Deep learning extended depth-of-field microscope for fast and slide-free histology

Microscopic evaluation of resected tissue plays a central role in the surgical management of cancer. Because optical microscopes have a limited depth-of-field (DOF), resected tissue is either frozen or preserved with chemical fixatives, sliced into thin sections placed on microscope slides, stained, and imaged to determine whether surgical margins are free of tumor cells—a costly and time- and labor-intensive procedure. Here, we introduce a deep-learning extended DOF (DeepDOF) microscope to quickly image large areas of freshly resected tissue to provide histologic-quality images of surgical margins without physical sectioning. The DeepDOF microscope consists of a conventional fluorescence microscope with the simple addition of an inexpensive (less than $10) phase mask inserted in the pupil plane to encode the light field and enhance the depth-invariance of the point-spread function. When used with a jointly optimized image-reconstruction algorithm, diffraction-limited optical performance to resolve subcellular features can be maintained while significantly extending the DOF (200 µm). Data from resected oral surgical specimens show that the DeepDOF microscope can consistently visualize nuclear morphology and other important diagnostic features across highly irregular resected tissue surfaces without serial refocusing. With the capability to quickly scan intact samples with subcellular detail, the DeepDOF microscope can improve tissue sampling during intraoperative tumor-margin assessment, while offering an affordable tool to provide histological information from resected tissue specimens in resource-limited settings.

