Introduction: A whole slide image (WSI) is typically comprised of a macro image (low-power snapshot of the entire glass slide) and stacked tiles in a pyramid structure (with the lowest resolution thumbnail at the top). The macro image shows the label and all pieces of tissue on the slide. Many whole slide scanner vendors do not readily show the macro overview to pathologists. We demonstrate that failure to do so may result in a serious misdiagnosis. Materials and Methods: Various examples of errors were accumulated that occurred during the digitization of glass slides where the virtual slide differed from the macro image of the original glass slide. Such examples were retrieved from pathology laboratories using different types of scanners in the USA, Canada, Europe, and Asia. Results: The reasons for image errors were categorized into technical problems (e.g., automatic tissue finder failure, image mismatches, and poor scan coverage) and human operator mistakes (e.g., improper manual region of interest selection). These errors were all detected because they were highlighted in the macro image. Conclusion: Our experience indicates that WSI can be subject to inadvertent errors related to glitches in scanning slides, corrupt images, or mistakes made by humans when scanning slides. Displaying the macro image that accompanies WSIs is critical from a quality control perspective in digital pathology practice as this can help detect these serious image-related problems and avoid compromised diagnoses.

Background: In an effort to provide improved user experience and system reliability at a moderate cost, our department embarked on targeted upgrades of a total of 87 computers over a period of 3 years. Upgrades came in three forms: (i) replacement of the computer with newer architecture, (ii) replacement of the computer's hard drive with a solid-state drive (SSD), or (iii) replacement of the computer with newer architecture and a SSD. Methods: We measured the impact of each form of upgrade on a set of pathology-relevant tasks that fell into three categories: standard use, whole-slide navigation, and whole-slide analysis. We used time to completion of a task as the primary variable of interest. Results: We found that for most tasks, the SSD upgrade had a greater impact than the upgrade in architecture. This effect was especially prominent for whole-slide viewing, likely due to the way in which most whole-slide viewers cached image tiles. However, other tasks, such as whole-slide image analysis, often relied less on disk input or output and were instead more sensitive to the computer architecture. Conclusions: Based on our experience, we suggest that SSD upgrades are viewed in some settings as a viable alternative to complete computer replacement and recommend that computer replacements in a digital pathology setting are accompanied by an upgrade to SSDs.

Introduction: Fine-needle aspiration cytology (FNAC) for identification of papillary carcinoma thyroid is a moderately sensitive and specific modality. The present machine learning tools can correctly classify images into broad categories. Training software for recognition of papillary thyroid carcinoma on FNAC smears will be a decisive step toward automation of cytopathology. Aim: The aim of this study is to develop an artificial neural network (ANN) for the purpose of distinguishing papillary carcinoma thyroid and nonpapillary carcinoma thyroid on microphotographs from thyroid FNAC smears. Subjects and Methods: An ANN was developed in the Python programming language. In the training phase, 186 microphotographs from Romanowsky/Pap-stained smears of papillary carcinoma and 184 microphotographs from smears of other thyroid lesions (at ×10 and ×40 magnification) were used for training the ANN. After completion of training, performance was evaluated with a set of 174 microphotographs (66 – nonpapillary carcinoma and 21 – papillary carcinoma, each photographed at two magnifications ×10 and ×40). Results: The performance characteristics and limitations of the neural network were assessed, assuming FNAC diagnosis as gold standard. Combined results from two magnifications showed good sensitivity (90.48%), moderate specificity (83.33%), and a very high negative predictive value (96.49%) and 85.06% diagnostic accuracy. However, vague papillary formations by benign follicular cells identified wrongly as papillary carcinoma remain a drawback. Conclusion: With further training with a diverse dataset and in conjunction with automated microscopy, the ANN has the potential to develop into an accurate image classifier for thyroid FNACs.

