VSGD-Net: Virtual Staining Guided Melanocyte Detection on Histopathological Images

Detection of melanocytes serves as a critical prerequisite in assessing melanocytic growth patterns when diagnosing melanoma and its precursor lesions on skin biopsy specimens. However, this detection is challenging due to the visual similarity of melanocytes to other cells in routine Hematoxylin and Eosin (H&E) stained images, leading to the failure of current nuclei detection methods. Stains such as Sox10 can mark melanocytes, but they require an additional step and expense and thus are not regularly used in clinical practice. To address these limitations, we introduce VSGD-Net, a novel detection network that learns melanocyte identification through virtual staining from H&E to Sox10. The method takes only routine H&E images during inference, resulting in a promising approach to support pathologists in the diagnosis of melanoma. To the best of our knowledge, this is the first study that investigates the detection problem using image synthesis features between two distinct pathology stainings. Extensive experimental results show that our proposed model outperforms state-of-the-art nuclei detection methods for melanocyte detection. The source code and pre-trained model are available at: https://github.com/kechunl/VSGD-Net.

Segmenting Skin Biopsy Images with Coarse and Sparse Annotations using U-Net

The number of melanoma diagnoses has increased dramatically over the past three decades, outpacing almost all other cancers. Nearly 1 in 4 skin biopsies is of melanocytic lesions, highlighting the clinical and public health importance of correct diagnosis. Deep learning image analysis methods may improve and complement current diagnostic and prognostic capabilities. The histologic evaluation of melanocytic lesions, including melanoma and its precursors, involves determining whether the melanocytic population involves the epidermis, dermis, or both. Semantic segmentation of clinically important structures in skin biopsies is a crucial step towards an accurate diagnosis. While training a segmentation model requires ground-truth labels, annotation of large images is a labor-intensive task. This issue becomes especially pronounced in a medical image dataset in which expert annotation is the gold standard. In this paper, we propose a two-stage segmentation pipeline using coarse and sparse annotations on a small region of the whole slide image as the training set. Segmentation results on whole slide images show promising performance for the proposed pipeline.

Machine learning techniques for mitoses classification

BACKGROUND

Pathologists analyze biopsy material at both the cellular and structural level to determine diagnosis and cancer stage. Mitotic figures are surrogate biomarkers of cellular proliferation that can provide prognostic information; thus, their precise detection is an important factor for clinical care. Convolutional Neural Networks (CNNs) have shown remarkable performance on several recognition tasks. Utilizing CNNs for mitosis classification may aid pathologists to improve the detection accuracy.

METHODS

We studied two state-of-the-art CNN-based models, ESPNet and DenseNet, for mitosis classification on six whole slide images of skin biopsies and compared their quantitative performance in terms of sensitivity, specificity, and F-score. We used raw RGB images of mitosis and non-mitosis samples with their corresponding labels as training input. In order to compare with other work, we studied the performance of these classifiers and two other architectures, ResNet and ShuffleNet, on the publicly available MITOS breast biopsy dataset and compared the performance of all four in terms of precision, recall, and F-score (which are standard for this data set), architecture, training time and inference time.

RESULTS

The ESPNet and DenseNet results on our primary melanoma dataset had a sensitivity of 0.976 and 0.968, and a specificity of 0.987 and 0.995, respectively, with F-scores of .968 and .976, respectively. On the MITOS dataset, ESPNet and DenseNet showed a sensitivity of 0.866 and 0.916, and a specificity of 0.973 and 0.980, respectively. The MITOS results using DenseNet had a precision of 0.939, recall of 0.916, and F-score of 0.927. The best published result on MITOS (Saha et al. 2018) reported precision of 0.92, recall of 0.88, and F-score of 0.90. In our architecture comparisons on MITOS, we found that DenseNet beats the others in terms of F-Score (DenseNet 0.927, ESPNet 0.890, ResNet 0.865, ShuffleNet 0.847) and especially Recall (DenseNet 0.916, ESPNet 0.866, ResNet 0.807, ShuffleNet 0.753), while ResNet and ESPNet have much faster inference times (ResNet 6 s, ESPNet 8 s, DenseNet 31 s). ResNet is faster than ESPNet, but ESPNet has a higher F-Score and Recall than ResNet, making it a good compromise solution.