Histopathology, or microscopic examination of thinly sectioned and stained tissue slices on glass slides, is the gold standard to diagnose and guide surgical management of cancer. To prepare histopathology slides, biopsies or surgical specimens are typically formalin-fixed and paraffin-embedded (FFPE), sliced with a microtome to around 5 µm, stained with hematoxylin and eosin (H&E) dyes, and evaluated under a light microscope. For intraoperative assessment, resected surgical specimens will be first cut with a scalpel into 3- to 4-mm-thick slices to access potential tumor margins on cross-sectional surfaces (Fig. 1A); the thick slices can then be quickly frozen to acquire thin (∼5 µm) transverse tissue sections for staining and microscopic examination. While frozen sections can reduce the processing time, a cryostat microtome is required to cut thin sections of frozen tissue, which still must be fixed and stained. Despite the central role of histopathology in cancer diagnosis, the time- and labor-intensive sample-preparation steps require specialized personnel and expensive equipment, while allowing for only limited sampling of resected tissue. In addition, these destructive procedures are susceptible to tissue-processing artifacts (1, 2) and can also interfere with downstream molecular or genetic analysis.Open in a separate windowFig. 1.DeepDOF microscope schematic and imaging performance in comparison to conventional microscopes for fluorescence imaging of intact tissue specimens. (A) Prior to imaging, the resected specimen is bread-loafed by using a pathology scalpel, and the cross-section surface can be evaluated for tumor-margin assessment. (B) Variations in the surface topology of intact tissue specimens exceed the DOF of a conventional microscope with subcellular resolution. In comparison, with the simple addition of an inexpensive phase mask, the end-to-end optimized DeepDOF microscope allows subcellular imaging of large areas of intact tissue samples at 5.4 cm²/min. (C) Based on a standard 4× objective (obj), the DeepDOF microscope combines wavefront encoding with deep-learning-enabled image reconstruction to significantly improve the DOF and, thus, the volumetric FOV while maintaining subcellular resolution. As a result, the DeepDOF microscope offers fast scanning of the cross-sectional surface of tissue slices without need for refocusing.In view of the challenges associated with standard histopathology, the ability to image cross-sectional surfaces of thick tissue slices (Fig. 1A) directly and nondestructively is highly desired. Recent studies have demonstrated successful imaging of large areas of intact specimens using fluorescence microscopy, including approaches based on confocal scanning (3, 4), structured illumination (5), and ultraviolet (UV) excitation (6). Clinical application of these techniques, however, is largely hindered by the shallow depth-of-field (DOF). In conventional microscopy, DOF is fundamentally coupled to lateral resolution:DOF∝λNA2∝resolution2λ.[1]As shown in Fig. 1C, in conventional microscopy with standard objectives, achieving subcellular lateral resolution (∼2 to 3 µm) restricts the DOF to ∼30 µm. This is almost one order of magnitude smaller than that needed to accommodate the variations in surface topography of freshly resected tissue surfaces, which can extend to up to 200 µm (7). As an example, Fig. 1 B, Upper Right shows a fluorescence image of an ex vivo porcine esophageal sample, stained with proflavine to highlight epithelial cell nuclei. In the image acquired with a conventional fluorescence microscope, the resulting defocus blur compromises the ability to visualize detailed cellular structures across a large field of view (FOV) without serial refocusing.To overcome the intrinsic optical constraints of conventional fluorescence microscopy for extended DOF, different approaches have been employed, such as decoupled illumination and detection in light-sheet microscopy (7), dynamic remote focusing (8, 9), and spatial and spectral multiplexing (10, 11); nonetheless, they usually require customized and expensive optics or complicated geometrical configurations. Alternatively, reflectance-based label-free modalities, including reflectance confocal microscopy and full-field optical coherence tomography, have been demonstrated for cancer-lesion characterization in skin and different types of epithelium (12–15). While initial results are promising, these systems are significantly more expensive (more than $100,000) than conventional microscopes due to their optomechanical complexity (16, 17). Computationally, extended DOF has also been demonstrated by using Fourier ptychographic microscopy (18), but the image reconstruction assumes a thin sample target transilluminated with oblique plane waves and is not suited for clinical fluorescence imaging.Wavefront encoding, when combined with computational methods, offers a convenient and inexpensive route to enhance imaging performance (19, 20). Wavefront modulating elements, such as cubic phase masks, annular phase masks, and other adaptive optics components, have been employed in photography, microscopy, and optical coherent microscopy to extend the DOF and to correct other forms of aberrations (21–32). Despite their adoption in different modalities, phase masks usually cause image degradation, thus necessitating a separately designed reconstruction algorithm to retrieve original features. Recently, deep learning is emerging as a powerful tool to complement microscopy for analysis of complex microscopic data (33–36). In this work, we integrate a wave-optics model with deep learning to develop a physics-informed, end-to-end optimization framework for extended DOF. In contrast to conventional approaches, the deep-learning framework optimizes the phase-mask design with large realistic data, while codesigning the reconstruction algorithm. Using this data-driven approach, we design, optimize, and experimentally validate the deep-learning extended DOF microscope (DeepDOF microscope), a low-cost (less than $6,000) computational microscope for fast and slide-free histology of surgical specimens. The DeepDOF microscope consists of two key co-optimized components: the phase mask and the image-reconstruction algorithm (Fig. 1 B, Upper Left). As shown in Fig. 1 B, Lower Left, and C, by jointly optimizing the phase mask and reconstruction algorithm, the DOF of the DeepDOF microscope is significantly extended to 200 µm, accommodating for variations in surface topology of thick cross-sectional tissue slices. Thanks to its capability to map irregular surfaces in a high-volumetric FOV (6.9 mm³ in DeepDOF microscope vs. 1.2 mm³ in a conventional microscope) with subcellular resolution, the DeepDOF microscope can image large areas of bread-loafed tissue slices without refocusing. Importantly, this is achieved by using an inexpensive phase-modulating element (less than $10 at production volume of 500 masks) that does not sacrifice optical throughput, making the DeepDOF microscope design readily adaptable to image fluorophores with low brightness.Here, we describe key components of the end-to-end optimized DeepDOF microscope from initial numerical simulation, to optical design, to subsequent experimental validation. We first present simulated results to jointly design and optimize the DeepDOF microscope optics and algorithm using a deep neural network. We then report characterization of the optimized DeepDOF phase mask, with simulated and experimental data. Furthermore, imaging of resected surgical specimens from the oral cavity is provided to validate clinical performance. We show that, using the current economical sample stage, DeepDOF can scan large specimens at 5.4 cm²/min, offering a fast, easy-to-use, and inexpensive alternative to standard histopathology for assessment of intact biopsies and surgical specimens with cellular detail.     

