CohortFinder: an open-source tool for data-driven partitioning of digital pathology and imaging cohorts to yield robust machine-learning models.

Batch effects (BEs) refer to systematic technical differences in data collection unrelated to biological variations whose noise is shown to negatively impact machine learning (ML) model generalizability. Here we release CohortFinder (http://cohortfinder.com), an open-source tool aimed at mitigating BEs via data-driven cohort partitioning. We demonstrate CohortFinder improves ML model performance in downstream digital pathology and medical image processing tasks. CohortFinder is freely available for download at cohortfinder.com.

Clinical Relevance of Computationally Derived Tubular Features: Spatial Relationships and the Development of Tubulointerstitial Scarring in MCD/FSGS.

Background: Visual scoring of tubular damage has limitations in capturing the full spectrum of structural changes and prognostic potential. We investigate if computationally quantified tubular features can enhance prognostication and reveal spatial relationships with interstitial fibrosis. Methods: Deep-learning and image-processing-based segmentations were employed in N=254/266 PAS-WSIs from the NEPTUNE/CureGN datasets (135/153 focal segmental glomerulosclerosis and 119/113 minimal change disease) for: cortex, tubular lumen (TL), epithelium (TE), nuclei (TN), and basement membrane (TBM). N=104 pathomic features were extracted from these segmented tubular substructures and summarized at the patient level using summary statistics. The tubular features were quantified across the biopsy and in manually segmented regions of mature interstitial fibrosis and tubular atrophy (IFTA), pre-IFTA and non-IFTA in the NEPTUNE dataset. Minimum Redundancy Maximum Relevance was used in the NEPTUNE dataset to select features most associated with disease progression and proteinuria remission. Ridge-penalized Cox models evaluated their predictive discrimination compared to clinical/demographic data and visual-assessment. Models were evaluated in the CureGN dataset. Results: N=9 features were predictive of disease progression and/or proteinuria remission. Models with tubular features had high prognostic accuracy in both NEPTUNE and CureGN datasets and increased prognostic accuracy for both outcomes (5.6%-7.7% and 1.6%-4.6% increase for disease progression and proteinuria remission, respectively) compared to conventional parameters alone in the NEPTUNE dataset. TBM thickness/area and TE simplification progressively increased from non- to pre- and mature IFTA. Conclusions: Previously under-recognized, quantifiable, and clinically relevant tubular features in the kidney parenchyma can enhance understanding of mechanisms of disease progression and risk stratification.

PatchSorter: a high throughput deep learning digital pathology tool for object labeling.

The discovery of patterns associated with diagnosis, prognosis, and therapy response in digital pathology images often requires intractable labeling of large quantities of histological objects. Here we release an open-source labeling tool, PatchSorter, which integrates deep learning with an intuitive web interface. Using >100,000 objects, we demonstrate a >7x improvement in labels per second over unaided labeling, with minimal impact on labeling accuracy, thus enabling high-throughput labeling of large datasets.

An end-to-end workflow for nondestructive 3D pathology.

Recent advances in 3D pathology offer the ability to image orders of magnitude more tissue than conventional pathology methods while also providing a volumetric context that is not achievable with 2D tissue sections, and all without requiring destructive tissue sectioning. Generating high-quality 3D pathology datasets on a consistent basis, however, is not trivial and requires careful attention to a series of details during tissue preparation, imaging and initial data processing, as well as iterative optimization of the entire process. Here, we provide an end-to-end procedure covering all aspects of a 3D pathology workflow (using light-sheet microscopy as an illustrative imaging platform) with sufficient detail to perform well-controlled preclinical and clinical studies. Although 3D pathology is compatible with diverse staining protocols and computationally generated color palettes for visual analysis, this protocol focuses on the use of a fluorescent analog of hematoxylin and eosin, which remains the most common stain used for gold-standard pathological reports. We present our guidelines for a broad range of end users (e.g., biologists, clinical researchers and engineers) in a simple format. The end-to-end workflow requires 3-6 d to complete, bearing in mind that data analysis may take longer.

Engineering the future of 3D pathology.