CONCLUSION

We studied several state-of-the-art CNNs for detecting mitotic figures in whole slide biopsy images. We evaluated two CNNs on a melanoma cancer dataset and then compared four CNNs on a public breast cancer data set, using the same methodology on both. Our methodology and architecture for mitosis finding in both melanoma and breast cancer whole slide images has been thoroughly tested and is likely to be useful for finding mitoses in any whole slide biopsy images.

Classifying Breast Histopathology Images with a Ductal Instance-Oriented Pipeline

In this study, we propose the Ductal Instance-Oriented Pipeline (DIOP) that contains a duct-level instance segmentation model, a tissue-level semantic segmentation model, and three-levels of features for diagnostic classification. Based on recent advancements in instance segmentation and the Mask RCNN model, our duct-level segmenter tries to identify each ductal individual inside a microscopic image; then, it extracts tissue-level information from the identified ductal instances. Leveraging three levels of information obtained from these ductal instances and also the histopathology image, the proposed DIOP outperforms previous approaches (both feature-based and CNN-based) in all diagnostic tasks; for the four-way classification task, the DIOP achieves comparable performance to general pathologists in this unique dataset. The proposed DIOP only takes a few seconds to run in the inference time, which could be used interactively on most modern computers. More clinical explorations are needed to study the robustness and generalizability of this system in the future.

Scale-Aware Transformers for Diagnosing Melanocytic Lesions

Diagnosing melanocytic lesions is one of the most challenging areas of pathology with extensive intra- and inter-observer variability. The gold standard for a diagnosis of invasive melanoma is the examination of histopathological whole slide skin biopsy images by an experienced dermatopathologist. Digitized whole slide images offer novel opportunities for computer programs to improve the diagnostic performance of pathologists. In order to automatically classify such images, representations that reflect the content and context of the input images are needed. In this paper, we introduce a novel self-attention-based network to learn representations from digital whole slide images of melanocytic skin lesions at multiple scales. Our model softly weighs representations from multiple scales, allowing it to discriminate between diagnosis-relevant and -irrelevant information automatically. Our experiments show that our method outperforms five other state-of-the-art whole slide image classification methods by a significant margin. Our method also achieves comparable performance to 187 practicing U.S. pathologists who interpreted the same cases in an independent study. To facilitate relevant research, full training and inference code is made publicly available at https://github.com/meredith-wenjunwu/ScATNet.

Assessment of Second-Opinion Strategies for Diagnoses of Cutaneous Melanocytic Lesions

Importance

Histopathologic criteria have limited diagnostic reliability for a range of cutaneous melanocytic lesions.

Objective

To evaluate the association of second-opinion strategies by general pathologists and dermatopathologists with the overall reliability of diagnosis of difficult melanocytic lesions.

Design, Setting, and Participants

This diagnostic study used samples from the Melanoma Pathology Study, which comprises 240 melanocytic lesion samples selected from a dermatopathology laboratory in Bellevue, Washington, and represents the full spectrum of lesions from common nevi to invasive melanoma. Five sets of 48 samples were evaluated independently by 187 US pathologists from July 15, 2013, through May 23, 2016. Data analysis was performed from April 2016 through November 2017.

Main Outcomes and Measures

Accuracy of diagnosis, defined as concordance with an expert consensus diagnosis of 3 experienced pathologists, was assessed after applying 10 different second-opinion strategies.

Results

Among the 187 US pathologists examining the 24 lesion samples, 113 were general pathologists (65 men [57.5%]; mean age at survey, 53.7 years [range, 33.0-79.0 years]) and 74 were dermatopathologists (49 men [66.2%]; mean age at survey, 46.4 years [range, 33.0-77.0 years]). Among the 8976 initial case interpretations, physicians desired second opinions for 3899 (43.4%), most often for interpretation of severely dysplastic nevi. The overall misclassification rate was highest when interpretations did not include second opinions and initial reviewers were all general pathologists lacking subspecialty training (52.8%; 95% CI, 51.3%-54.3%). When considering different second opinion strategies, the misclassification of melanocytic lesions was lowest when the first, second, and third consulting reviewers were subspecialty-trained dermatopathologists and when all lesions were subject to second opinions (36.7%; 95% CI, 33.1%-40.7%). When the second opinion strategies were compared with single interpretations without second opinions, the reductions in misclassification rates for some of the strategies were statistically significant, but none of the strategies eliminated diagnostic misclassification. Melanocytic lesions in the middle of the diagnostic spectrum had the highest misclassification rates (eg, moderately or severely dysplastic nevus, Spitz nevus, melanoma in situ, and pathologic stage [p]T1a invasive melanoma). Variability of in situ and thin invasive melanoma was relatively intractable to all examined strategies.

