Fine needle aspiration cytology of metastatic carcinomas with papillary architecture: A systemic assessment of clinical, cytological and immunohistochemical parameters.

BACKGROUND: Papillary structures are frequently encountered in metastatic carcinomas from various organs and tumours of different histotypes. This study aims to investigate the predictive value of fine needle aspiration cytology (FNAC), immunohistochemistry (IHC) and the clinical parameters which can be examined in the assessment of the primary sites of metastatic carcinomas with papillary architecture. METHODS: FNAC samples of metastatic carcinomas with papillary architecture were evaluated for overall cellularity, epithelial cohesion, background features, papillary architecture, cytology and IHC. The corresponding clinical information was also reviewed. RESULTS: A total of 130 cases were included. The most common primary sites were thyroid (38.5%), lung (30.8%) and gynecological organs (22.3%); the others were pancreaticobiliary, urothelial, colorectal, and esophageal. Age (P = 0.039), biopsy site (P < 0.001) and laterality (P = 0.006) correlated with primary site. Papillary structures were confirmed on biopsy/excision of most cases (n = 85/87, 97.7%). Thyroid primaries exhibited broad papillary stalks, thin lining epithelium, fewer epithelial polymorphs, and the presence of background giant cells and histiocytes (P = 0.021- < 0.001). Low-grade cytological features, nuclear grooves and inclusions (P < 0.001) were seen in thyroid primaries. High-grade features (P < 0.001-0.49), multinucleated tumour cells, apoptotic bodies and mitoses (P < 0.001-0.49) were more common in lung/gynecological primaries. Multivariate analysis identified nuclear/cytoplasmic ratio, chromatin character, the presence of nuclear grooves and mitosis as independent features (P = 0.001-0.024). TTF1/TGB/PAX8 panel results showed good agreement with the cytological assessment and site of primary. CONCLUSION: Papillary structures and cytological features are reproducible in FNAC assessment of metastases and their corresponding primary sites. Cytological features, IHC and clinical information are invaluable in determining the primary site.

Noninvasive Detection of Bladder Cancer by Shallow-Depth Genome-Wide Bisulfite Sequencing of Urinary Cell-Free DNA for Methylation and Copy Number Profiling.

BACKGROUND: The current diagnosis and monitoring of bladder cancer are heavily reliant on cystoscopy, an invasive and costly procedure. Previous efforts in urine-based detection of bladder cancer focused on targeted approaches that are predicated on the tumor expressing specific aberrations. We aimed to noninvasively detect bladder cancer by the genome-wide assessment of methylomic and copy number aberrations (CNAs). We also investigated the size of tumor cell-free (cf)DNA fragments. METHODS: Shallow-depth paired-end genome-wide bisulfite sequencing of urinary cfDNA was done for 46 bladder cancer patients and 39 cancer-free controls with hematuria. We assessed (a) proportional contribution from different tissues by methylation deconvolution, (b) global hypomethylation, (c) CNA, and (d) cfDNA size profile. RESULTS: Methylomic and copy number approaches were synergistically combined to detect bladder cancer with a sensitivity of 93.5% (84.2% for low-grade nonmuscle-invasive disease) and a specificity of 95.8%. The prevalence of methylomic and CNAs reflected disease stage and tumor size. Sampling over multiple time points could assess residual disease and changes in tumor load. Muscle-invasive bladder cancer was associated with a higher proportion of long cfDNA, as well as longer cfDNA fragments originating from genomic regions enriched for tumor DNA. CONCLUSIONS: Bladder cancer can be detected noninvasively in urinary cfDNA by methylomic and copy number analysis without previous knowledge or assumptions of specific aberrations. Such analysis could be used as a liquid biopsy to aid diagnosis and for potential longitudinal monitoring of tumor load. Further understanding of the differential size and fragmentation of cfDNA could improve the detection of bladder cancer.

New simple prognostic score for primary biliary cirrhosis: Albumin-bilirubin score.