In recent years, technological advances in tissue preparation, high-throughput volumetric microscopy, and computational infrastructure have enabled rapid developments in nondestructive 3D pathology, in which high-resolution histologic datasets are obtained from thick tissue specimens, such as whole biopsies, without the need for physical sectioning onto glass slides. While 3D pathology generates massive datasets that are attractive for automated computational analysis, there is also a desire to use 3D pathology to improve the visual assessment of tissue histology. In this perspective, we discuss and provide examples of potential advantages of 3D pathology for the visual assessment of clinical specimens and the challenges of dealing with large 3D datasets (of individual or multiple specimens) that pathologists have not been trained to interpret. We discuss the need for artificial intelligence triaging algorithms and explainable analysis methods to assist pathologists or other domain experts in the interpretation of these novel, often complex, large datasets.

Swiss digital pathology recommendations: results from a Delphi process conducted by the Swiss Digital Pathology Consortium of the Swiss Society of Pathology.

Integration of digital pathology (DP) into clinical diagnostic workflows is increasingly receiving attention as new hardware and software become available. To facilitate the adoption of DP, the Swiss Digital Pathology Consortium (SDiPath) organized a Delphi process to produce a series of recommendations for DP integration within Swiss clinical environments. This process saw the creation of 4 working groups, focusing on the various components of a DP system (1) scanners, quality assurance and validation of scans, (2) integration of Whole Slide Image (WSI)-scanners and DP systems into the Pathology Laboratory Information System, (3) digital workflow-compliance with general quality guidelines, and (4) image analysis (IA)/artificial intelligence (AI), with topic experts for each recruited for discussion and statement generation. The work product of the Delphi process is 83 consensus statements presented here, forming the basis for "SDiPath Recommendations for Digital Pathology". They represent an up-to-date resource for national and international hospitals, researchers, device manufacturers, algorithm developers, and all supporting fields, with the intent of providing expectations and best practices to help ensure safe and efficient DP usage.

National digital pathology projects in Switzerland: A 2023 update.

The Swiss Digital Pathology Consortium (SDiPath) was founded in 2018 as a working group of the Swiss Society for Pathology with the aim of networking, training, and promoting digital pathology (DP) at a national level. Since then, two national surveys have been carried out on the level of knowledge, dissemination, use, and needs in DP, which have resulted in clear fields of action. In addition to organizing symposia and workshops, national guidelines were drawn up and an initiative for a national DP platform actively codesigned. With the growing use of digital image processing and artificial intelligence tools, continuous monitoring, evaluation, and exchange of experiences will be pursued, along with best practices.

Digital image analysis and artificial intelligence in pathology diagnostics-the Swiss view.

Digital pathology (DP) is increasingly entering routine clinical pathology diagnostics. As digitization of the routine caseload advances, implementation of digital image analysis algorithms and artificial intelligence tools becomes not only attainable, but also desirable in daily sign out. The Swiss Digital Pathology Consortium (SDiPath) has initiated a Delphi process to generate best-practice recommendations for various phases of the process of digitization in pathology for the local Swiss environment, encompassing the following four topics: i) scanners, quality assurance, and validation of scans; ii) integration of scanners and systems into the pathology laboratory information system; iii) the digital workflow; and iv) digital image analysis (DIA)/artificial intelligence (AI). The current article focuses on the DIA-/AI-related recommendations generated and agreed upon by the working group and further verified by the Delphi process among the members of SDiPath. Importantly, they include the view and the currently perceived needs of practicing pathologists from multiple academic and cantonal hospitals as well as private practices.

The prognostic value of tumor-stroma ratio and a newly developed computer-aided quantitative analysis of routine H&E slides in high-grade serous ovarian cancer.

