Challenges in Ki-67 assessments in pulmonary large-cell neuroendocrine carcinomas

Learning with limited target data to detect cells in cross-modality images

Deep neural networks have achieved excellent cell or nucleus quantification performance in microscopy images, but they often suffer from performance degradation when applied to cross-modality imaging data. Unsupervised domain adaptation (UDA) based on generative adversarial networks (GANs) has recently improved the performance of cross-modality medical image quantification. However, current GAN-based UDA methods typically require abundant target data for model training, which is often very expensive or even impossible to obtain for real applications. In this paper, we study a more realistic yet challenging UDA situation, where (unlabeled) target training data is limited and previous work seldom delves into cell identification. We first enhance a dual GAN with task-specific modeling, which provides additional supervision signals to assist with generator learning. We explore both single-directional and bidirectional task-augmented GANs for domain adaptation. Then, we further improve the GAN by introducing a differentiable, stochastic data augmentation module to explicitly reduce discriminator overfitting. We examine source-, target-, and dual-domain data augmentation for GAN enhancement, as well as joint task and data augmentation in a unified GAN-based UDA framework. We evaluate the framework for cell detection on multiple public and in-house microscopy image datasets, which are acquired with different imaging modalities, staining protocols and/or tissue preparations. The experiments demonstrate that our method significantly boosts performance when compared with the reference baseline, and it is superior to or on par with fully supervised models that are trained with real target annotations. In addition, our method outperforms recent state-of-the-art UDA approaches by a large margin on different datasets.

Lung carcinoid tumours: histology and Ki-67, the eternal rivalry

WHO classification of Thoracic Tumours defines lung carcinoid tumours (LCTs) as well-differentiated neuroendocrine neoplasms (NENs) classified in low grade typical (TC) and intermediate grade atypical carcinoids (AC). Limited data exist concerning protein expression and morphologic factors able to predict disease aggressiveness. Though Ki-67 has proved to be a powerful diagnostic and prognostic factor for Gastro-entero-pancreatic NENs, its role in lung NENs is still debated. A retrospective series of 370 LCT from two oncology centers was centrally reviewed. Morphology and immunohistochemical markers (Ki-67, TTF-1, CD44, OTP, SSTR-2A, Ascl1, and p53) were studied and correlated with Overall Survival (OS), Cancer-specific survival (CSS) and Disease-free survival (DFS). Carcinoid histology was confirmed in 355 patients: 297 (83.7%) TC and 58 (16.3%) AC. Ki-67 at 3% was the best value in predicting DFS. Ki-67 ≥ 3% tumours were significantly associated with AC histology, stage III-IV, smoking, vascular invasion, tumour spread through air spaces OTP negativity, and TTF-1, Ascl1 and p53 positivity. After adjustment for center and period of diagnosis, both Ki-67 (≥3 versus <3) and histology (AC versus TC) alone significantly added prognostic information to OS and CSS multivariable model with age, stage and OTP; addition of both variables did not provide further prognostic information. Conversely, an improved significance of the DFS prediction model at multivariate analysis was seen by adding Ki-67 (≥3 versus <3, P adj = 0.01) to TC and AC histological distinction, age, lymph node involvement, residual tumour and OTP. Ki-67 ≥ 3% plays a potentially pivotal role in LCT prognosis, irrespective of histological grade.

Proposing Specific Neuronal Epithelial-to-Mesenchymal Transition Genes as an Ancillary Tool for Differential Diagnosis among Pulmonary Neuroendocrine Neoplasms

Pulmonary neuroendocrine neoplasms (PNENs) are currently classified into four major histotypes, including typical carcinoid (TC), atypical carcinoid (AC), large cell neuroendocrine carcinoma (LCNEC), and small cell lung carcinoma (SCLC). This classification was designed to be applied to surgical specimens mostly anchored in morphological parameters, resulting in considerable overlapping among PNENs, which may result in important challenges for clinicians' decisions in the case of small biopsies. Since PNENs originate from the neuroectodermic cells, epithelial-to-mesenchymal transition (EMT) gene expression shows promise as biomarkers involved in the genotypic transformation of neuroectodermic cells, including mutation burden with the involvement of chromatin remodeling genes, apoptosis, and mitosis rate, leading to modification in final cellular phenotype. In this situation, additional markers also applicable to biopsy specimens, which correlate PNENs subtypes with systemic treatment response, are much needed, and current potential candidates are neurogenic EMT genes. This study investigated EMT genes expression and its association with PNENs histotypes in tumor tissues from 24 patients with PNENs. PCR Array System for 84 EMT-related genes selected 15 differentially expressed genes among the PNENs, allowing to discriminate TC from AC, LCNEC from AC, and SCLC from AC. Functional enrichment analysis of the EMT genes differentially expressed among PNENs subtypes showed that they are involved in cellular proliferation, extracellular matrix degradation, regulation of cell apoptosis, oncogenesis, and tumor cell invasion. Interestingly, four EMT genes (MAP1B, SNAI2, MMP2, WNT5A) are also involved in neurological diseases, in brain metastasis, and interact with platinum-based chemotherapy and tyrosine-kinase inhibitors. Collectively, these findings emerge as an important ancillary tool to improve the strategies of histologic diagnosis in PNENs and unveil the four EMT genes that can play an important role in driving chemical response in PNENs.