BACKGROUND AND AIM: Serum albumin and bilirubin are the most significant independent prognostic factors to predict hepatic events in patients with primary biliary cirrhosis (PBC). We aimed to investigate the prognostic significance of a new prognostic score, the albumin-bilirubin (ALBI) score, among PBC patients. METHODS: In a retrospective longitudinal cohort of 61 Chinese PBC patients with follow-up period up to 18.3 years, the prognostic performance of the ALBI in prediction of hepatic events was compared with other well-established prognostic scores: Child-Pugh score, model of end-stage liver disease, Mayo risk score, Yale, European, and Newcastle models. RESULTS: Fifteen patients (24.6%) developed hepatic events during follow-up. The c-index (0.894) and χ(2) by likelihood ratio test (36.34) of the ALBI score were highest in comparison to other models. The ALBI score was the only independent prognostic factor by multivariate analysis and its adjusted hazard ratio of developing hepatic event was 27.8 (P < 0.001). There were three prognostically different groups stratified by the ALBI score: ALBI grade 1 (≤ -2.60), grade 2 (> -2.60 to -1.39), and grade 3 (> -1.39) groups. The 2-, 5-, and 10-year event-free survivals for grade 1, grade 2, and grade 3 groups were 100.0% versus 100.0% versus 57.1%, 100.0% versus 88.5% versus 14.3%, and 100.0% versus 81.7% versus 0.0%, respectively (P < 0.001). CONCLUSION: The ALBI score is readily derived from a blood test without using those factors evaluated subjectively or obtained by invasive procedures. It is an independent prognostic factor for PBC patients and provides better/similar prognostic performance compared with other prognostic scores.

Evaluation of histological staging systems for primary biliary cirrhosis: correlation with clinical and biochemical factors and significance of pathological parameters in prognostication.

AIMS: A new Japanese histological staging system for primary biliary cirrhosis (PBC) has been proposed. We aimed to evaluate the efficacies of the Scheuer, Ludwig and Japanese staging systems, with emphasis on their clinical and biochemical correlations and prognostic significances. METHODS AND RESULTS: We conducted a retrospective review of a cohort of 58 Chinese PBC patients, with follow-up of up to 16.9 years. All three systems correlated well with prognostically significant parameters, namely serum bilirubin, Mayo scores and model for end-stage liver disease (MELD) score. Only the Japanese staging system was associated with Child-Pugh score, which was the single independent prognostic factor for liver-related events (log-rank P < 0.001; Cox proportional hazard ratio (HR) 6.723, P < 0.001). The Japanese system (log-rank P = 0.007; Cox proportional HR 10.400, P = 0.025) predicted liver-related events, while Scheuer (log-rank P = 0.112) and Ludwig (log-rank P = 0.147) systems did not. The copper-associated protein (CAP) deposition score, a component of the Japanese system, was the most powerful histological prognostic parameter (log-rank P < 0.001; Cox proportional HR 99.534, P = 0.049) and provided extra prognostic values in additional to serum albumin, serum bilirubin, Child-Pugh score, Mayo scores and MELD score. CONCLUSION: The Japanese staging system is more effective than classical systems. The degree of CAP deposition is an essential prognostic histological parameter.

Deep learning for computational cytology: A survey.

Computational cytology is a critical, rapid-developing, yet challenging topic in medical image computing concerned with analyzing digitized cytology images by computer-aided technologies for cancer screening. Recently, an increasing number of deep learning (DL) approaches have made significant achievements in medical image analysis, leading to boosting publications of cytological studies. In this article, we survey more than 120 publications of DL-based cytology image analysis to investigate the advanced methods and comprehensive applications. We first introduce various deep learning schemes, including fully supervised, weakly supervised, unsupervised, and transfer learning. Then, we systematically summarize public datasets, evaluation metrics, versatile cytology image analysis applications including cell classification, slide-level cancer screening, nuclei or cell detection and segmentation. Finally, we discuss current challenges and potential research directions of computational cytology.