Introduction: Tumor-stroma ratio (TSR) is prognostic in multiple cancers, while its role in high-grade serous ovarian cancer (HGSOC) remains unclear. Despite the prognostic insight gained from genetic profiles and tumor-infiltrating lymphocytes (TILs), the prognostic use of histology slides remains limited, while it enables the identification of tumor characteristics via computational pathology reducing scoring time and costs. To address this, this study aimed to assess TSR's prognostic role in HGSOC and its association with TILs. We additionally developed an algorithm, Ovarian-TSR (OTSR), using deep learning for TSR scoring, comparing it to manual scoring. Methods: 340 patients with advanced-stage who underwent primary debulking surgery (PDS) or neo-adjuvant chemotherapy (NACT) with interval debulking (IDS). TSR was assessed in both the most invasive (MI) and whole tumor (WT) regions through manual scoring by pathologists and quantification using OTSR. Patients were categorized as stroma-rich (≥ 50% stroma) or stroma-poor (< 50%). TILs were evaluated via immunohistochemical staining. Results: In PDS, stroma-rich tumors were significantly associated with a more frequent papillary growth pattern (60% vs 34%), while In NACT stroma-rich tumors had a lower Tumor Regression Grading (TRG 4&5, 21% vs 57%) and increased pleural metastasis (25% vs 16%). Stroma-rich patients had significantly shorter overall and progression-free survival compared to stroma-poor (31 versus 45 months; P < 0.0001, and 15 versus 17 months; P = 0.0008, respectively). Combining stromal percentage and TILs led to three distinct survival groups with good (stroma-poor, high TIL), medium (stroma-rich, high TIL, or; stroma-poor, Low TIL), and poor(stroma-rich, low TIL) survival. These survival groups remained significant in CD8 and CD103 in multivariable analysis (Hazard ratio (HR) = 1.42, 95% Confidence-interval (CI) = 1.02-1.99; HR = 1.49, 95% CI = 1.01-2.18, and HR = 1.48, 95% CI = 1.05-2.08; HR = 2.24, 95% CI = 1.55-3.23, respectively). OTSR was able to recapitulate these results and demonstrated high concordance with expert pathologists (correlation = 0.83). Conclusions: TSR is an independent prognostic factor for survival assessment in HGSOC. Stroma-rich tumors have a worse prognosis and, in the case of NACT, a higher likelihood of pleural metastasis. OTSR provides a cost and time-efficient way of determining TSR with high reproducibility and reduced inter-observer variability.

Visual Assessment of 2-Dimensional Levels Within 3-Dimensional Pathology Data Sets of Prostate Needle Biopsies Reveals Substantial Spatial Heterogeneity.

Prostate cancer prognostication largely relies on visual assessment of a few thinly sectioned biopsy specimens under a microscope to assign a Gleason grade group (GG). Unfortunately, the assigned GG is not always associated with a patient's outcome in part because of the limited sampling of spatially heterogeneous tumors achieved by 2-dimensional histopathology. In this study, open-top light-sheet microscopy was used to obtain 3-dimensional pathology data sets that were assessed by 4 human readers. Intrabiopsy variability was assessed by asking readers to perform Gleason grading of 5 different levels per biopsy for a total of 20 core needle biopsies (ie, 100 total images). Intrabiopsy variability (Cohen κ) was calculated as the worst pairwise agreement in GG between individual levels within each biopsy and found to be 0.34, 0.34, 0.38, and 0.43 for the 4 pathologists. These preliminary results reveal that even within a 1-mm-diameter needle core, GG based on 2-dimensional images can vary dramatically depending on the location within a biopsy being analyzed. We believe that morphologic assessment of whole biopsies in 3 dimension has the potential to enable more reliable and consistent tumor grading.

Pathologist Computer-Aided Diagnostic Scoring of Tumor Cell Fraction: A Swiss National Study.