Background: Telepathology can potentially be utilized as an alternative to having on-site pathology services for rural and regional hospitals. The goal of the study was to validate two small-footprint desktop telepathology systems for remote parathyroid frozen sections. Subjects and Methods: Three pathologists retrospectively diagnosed 76 parathyroidectomy frozen sections of 52 patients from three pathology services in Australia using the “live-view mode” of MikroScan D2 and Aperio LV1 and in-house direct microscopy. The final paraffin section diagnosis served as the “gold standard” for accuracy evaluation. Concordance rates of the telepathology systems with direct microscopy, inter-pathologist and intra-pathologist agreement, and the time taken to report each slide were analyzed. Results: Both telepathology systems showed high diagnostic accuracy (>99%) and high concordance (>99%) with direct microscopy. High inter-pathologist agreement for telepathology systems was demonstrated by overall kappa values of 0.92 for Aperio LV1 and 0.85 for MikroScan D2. High kappa values (from 0.85 to 1) for intra-pathologist agreement within the three systems were also observed. The time taken per slide by Aperio LV1 and MicroScan D2 within three pathologists was about 3.0 times (P < 0.001, 95% confidence interval [CI]: 2.8–3.2) and 7.7 times (P < 0.001, 95% CI: 7.1–8.3) as long as direct microscopy, respectively, while MikroScan D2 took about 2.6 times as long as Aperio LV1 (P < 0.001, 95% CI: 2.4–2.7). All pathologists evaluated Aperio LV1 as being more user-friendly. Conclusions: Telepathology diagnosis of parathyroidectomy frozen sections through small-footprint desktop systems is accurate, reliable, and comparable with in-house direct microscopy. Telepathology systems take longer than direct microscopy; however, the time taken is within clinically acceptable limits. Aperio LV1 takes shorter time than MikroScan D2 and is more user-friendly.

Background: Digital Imaging and Communications in Medicine (DICOM^®) is the standard for the representation, storage, and communication of medical images and related information. A DICOM file format and communication protocol for pathology have been defined; however, adoption by vendors and in the field is pending. Here, we implemented the essential aspects of the standard and assessed its capabilities and limitations in a multisite, multivendor healthcare network. Methods: We selected relevant DICOM attributes, developed a program that extracts pixel data and pixel-related metadata, integrated patient and specimen-related metadata, populated and encoded DICOM attributes, and stored DICOM files. We generated the files using image data from four vendor-specific image file formats and clinical metadata from two departments with different laboratory information systems. We validated the generated DICOM files using recognized DICOM validation tools and measured encoding, storage, and access efficiency for three image compression methods. Finally, we evaluated storing, querying, and retrieving data over the web using existing DICOM archive software. Results: Whole slide image data can be encoded together with relevant patient and specimen-related metadata as DICOM objects. These objects can be accessed efficiently from files or through RESTful web services using existing software implementations. Performance measurements show that the choice of image compression method has a major impact on data access efficiency. For lossy compression, JPEG achieves the fastest compression/decompression rates. For lossless compression, JPEG-LS significantly outperforms JPEG 2000 with respect to data encoding and decoding speed. Conclusion: Implementation of DICOM allows efficient access to image data as well as associated metadata. By leveraging a wealth of existing infrastructure solutions, the use of DICOM facilitates enterprise integration and data exchange for digital pathology.