Automatic classification of prostate cancer Gleason scores from multiparametric magnetic resonance images

Noninvasive, radiological image-based detection and stratification of Gleason patterns can impact clinical outcomes, treatment selection, and the determination of disease status at diagnosis without subjecting patients to surgical biopsies. We present machine learning-based automatic classification of prostate cancer aggressiveness by combining apparent diffusion coefficient (ADC) and T2-weighted (T2-w) MRI-based texture features. Our approach achieved reasonably accurate classification of Gleason scores (GS) 6(3 + 3) vs.  ≥ 7 and 7(3 + 4) vs. 7(4 + 3) despite the presence of highly unbalanced samples by using two different sample augmentation techniques followed by feature selection-based classification. Our method distinguished between GS 6(3 + 3) and  ≥ 7 cancers with 93% accuracy for cancers occurring in both peripheral (PZ) and transition (TZ) zones and 92% for cancers occurring in the PZ alone. Our approach distinguished the GS 7(3 + 4) from GS 7(4 + 3) with 92% accuracy for cancers occurring in both the PZ and TZ and with 93% for cancers occurring in the PZ alone. In comparison, a classifier using only the ADC mean achieved a top accuracy of 58% for distinguishing GS 6(3 + 3) vs. GS  ≥ 7 for cancers occurring in PZ and TZ and 63% for cancers occurring in PZ alone. The same classifier achieved an accuracy of 59% for distinguishing GS 7(3 + 4) from GS 7(4 + 3) occurring in the PZ and TZ and 60% for cancers occurring in PZ alone. Separate analysis of the cancers occurring in TZ alone was not performed owing to the limited number of samples. Our results suggest that texture features derived from ADC and T2-w MRI together with sample augmentation can help to obtain reasonably accurate classification of Gleason patterns.Prostate cancer (PCa) is among the most common cancers and a leading cause of cancer-related death in men in the United States (1). In general, patients diagnosed with PCa with a Gleason score (GS) ( ≤ 6) have better 5- and 10-y survival rates, lower biochemical recurrence rates, and lower prostate cancer-specific mortality than do patients with GS  ≥ 7 (2). Similarly, compared with patients with GS 7(4 + 3), those with GS 7(3 + 4) have better outcomes (2). The GS and prostate specific antigen (PSA) level are clinically used to determine PCa aggressiveness (3). GS is a well-validated factor and known to be a powerful predictor of disease progression, mortality, and outcomes (4, 5). However, owing to random sampling, the GS determined through biopsies is known to differ from those determined following radical prostatectomy (6, 7), as well as between immediate repeat biopsies (8). Therefore, the ability to automatically detect the GS with high accuracy from the diagnostic MRIs would have a significant impact on clinical decision making, treatment selection, and prediction of outcomes for patients and spare them from painful biopsies and their accompanying risk of complications. Noninvasive and accurate techniques that determine the aggressiveness of PCa are needed to enhance the quality of patient care.Previously, MRI has been investigated (9) as a modality for determining PCa aggressiveness. Although MRI has been shown to be a valuable tool for PCa detection (10–13), there is no clear consensus on the specific imaging biomarker that is most effective in distinguishing the aggressiveness of PCa lesions. In addition to MR spectroscopic and T2-weighted (T2-w) MR imaging, the apparent diffusion coefficient (ADC) from diffusion-weighted MRI has been confirmed to be valuable for differentiating PCa aggressiveness (14–17). However, studies differ in the specific ADC value used to distinguish between the cancers. The features used have included ADC mean computed from a single slice region of interest (ROI) (15, 16, 18), ADC mean computed from the entire volume using different sets of diffusion b-values (all vs. fast vs. slow) (19), 10th percentile of the ADC computed from the entire lesion (20), 10th percentile and ADC mean (21), and ADC mean computed over the entire lesion (22). Furthermore, none of the aforementioned studies used more than five imaging features for the analysis.Texture-based imaging features in conjunction with machine learning-based classification have predominantly been applied for classifying malignant from noncancerous prostate tissues (23–25) with one exception (26). Linear discriminant analysis (LDA)-based classification of various histogram-based ADC measures, namely, ADC mean, 10th percentile, T2-w histogram-based skewness, and k-trans were used to distinguish between cancer vs. benign and between cancer GSs (21). Our work builds on the aforementioned work by classifying GS 7(3 + 4) vs. 7(4 + 3) in addition to classifying cancer vs. noncancerous prostate and GS 6(3 + 3) vs. GS  ≥ 7 with texture-based features derived from ADC and T2-w MR images. Furthermore, our work addresses an important problem of obtaining highly accurate machine learning despite severe class imbalance between the different groups of cancers by using sample augmentation with feature selection.Our work demonstrates that PCa diagnosis can be improved by combining data-augmented classification together with more of the latent information in standard MRIs (the so-called “radiomics hypothesis”) (27, 28) compared with using ADC mean or T2 signal intensities alone, thereby reducing the potential for under- or overdiagnosis. Fig. 1 A and B show the ADC energy, ADC entropy, T2 energy, and T2 entropy overlaid on a slice of the ADC and corresponding T2-w MR image for two different patients: one with a tumor of GS 6(3 + 3) and the other with a tumor of GS 9(4 + 5). As shown in Fig. 1 A and B, the energy and entropy values computed from different tumor types appear to be very different, which suggests that textures, in combination with ADC, can help to differentiate between the cancer types.Open in a separate windowFig. 1.Example of (A) a GS 6(3 + 3) tumor and (B) a GS 9(4 + 5) tumor. The top row shows the ADC image with the computed energy and entropy values overlaid on the tumor. The bottom row shows the T2-w MR image with the computed energy and entropy values overlaid on the same tumor on the corresponding slice. The texture features were computed per voxel by using a 5 × 5 × 5 patch centered at each voxel.     

