Weakly-supervised learning-based pathology detection and localization in 3D chest CT scans.

BACKGROUND: Recent advancements in anomaly detection have paved the way for novel radiological reading assistance tools that support the identification of findings, aimed at saving time. The clinical adoption of such applications requires a low rate of false positives while maintaining high sensitivity. PURPOSE: In light of recent interest and development in multi pathology identification, we present a novel method, based on a recent contrastive self-supervised approach, for multiple chest-related abnormality identification including low lung density area ("LLDA"), consolidation ("CONS"), nodules ("NOD") and interstitial pattern ("IP"). Our approach alerts radiologists about abnormal regions within a computed tomography (CT) scan by providing 3D localization. METHODS: We introduce a new method for the classification and localization of multiple chest pathologies in 3D Chest CT scans. Our goal is to distinguish four common chest-related abnormalities: "LLDA", "CONS", "NOD", "IP" and "NORMAL". This method is based on a 3D patch-based classifier with a Resnet backbone encoder pretrained leveraging recent contrastive self supervised approach and a fine-tuned classification head. We leverage the SimCLR contrastive framework for pretraining on an unannotated dataset of randomly selected patches and we then fine-tune it on a labeled dataset. During inference, this classifier generates probability maps for each abnormality across the CT volume, which are aggregated to produce a multi-label patient-level prediction. We compare different training strategies, including random initialization, ImageNet weight initialization, frozen SimCLR pretrained weights and fine-tuned SimCLR pretrained weights. Each training strategy is evaluated on a validation set for hyperparameter selection and tested on a test set. Additionally, we explore the fine-tuned SimCLR pretrained classifier for 3D pathology localization and conduct qualitative evaluation. RESULTS: Validated on 111 chest scans for hyperparameter selection and subsequently tested on 251 chest scans with multi-abnormalities, our method achieves an AUROC of 0.931 (95% confidence interval [CI]: [0.9034, 0.9557], p $ p$ -value < 0.001) and 0.963 (95% CI: [0.952, 0.976], p $ p$ -value < 0.001) in the multi-label and binary (i.e., normal versus abnormal) settings, respectively. Notably, our method surpasses the area under the receiver operating characteristic (AUROC) threshold of 0.9 for two abnormalities: IP (0.974) and LLDA (0.952), while achieving values of 0.853 and 0.791 for NOD and CONS, respectively. Furthermore, our results highlight the superiority of incorporating contrastive pretraining within the patch classifier, outperforming Imagenet pretraining weights and non-pretrained counterparts with uninitialized weights (F1 score = 0.943, 0.792, and 0.677 respectively). Qualitatively, the method achieved a satisfactory 88.8% completeness rate in localization and maintained an 88.3% accuracy rate against false positives. CONCLUSIONS: The proposed method integrates self-supervised learning algorithms for pretraining, utilizes a patch-based approach for 3D pathology localization and develops an aggregation method for multi-label prediction at patient-level. It shows promise in efficiently detecting and localizing multiple anomalies within a single scan.

Application of Innovative 3D Pathological Tactic for Diagnosis of Organizing Pneumonia.

BACKGROUND/AIM: The pathological diagnosis of organizing pneumonia (OP) relies on conventional traditional histopathological analysis, which involves examining stained thin slices of tissue. However, this method often results in suboptimal diagnostic objectivity due to low tissue sampling rates. This study aimed to assess the efficacy of tissue-clearing and infiltration-enhanced 3D spatial imaging techniques for elucidating the tissue architecture of OP. MATERIALS AND METHODS: H&E staining, 3D imaging technology, and AI-assisted analysis were employed to facilitate the construction of a multidimensional tissue architecture using six OP patient specimens procured from Taichung Veterans General Hospital, enabling a comprehensive morphological assessment. RESULTS: Specimens underwent H&E staining and exhibited Masson bodies and varying degrees of interstitial fibrosis. Furthermore, we conducted a comprehensive study of 3D images of the pulmonary histology reconstructed through an in-depth pathology analysis, and uncovered heterogenous distributions of fibrosis and Masson bodies across different depths of the OP specimens. CONCLUSION: Integrating 3D imaging for OP with AI-assisted analysis permits a substantially enhanced visualization and delineation of complex histological pulmonary disorders such as OP. The synergistic application of conventional histopathology with novel 3D imaging elucidated the sophisticated spatial configuration of OP, revealing the presence of Masson bodies and interstitial fibrosis. This methodology transcends conventional pathology constraints and paves the way for advanced algorithmic approaches to enhance precision in the detection, classification, and clinical management of lung pathologies.