Prognostic Factors across Poorly Differentiated Neuroendocrine Neoplasms: A Pooled Analysis

INTRODUCTION

Poorly differentiated neuroendocrine carcinomas (NECs) are characterized by aggressive clinical course and poor prognosis. No reliable prognostic markers have been validated to date; thus, the definition of a specific NEC prognostic algorithm represents a clinical need. This study aimed to analyze a large NEC case series to validate the specific prognostic factors identified in previous studies on gastro-entero-pancreatic and lung NECs and to assess if further prognostic parameters can be isolated.

METHODS

A pooled analysis of four NEC retrospective studies was performed to evaluate the prognostic role of Ki-67 cut-off, the overall survival (OS) according to primary cancer site, and further prognostic parameters using multivariable Cox proportional hazards model and machine learning random survival forest (RSF).

RESULTS

422 NECs were analyzed. The most represented tumor site was the colorectum (n = 156, 37%), followed by the lungs (n = 111, 26%), gastroesophageal site (n = 83, 20%; 66 gastric, 79%) and pancreas (n = 42, 10%). The Ki-67 index was the most relevant predictor, followed by morphology (pure or mixed/combined NECs), stage, and site. The predicted RSF response for survival at 1, 2, or 3 years showed decreasing survival with increasing Ki-67, pure NEC morphology, stage III-IV, and colorectal NEC disease. Patients with Ki-67 <55% and mixed/combined morphology had better survival than those with pure morphology. Morphology pure or mixed/combined became irrelevant in NEC survival when Ki-67 was ≥55%. The prognosis of metastatic patients who did not receive any treatment tended to be worse compared to that of the treated group. The prognostic impact of Rb1 immunolabeling appears to be limited when multiple risk factors are simultaneously assessed.

CONCLUSION

The most effective parameters to predict OS for NEC patients could be Ki-67, pure or mixed/combined morphology, stage, and site.

Integration of lncRNAs, Protein-Coding Genes and Pathology Images for Detecting Metastatic Melanoma

Melanoma is a lethal skin disease that develops from moles. This study aimed to integrate multimodal data to predict metastatic melanoma, which is highly aggressive and difficult to treat. The proposed EnsembleSKCM method evaluated the prediction performances of long noncoding RNAs (lncRNAs), protein-coding messenger genes (mRNAs) and pathology images (images) for metastatic melanoma. Feature selection was used to screen for metastatic biomarkers in the lncRNA and mRNA datasets. The integrated EnsembleSKCM model was built based on the weighted results of the lncRNA-, mRNA- and image-based models. EnsembleSKCM achieved 0.9444 in the prediction accuracy of metastatic melanoma and outperformed the single-modal prediction models based on the lncRNA, mRNA and image data. The experimental data suggest the importance of integrating the complementary information from the three data modalities. WGCNA was used to analyze the relationship of molecular-level features and image features, and the results show connections between them. Another cohort was used to validate our prediction.

Diagnostic criteria and evolving molecular characterization of pulmonary neuroendocrine carcinomas

Neuroendocrine carcinomas of the lung are currently classified into two categories: small cell lung carcinoma and large cell neuroendocrine carcinoma. Diagnostic criteria for small cell- and large cell neuroendocrine carcinoma are based solely on tumor morphology; however, overlap in histologic and immunophenotypic features between the two types of carcinoma can potentially make their classification challenging. Accurate diagnosis of pulmonary neuroendocrine carcinomas is paramount for patient management, as clinical course and treatment differ between small cell and large cell neuroendocrine carcinoma. Molecular-genetic, transcriptomic, and proteomic data published over the past decade suggest that small cell and large cell neuroendocrine carcinomas are not homogeneous categories but rather comprise multiple groups of distinctive malignancies. Nuances in the susceptibility of small cell lung carcinoma subtypes to different chemotherapeutic regimens and the discovery of targetable mutations in large cell neuroendocrine carcinoma suggest that classification and treatment of neuroendocrine carcinomas may be informed by ancillary molecular and protein expression testing going forward. This review summarizes current diagnostic criteria, prognostic and predictive correlates of classification, and evidence of previously unrecognized subtypes of small cell and large cell neuroendocrine carcinoma.