A comparative study of diagnostic accuracy in 3026 pleural biopsies and matched pleural effusion cytology with clinical correlation.

BACKGROUND: Pleural effusion can be caused by a wide range of benign and malignant conditions. Pleural biopsy and effusion cytology represent two key methods of pathological diagnosis. To compare the performance these two methods, a large cohort of matched pleural biopsy and effusion cytology with clinical follow-up was reviewed. METHODS: Pleural biopsies and effusion cytology specimens over a period of 18 years were retrieved. Cytology specimens collected within 7 days of pleural biopsy were matched. Reports were reviewed, and the cause for pleural effusion was determined by hospital disease coding and clinical data. RESULTS: Totally, 3026 cases were included. The leading cause of benign effusion was tuberculosis (n = 650). Malignant pleural effusion (MPE) was more common in older females (p < 0.001) and mostly due to lung cancer (n = 959), breast cancer (n = 64), and mesothelioma (n = 48). The inadequate/insufficient (B1/C1) rate of biopsy was higher than cytology (15.6% vs. 0.3%) but the rates for other diagnostic categories were similar. Biopsy and cytology showed a correlation coefficient of 0.315, improving to 0.449 when inadequate/insufficient (B1/C1) cases were excluded. The ROM for benign cytology (C2) was lower than biopsy (B2) (p < 0.001). Compared with biopsy, the diagnostic accuracy was higher in cytology overall and for metastatic carcinomas (p < 0.001) but lower for hematolymphoid malignancies (p = 0.014) and mesotheliomas (p = 0.002). CONCLUSIONS: These results suggest that effusion cytology may be better for confirming benignity and diagnosing carcinomatous MPE. In these cases, pleural biopsy may be withheld to reduce procedural risks. However, for suspected hematolymphoid malignancies and mesothelioma, biopsy should be considered.

Number Needed to Diagnose in Malignant Ascites: Effects of Volume, Repeating Collection, and Primary Malignancy on Diagnostic Performance in Peritoneal Fluid Cytology.

INTRODUCTION: Volume recommendations of 80-200 mL have been proposed for peritoneal fluid cytology. While cutoffs are impractical when volume is limited by the amount present and disease factors, collections, however, can be repeated. This study addresses adequacy and number needed to diagnose by comparing diagnostic agreement to volumes in single specimens, total volumes collected daily, and within admissions. The diagnostic yield of repeating collection within a single day, admission, and throughout admissions of a patient's lifetime was also investigated. METHODS: Peritoneal fluid cytology specimens over a 27-year period were retrieved and matched by collection date, admission number, and patient number. Case notes were reviewed to establish all cases of malignant ascites. RESULTS: In total, 19,392 specimens from 14,327 admissions and 11,089 patients were retrieved, with 1,531 patients confirmed with malignant ascites. Agreements between cytologic diagnoses within the same day and admission were high (κ > 0.8). Fluid volume increased with grade of cytologic diagnosis (p < 0.001), and greater volume was associated with higher discordance (p < 0.05). Specimens of 60-100 mL showed the best diagnostic concordance. To achieve a 99.5% diagnostic rate, three sequential aliquots, collections from two different days in an admission, or three admissions within a lifetime are required. The diagnostic yield of one aliquot within batches from the same day was only 88.9%. Gastrointestinal (p = 0.040), gynecologic (p = 0.005), and lung (p < 0.001) malignancies required the least repeats for diagnosis. CONCLUSIONS: Omission of any fluid from laboratory submission is strongly discouraged. As a simple rule, three repeats are necessary for excluding malignant ascites.

Idiopathic granulomatous mastitis and cystic neutrophilic granulomatous mastitis: two sides of the same coin or distinct entities?