BAL findings in idiopathic nonspecific interstitial pneumonia and usual interstitial pneumonia.

Idiopathic pulmonary fibrosis (IPF), which has the histological pattern of usual interstitial pneumonia (UIP), is a progressive interstitial lung disease with a poor prognosis. Idiopathic interstitial pneumonias with a histological pattern of nonspecific interstitial pneumonia (NSIP) have a better prognosis than UIP, and may present with a clinical picture identical to IPF. The authors hypothesised that bronchoalveolar lavage (BAL) findings may distinguish between UIP and NSIP, and have prognostic value within disease subgroups. BAL findings were studied retrospectively in 54 patients with histologically proven (surgical biopsy) idiopathic UIP (n=35) or fibrotic NSIP (n=19), all presenting clinically as IPF. These findings were also compared with the BAL profile of patients with other categories of idiopathic interstitial pneumonias. BAL total and differential cell counts did not differ between the two groups. Survival was better in NSIP. In neither group were BAL findings predictive of survival or changes in lung function at 1 yr, even after adjustment for disease severity, smoking and treatment. BAL differential counts in fibrotic NSIP differed from respiratory bronchiolitis-associated interstitial lung disease, but not from desquamative interstitial pneumonia or cellular NSIP. The authors conclude that bronchoalveolar lavage findings do not discriminate between usual interstitial pneumonia and nonspecific interstitial pneumonia in patients presenting with clinical features of idiopathic pulmonary fibrosis, and have no prognostic value, once the distinction between the two has been made histologically.     

The perinexus: Sign-post on the path to a new model of cardiac conduction?

The perinexus is a recently identified microdomain surrounding the cardiac gap junction that contains elevated levels of connexin43 and the sodium channel protein, Na_v1.5. Ongoing work has established a role for the perinexus in regulating gap junction aggregation. However, recent studies have raised the possibility of a perinexal contribution at the gap junction cleft to intercellular propagation of action potential via non-electrotonic mechanisms. The latter possibility could modify the current theoretical understanding of cardiac conduction, help explain paradoxical experimental findings, and open up entirely new avenues for antiarrhythmic therapy. We review recent structural insights into the perinexus and its potential novel functional role in cardiac-excitation spread, highlighting presently unanswered questions, the evidence for ephaptic conduction in the heart and how structural insights may help complete this picture.     