Conclusions and Relevance

The results of this study suggest that second opinions rendered by dermatopathologists improve reliability of melanocytic lesion diagnosis. However, discordance among pathologists remained high.

Influence of variability in assessment of Breslow thickness, mitotic rate and ulceration among US pathologists interpreting invasive melanoma, for the purpose of AJCC staging

BACKGROUND

Melanoma staging has depended on depth of invasion (Breslow thickness, BT), mitotic rate (MR) and ulceration. In anticipation of the AJCC's eighth edition, variability in pathologists' assessment of these factors and consequently in tumor staging was assessed.

METHODS

One-hundred and fifteen cases of invasive melanoma, established by a consensus panel, were assessed by 187 pathologists. Variation was studied in BT, the detection of mitotic figures, and ulceration. The sources of this variation and its effect on tumor staging are considered.

RESULTS

On average, participant assessments closely approached consensus BT. Greater variation was identified in the classification of mitogenicity, which (like ulceration) upstages a T1 melanoma from T1a to T1b in the seventh but not eighth edition. In cases with a T1a diagnosis by the consensus panel, 15.6% of participants identified one or more mitotic figures (indicative of a false positive); and in cases diagnosed asT1b by the consensus panel, 32.0% of participants failed to find mitotic figures (false negative).

CONCLUSION

Variability in the staging of T1 melanoma among pathologists when using the AJCC seventh edition criteria is closely related to the detection of mitotic figures, with BT playing a less prominent role. Decreased variability is expected after implementation of the eighth edition.

Achieving consensus for the histopathologic diagnosis of melanocytic lesions: use of the modified Delphi method

OBJECTIVE

To understand the sophisticated nature of coming to consensus when diagnosing complex melanocytic lesions among a panel of experienced dermatopathologists.

METHODS

A total of 240 melanocytic lesions were assessed independently by three experienced dermatopathologists with their diagnoses mapped into one of five Melanocytic Pathology Assessment Tool and Hierarchy for Diagnosis (MPATH-DX) categories: (I) nevus/mild atypia, (II) moderate atypia, (III) severe atypia/melanoma in situ, (IV) T1a invasive melanoma and (V) ≥ T1b invasive melanoma. The dermatopathologists then discussed the cases, using a modified Delphi method to facilitated consensus building for cases with discordant diagnoses.

RESULTS

For most cases, a majority of interpretations (two or three of three) agreed with the consensus diagnosis in 95% of Category I, 64% of Category II, 84% of Category III, 88% for Category IV and 100% of Category V cases. Disagreements were typically due to diagnostic threshold differences (64.5%), differing contents on slides even though the slides were sequential cuts (18.5%), and missed findings (15.3%). Disagreements were resolved via discussion of histopathologic features and their significance while reviewing the slides using a multi-headed microscope, considering treatment recommendations, citing existing literature, reviewing additional slides for a case, and choosing a provisional/borderline diagnosis to capture diverse opinions. All experienced pathologists participating in this study reported that the process of coming to consensus was challenging for borderline cases and may have represented compromise rather than consensus. They also reported the process changed their approaches to diagnosing complex melanocytic lesions.

CONCLUSIONS

The most frequent reason for disagreement of experienced dermatopathologists was differences in diagnostic thresholds related to observer viewpoints. A range of approaches was needed to come to consensus, and this may guide pathology groups who do not currently hold consensus conferences.

The self-reported use of immunostains and cytogenetic testing in the diagnosis of melanoma by practicing U.S. pathologists of 10 selected states

BACKGROUNDS

The diagnosis of melanoma can be challenging, especially in lesions for which the histopathologic criteria bridge two or more taxonomic categories. Newer genomic analytical methods of fluorescence in situ hybridization (FISH) and comparative genomic hybridization (CGH) have been introduced as ancillary techniques to differentiate benign and malignant melanocytic proliferations.