Background: Validation studies in digital pathology addressed so far diverse aspects of the routine work. We aimed to establish a complete remote digital pathology service. Methods: Altogether 2295 routine cases (8640 slides) were reported in our studies on digital versus microscopic diagnostics, remote reporting, diagnostic time, fine-needle aspiration cytology (FNAC) clinics, frozen sections, and diagnostic sessions with residents. The same senior pathologist was involved in all studies. Slides were scanned by ScanScope AT Turbo (Aperio). Digital images were accessed through the laboratory system (LS) on either 14” laptops or desktop computers with double 23” displays for the remote and on-site digital reporting. Larger displays were used when available for remote reporting. First diagnosis was either microscopic, digital, or remote digital only (6 months washout period). Both diagnoses were recorded separately and compared. Turnaround was measured from the registration to sign off or scanning to diagnosis. A diagnostic time was measured from the point slides were made available to the point of diagnosis or additional investigations were necessary, recorded independently in minutes/session, and compared. Jabber Video (Cisco) and Lync (Microsoft) were interchangeably used for the secure, video supervision of activities. Mobile phone, broadband, broadband over Wi-Fi, and mobile broadband were tested for internet connections. Nine autopsies were performed remotely involving three staff pathologists, one autopsy technician, and one resident over the secure video link. Remote and on-site pathologists independently interpreted and compared gross findings. Diverse benefits and technical aspects were studied using logs or information recorded in LS. Satisfaction surveys on diverse technical and professional aspects of the studies were conducted. Results: The full concordance between digital and light microscopic diagnosis was 99% (594/600 cases). A minor discordance, without clinical implications, was 1% (6/600 cases). The instant upload of digital images was achieved at 20 Mbps. Deference to microscopic slides and rescanning were under 1%. Average turnaround was shorter and percentage of cases reported up to 3 days higher for remote digital reporting. Larger displays improved the most user experience at magnifications over ×20. A digital diagnostic time was shorter than microscopic in 13 sessions. Four sessions with shorter microscopic diagnostic time included more cases requiring extensive use of magnifications over ×20. Independent interpretations of gross findings between remote and on-site pathologists yielded full agreement in the remote autopsies. Delays in reporting of frozen sections and FNAC due to scanning were clinically insignificant. Satisfaction levels with diverse technical and/or professional aspects of all studies were high. Conclusions: Complete routine remote digital pathology services are found feasible in hands of experienced staff. The introduction of digital pathology has improved provisions and organizations of our pathology services in histology, cytology, and autopsy including teaching and interdepartmental collaboration.

Background: Autopsies usually serve to inform specific “causes of death” and associated mechanisms. However, multiple diseases can co-exist and interact leading to a final demise. We approached autopsy-produced data using network analysis in an unbiased fashion to inform about interaction among different diseases and identify possible targets of system-level health care. Methods: Reports of 261 full autopsies from one institution between 2011 and 2013 were reviewed. Comorbidities were recorded and their Spearman's association coefficients were calculated. Highly associated comorbidities (P < 0.01) were selected to construct a network in which each disease is represented by a node, and each link between the nodes represents significant co-occurrence. Results: The network comprised 140 diseases connected by 419 links. The mean number of connections per node was 6. The most highly connected nodes (“hubs”) represented infectious processes, whereas less connected nodes represented neoplasms and other chronic diseases. Eight clusters of biologically plausible associated diseases were identified. Conclusions: There is an unbiased relationship among autopsy-identified diseases. There were “hubs” (primarily infectious) with significantly more associations than others that could represent obligatory or important modulators of the final expression of other diseases. Clusters of co-occurring diseases, or “modules,” suggest the presence of clinically relevant presentations of pathobiologically related entities which are until now considered individual diseases. These modules may occur together prior to death and be amenable to interventions during life.

Introduction: Whole-slide imaging (WSI) technology can be used for primary diagnosis and consultation, including intraoperative (IO) frozen section (FS). We aimed to implement and validate a digital system for the FS evaluation of cancer and transplant specimens following recommendations of the College of American Pathologists. Materials and Methods: FS cases were routinely scanned at ×20 employing the “Navigo” scanner system. IO diagnoses using glass versus digital slides after a 3-week washout period were recorded. Intraobserver concordance was evaluated using accuracy rate and kappa statistics. Feasibility of WSI diagnoses was assessed by the way of sensitivity, specificity, as well as positive and negative predictive values. Participants also completed a survey denoting scan time, time spent viewing cases, preference for glass versus WSI, image quality, interface experience, and any problems encountered. Results: Of the 125 cases submitted, 121 (436 slides) were successfully scanned including 93 oncological and 28 donor-organ FS biopsies. Four cases were excluded because of failed digitalization due to scanning problems or sample preparation artifacts. Full agreement between glass and digital-slide diagnosis was obtained in 90 of 93 (97%, κ = 0.96) oncology and in 24 of 28 (86%, κ = 0.91) transplant cases. There were two major and one minor discrepancy for cancer cases (sensitivity 100%, specificity 96%) and two major and two minor disagreements for transplant cases (sensitivity 96%, specificity 75%). Average scan and viewing/reporting time were 12 and 3 min for cancer cases, compared to 18 and 5 min for transplant cases. A high diagnostic comfort level among pathologists emerged from the survey. Conclusions: These data demonstrate that the “Navigo” digital WSI system can reliably support an IO FS service involving complicated cancer and transplant cases.