Tumor cell fraction (TCF) estimation is a common clinical task with well-established large interobserver variability. It thus provides an ideal test bed to evaluate potential impacts of employing a tumor cell fraction computer-aided diagnostic (TCFCAD) tool to support pathologists' evaluation. During a National Slide Seminar event, pathologists (n = 69) were asked to visually estimate TCF in 10 regions of interest (ROIs) from hematoxylin and eosin colorectal cancer images intentionally curated for diverse tissue compositions, cellularity, and stain intensities. Next, they re-evaluated the same ROIs while being provided a TCFCAD-created overlay highlighting predicted tumor vs nontumor cells, together with the corresponding TCF percentage. Participants also reported confidence levels in their assessments using a 5-tier scale, indicating no confidence to high confidence, respectively. The TCF ground truth (GT) was defined by manual cell-counting by experts. When assisted, interobserver variability significantly decreased, showing estimates converging to the GT. This improvement remained even when TCFCAD predictions deviated slightly from the GT. The standard deviation (SD) of the estimated TCF to the GT across ROIs was 9.9% vs 5.8% with TCFCAD (P < .0001). The intraclass correlation coefficient increased from 0.8 to 0.93 (95% CI, 0.65-0.93 vs 0.86-0.98), and pathologists stated feeling more confident when aided (3.67 ± 0.81 vs 4.17 ± 0.82 with the computer-aided diagnostic [CAD] tool). TCFCAD estimation support demonstrated improved scoring accuracy, interpathologist agreement, and scoring confidence. Interestingly, pathologists also expressed more willingness to use such a CAD tool at the end of the survey, highlighting the importance of training/education to increase adoption of CAD systems.

An end-to-end workflow for non-destructive 3D pathology.

Recent advances in 3D pathology offer the ability to image orders-of-magnitude more tissue than conventional pathology while providing a volumetric context that is lacking with 2D tissue sections, all without requiring destructive tissue sectioning. Generating high-quality 3D pathology datasets on a consistent basis is non-trivial, requiring careful attention to many details regarding tissue preparation, imaging, and data/image processing in an iterative process. Here we provide an end-to-end protocol covering all aspects of a 3D pathology workflow (using light-sheet microscopy as an illustrative imaging platform) with sufficient detail to perform well-controlled preclinical and clinical studies. While 3D pathology is compatible with diverse staining protocols and computationally generated color palettes for visual analysis, this protocol will focus on a fluorescent analog of hematoxylin and eosin (H&E), which remains the most common stain for gold-standard diagnostic determinations. We present our guidelines for a broad range of end-users (e.g., biologists, clinical researchers, and engineers) in a simple tutorial format.

Nondestructive 3D pathology with analysis of nuclear features for prostate cancer risk assessment.

Prostate cancer treatment decisions rely heavily on subjective visual interpretation [assigning Gleason patterns or International Society of Urological Pathology (ISUP) grade groups] of limited numbers of two-dimensional (2D) histology sections. Under this paradigm, interobserver variance is high, with ISUP grades not correlating well with outcome for individual patients, and this contributes to the over- and undertreatment of patients. Recent studies have demonstrated improved prognostication of prostate cancer outcomes based on computational analyses of glands and nuclei within 2D whole slide images. Our group has also shown that the computational analysis of three-dimensional (3D) glandular features, extracted from 3D pathology datasets of whole intact biopsies, can allow for improved recurrence prediction compared to corresponding 2D features. Here we seek to expand on these prior studies by exploring the prognostic value of 3D shape-based nuclear features in prostate cancer (e.g. nuclear size, sphericity). 3D pathology datasets were generated using open-top light-sheet (OTLS) microscopy of 102 cancer-containing biopsies extracted ex vivo from the prostatectomy specimens of 46 patients. A deep learning-based workflow was developed for 3D nuclear segmentation within the glandular epithelium versus stromal regions of the biopsies. 3D shape-based nuclear features were extracted, and a nested cross-validation scheme was used to train a supervised machine classifier based on 5-year biochemical recurrence (BCR) outcomes. Nuclear features of the glandular epithelium were found to be more prognostic than stromal cell nuclear features (area under the ROC curve [AUC] = 0.72 versus 0.63). 3D shape-based nuclear features of the glandular epithelium were also more strongly associated with the risk of BCR than analogous 2D features (AUC = 0.72 versus 0.62). The results of this preliminary investigation suggest that 3D shape-based nuclear features are associated with prostate cancer aggressiveness and could be of value for the development of decision-support tools. © 2023 The Pathological Society of Great Britain and Ireland.

Computational pathology improves risk stratification of a multi-gene assay for early stage ER+ breast cancer.

