Artificial intelligence (AI) has driven major disruption across multiple domains of clinical medicine and patient care and is fundamentally redrawing the landscape of modern medicine. In this context, cytopathology stands at a critical crossroads, where traditional microscopic evaluation meets the frontier of computational medicine. Conventionally, a successful cytopathology workflow entails intensive manual effort performed under the close supervision of an expert cytopathologist and an experienced, highly competent team of cytotechnologists. With the advancements in medical science driven by the demand for precision and personalized medicine, workload of the cytopathology laboratory is ever increasing by many folds, while there is an alarming decreasing trend in the availability of skilled human resource. A new era of diagnostic precision is emerging, as machine learning and deep learning algorithms take center stage in the laboratory. These systems tend to streamline the workflow by reviewing high-volume slide sets, prioritizing high-risk cases, and offering prognostic insights. These processes are highly dependent on meticulous digitization of cytology smears using whole slide imaging pathology scanners, which in turn enables telecytology, large-scale data sharing, and the development of robust training datasets, all of which accelerate AI innovation. This review provides an overview of the technical processes and applications of AI-based cytopathology algorithms across different organ systems, workflow transformation (from preanalytical to quality control, telecytopathology, and integration with molecular diagnostics), and the various challenges and limitations in their adoption in the routine diagnostic workflow for patient care.
Context:Pancreatic lesions range from benign cystic to premalignant and frankly malignant. The Papanicolaou Society of Cytopathology (PSC) was established in 2014 for pancreaticobiliary cytology, with a six-tiered system. Recently, the World Health Organization (WHO) proposed a reporting system for pancreaticobiliary cytopathology.Aims:This study aimed to apply PSC and WHO systems to pancreatic cytology samples and compare their risk assessment.Materials and Methods:The Institutional Hospital Information System was searched for all pancreatic endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA), transabdominal ultrasound-guided FNA (TUS-FNA), and pancreatic fluid cytology samples from January 2016 to April 2025, and cases were reassigned as per the PSC and WHO diagnostic categories by two cytopathologists. Risk of malignancy (ROM) in each diagnostic category was assessed using available histopathological, clinical, radiological, biochemical, and follow-up data.Results:Of 281 pancreatic cytology cases via EUS (64.1%) and TUS-FNA (35.9%), 197 had complete follow-up records. The ROM for the PSC categories I-VI was 58.3%, 0%, 75%, 64.7%, 100%, and 100%. For WHO system categories I-IV, VI-VII, ROM values were 66.7%, 0%, 75%, 0%, 100%, and 100%. These findings indicate that the WHO system offers more effective risk stratification. Furthermore, compared with the PSC system, the WHO system showed an increase in sensitivity (89.2% vs. 81.4%), negative predictive value (NPV; 82.7% vs. 73.9%), and accuracy (92.9% vs. 87.8%), while specificity and positive predictive value (PPV) stayed consistent at 100%.Conclusion:The WHO reporting system for pancreatic cytology offers superior risk categorization, sensitivity, NPV, and accuracy compared with the PSC system, while maintaining comparable specificity and PPV.
Introduction: Cervical cancer screening using cervical Papanicolaou (Pap) smears has been well established since the mid-20th century. Specimen adequacy is considered the most relevant quality assurance component since the invention of the Bethesda system for reporting cervical cytology. Cervical smears that are reported as unsatisfactory (US) can significantly delay diagnosis and increase the risk of cancer in follow-up tests. They also lead to unnecessary use of resources. Materials and Methods: The cases signed out as US for one and half year duration in the cytology laboratory were retrieved, and demographic details were collected. Results: Among the 4,139 Pap tests accepted and reported for the period, 240 (5.8%) were signed out as US. The ratio of conventional to ThinPrep smears was 2.4:1. 70% cases were conventional Pap smears. 66% (159/ 240) cases of the patients were in the fourth and fifth decade age group. Low squamous cellularity was the most common cause (80%) of US smears, obscured by inflammation (12%); other causes were obscuration by blood (0.8%) and mucus (0.4%), drying artifacts (1.6%) and cytolysis (0.4%). Discussion and Conclusion: US smears largely depend on various factors, that is, the expertise of the person taking a sample, the date of collection in the menstrual cycle, the method of sampling, sampling devices and staining procedures. The variability in sample collection, including occasional collection by personnel with inadequate training, may also have contributed to the US smears. The result of our study is expected to have a significant impact on proper sampling and the use of adequate quality control measures. In addition, through our follow-up data, we could emphasize the importance of follow-up and human papilloma virus co-testing in patients with US smears.