Analysis of 3D pathology samples using weakly supervised AI.

Human tissue, which is inherently three-dimensional (3D), is traditionally examined through standard-of-care histopathology as limited two-dimensional (2D) cross-sections that can insufficiently represent the tissue due to sampling bias. To holistically characterize histomorphology, 3D imaging modalities have been developed, but clinical translation is hampered by complex manual evaluation and lack of computational platforms to distill clinical insights from large, high-resolution datasets. We present TriPath, a deep-learning platform for processing tissue volumes and efficiently predicting clinical outcomes based on 3D morphological features. Recurrence risk-stratification models were trained on prostate cancer specimens imaged with open-top light-sheet microscopy or microcomputed tomography. By comprehensively capturing 3D morphologies, 3D volume-based prognostication achieves superior performance to traditional 2D slice-based approaches, including clinical/histopathological baselines from six certified genitourinary pathologists. Incorporating greater tissue volume improves prognostic performance and mitigates risk prediction variability from sampling bias, further emphasizing the value of capturing larger extents of heterogeneous morphology.

Ductal carcinoma in situ develops within clonal fields of mutant cells in morphologically normal ducts.

Mutations are abundantly present in tissues of healthy individuals, including the breast epithelium. Yet it remains unknown whether mutant cells directly induce lesion formation or first spread, leading to a field of mutant cells that is predisposed towards lesion formation. To study the clonal and spatial relationships between morphologically normal breast epithelium adjacent to pre-cancerous lesions, we developed a three-dimensional (3D) imaging pipeline combined with spatially resolved genomics on archival, formalin-fixed breast tissue with the non-obligate breast cancer precursor ductal carcinoma in situ (DCIS). Using this 3D image-guided characterization method, we built high-resolution spatial maps of DNA copy number aberration (CNA) profiles within the DCIS lesion and the surrounding normal mammary ducts. We show that the local heterogeneity within a DCIS lesion is limited. However, by mapping the CNA profiles back onto the 3D reconstructed ductal subtree, we find that in eight out of 16 cases the healthy epithelium adjacent to the DCIS lesions has overlapping structural variations with the CNA profile of the DCIS. Together, our study indicates that pre-malignant breast transformations frequently develop within mutant clonal fields of morphologically normal-looking ducts. © 2024 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

Open-top Bessel beam two-photon light sheet microscopy for three-dimensional pathology.

Nondestructive pathology based on three-dimensional (3D) optical microscopy holds promise as a complement to traditional destructive hematoxylin and eosin (H&E) stained slide-based pathology by providing cellular information in high throughput manner. However, conventional techniques provided superficial information only due to shallow imaging depths. Herein, we developed open-top two-photon light sheet microscopy (OT-TP-LSM) for intraoperative 3D pathology. An extended depth of field two-photon excitation light sheet was generated by scanning a nondiffractive Bessel beam, and selective planar imaging was conducted with cameras at 400 frames/s max during the lateral translation of tissue specimens. Intrinsic second harmonic generation was collected for additional extracellular matrix (ECM) visualization. OT-TP-LSM was tested in various human cancer specimens including skin, pancreas, and prostate. High imaging depths were achieved owing to long excitation wavelengths and long wavelength fluorophores. 3D visualization of both cells and ECM enhanced the ability of cancer detection. Furthermore, an unsupervised deep learning network was employed for the style transfer of OT-TP-LSM images to virtual H&E images. The virtual H&E images exhibited comparable histological characteristics to real ones. OT-TP-LSM may have the potential for histopathological examination in surgical and biopsy applications by rapidly providing 3D information.

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.

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.

Artificial Intelligence-Triaged 3-Dimensional Pathology to Improve Detection of Esophageal Neoplasia While Reducing Pathologist Workloads.