Risk factors, survival analysis, and nomograms for distant metastasis in patients with primary pulmonary large cell neuroendocrine carcinoma: A population-based study

INTRODUCTION

Pulmonary large cell neuroendocrine carcinoma (LCNEC) is a rapidly progressive and easily metastatic high-grade lung cancer, with a poor prognosis when distant metastasis (DM) occurs. The aim of our study was to explore risk factors associated with DM in LCNEC patients and to perform survival analysis and to develop a novel nomogram-based predictive model for screening risk populations in clinical practice.

METHODS

The study cohort was derived from the Surveillance, Epidemiology, and End Results database, from which we selected patients with LCNEC between 2004 to 2015 and formed a diagnostic cohort (n = 959) and a prognostic cohort (n = 272). The risk and prognostic factors of DM were screened by univariate and multivariate analyses using logistic and Cox regressions, respectively. Then, we established diagnostic and prognostic nomograms using the data in the training group and validated the accuracy of the nomograms in the validation group. The diagnostic nomogram was evaluated using receiver operating characteristic curves, decision curve analysis curves, and the GiViTI calibration belt. The prognostic nomogram was evaluated using receiver operating characteristic curves, the concordance index, the calibration curve, and decision curve analysis curves. In addition, high- and low-risk groups were classified according to the prognostic monogram formula, and Kaplan-Meier survival analysis was performed.

RESULTS

In the diagnostic cohort, LCNEC close to bronchus, with higher tumor size, and with higher N stage indicated higher likelihood of DM. In the prognostic cohort (patients with LCNEC and DM), men with higher N stage, no surgery, and no chemotherapy had poorer overall survival. Patients in the high-risk group had significantly lower median overall survival than the low-risk group.

CONCLUSION

Two novel established nomograms performed well in predicting DM in patients with LCNEC and in evaluating their prognosis. These nomograms could be used in clinical practice for screening of risk populations and treatment planning.

Bidirectional Mapping-Based Domain Adaptation for Nucleus Detection in Cross-Modality Microscopy Images

Cell or nucleus detection is a fundamental task in microscopy image analysis and has recently achieved state-of-the-art performance by using deep neural networks. However, training supervised deep models such as convolutional neural networks (CNNs) usually requires sufficient annotated image data, which is prohibitively expensive or unavailable in some applications. Additionally, when applying a CNN to new datasets, it is common to annotate individual cells/nuclei in those target datasets for model re-learning, leading to inefficient and low-throughput image analysis. To tackle these problems, we present a bidirectional, adversarial domain adaptation method for nucleus detection on cross-modality microscopy image data. Specifically, the method learns a deep regression model for individual nucleus detection with both source-to-target and target-to-source image translation. In addition, we explicitly extend this unsupervised domain adaptation method to a semi-supervised learning situation and further boost the nucleus detection performance. We evaluate the proposed method on three cross-modality microscopy image datasets, which cover a wide variety of microscopy imaging protocols or modalities, and obtain a significant improvement in nucleus detection compared to reference baseline approaches. In addition, our semi-supervised method is very competitive with recent fully supervised learning models trained with all real target training labels.

Clinical-Pathologic Challenges in the Classification of Pulmonary Neuroendocrine Neoplasms and Targets on the Horizon for Future Clinical Practice