Background: Artificial intelligence is advancing at an accelerated pace into clinical applications, providing opportunities for increased efficiency, improved accuracy, and cost savings through computer-aided diagnostics. Dermatopathology, with emphasis on pattern recognition, offers a unique opportunity for testing deep learning algorithms. Aims: This study aims to determine the accuracy of deep learning algorithms to diagnose three common dermatopathology diagnoses. Methods: Whole slide images (WSI) of previously diagnosed nodular basal cell carcinomas (BCCs), dermal nevi, and seborrheic keratoses were annotated for areas of distinct morphology. Unannotated WSIs, consisting of five distractor diagnoses of common neoplastic and inflammatory diagnoses, were included in each training set. A proprietary fully convolutional neural network was developed to train algorithms to classify test images as positive or negative relative to ground truth diagnosis. Results: Artificial intelligence system accurately classified 123/124 (99.45%) BCCs (nodular), 113/114 (99.4%) dermal nevi, and 123/123 (100%) seborrheic keratoses. Conclusions: Artificial intelligence using deep learning algorithms is a potential adjunct to diagnosis and may result in improved workflow efficiencies for dermatopathologists and laboratories.

Background: Interest has grown into the use of multidetector computed tomography (CT) and magnetic resonance imaging as an adjunct or alternative to the invasive autopsy. We sought to investigate these possibilities in postmortem CT scan using an innovative virtual autopsy approach. Methods: Twenty-five postmortem cases were scanned with the Philips Brilliance CT-64 and then underwent traditional autopsy. The images were interpreted by two blinded forensic pathologists assisted by a radiologist with the INFOPSY^® Digital Autopsy Software System which provides three-dimensional images in Digital Imaging and Communications in Medicine format. Diagnostic validity of virtual autopsy (accuracy rate, sensitivity, specificity, and predictive values) and concordance between the two forensic pathologists (kappa intraobserver coefficients) were determined. Results: The causes of death at traditional autopsies were hemorrhage due to traumatic injuries (n = 8), respiratory failure (5), asphyxia due to drowning (4), asphyxia due to hanging or strangulation (2), heart failure (2), nontraumatic hemorrhage (1), and severe burns (1). In two cases, the cause of death could not be ascertained. In 15/23 (65%) cases, the cause of death diagnosed after virtual autopsy matched the diagnosis reported after traditional autopsy. In 8/23 cases (35%), traditional autopsy was necessary to establish the cause of death. Digital data provided relevant information for inferring both cause and manner of death in nine traumatic cases. The validity of virtual autopsy as a diagnostic tool was higher for traumatic deaths than other causes of death (accuracy 84%, sensitivity 82%, and specificity 86%). The concordance between the two forensic pathologists was almost perfect (>0.80). Conclusions: Our experience supports the use of virtual autopsy in postmortem investigations as an alternative diagnostic practice and does suggest a potential role as a screening test among traumatic deaths.