METHODS

We evaluated how pathologists perceive and are incorporating these new cytogenetic testing technologies into their practices. We conducted a study of 207 U.S. pathologists who interpret melanocytic lesions in clinical practice in 10 SEER states. Pathologists were surveyed regarding perceptions and utilization of FISH and/or CGH in their clinical practices.

RESULTS

Results showed that 38% of pathologists use FISH and/or CGH in interpreting melanocytic lesions. Pathologists reporting FISH and/or CGH use were significantly younger (p < 0.05), were fellowship trained or board certified in dermatopathology (p < 0.001) and were affiliated with an academic institute (p < 0.001). Pathologists reporting that their colleagues consider them an expert in the assessment of melanocytic lesions were more likely to employ FISH and/or CGH in their practices than non-experts.

CONCLUSIONS

Early users of cytogenetic testing technologies in cutaneous pathology are more likely to be younger, affiliated with an academic institution and fellowship trained or board certified in dermatopathology.

How concerns and experiences with medical malpractice affect dermatopathologists' perceptions of their diagnostic practices when interpreting cutaneous melanocytic lesions

OBJECTIVE

We sought to identify characteristics associated with past malpractice lawsuits and how malpractice concerns may affect interpretive practices.

METHODS

We surveyed 207 of 301 (68.8%) eligible dermatopathologists who interpret melanocytic skin lesions in 10 states. The survey assessed dermatopathologists' demographic and clinical practice characteristics, perceptions of how medical malpractice concerns could influence their interpretive practices, and past malpractice lawsuits.

RESULTS

Of dermatopathologists, 33% reported past malpractice experiences. Factors associated with being sued included older age (57 vs 48 years, P < .001), lack of board certification or fellowship training in dermatopathology (76.5% vs 53.2%, P = .001), and greater number of years interpreting melanocytic lesions (>20 years: 52.9% vs 20.1%, P < .001). Of participants, 64% reported being moderately or extremely confident in their melanocytic interpretations. Although most dermatopathologists believed that malpractice concerns increased their likelihood of ordering specialized pathology tests, obtaining recuts, and seeking a second opinion, none of these practices were associated with past malpractice. Most dermatopathologists reported concerns about potential harms to patients that may result from their assessments of melanocytic lesions.

LIMITATIONS

Limitations of this study include lack of validation of and details about the malpractice suits experienced by participating dermatopathologists. In addition, the study assessed perceptions of practice rather than actual practices that might be associated with malpractice incidents.

CONCLUSIONS

Most dermatopathologists reported apprehension about how malpractice affects their clinical practice and are concerned about patient safety irrespective of whether they had actually experienced a medical malpractice suit.

Teledermatology in the diagnosis of melanoma

Summary

We conducted a retrospective chart review of US Veterans in the Pacific Northwest area to compute melanoma incidence and Breslow depth at diagnosis. We compared Veterans with access to teledermatology (TD) and those without (non-TD). We identified pathology-confirmed primary melanomas in Veterans who had had at least one encounter at a VA facility during a 3-year study period. The age-adjusted melanoma incidence for all, TD and non-TD Veterans was 36, 15 and 57 per 100,000, respectively. The mean Breslow depth was significantly greater in the TD group ( P = 0.03). Although a higher proportion of thin (Breslow depth ≤1 mm) TD melanomas were mitotically active, this difference was not significant. We also found that 180 (40%) of the non-TD (face-to-face) diagnosed melanomas were from Veterans living in areas where TD was available. This suggests that the higher melanoma incidence in the non-TD group was mainly due to under-utilization of TD services. The study demonstrated that the TD service was not fully utilized in the VISN20 region, although the reasons for this are not clear. Where TD was utilized it tended to diagnose more advanced melanomas with worse initial prognosis.

