (ROM) in malignant/category V. In conclusion, our study demonstrates that LN-FNAC offers high diagnostic accuracy for lymph node (LN) aspirates, with the Sydney approach potentially aiding risk stratification and achieving consistency in cytologic diagnosis, but further multi-centric research is needed. Full article

Pulmonary actinomycosis is an uncommon clinical entity that can be challenging to diagnose due to its non-specific symptomatology. Misdiagnosis and delayed treatment may result in invasive procedures and extended antimicrobial treatment courses. We report a case involving a 65-year-old female with poor oral dentition admitted for acute respiratory failure subsequently found to have a left-sided pleural effusion and perihepatic abscess formation. Cytopathology examination and microbiology studies confirmed the diagnosis of pulmonary actinomycosis. Full article

Over the past decade, molecular cytopathology has emerged as a relevant area of modern pathology. Notably, in patients with advanced-stage cancer, cytological samples could be the only material available for diagnosis and molecular biomarker testing to identify patients suitable for targeted therapies. As a result, the contemporary cytopathologist’s role extends beyond morphological assessments to include critical skills such as evaluating the adequacy of the cytological samples and managing these specimens for molecular testing. This case collection can be a valuable source of insight, especially for young pathologists, who should learn to combine the opportunities offered by molecular biology with the basis of morphological evaluation. Full article

Background and Objectives: This study aimed to elucidate the cytologic characteristics and diagnostic usefulness of endoscopic ultrasonography–fine needle aspiration cytology (EUS-FNAC) by comparing it with liquid-based preparation (LBP) and conventional smear (CS) in pancreas. Methods: The diagnostic categories (I through VII) were classified according to the World Health Organization Reporting System for Pancreaticobiliary Cytopathology. Ten cytologic features, including nuclear and additional features, were evaluated in 53 cases subjected to EUS-FNAC. Nuclear features comprised irregular nuclear contours, nuclear enlargement, hypochromatic nuclei with parachromatin clearing, and nucleoli. Additional cellular features included isolated atypical cells, mucinous cytoplasm, drunken honeycomb architecture, mitosis, necrotic background, and cellularity. A decision tree analysis was conducted to assess diagnostic efficacy. Results: The diagnostic concordance rate between LBP and CS was 49.1% (26 out of 53 cases). No significant differences in nuclear features were observed between categories III (atypical), VI (suspicious for malignancy), and VII (malignant). The decision tree analysis of LBP indicated that cases with moderate or high cellularity and mitosis could be considered diagnostic for those exhibiting nuclear atypia. Furthermore, in CS, mitosis, isolated atypical cells, and necrotic background exerted a more significant impact on the diagnosis of EUS-FNAC. Conclusions: Significant parameters for interpreting EUS-FNAC may differ between LBP and CS. While nuclear atypia did not influence the diagnosis of categories III, VI, and VII, other cytopathologic features, such as cellularity, mitosis, and necrotic background, may present challenges in diagnosing EUS-FNAC. Full article

Pancreatic cancer is one of the deadliest malignancies, characterized by late-stage diagnosis and limited treatment options. Comprehensive genomic profiling plays an important role in understanding the molecular mechanisms underlying the disease and identifying potential therapeutic targets. Cell blocks (CBs), derived from EUS-FNA, have become valuable resources for diagnosis and genomic analysis. We examine the molecular profile of pancreatic ductal adenocarcinoma (PDAC) using specimens obtained from CB EUS-FNA, across a large gene panel, within the framework of next-generation sequencing (NGS). Our findings revealed that over half (55%) of PDAC CB cases provided adequate nucleic acid for next-generation sequencing, with tumor cell percentages averaging above 30%. Despite challenges such as low DNA quantification and degraded DNA, sequencing reads showed satisfactory quality control statistics, demonstrating the detection of genomic alterations. Most cases (84.6%) harbored at least one gene variant, including clinically significant gene mutation variants such as KRAS, TP53, and CDKN2A. Even at minimal concentrations, as long as the extracted DNA is of high quality, performing comprehensive molecular profiling on PDAC samples from cell blocks has remained feasible. This strategy has yielded valuable information about the diagnosis, genetic landscape, and potential therapeutic targets, aligning closely with a precision cytopathology approach. Full article