Background: The aim of this study is to derive and to validate a cohort of rectal cancer surgical patients within administrative datasets using text-search analysis of pathology reports. Materials and Methods: A text-search algorithm was developed and validated on pathology reports from 694 known rectal cancers, 1000 known colon cancers, and 1000 noncolorectal specimens. The algorithm was applied to all pathology reports available within the Ottawa Hospital Data Warehouse from 1996 to 2010. Identified pathology reports were validated as rectal cancer specimens through manual chart review. Sensitivity, specificity, and positive predictive value (PPV) of the text-search methodology were calculated. Results: In the derivation cohort of pathology reports (n = 2694), the text-search algorithm had a sensitivity and specificity of 100% and 98.6%, respectively. When this algorithm was applied to all pathology reports from 1996 to 2010 (n = 284,032), 5588 pathology reports were identified as consistent with rectal cancer. Medical record review determined that 4550 patients did not have rectal cancer, leaving a final cohort of 1038 rectal cancer patients. Sensitivity and specificity of the text-search algorithm were 100% and 98.4%, respectively. PPV of the algorithm was 18.6%. Conclusions: Text-search methodology is a feasible way to identify all rectal cancer surgery patients through administrative datasets with high sensitivity and specificity. However, in the presence of a low pretest probability, text-search methods must be combined with a validation method, such as manual chart review, to be a viable approach.

Background: The alumni of today's Pathology Informatics and Clinical Informatics fellowships fill diverse roles in academia, large health systems, and industry. The evolving training tracks and curriculum of Pathology Informatics fellowships have been well documented. However, less attention has been given to the posttraining experiences of graduates from informatics training programs. Here, we examine the career paths of subspecialty fellowship-trained pathology informaticians. Methods: Alumni from four Pathology Informatics fellowship training programs were contacted for their voluntary participation in the study. We analyzed various components of training, and the subsequent career paths of Pathology Informatics fellowship alumni using data extracted from alumni provided curriculum vitae. Results: Twenty-three out of twenty-seven alumni contacted contributed to the study. A majority had completed undergraduate study in science, technology, engineering, and math fields and combined track training in anatomic and clinical pathology. Approximately 30% (7/23) completed residency in a program with an in-house Pathology Informatics fellowship. Most completed additional fellowships (15/23) and many also completed advanced degrees (10/23). Common primary posttraining appointments included chief medical informatics officer (3/23), director of Pathology Informatics (10/23), informatics program director (2/23), and various roles in industry (3/23). Many alumni also provide clinical care in addition to their informatics roles (14/23). Pathology Informatics alumni serve on a variety of institutional committees, participate in national informatics organizations, contribute widely to scientific literature, and more than half (13/23) have obtained subspecialty certification in Clinical Informatics to date. Conclusions: Our analysis highlights several interesting phenomena related to the training and career trajectory of Pathology Informatics fellowship alumni. We note the long training track alumni complete in preparation for their careers. We believe flexible training pathways combining informatics and clinical training may help to alleviate the burden. We highlight the importance of in-house Pathology Informatics fellowships in promoting interest in informatics among residents. We also observe the many important leadership roles in academia, large community health systems, and industry available to early career alumni and believe this reflects a strong market for formally trained informaticians. We hope this analysis will be useful as we continue to develop the informatics fellowships to meet the future needs of our trainees and discipline.

Background: The tagging system is based on a small, electronic, wireless, laser-light-activated microtransponder named “p-Chip.” The p-Chip is a silicon integrated circuit, the size of which is 600 μm × 600 μm × 100 μm. Each p-Chip contains a unique identification code stored within its electronic memory that can be retrieved with a custom reader. These features allow the p-Chip to be used as an unobtrusive and scarcely noticeable ID tag on glass slides and tissue cassettes. Methods: The system is comprised of p-Chip-tagged sample carriers, a dedicated benchtop p-Chip ID reader that can accommodate both objects, and an additional reader (the Wand), with an adapter for reading IDs of glass slides stored vertically in drawers. On slides, p-Chips are attached with adhesive to the center of the short edge, and on cassettes – embedded directly into the plastic. ID readout is performed by bringing the reader to the proximity of the chip. Standard histopathology laboratory protocols were used for testing. Results: Very good ID reading efficiency was observed for both glass slides and cassettes. When processed slides are stored in vertical filing drawers, p-Chips remain readable without the need to remove them from the storage location, thereby improving the speed of searches in collections. On the cassettes, the ID continues to be readable through a thin layer of paraffin. Both slides and tissue cassettes can be read with the same reader, reducing the need for redundant equipment. Conclusions: The p-Chip is stable to all chemical challenges commonly used in the histopathology laboratory, tolerates temperature extremes, and remains durable in long-term storage. The technology is compatible with laboratory information management systems software systems. The p-Chip system is very well suited for identification of glass slides and cassettes in the histopathology laboratory.