Diagnosing a pulmonary neuroendocrine neoplasm (NEN) may be difficult, challenging clinical decision making. In this review, the following key clinical and pathologic issues and informative molecular markers are being discussed: (1) What is the preferred outcome parameter for curatively resected low-grade NENs (carcinoid), for example, overall survival or recurrence-free interval? (2) Does the WHO classification combined with a Ki-67 proliferation index and molecular markers, such as OTP and CD44, offer improved prognostication in low-grade NENs? (3) What is the value of a typical versus atypical carcinoid diagnosis on a biopsy specimen in local and metastatic disease? Diagnosis is difficult in biopsy specimens and recent observations of an increased mitotic rate in metastatic carcinoid from typical to atypical and high-grade NEN can further complicate diagnosis. (4) What is the (ir)relevance of morphologically separating large cell neuroendocrine carcinoma (LCNEC) SCLC and the value of molecular markers (RB1 gene and pRb protein or transcription factors NEUROD1, ASCL1, POU2F3, or YAP1 [NAPY]) to predict systemic treatment outcome? (5) Are additional diagnostic criteria required to accurately separate LCNEC from NSCLC in biopsy specimens? Neuroendocrine morphology can be absent owing to limited sample size leading to missed LCNEC diagnoses. Evaluation of genomic studies on LCNEC and marker studies have identified that a combination of napsin A and neuroendocrine markers could be helpful. Hence, to improve clinical practice, we should consider to adjust our NEN classification incorporating prognostic and predictive markers applicable on biopsy specimens to inform a treatment outcome-driven classification.

Challenging, Accurate and Feasible: CAF-1 as a Tumour Proliferation Marker of Diagnostic and Prognostic Value

Simple Summary

There is an emerging need for new weapons in the battle against cancer; therefore, the discovery of new biomarkers with diagnostic, prognostic, and therapeutic value is a priority of current cancer research. An important task is to identify how quickly a tumour proliferates. A tumour’s proliferation rate is critical for grading and clinical decision-making; hence, there is an imperative need for accurate proliferation markers. Here, we review evidence demonstrating that chromatin assembly factor 1 (CAF-1) is a proliferation marker of clinical value. CAF-1 is selectively expressed in proliferating cells and its expression can be evaluated by immunohistochemistry in cytology smears and biopsies. CAF-1 expression is increased in almost all cancers and correlates strongly with the expression of Ki-67, the current routine proliferation marker. Overexpression of CAF-1 is associated with poor clinical outcome (advanced cancer stage, recurrence, metastasis, and decreased survival). CAF-1 is a robust, reproducible, and feasible proliferation marker of prognostic importance and may represent an attractive alternative or complementary to Ki-67 for cancer stratification and clinical guidance.

Abstract

The discovery of novel biomarkers of diagnostic, prognostic, and therapeutic value is a major challenge of current cancer research. The assessment of tumour cell proliferative capacity is pivotal for grading and clinical decision-making, highlighting the importance of proliferation markers as diagnostic and prognostic tools. Currently, the immunohistochemical analysis of Ki-67 expression levels is routinely used in clinical settings to assess tumour proliferation. Inasmuch as the function of Ki-67 is not fully understood and its evaluation lacks standardization, there is interest in chromatin regulator proteins as alternative proliferation markers of clinical value. Here, we review recent evidence demonstrating that chromatin assembly factor 1 (CAF-1), a histone chaperone selectively expressed in cycling cells, is a proliferation marker of clinical value. CAF-1 expression, when evaluated by immunocytochemistry in breast cancer cytology smears and immunohistochemistry in cancer biopsies from several tissues, strongly correlates with the expression of Ki-67 and other proliferation markers. Notably, CAF-1 expression is upregulated in almost all cancers, and CAF-1 overexpression is significantly associated, in most cancer types, with high histological tumour grade, advanced stage, recurrence, metastasis, and decreased patient survival. These findings suggest that CAF-1 is a robust, reproducible, and feasible proliferation marker of prognostic importance. CAF-1 may represent an attractive alternative or complementary to Ki-67 for cancer stratification and clinical guidance.

Molecular Pathology of Pulmonary Large Cell Neuroendocrine Carcinoma: Novel Concepts and Treatments

Pulmonary large cell neuroendocrine carcinoma (LCNEC) is an aggressive neoplasm with poor prognosis. Histologic diagnosis of LCNEC is not always straightforward. In particular, it is challenging to distinguish small cell lung carcinoma (SCLC) or poorly differentiated carcinoma from LCNEC. However, histological classification for LCNEC as well as their therapeutic management has not changed much for decades. Recently, genomic and transcriptomic analyses have revealed different molecular subtypes raising hopes for more personalized treatment. Two main molecular subtypes of LCNEC have been identified by studies using next generation sequencing, namely type I with TP53 and STK11/KEAP1 alterations, alternatively called as non-SCLC type, and type II with TP53 and RB1 alterations, alternatively called as SCLC type. However, there is still no easy way to classify LCNEC subtypes at the actual clinical level. In this review, we have discussed histological diagnosis along with the genomic studies and molecular-based treatment for LCNEC.