Cervical cancer screening has evolved immensely with advances in molecular diagnostics and improved understanding of human papillomavirus (HPV)-mediated carcinogenesis. While conventional cervical cytology has historically played a pivotal role in reductions in cervical cancer incidence and mortality, its limitations have led to the development of liquid-based cytology (LBC) and HPV-based testing. LBC offers improved sample adequacy, better preservation of cellular morphology, and permits the use of residual material for ancillary testing. HPV DNA testing, particularly as a primary screening modality, provides high sensitivity for detecting high-grade cervical intraepithelial neoplasia. In addition, emerging biomarkers such as p16INK4a and Ki-67 dual staining enhance risk stratification and triage of HPV-positive women. Despite the growing adoption of primary HPV screening, cytology remains indispensable for morphologic assessment, detection of HPV-independent lesions, and clinical correlation. An integrated approach combining HPV testing, cytology, and biomarkers represents the most effective strategy for cervical cancer screening. Such a balanced model is essential for optimizing diagnostic accuracy, guiding management, and supporting global efforts toward cervical cancer elimination.
Introduction: Thyroid swellings are among the most common clinical presentations, yet not all cases necessitate surgical intervention. Fine-needle aspiration cytology (FNAC) is a widely accepted, minimally invasive diagnostic method that reduces the need for unnecessary surgeries and guides appropriate clinical management. This study is important as it provides population-specific data on the diagnostic accuracy of FNAC and evaluates the risk of malignancy (ROM) and risk of neoplasm (RON) in thyroid nodules (TNs) using the latest Bethesda System for Reporting Thyroid Cytopathology (2023 edition). By systematically categorizing thyroid lesions, the study offers insights into effective risk stratification and supports evidence-based, cost-efficient approaches to patient care. Materials and Methods: A retro-prospective study was conducted over 2 years, enrolling 111 patients with TNs, after obtaining approval from the institutional ethical committee. FNAC was performed on all cases. Smears were stained with Papanicolaou and Giemsa stains and categorized according to the Bethesda classification. Statistical analysis was conducted using IBM Statistical Package for the Social Sciences statistics (version 21.0), and Fisher’s exact test was applied. Results: This study demonstrated a sensitivity of 83.33%, specificity of 100%, and diagnostic accuracy of 90.24%. The positive predictive value was 100%, and the negative predictive value was 80.95%. ROM was 0% in Category I, 19.4% in Category II, and 100% in Categories IV, V, and VI. The overall RON was 21.62%. Papillary thyroid carcinoma was the most common malignancy observed. Conclusion: FNAC remains a highly effective and accessible tool for diagnosing thyroid lesions, particularly when combined with rapid on-site evaluation and cell block preparation. It enables accurate categorization, risk stratification, and a more individualized, cost-effective approach to the management of TNs.