Idiopathic granulomatous mastitis (IGM) is a benign mimic of breast carcinomas. It is defined histologically by the presence of granulomas and inflammation. The closely related cystic neutrophilic granulomatous mastitis (CNGM) shows lipogranulomas, with a reported association with corynebacteria. A large cohort of IGM was reviewed to compare clinical, microbiological and histological features between non-CNGM IGM and CNGM. Cases of IGM were reviewed for histological parameters including the presence of lipogranulomas and composition of inflammatory cells. Clinical data were obtained through hospital records. The cohort included 79 cases, including 51 non-CNGM IGM and 28 CNGM. Comparing non-CNGM IGM and CNGM, there were no differences in clinical or demographical data, other than a younger age of presentation (36.2 vs 41.5 years, p=0.012) for CNGM. Most IGM resolved within the follow-up period (n=57/64, 89.1%), with similar outcomes regardless of treatment (p>0.05). In CNGM, there were more infiltrates of neutrophils (p=0.001), histiocytes (p=0.047), and multinucleated giant cells (p=0.006), but less lymphocytes (p=0.008). Corynebacteria was cultured in two (25%) cases of CNGM, and one non-CNGM IGM (14.3%). Gram-positive bacilli were identified in two cases of CNGM. 'Early' lipogranulomas were observed closely associated to inflamed ducts in three cases of CNGM. Apart from age, there was no distinct clinical or microbiological feature for CNGM. These findings do not support CNGM as a distinct entity. Rather, CNGM-pattern may represent a continuum of IGM, possibly age-related and attributable to ductal inflammation and cystic changes in the breast parenchyma.

Deep learning-derived spatial organization features on histology images predicts prognosis in colorectal liver metastasis patients after hepatectomy.

Histopathological images of colorectal liver metastases (CRLM) contain rich morphometric information that may predict patients' outcomes. However, to our knowledge, no study has reported any practical deep learning framework based on the histology images of CRLM, and their direct association with prognosis remains largely unknown. In this study, we developed a deep learning-based framework for fully automated tissue classification and quantification of clinically relevant spatial organization features (SOFs) in H&E-stained images of CRLM. The SOFs based risk-scoring system demonstrated a strong and robust prognostic value that is independent of the current clinical risk score (CRS) system in independent clinical cohorts. Our framework enables fully automated tissue classification of H&E images of CRLM, which could significantly reduce assessment subjectivity and the workload of pathologists. The risk-scoring system provides a time- and cost-efficient tool to assist clinical decision-making for patients with CRLM, which could potentially be implemented in clinical practice.

High-Throughput, Label-Free and Slide-Free Histological Imaging by Computational Microscopy and Unsupervised Learning.

Rapid and high-resolution histological imaging with minimal tissue preparation has long been a challenging and yet captivating medical pursuit. Here, the authors propose a promising and transformative histological imaging method, termed computational high-throughput autofluorescence microscopy by pattern illumination (CHAMP). With the assistance of computational microscopy, CHAMP enables high-throughput and label-free imaging of thick and unprocessed tissues with large surface irregularity at an acquisition speed of 10 mm2 /10 s with 1.1-µm lateral resolution. Moreover, the CHAMP image can be transformed into a virtually stained histological image (Deep-CHAMP) through unsupervised learning within 15 s, where significant cellular features are quantitatively extracted with high accuracy. The versatility of CHAMP is experimentally demonstrated using mouse brain/kidney and human lung tissues prepared with various clinical protocols, which enables a rapid and accurate intraoperative/postoperative pathological examination without tissue processing or staining, demonstrating its great potential as an assistive imaging platform for surgeons and pathologists to provide optimal adjuvant treatment.

Identification of Tissue Types and Gene Mutations From Histopathology Images for Advancing Colorectal Cancer Biology.

Objective: Colorectal cancer (CRC) patients respond differently to treatments and are sub-classified by different approaches. We evaluated a deep learning model, which adopted endoscopic knowledge learnt from AI-doscopist, to characterise CRC patients by histopathological features. Results: Data of 461 patients were collected from TCGA-COAD database. The proposed framework was able to 1) differentiate tumour from normal tissues with an Area Under Receiver Operating Characteristic curve (AUROC) of 0.97; 2) identify certain gene mutations (MYH9, TP53) with an AUROC > 0.75; 3) classify CMS2 and CMS4 better than the other subtypes; and 4) demonstrate the generalizability of predicting KRAS mutants in an external cohort. Conclusions: Artificial intelligent can be used for on-site patient classification. Although KRAS mutants were commonly associated with therapeutic resistance and poor prognosis, subjects with predicted KRAS mutants in this study have a higher survival rate in 30 months after diagnoses.