Clinical evaluation of a visible light-cured indirect composite for long-term provisionalization

OBJECTIVE: To clinically evaluate a visible light-cured (VLC) resin composite system for long-term provisional and esthetic diagnostic restorations, fabricated using indirect techniques. METHODS: One-hundred and nine teeth were restored in 31 patients. Preoperational impressions were used to create VLC resin composite restorations (Radica) using indirect techniques. Restorations were relined as necessary and placed using various provisional cements at a follow-up appointment, subsequent to preparation of the teeth. Both fabricating laboratory technicians and placing dentists rated the restorations for acceptability in esthetics, marginal fit, occlusion, and functionality in various stages of provisionalization. RESULTS: All restorations (100%) were rated acceptable for esthetics prior to relining. After relining, a majority (93-100%) of restorations were rated acceptable in esthetic and functional criteria. At the placement of the permanent restoration, a majority (96-100%) of restorations were rated acceptable in esthetic and functional criteria. Terms of service ranged from two to seventy-six days. CONCLUSION: In combination with in vitro results, the clinical performance of the Radica VLC system for provisionalization and esthetic diagnostic restorations was judged to be acceptable. The system offers esthetics that are superior to conventional provisional restorations, and should be a valuable option to practitioners considering longer-term provisionalization in complex cases.     

Effect of acute and chronic fluoxetine on extracellular dopamine levels in the caudate-putamen and nucleus accumbens of rat

Recent studies indicate that an increase in serotonergic (5-HT) activity in the nucleus accumbens (NAc) produces an increase in dopamine (DA) release, providing a possible mechanism for the involvement of DA in the therapeutic action of selective serotonin reuptake inhibitor (SSRI) antidepressants. However, acutely administered fluoxetine (2.5, 5.0, or 10.0 mg/kg, i.p.) failed to elevate extracellular levels of DA, or its metabolites in the NAc or caudate-putamen (CP). In fact, the highest dose produced a small (20%) decrease in DA levels in the NAc. Extracellular levels of the 5-HT metabolite 5HIAA were consistently decreased at all doses of fluoxetine in both structures. Since SSRIs generally require several weeks of treatment to be effective clinically, a second experiment examined the effect of chronic administration of fluoxetine. Chronic (21 day) daily treatment with 5 mg/kg had no effect on NAc basal levels of DA, DA metabolites, or 5HIAA, relative to a saline-treated control group. Finally, pretreatment with fluoxetine appeared to slightly enhance the elevation of NAc DA induced by an injection of cocaine (10 mg/kg, i.p.), an effect that was not quite significant (P < .06). In conclusion, the 5-HT-induced facilitation of NAc DA neurotransmission described in the literature may not be relevant to the therapeutic action of fluoxetine. © 1996 Wiley-Liss, Inc.     

A novel ItrA4 d-galactosyl 1-phosphate transferase is predicted to initiate synthesis of an amino sugar-lacking K92 capsular polysaccharide of Acinetobacter baumannii B8300

The K92 capsular polysaccharide (CPS) from Acinetobacter baumannii B8300 was studied by sugar analysis, Smith degradation, and one- and two-dimensional ¹H and ¹³C NMR spectroscopy. The elucidated CPS includes a branched pentasaccharide repeat unit containing one d-Galp and four l-Rhap residues; an atypical composition given that all A. baumannii CPS structures determined to date contain at least one amino sugar. Accordingly, biosynthesis of A. baumannii CPS types are initiated by initiating transferases (Itrs) that transfer 1-phosphate of either a 2-acetamido-2-deoxy-d-hexose, a 2-acetamido-2,6-dideoxy-d-hexose or a 2-acetamido-4-acylamino-2,4,6-trideoxy-d-hexose to an undecaprenyl phosphate (UndP) carrier. However, the KL92 capsule biosynthesis gene cluster in the B8300 genome sequence includes a gene for a novel Itr type, ItrA4, which is predicted to begin synthesis of the K92 CPS by transferring D-Galp 1-phosphate to the UndP lipid carrier. The itrA4 gene was found in a module transcribed in the opposite direction to the majority of the K locus. This module also includes an unknown open reading frame (orf_KL92), a gtr166 glycosyltransferase gene, and a wzi gene predicted to be involved in the attachment of CPS to the cell surface. Investigation into the origins of orf_KL92-gtr166-itrA4-wzi_KL92 revealed it might have originated from Acinetobacter junii.