Background: Crush/imprint cytology, a diagnostic technique within the realm of cytological examinations, has emerged as a powerful tool for rapid and minimally invasive tissue evaluation. Its simplicity, cost-effectiveness, and rapidity make it indispensable in pathology practice. Aims and Objectives: The aims and objectives were as follows: (1) to evaluate the diagnostic utility of crush/imprint cytology and (2) to correlate cytomorphological and histopathological (HP) findings for comprehensive diagnostic evaluation. Materials and Methods: This prospective study was carried out in a tertiary health care hospital for a period of 1 year, and 56 cases were included in this study. Both benign and malignant lesions from different organ systems were included in this study. The cytological diagnosis was correlated with HP diagnosis to evaluate the sensitivity and specificity of crush cytology. Results: This study showed a male predominance (69%). Most lesions were from the respiratory system (33.33%), followed by head and neck (24.07%), lymphoreticular system (16.66%), gastrointestinal system (11.11%), skin (9.25%), and genitourinary system (5.55%). Malignant lesions showed good (44%) cell yield on crush/imprint cytology. The sensitivity was 100%, with 94.12% specificity for detecting benign and malignant lesions, and the overall diagnostic accuracy was 98.15%. Conclusion: The crush/imprint cytology stands as a cornerstone in the realm of cellular diagnosis, offering a rapid, cost-effective, and minimally invasive approach to evaluating tissue specimens. It is an indispensable tool in intraoperative diagnostics, offering rapid and accurate assessments of tissue pathology. It acts as a good complement to histopathology and can be of benefit for rapid preliminary diagnosis and intraoperative surgical management planning. Its use is highly recommended routinely.
The fine-needle aspiration (FNA) technique has emerged as a rapid diagnostic modality in diagnosis of metastatic tumors of unknown primary. Renal cell carcinoma (RCC) is well known for its potential to metastasize to remote sites. However, RCC primarily presenting with cervical lymph node metastasis is an unusual phenomenon. Herein, we describe a case of RCC with cervical lymph node metastasis as the initial clinical presentation, which was diligently diagnosed in FNA smears by appreciating the classic cytomorphological features and with wise use of an immunocytochemical panel.
Introduction:Fine-needle aspiration biopsy (FNAB) is a valuable technique for diagnosing axillary breast cancer metastasis. However, its sensitivity can sometimes be affected by sampling errors and technical artifacts. Aim:We examined whether cytokeratin fragment antigen 21-1 (CYFRA) estimation by manual enzyme-linked immunosorbent assay (ELISA) on FNAB needle washouts could add diagnostic information to FNAB. Materials and Methods:This cross-sectional study involved FNAB from 138 lymph nodes in 108 suspected breast cancer patients. Papanicolaou and May-Grünwald-Giemsa smears were prepared for cytological examination. CYFRA was estimated by manual ELISA from fine-needle washouts after smear preparation. A cut-off of 1.93 ng/mL was considered to be a positive test for CYFRA. Results:The CYFRA estimates were significantly associated with the FNAB results (P < 0.001 by Kruskal-Wallis test). On FNAB, 84, 5, 36, and 13 cases were classified as malignant, suspicious, benign, and inadequate, respectively. CYFRA was reported as positive in 92 cases (75 in malignant FNAB, 3 in suspicious FNAB, 11 in benign FNAB, and 3 in inadequate FNAB). Six of 11 CYFRA positives in benign FNAB were confirmed by re-examination of cytology slides; the remaining 5 CYFRA-positive cases did not have further biopsy. Only two inadequate FNAB had histopathology follow-up; both revealed metastasis, of which one was CYFRA positive. Sensitivity and specificity with 95% confidence intervals of CYFRA for detecting axillary metastasis were 0.876 [0.794, 0.934] and 0.833 [0.653, 0.954], respectively. Conclusion:CYFRA estimation by manual ELISA is a simple and accurate test which supplements and enhances the diagnostic usefulness of axillary FNAB in suspected breast cancer metastasis.
The art of cytopathology lies not only in the interpretation of the slide but also careful consideration of demographic, clinical, radiological, and other relevant information of individual patients. While deep image learning may perform well in specific cytology image sets or uncover novel, significant cytomorphologic parameters, the lack of clinical context puts prediction models at a dangerous position where irrelevant and potentially harmful diagnoses may be issued. The increment in prediction accuracy by expanding the size of image sets diminishes and plateaus after reaching certain thresholds. This can be overcome by introducing multimodal data to deep learning, which has seen success in resection and small biopsy specimens from histopathological specimens and early adoption in cytopathology. In this study, a qualitative review on the current state of the potential sources of multimodal input—demographics, clinical investigations, and radiology; the diagnostic data that can be obtained in a cytology specimen—stains and preparations, immunocytochemistry, and molecular testing will be discussed with reference to the potential of multimodal deep learning in cytopathology.