Papillary lesions of the breast: A systematic evaluation of cytologic parameters.

BACKGROUND: The cytologic diagnosis of papillary lesions of the breast is challenging because of the diverse morphology, including epithelial hyperplasia, atypia, low-grade malignancy, and neuroendocrine differentiation; also, traditional malignant features such as necrosis and myoepithelial cell loss can be lacking. Thus, the diagnostic criteria for papillary lesions may differ from those for other breast lesions. This study evaluated various cytologic parameters in a large cohort to identify useful diagnostic features. METHODS: Cytologic preparations of papillary lesions with histologic follow-up were reviewed for features related to cellularity, epithelial cohesiveness, cellular and stromal architecture, cytomorphology, and background. Corresponding histologic slides were also reviewed. RESULTS: In all, 153 cases were included. Epithelial discohesion, solid and cribriform patterns, atypical nuclear features, and mitoses (P ≤ .001 to P = .017) were associated with malignancy. Cell balls, monolayer sheets, and features of cystic change (P < .001 to P = .016) were associated with benign lesions. Complex (P = .031) and slender (P = .026) papillae and neuroendocrine features (P < .001) were associated with malignancy. Hemorrhage, background, and infiltrating neutrophils (P < .001 to P = .025) were associated with malignancy; fibrotic broad papillary stromal fragments (naked papillary fronds [NPFs]; P = .043) were associated with benignity. The presence of any single parameter, including the absence of myoepithelial cells within epithelial structure, the presence of cytoplasmic granules, an increased amount of cytoplasm, and a nuclear to cytoplasmic (N/C) ratio greater than 0.7, which were identified by principal component analysis, yielded a sensitivity of 95.1% and a specificity of 100.0% in predicting malignancy. CONCLUSIONS: Methodological assessment of multiple features is recommended. Myoepithelial cells, cytoplasmic granules, the amount of cytoplasm, and the N/C ratio are key features for diagnosis.

NOTCH3, a crucial target of miR-491-5p/miR-875-5p, promotes gastric carcinogenesis by upregulating PHLDB2 expression and activating Akt pathway.

Aberrant Notch activation has been implicated in multiple malignancies and the identification of NOTCH receptors and related pathways is critical for targeted therapy. In this study, we aim to delineate the most prominent dysregulated NOTCH receptor and comprehensively reveal its deregulation in gastric cancer (GC). In the four Notch members, NOTCH3 was found uniformly upregulated and associated with poor clinical outcomes in multiple GC datasets. siRNA-mediated NOTCH3 knockdown demonstrated antitumor effects by suppressing cell proliferation, inhibiting monolayer formation, and impairing cell invasion abilities. Its depletion also induced early and late apoptosis. NOTCH3 was confirmed to be a direct target of two tumor suppressor microRNAs (miRNAs), namely miR-491-5p and miR-875-5p. The activation of NOTCH3 is partly due to the silence of these two miRNAs. Through RNA-seq profiling and functional validation, PHLDB2 was identified as a potent functional downstream modulator for NOTCH3 in gastric carcinogenesis. PHLDB2 expression demonstrated a positive correlation with NOTCH3, but was negatively correlated with miR-491-5p. Akt-mTOR was revealed as the downstream signaling of PHLDB2. The NOTCH3-PHLDB2-Akt co-activation was found in 33.7% GC patients and the activation of this axis predicted poor clinical outcome. GC cells treated with siNOTCH3, siPHLDB2, miR-491-5p, miR-875-5p, were more sensitive to Cisplatin and 5-FU. Taken together, the NOTCH3-PHLDB2-Akt cascade plays oncogenic role in gastric carcinogenesis and serves as a therapeutic target. Our study provided insights into Notch-mediated underlying molecular mechanisms and implied translational potential.