Early detection of esophageal neoplasia via evaluation of endoscopic surveillance biopsies is the key to maximizing survival for patients with Barrett's esophagus, but it is hampered by the sampling limitations of conventional slide-based histopathology. Comprehensive evaluation of whole biopsies with 3-dimensional (3D) pathology may improve early detection of malignancies, but large 3D pathology data sets are tedious for pathologists to analyze. Here, we present a deep learning-based method to automatically identify the most critical 2-dimensional (2D) image sections within 3D pathology data sets for pathologists to review. Our method first generates a 3D heatmap of neoplastic risk for each biopsy, then classifies all 2D image sections within the 3D data set in order of neoplastic risk. In a clinical validation study, we diagnose esophageal biopsies with artificial intelligence-triaged 3D pathology (3 images per biopsy) vs standard slide-based histopathology (16 images per biopsy) and show that our method improves detection sensitivity while reducing pathologist workloads.

Panoramic Tissue Examination That Integrates 3-Dimensional Pathology Imaging and Gene Mutation: Potential Utility in Non-Small Cell Lung Cancer.

Novel therapeutics have significantly improved the survival and quality of life of patients with malignancies in this century. Versatile precision diagnostic data were used to formulate personalized therapeutic strategies for patients. However, the cost of extensive information depends on the consumption of the specimen, raising the challenges of effective specimen utilization, particularly in small biopsies. In this study, we proposed a tissue-processing cascaded protocol that obtains 3-dimensional (3D) protein expression spatial distribution and mutation analysis from an identical specimen. In order to reuse the thick section tissue evaluated after the 3D pathology technique, we designed a novel high-flatness agarose-embedded method that could improve tissue utilization rate by 1.52 fold, whereas it reduced the tissue-processing time by 80% compared with the traditional paraffin-embedding method. In animal studies, we demonstrated that the protocol would not affect the results of DNA mutation analysis. Furthermore, we explored the utility of this approach in non-small cell lung cancer because it is a compelling application for this innovation. We used 35 cases including 7 cases of biopsy specimens of non-small cell lung cancer to simulate future clinical application. The cascaded protocol consumed 150-µm thickness of formalin-fixed, paraffin-embedded specimens, providing 3D histologic and immunohistochemical information approximately 38 times that of the current paraffin-embedding protocol, and 3 rounds of DNA mutation analysis, offering both essential guidance for routine diagnostic evaluation and advanced information for precision medicine. Our designed integrated workflow provides an alternative way for pathological examination and paves the way for multidimensional tumor tissue assessment.

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.

The Role of New 3D Pathology and Lymphocyte Expression of Interstitial Inflammation in Pediatric-Onset Lupus Nephritis.

Lupus nephritis (LN) is a common and severe manifestation of pediatric-onset systemic lupus erythematosus (pSLE). It is one of the major causes of long-term glucocorticoid/immune suppressants use in pSLE. It causes long-term glucocorticoid/immune suppressants use and even end-stage renal disease (ESRD) in pSLE. It is now well known that high chronicity, especially the tubulointerstitial components in the renal biopsy, predicts a poor renal outcome. Interstitial inflammation (II), a component of activity in LN pathology, can be an early predictor for the renal outcome. With the advent of 3D pathology and CD19-targeted CAR-T cell therapy in the 2020s, the present study focuses on detailed pathology and B cell expression in II. We recruited 48 pSLE patients with class III/IV LN to analyze the risk of ESRD based on different II scores. We also studied 3D renal pathology and immunofluorescence (IF) staining of CD3, 19, 20, and 138 in patients with a high II score but low chronicity. Those pSLE LN patients with II scores of 2 or 3 showed a higher risk for ESRD (p = 0.003) than those with II scores of 0 or 1. Excluding patients with chronicity >3, high II scores still carried a higher risk for ESRD (p = 0.005). Checking the average scores from the renal specimens from different depths, the II, and chronicity showed good consistency between 3D and 2D pathology (interclass correlation coefficient [ICC], II = 0.91, p = 0.0015; chronicity = 0.86, p = 0.024). However, the sum of tubular atrophy plus interstitial fibrosis showed no good consistency (ICC = 0.79, p = 0.071). The selected LN patients with negative CD19/20 IF stains showed scattered CD3 infiltration and a different IF pattern of Syndecan-1 expression. Our study provides unique data in LN, including 3D pathology and different in situ Syndecan-1 patterns in LN patients.