Prognostic markers currently utilized in clinical practice for estrogen receptor-positive (ER+) and lymph node-negative (LN-) invasive breast cancer (IBC) patients include the Nottingham grading system and Oncotype Dx (ODx). However, these biomarkers are not always optimal and remain subject to inter-/intra-observer variability and high cost. In this study, we evaluated the association between computationally derived image features from H&E images and disease-free survival (DFS) in ER+ and LN- IBC. H&E images from a total of n = 321 patients with ER+ and LN- IBC from three cohorts were employed for this study (Training set: D1 (n = 116), Validation sets: D2 (n = 121) and D3 (n = 84)). A total of 343 features relating to nuclear morphology, mitotic activity, and tubule formation were computationally extracted from each slide image. A Cox regression model (IbRiS) was trained to identify significant predictors of DFS and predict a high/low-risk category using D1 and was validated on independent testing sets D2 and D3 as well as within each ODx risk category. IbRiS was significantly prognostic of DFS with a hazard ratio (HR) of 2.33 (95% confidence interval (95% CI) = 1.02-5.32, p = 0.045) on D2 and a HR of 2.94 (95% CI = 1.18-7.35, p = 0.0208) on D3. In addition, IbRiS yielded significant risk stratification within high ODx risk categories (D1 + D2: HR = 10.35, 95% CI = 1.20-89.18, p = 0.0106; D1: p = 0.0238; D2: p = 0.0389), potentially providing more granular risk stratification than offered by ODx alone.

Multi-site cross-organ calibrated deep learning (MuSClD): Automated diagnosis of non-melanoma skin cancer.

Although deep learning (DL) has demonstrated impressive diagnostic performance for a variety of computational pathology tasks, this performance often markedly deteriorates on whole slide images (WSI) generated at external test sites. This phenomenon is due in part to domain shift, wherein differences in test-site pre-analytical variables (e.g., slide scanner, staining procedure) result in WSI with notably different visual presentations compared to training data. To ameliorate pre-analytic variances, approaches such as CycleGAN can be used to calibrate visual properties of images between sites, with the intent of improving DL classifier generalizability. In this work, we present a new approach termed Multi-Site Cross-Organ Calibration based Deep Learning (MuSClD) that employs WSIs of an off-target organ for calibration created at the same site as the on-target organ, based off the assumption that cross-organ slides are subjected to a common set of pre-analytical sources of variance. We demonstrate that by using an off-target organ from the test site to calibrate training data, the domain shift between training and testing data can be mitigated. Importantly, this strategy uniquely guards against potential data leakage introduced during calibration, wherein information only available in the testing data is imparted on the training data. We evaluate MuSClD in the context of the automated diagnosis of non-melanoma skin cancer (NMSC). Specifically, we evaluated MuSClD for identifying and distinguishing (a) basal cell carcinoma (BCC), (b) in-situ squamous cell carcinomas (SCC-In Situ), and (c) invasive squamous cell carcinomas (SCC-Invasive), using an Australian (training, n = 85) and a Swiss (held-out testing, n = 352) cohort. Our experiments reveal that MuSCID reduces the Wasserstein distances between sites in terms of color, contrast, and brightness metrics, without imparting noticeable artifacts to training data. The NMSC-subtyping performance is statistically improved as a result of MuSCID in terms of one-vs. rest AUC: BCC (0.92 vs 0.87, p = 0.01), SCC-In Situ (0.87 vs 0.73, p = 0.15) and SCC-Invasive (0.92 vs 0.82, p = 1e-5). Compared to baseline NMSC-subtyping with no calibration, the internal validation results of MuSClD (BCC (0.98), SCC-In Situ (0.92), and SCC-Invasive (0.97)) suggest that while domain shift indeed degrades classification performance, our on-target calibration using off-target tissue can safely compensate for pre-analytical variabilities, while improving the robustness of the model.

MMO-Net (Multi-Magnification Organ Network): A use case for Organ Identification using Multiple Magnifications in Preclinical Pathology Studies.