Virtual multiplex immunohistochemistry: Application on cell block of effusion and aspiration cytology.

BACKGROUND: With the continuous development of new antibodies, the use of immunohistochemistry (IHC) is becoming more often a requirement. IHC is frequently necessary for establishing cancer diagnosis and making therapeutic decisions. However, cytology specimens such as effusion fluid and fine-needle aspiration are highly variable in cellularity and contain background inflammatory and mesothelial cells. Compared to biopsy or excision specimens, tissue exhaustion and levels of sections not matching are more commonly encountered. We present a method of integrating whole-slide cell block image of multiple antibodies by digital image processing and reconstructing a virtual multiplex IHC image for enhanced interpretation. METHODS: Historic archived cell block preparations of carcinomas (n = 19) and melanoma (n = 1) and IHC performed with 3,3'-diaminobenzidine chromogen were reviewed. The slides were digitized by a whole-slide image scanner. Using ImageJ and self-developed code, the slides were aligned by image registration, layered, and recolored to reconstruct a virtual multiplex image, simulating a multiplex preparation with multiple chromogens. To quantity the performance of the image registration, the mean distance between the same cell clusters in aligned images were measured. RESULTS: All 20 cases were successfully registered. The mean distance between cell clusters after image registration was 8.40 +/- 5.52 µm. The reconstructed images were able to demonstrate coexpression of membranous, cytoplasmic, and nuclear antibodies. CONCLUSION: With virtual multiplex IHC, we were able to visualize coexpression of multiple antibodies without the added cost of multiplex IHC. Routine and historic slides, without additional tissue consumption, can be retrieved for image reconstruction. This technique is a low-cost adjunct to diagnosis in cytology for more efficient and accurate assessment of antibody coexpression and histochemical stains.

FGF18-FGFR2 signaling triggers the activation of c-Jun-YAP1 axis to promote carcinogenesis in a subgroup of gastric cancer patients and indicates translational potential.

Fibroblast growth factor receptor type 2 (FGFR2) has emerged as a key oncogenic factor that regulates gastric cancer (GC) progression, but the underlying mechanism of FGF-FGFR2 signaling pathway remains largely unknown. To identify the potential molecular mechanisms of the oncogenic FGFR2 in gastric carcinogenesis and convey a novel therapeutic strategy, we profiled the FGFR alterations and analyzed their clinical associations in TCGA and Hong Kong GC cohorts. We found that FGFR2 overexpression in GC cell lines and primary tumors predicted poor survival and was associated with advanced stages of GC. Functionally, growth abilities and cell cycle progression of GC were inhibited by inactivation of ERK-MAPK signal transduction after FGFR2 knockdown, while apoptosis was promoted. Meanwhile, the first-line anti-cancer drug sensitivity was enhanced. RNA-seq analysis further revealed that YAP1 signaling serves as a significant downstream modulator and mediates the oncogenic signaling of FGFR2. When stimulating FGFR2 by rhFGF18, we observed intensified F-actin, nuclear accumulation of YAP1, and overexpression of YAP1 targets, but these effects were attenuated by either FGFR2 depletion or AZD4547 administration. Additionally, the FGF18-FGFR2 signaling upregulated YAP1 expression through activating c-Jun, an effector of MAPK signaling. In our cohort, 28.94% of GC cases were characterized as FGFR2, c-Jun, and YAP1 co-positive and demonstrated worse clinical outcomes. Remarkably, we also found that co-targeting FGFR2 and YAP1 by AZD4547 and Verteporfin synergistically enhanced the antitumor effects in vitro and in vivo. In conclusion, we have identified the oncogenic FGF-FGFR2 regulates YAP1 signaling in GC. The findings also highlight the translational potential of FGFR2-c-Jun-YAP1 axis, which may serve as a prognostic biomarker and therapeutic target for GC.