Background: Fine-needle aspiration cytology (FNAC) is a simple and reliable technique for diagnosing various inflammatory lesions. Some of these lesions are caused by fungal organisms, which are often missed clinically and thus remain untreated. Most of these infections require prompt diagnosis as they respond to early treatment, and delayed diagnosis may lead to unnecessary complications. Aim: To study the clinico-cytomorphological spectrum of fungal infections in a tertiary care center. Materials and Methods: This is a retrospective study and includes a total of 25 cases of various fungal infections diagnosed on FNAC over 5 years. Clinical details of these cases were retrieved from cytopathology records, and cytomorphological features were studied in detail. Results: All 25 cases were analyzed and included rhino-orbital cerebral mucormycosis (15), histoplasmosis (1), cryptococcosis (1), phaeohyphomycosis (1), and aspergillosis (1), along with six cases broadly classified as fungal abscesses. Conclusion: FNAC is an efficient and unswerving tool, helpful in diagnosing a plethora of fungal infections, especially in clinically unsuspected cases, and thereby facilitates early diagnosis and appropriate treatment.
Background:The International Academy of Cytology Yokohama System for Reporting Breast Fine Needle Aspiration Cytopathology (IAC YSRB) aims to standardize breast fine needle aspiration cytology (FNAC) reporting by categorizing cytologic findings into five diagnostic groups. IAC YSRB is widely adopted; however, its reproducibility among cytopathologists with varying experience, especially in intermediate categories, remains a concern. Objective:To assess the interobserver reproducibility of the IAC YSRB among four cytopathologists using archived breast FNAC cases. Materials and Methods:This retrospective study included 239 breast FNAC cases reported in 2022 at a tertiary care center. Four cytopathologists with 20, 14, 5, and 3 years of experience independently reviewed smears and assigned each case to one of five IAC YSRB diagnostic categories. Observers were blinded to prior diagnoses and one another's assessments. Interobserver agreement was analyzed using Fleiss' kappa (κ) and pairwise κ values, with category-wise and overall concordance calculated. Results:Overall interobserver agreement was substantial (κ = 0.69). The highest agreement was observed for malignant (κ = 0.95) and benign (κ = 0.72) categories. Atypical (κ = 0.22) and suspicious (κ = 0.26) categories showed fair agreement. Pairwise comparisons revealed higher concordance between more experienced cytopathologists (κ = 0.80) than less experienced ones. Complete agreement among all four raters was achieved in 81% of cases. Disagreements were mainly concentrated in intermediate categories. Conclusion:IAC YSRB shows good interobserver reproducibility in breast FNAC reporting, particularly in benign and malignant categories. Interpretive variability in intermediate categories highlights the need for training, standardization, and ancillary tools to improve reliability.
Follicular thyroid carcinoma (FTC) is a malignant follicular cell-derived neoplasm with prominent hematogenous spread. Metastases may be the presenting symptom; however, scalp metastasis is uncommon. A 66-year-old woman presented with progressive scalp swelling since 4 years accompanied by bilateral lower limb weakness and incontinence. On examination, there was a large 10-cm parietal scalp mass. Fine-needle aspiration smears prepared from the pyramidal scalp swelling showed microfollicular arrangement of uniform tumor cells. Diffuse positivity for pancytokeratin, PAX 8, and thyroglobulin on cell block immunocytochemistry confirmed the diagnosis of metastatic FTC. fluorodeoxyglucose avid lesions in the right lobe of the thyroid and multiple skeletal lesions were seen on positron emission tomography-computed tomography done subsequently. Despite I-131 treatment, she succumbed to death 6 months post-diagnosis. The index case highlights the utility of fine-needle aspiration cytology and the application of immunocytochemistry on cell blocks changing the norms for diagnosing carcinomas with an unknown primary.