Loss of the p53/p63 regulated desmosomal protein Perp promotes tumorigenesis

Dysregulated cell–cell adhesion plays a critical role in epithelial cancer development. Studies of human and mouse cancers have indicated that loss of adhesion complexes known as adherens junctions contributes to tumor progression and metastasis. In contrast, little is known regarding the role of the related cell–cell adhesion junction, the desmosome, during cancer development. Studies analyzing expression of desmosome components during human cancer progression have yielded conflicting results, and therefore genetic studies using knockout mice to examine the functional consequence of desmosome inactivation for tumorigenesis are essential for elucidating the role of desmosomes in cancer development. Here, we investigate the consequences of desmosome loss for carcinogenesis by analyzing conditional knockout mice lacking Perp, a p53/p63 regulated gene that encodes an important component of desmosomes. Analysis of Perp-deficient mice in a UVB-induced squamous cell skin carcinoma model reveals that Perp ablation promotes both tumor initiation and progression. Tumor development is associated with inactivation of both of Perp's known functions, in apoptosis and cell–cell adhesion. Interestingly, Perp-deficient tumors exhibit widespread downregulation of desmosomal constituents while adherens junctions remain intact, suggesting that desmosome loss is a specific event important for tumorigenesis rather than a reflection of a general change in differentiation status. Similarly, human squamous cell carcinomas display loss of PERP expression with retention of adherens junctions components, indicating that this is a relevant stage of human cancer development. Using gene expression profiling, we show further that Perp loss induces a set of inflammation-related genes that could stimulate tumorigenesis. Together, these studies suggest that Perp-deficiency promotes cancer by enhancing cell survival, desmosome loss, and inflammation, and they highlight a fundamental role for Perp and desmosomes in tumor suppression. An understanding of the factors affecting cancer progression is important for ultimately improving the diagnosis, prognostication, and treatment of cancer.

Changes in tissue architecture, such as loss of adhesion between cells, have been shown to facilitate cancer development, especially metastasis where cells can detach from a tumor and spread throughout the body. While various studies have demonstrated that inactivation of an adhesion complex known as the adherens junction promotes cancer development and metastasis, little is known about the role of the desmosome—a related cell–cell adhesion complex—in tumorigenesis. Here we examine the consequence of desmosome-deficiency for tumor development by studying mice lacking a key component of desmosomes in the skin, a protein known as Perp. Using a mouse model for human skin cancer, in which ultraviolet light promotes skin cancer development, we demonstrate that Perp-deficiency indeed leads to accelerated skin tumorigenesis. We similarly observe that PERP is lost during human skin cancer development, suggesting that PERP is also important as a tumor suppressor in humans. These findings demonstrate that desmosome-deficiency achieved by Perp inactivation can promote cancer and suggest the potential utility of monitoring PERP status for staging, prognostication, or treatment of human cancers.

Functional health status of adolescents after the Fontan procedure -- comparison with their siblings

BACKGROUND

Studies have suggested that patients who have undergone the Fontan procedure experience lower functional health status and diminished exercise capacity compared with other children.

OBJECTIVES

To compare the functional health status of Fontan patients with and without siblings, assess whether there are any differences between Fontan patients and their siblings, and determine associated factors.

METHODS

A cross-sectional, single-centre, observational study was performed on Fontan patients 10 to 20 years of age, and their sibling closest in age, followed in a tertiary pediatric hospital. Functional health status was measured by the Child Health Questionnaire Child Form and the Pediatric Quality of Life Inventory.

RESULTS

A total of 68 patients and 38 siblings were enrolled. Patients with siblings scored significantly lower on numerous domains of physical functional status than those without siblings. Compared with their matched siblings, Fontan patients reported significantly lower scores in all domains of the Pediatric Quality of Life Inventory and on physical (but not psychosocial) domains of the Child Health Questionnaire Child Form. Factors associated with increased patient-sibling differences included younger patient age, female sex, intracardiac lateral tunnel Fontan connection and lower ejection fraction at the time of study enrollment.

CONCLUSIONS

Adolescents with Fontan physiology reported a lower functional health status in physical domains than their siblings, but had similar status in psychosocial domains. Having a sibling was associated with lower reported functional health status, suggesting an important effect of self-perceived physical limitations over true limitations.

DNA sequence of the translocation breakpoints in undifferentiated embryonal sarcoma arising in mesenchymal hamartoma of the liver harboring the t(11;19)(q11;q13.4) translocation