Identifying organs within histology images is a fundamental and non-trivial step in toxicological digital pathology workflows as multiple organs often appear on the same whole slide image (WSI). Previous works in automated tissue classification have investigated the use of single magnifications, and demonstrated limitations when attempting to identify small and contiguous organs at low magnifications. In order to overcome these shortcomings, we present a multi-magnification convolutional neural network (CNN), called MMO-Net, which employs context and cellular detail from different magnifications to facilitate the recognition of complex organs. Across N=320 WSI from 3 contract research organization (CRO) laboratories, we demonstrate state-of-the-art organ detection and segmentation performance of 7 rat organs with and without lesions: liver, kidney, thyroid gland, parathyroid gland, urinary bladder, salivary gland, and mandibular lymph node (AUROC=0.99-1.0 for all organs, Dice≥0.9 except parathyroid (0.73)). Evaluation takes place at both inter- and intra CRO levels, suggesting strong generalizability performance. Results are qualitatively reviewed using visualization masks to ensure separation of organs in close proximity (e.g., thyroid vs parathyroid glands). MMO-Net thus offers organ localization that serves as a potential quality control tool to validate WSI metadata and as a preprocessing step for subsequent organ-specific artificial intelligence (AI) use cases. To facilitate research in this area, all associated WSI and metadata used for this study are being made freely available, forming a first of its kind dataset for public use.

Impact of scanner variability on lymph node segmentation in computational pathology.

Computer-aided diagnostics in histopathology are based on the digitization of glass slides. However, heterogeneity between the images generated by different slide scanners can unfavorably affect the performance of computational algorithms. Here, we evaluate the impact of scanner variability on lymph node segmentation due to its clinical importance in colorectal cancer diagnosis. 100 slides containing 276 lymph nodes were digitized using 4 different slide scanners, and 50 of the lymph nodes containing metastatic cancer cells. These 400 scans were subsequently annotated by 2 experienced pathologists to precisely label lymph node boundary. Three different segmentation methods were then applied and compared: Hematoxylin-channel-based thresholding (HCT), Hematoxylin-based active contours (HAC), and a convolution neural network (U-Net). Evaluation of U-Net trained from both a single scanner and an ensemble of all scanners was completed. Mosaic images based on representative tiles from a scanner were used as a reference image to normalize the new data from different test scanners to evaluate the performance of a pre-trained model. Fine-tuning was carried out by using weights of a model trained on one scanner to initialize model weights for other scanners. To evaluate the domain generalization, domain adversarial learning and stain mix-up augmentation were also implemented. Results show that fine-tuning and domain adversarial learning decreased the impact of scanner variability and greatly improved segmentation across scanners. Overall, U-Net with stain mix-up (Matthews correlation coefficient (MCC) = 0.87), domain adversarial learning (MCC = 0.86), and HAC (MCC = 0.87) were shown to outperform HCT (MCC = 0.81) for segmentation of lymph nodes when compared against the ground truth. The findings of this study should be considered for future algorithms applied in diagnostic routines.

Towards a national strategy for digital pathology in Switzerland.

Precision medicine is entering a new era of digital diagnostics; the availability of integrated digital pathology (DP) and structured clinical datasets has the potential to become a key catalyst for biomedical research, education and business development. In Europe, national programs for sharing of this data will be crucial for the development, testing, and validation of machine learning-enabled tools supporting clinical decision-making. Here, the Swiss Digital Pathology Consortium (SDiPath) discusses the creation of a Swiss Digital Pathology Infrastructure (SDPI), which aims to develop a unified national DP network bringing together the Swiss Personalized Health Network (SPHN) with Swiss university hospitals and subsequent inclusion of cantonal and private institutions. This effort builds on existing developments for the national implementation of structured pathology reporting. Opening this national infrastructure and data to international researchers in a sequential rollout phase can enable the large-scale integration of health data and pooling of resources for research purposes and clinical trials. Therefore, the concept of a SDPI directly synergizes with the priorities of the European Commission communication on the digital transformation of healthcare on an international level, and with the aims of the Swiss State Secretariat for Economic Affairs (SECO) for advancing research and innovation in the digitalization domain. SDPI directly addresses the needs of existing national and international research programs in neoplastic and non-neoplastic diseases by providing unprecedented access to well-curated clinicopathological datasets for the development and implementation of novel integrative methods for analysis of clinical outcomes and treatment response. In conclusion, a SDPI would facilitate and strengthen inter-institutional collaboration in technology, clinical development, business and research at a national and international scale, promoting improved patient care via precision medicine.