Aims and Objectives:To evaluate the performance of a soft-voting ensemble deep learning (DL) model incorporating five transfer learning (TL) architectures for categorizing cervical smears according to the Bethesda system.Materials and Methods:A total of 259 cervical cytology cases processed by SurePath liquid-based cytology were included, yielding 1016 representative microphotographs. Images were divided into training (70%), validation (15%), and test (15%) sets. An ensemble model comprising VGG16, ResNet50, InceptionV3, DenseNet121, and MobileNetV2 was trained using TL with soft-voting integration. Diagnostic performance was assessed using standard metrics, including 95% confidence intervals (CI).Results:In the test set, the model correctly classified 27/31 negative for intraepithelial lesions and malignancy (NILM) images and 56/58 carcinoma images. For NILM, the sensitivity was 0.871 (95% CI: 0.72-0.95), specificity 0.977 (95% CI: 0.94-0.99), accuracy 0.956 (95% CI: 0.92-0.98), and area under the curve of receiver operating characteristic (AUROC) 0.98 (95% CI: 0.96-1.00). For carcinoma, the sensitivity was 0.966 (95% CI: 0.88-1.00), specificity 0.960 (95% CI: 0.87-0.99), accuracy 0.962 (95% CI: 0.93-0.99), and AUROC 0.99 (95% CI: 0.98-1.00). Intermediate categories demonstrated variable performance, with lower sensitivity for low-grade squamous intraepithelial lesion (0.20), though the specificity remained high (1.00). Overall, AUROC values were high across all categories (0.93-0.99). Among all the models, the ensemble learning (EL) model had the best performance.Conclusions:The proposed EL approach demonstrates high diagnostic accuracy and excellent discrimination for clinically critical categories such as high-grade squamous intraepithelial lesion (HSIL) and carcinoma. This model highlights the potential of DL-based ensemble frameworks in improving the reliability of cervical cytology interpretation. Larger datasets and whole-slide image analysis may further refine the performance for intermediate-grade lesions.
Background: Peritoneal washing (PW) cytology has become a widely accepted procedure for staging of gynecological malignancies. Aims: To compare conventional cytology, liquid-based cytology (LBC) and cell block in the evaluation of PWs. Materials and Methods: A total of 50 clinically suspected cases of gynecological malignancies were enrolled, which included 41 ovarian neoplasms, eight uterine, and one case of cervical malignancy. The age of the women ranged from 13 to 74 years. The PW samples obtained in all cases were subjected to conventional cytology, LBC, and cell block preparations. Immunohistochemistry (IHC) was applied on cell blocks and analyzed. A comparative evaluation and statistical correlation between the three techniques was performed. Results: The number of cases reported as positive for malignancy on conventional smear and LBC were 9 (18%) and 11 (22%), respectively. Fifteen (30%) cases were reported as positive for malignancy on cell block after IHC analysis. All three cytological techniques showed a good correlation in diagnosing peritoneal metastasis on evaluation by Cohen’s kappa score (≤ 1.00). However, the overall sensitivity, specificity, and diagnostic accuracy of PW cytology was 43.8%, 94.4%, and 62%, respectively, taking histopathology as the gold standard. The most common malignancy which led to peritoneal fluid involvement was high-grade serous carcinoma of the ovary. Conclusion: The study demonstrates the usefulness of combining the three cytological techniques to diagnose peritoneal metastasis in gynecological malignancies. Cell block with IHC can help in reducing the false-negative cases and further adds to objectivity in ambiguous cases.
Background and Objectives: Papillary lesions of the breast represent a diverse group of neoplasms with varying biological behavior, ranging from benign to malignant. Accurate diagnosis is critical for appropriate clinical management. This study aims to analyze the cytopathological features of papillary breast lesions to improve diagnostic accuracy. Materials and Methods: A retrospective study was performed on cases of papillary neoplasm of the breast diagnosed on cytology followed by histopathology during the study period, after procuring institutional ethical committee clearance. The clinical, pathologic, and follow-up details were retrieved from the hospital records and the pathology database and were analyzed. Results: The study identified distinct benign and malignant papillary lesions, including intraductal papilloma, papillary ductal carcinoma in situ, encapsulated papillary carcinoma, and solid papillary carcinoma. The mean age of the study participants was 51.15 years. Benign neoplasms showed cohesive clusters and a clear background on cytology, while cases with atypical or malignant features demonstrated loss of myoepithelial cells, non-cohesive clusters, and necrotic/dirty background. Conclusion: Papillary breast lesions encompass a wide spectrum of pathology, necessitating precise cytologic and histopathologic evaluation. The presence or absence of myoepithelial markers aids in differentiating benign from malignant lesions. Identification of cytologic features of benign and malignant neoplasms can improve diagnostic accuracy, enhance patient management, and ensure timely intervention for malignant cases.