Undifferentiated embryonal sarcoma of the liver is a highly malignant and aggressive tumor that occasionally arises within mesenchymal hamartoma of the liver (MHL), a benign tumor that typically occurs in young children. Undifferentiated embryonal sarcoma arising in MHL, as well as uncomplicated MHL, frequently harbor rearrangements of band 19q13.4, including the translocation t(11;19)(q13;q13.4). In this study we report the cloning and DNA sequence analysis of the translocation breakpoints in an undifferentiated embryonal sarcoma arising in MHL known to harbor t(11;19). In this case, the breakpoint at 11q13 occurred in the MALAT1 gene, also known as ALPHA. MALAT1 is rearranged in renal tumors harboring the t(6;11)(p21;q13) translocation, and noncoding MALAT1 transcripts are overexpressed in a number of human carcinomas. The breakpoint at 19q13.4 occurs at a locus we refer to as MHLB1, for Mesenchymal Hamartoma of the Liver Breakpoint 1. Although the MHLB1 locus does not contain a known gene, several human ESTs map to the region (a subset of which show homology to the nuclear RNA export factor (NXF) gene family), and the region is conserved between many mammalian species.

Concurrent translocation of BCL2 and MYC with a single immunoglobulin locus in high-grade B-cell lymphomas

B-cell leukaemia or lymphoma with a combination of t(8;14)(q24;q32) of Burkitt leukaemia/lymphoma and t(14;18)(q32;q21) of follicular lymphoma may present clinically as de novo acute lymphoblastic leukaemia or transformation of follicular lymphoma to aggressive histology diffuse lymphoma. A number of cell lines have been reported with a complex t(8;14;18) with fusion of MYC, IGH and BCL2 on the same derivative 8 chromosome. The objective of this study was to determine the frequency and chromosomal features of this der(8)t(8;14;18) in a series of acute leukaemias and malignant lymphomas. A database of 1350 leukaemia and lymphoma karyotypes was searched for cases with structural alterations affecting both 8q24 and 18q21. A total of 55 cases were identified, of which eight revealed a complex der(8)t(8;14;18) with an MYC-IGH-BCL2 rearrangement resulting from translocation of BCL2 and MYC with a single disrupted IGH allele. Molecular cytogenetic investigation is essential to identify cases of high-grade leukaemia/lymphoma with concurrent translocations affecting the BCL2 and MYC loci.

N-methyl-D-aspartate-induced nitric oxide release: an in vivo microdialysis study

Increasing evidence indicates that nitric oxide acts as an intercellular signal transduction molecule in the nervous system. In particular, in vitro studies have demonstrated that nitric oxide is produced in the cerebellar cortex and is responsible for the increases in cyclic GMP seen in response to glutamate receptor activation. In this study, we have combined the technique of intracerebellar microdialysis with a sensitive assay for nitric oxide oxidation products nitrate and nitrite, to assess nitric oxide release directly in awake, freely moving animals. We have found that infusion of N-methyl-D-aspartate via the microdialysis probe results in a dose-dependent increase in cerebellar nitric oxide release. This increase was prevented by prior administration of an N-methyl-D-aspartate receptor antagonist, or the nitric oxide synthase inhibitor NG-nitroarginine. Both these pretreatments also reduced the basal extracellular nitrite and nitrate levels, suggesting that there is a tonic glutamate-induced nitric oxide production in the cerebellum of awake, freely moving animals. These results provide direct evidence for nitric oxide release in response to N-methyl-D-aspartate receptor activation in the adult cerebellar cortex, in vivo. This new approach, coupling microdialysis with the azo dye detection method of Griess, should thus prove useful for the in vivo study of nitric oxide release from various brain regions in response to pharmacological, physiological or behavioral manipulations.

Streptococcal toxic shocklike syndrome leading to bilateral lower extremity compartment syndrome and renal failure. Report of a case

Compartment syndrome is considered a true emergency in orthopedic practice. To reduce morbidity and mortality from this condition, prompt diagnosis and appropriate treatment are absolutely essential. An unusual bilateral lower extremity compartment syndrome leading to renal failure and crush syndrome occurred in a 13-year-old girl with Streptococcal toxic shocklike syndrome. This situation seems not to have been previously reported. Early diagnosis and expeditious treatment produced minimal sequelae of the condition.

Symptomatic scapholunatotriquetral carpal coalition with fusion of the capitatometacarpal joint. Report of a case

Carpal coalition is not an uncommon entity. It is most often asymptomatic and coincidentally noted on roentgenograms. The most common location is between the lunate and the triquetrum, although nearly every combination has been reported. Reported is a 33-year-old man with a symptomatic scapholunatotriquetral carpal coalition associated with capitatometacarpal coalition. This condition seems not to have been previously reported in the literature.