Background: Advances in image analysis and computational techniques have facilitated automatic detection of critical features in histopathology images. Detection of nuclei is critical for squamous epithelium cervical intraepithelial neoplasia (CIN) classification into normal, CIN1, CIN2, and CIN3 grades. Methods: In this study, a deep learning (DL)-based nuclei segmentation approach is investigated based on gathering localized information through the generation of superpixels using a simple linear iterative clustering algorithm and training with a convolutional neural network. Results: The proposed approach was evaluated on a dataset of 133 digitized histology images and achieved an overall nuclei detection (object-based) accuracy of 95.97%, with demonstrated improvement over imaging-based and clustering-based benchmark techniques. Conclusions: The proposed DL-based nuclei segmentation Method with superpixel analysis has shown improved segmentation results in comparison to state-of-the-art methods.

Background: Nine E-learning modules (ELMs) were developed in our program using Articulate software. This study assessed our cytotechnology (CT) students' perceptions on the content of the ELMs, and the perceived influence of the ELMs on students' performance during clinical rotations. Subjects and Methods: All CT students watched nine ELMs before the related classroom lecture and group discussion. Following that, students completed nine preclinical rotation surveys. After their clinical rotations, students completed nine postclinical rotation surveys. Results: Statements on the content of the ELMs regarding the quality of the video and audio, duration, navigation, and the materials presented, received positive responses from the majority of the students. While there were a few disagreements and neutral responses, most of the students responded positively saying that the ELMs better prepared them for their role, as well as helped them to better perform their roles during the clinical rotation. The majority of the students recommended developing more EMLs for cytology courses in the future Conclusions: This study has given hope that the ELMs have potential to enhance our online curriculum and benefit students, within the United States and internationally, who have no easy access to cytology clinical laboratories for hands-on training.

Context: Medical diagnosis and clinical decisions rely heavily on the histopathological evaluation of tissue samples, especially in oncology. Historically, classical histopathology has been the gold standard for tissue evaluation and assessment by pathologists. The most widely and commonly used dyes in histopathology are hematoxylin and eosin (H&E) as most malignancies diagnosis is largely based on this protocol. H&E staining has been used for more than a century to identify tissue characteristics and structures morphologies that are needed for tumor diagnosis. In many cases, as tissue is scarce in clinical studies, fluorescence imaging is necessary to allow staining of the same specimen with multiple biomarkers simultaneously. Since fluorescence imaging is a relatively new technology in the pathology landscape, histopathologists are not used to or trained in annotating or interpreting these images. Aims, Settings and Design: To allow pathologists to annotate these images without the need for additional training, we designed an algorithm for the conversion of fluorescence images to brightfield H&E images. Subjects and Methods: In this algorithm, we use fluorescent nuclei staining to reproduce the hematoxylin information and natural tissue autofluorescence to reproduce the eosin information avoiding the necessity to specifically stain the proteins or intracellular structures with an additional fluorescence stain. Statistical Analysis Used: Our method is based on optimizing a transform function from fluorescence to H&E images using least mean square optimization. Results: It results in high quality virtual H&E digital images that can easily and efficiently be analyzed by pathologists. We validated our results with pathologists by making them annotate tumor in real and virtual H&E whole slide images and we obtained promising results. Conclusions: Hence, we provide a solution that enables pathologists to assess tissue and annotate specific structures based on multiplexed fluorescence images.