Objectives: Interstitial lung diseases (ILDs) are a heterogeneous group of pulmonary disorders requiring a multidisciplinary diagnostic approach. Bronchoalveolar lavage (BAL) is a minimally invasive procedure that provides cellular insights into lung pathology. This study evaluates the role of BAL cytology in ILD cases. Materials and Methods: This observational study was conducted at a tertiary care health care center, from April 2023 to June 2024. A total of 100 ILD patients were enrolled, out of which 31 patients who required multidisciplinary diagnosis (MDD) for a final diagnosis of ILD were included. BAL fluid was processed using cytocentrifugation and stained with Giemsa and Papanicolaou stains, and differential cell count patterns (cellular analysis) were assessed in 400 cells. In cases of sarcoidosis, endobronchial biopsy was performed in five cases, and endobronchial ultrasound-guided transbronchial needle aspiration was performed in nine cases. The MDD approach incorporated clinical features, high-resolution computed tomography, biochemical, microbiological, biopsy, and cytology findings. Results: The mean age of patients was 41.2 ± 12.2 years, with a female predominance (74.2%). The distribution of ILD subtypes included sarcoidosis (45.2%), connective tissue disease-associated interstitial lung disease (CTD-ILD) (25.8%) idiopathic nonspecific interstitial pneumonia (6.5%), and hypersensitivity pneumonitis (HP) (22.6%). BAL cellular analysis showed lymphocytosis in sarcoidosis (41.8%) and chronic HP (56.7%) with neutrophilia (mean value >20%). The CTD-ILD cellular pattern is similar to that of sarcoidosis. BAL findings corroborated with the final diagnoses in cases with multiple differentials. Conclusion: BAL cellular analysis is a valuable tool in the MDD approach to ILD diagnosis. The presence of lymphocytosis in BAL commensurates with sarcoidosis and chronic HP diagnosis. The presence of concurrent neutrophilia in such cases may not necessarily be considered in lines of an infective etiology or acute exacerbation.
Background:Urinary cytology plays a pivotal role in the non-invasive diagnosis of upper tract urothelial carcinoma (UTUC), especially where histological sampling is limited or technically challenging. The second edition of the Paris System for Reporting Urinary Cytology standardizes for evaluating cytological specimens. However, in clinical practice, urinary samples obtained from the UT often contain residual contrast media (CM) due to prior imaging studies; hence, the effect of CM on urinary cytology accuracy remains unclear. Aims:The aim of this study was to examine the impact of CM on diagnostic performance of Paris system. Settings and Design:This was a controlled in vitro laboratory study using RT112 human bladder cancer cells to model UTUC. Materials and Methods:Cells were exposed to 100%, 50%, and 30% CM. Morphology, membrane integrity, and viability were assessed through time-lapse microphotography, Trypan blue, and MTT assay. Cytology was evaluated by a cytotechnologist. Results:CM exposure caused concentration- and time-dependent cellular degeneration. After 60 min, viability dropped to 2.0 ± 0.1% (100%), 29.0 ± 4.7% (50%), and 75.1 ± 2.0% (30%), while membrane integrity declined similarly. Trypan blue staining showed integrity of 1.6 ± 0.1%, 15.1 ± 1.6%, and 46.5 ± 2.8%, respectively. Cells treated with 100% or 50% CM were classified as atypical urothelial cells (AUC) after 1 h, while 30% CM preserved high-grade urothelial carcinoma (HGUC) features up to 6 h. However, 24-hour exposure to 30% CM led to nuclear degeneration and cytoplasmic loss, again resulting in AUC classification. Conclusion:CM at higher concentrations and longer durations significantly compromise cellular morphology and diagnostic accuracy in urinary cytology.