Whole Murine Brain Imaging Based on Optical Elastic Scattering.

Imaging whole brains is one of the central efforts of biophotonics. While the established imaging modalities used in radiology, such as MRI and CT, have enabled in vivo investigations of various cognitive and affective processes, the prevailing resolution of one-cubic-millimeter has limited their use in studying the "ground-truth" of neuronal activities. On the other hand, electron microscopy (EM) visualizes the finest anatomic structures at a resolution of around 30 nm. However, the extensive tissue preparation process and the required large-scale morphological reconstruction restrict this method to small sample volumes. Light microscopy (LM) has the potential to bridge the above two spatial scales, with a resolution ranging from a few hundred nanometers to a few micrometers. Recent advances in tissue clearing have paved the way for optical investigation of large intact tissue volumes. However, most of these LM studies rely on fluorescence-a nonlinear optical process to provide contrast. This chapter introduces an alternative type of LM that is solely based on a linear optical process-elastic scattering, which has some unique advantages over conventional LM methods for the investigation of large-scale biological systems, such as intact murine brains. Here, we will first lay out the background and the motivation of developing this scattering-based method. Then, the basic principle of this approach will be introduced, including controlling tissue scattering and coherent imaging. Next, we explore current implementation and practical considerations. Up-to-date results and the utility of this method will also be demonstrated. Finally, we discuss current limitations and future directions in this promising field.

Acquisition, Visualization and Potential Applications of 3D Data in Anatomic Pathology.

Although human anatomy and histology are naturally three-dimensional (3D), commonly used diagnostic and educational tools are technologically restricted to providing two-dimensional representations (e.g. gross photography and glass slides). This limitation may be overcome by employing techniques to acquire and display 3D data, which refers to the digital information used to describe a 3D object mathematically. There are several established and experimental strategies to capture macroscopic and microscopic 3D data. In addition, recent hardware and software innovations have propelled the visualization of 3D models, including virtual and augmented reality. Accompanying these advances are novel clinical and non-clinical applications of 3D data in pathology. Medical education and research stand to benefit a great deal from utilizing 3D data as it can change our understanding of complex anatomical and histological structures. Although these technologies are yet to be adopted in routine surgical pathology, forensic pathology has embraced 3D scanning and model reconstruction. In this review, we intend to provide a general overview of the technologies and emerging applications involved with 3D data.

Beyond isolated cells: microfluidic transport of large tissue for pancreatic cancer diagnosis.

For cancer diagnoses, core biopsies (CBs) obtained from patients using coring needles (CNs) are traditionally visualized and assessed on microscope slides by pathologists after samples are processed and sectioned. A fundamental gain in optical information (i.e., diagnosis/staging) may be achieved when whole, unsectioned CBs (L = 5-20, D = 0.5-2.0 mm) are analyzed in 3D. This approach preserves CBs for traditional pathology and maximizes the diagnostic potential of patient samples. To bridge CNs/CBs with imaging, our group developed a microfluidic device that performs biospecimen preparation on unsectioned CBs for pathology. The ultimate goal is an automated and rapid point-of-care system that aids pathologists by processing tissue for advanced 3D imaging platforms. An inherent, but essential device feature is the microfluidic transport of CBs, which has not been previously investigated. Early experiments demonstrated proof-of-concept: pancreas CBs (D = 0.3-2.0 mm) of set lengths were transported in straight/curved microchannels, but dimensional tolerance and flow rates were variable, and preservation of CB integrity was uncontrolled. A second study used metal cylinder substitutes (L = 10, D = 1 mm) in microchannels to understand the transport mechanism. However, CBs are imperfectly shaped, rough, porous and viscoelastic. In this study, fresh/formalin-fixed porcine and human pancreas CBs were deposited into our device through a custom interface using clinical CNs. CB integrity (i.e., sample viability) may be assessed at every stage using an optomechanical metric: physical breaks were determined when specimen intensity profile data deviated beyond xavg + 2σ. Flow rates for human CBs were determined for several CNs, and microfluidic transport of fresh and formalin-fixed CBs was analyzed.