Near real-time intraoperative brain tumor diagnosis using stimulated Raman histology and deep neural networks

Artificial-intelligence-based molecular classification of diffuse gliomas using rapid, label-free optical imaging

Molecular classification has transformed the management of brain tumors by enabling more accurate prognostication and personalized treatment. However, timely molecular diagnostic testing for patients with brain tumors is limited, complicating surgical and adjuvant treatment and obstructing clinical trial enrollment. In this study, we developed DeepGlioma, a rapid (<90 seconds), artificial-intelligence-based diagnostic screening system to streamline the molecular diagnosis of diffuse gliomas. DeepGlioma is trained using a multimodal dataset that includes stimulated Raman histology (SRH); a rapid, label-free, non-consumptive, optical imaging method; and large-scale, public genomic data. In a prospective, multicenter, international testing cohort of patients with diffuse glioma (n = 153) who underwent real-time SRH imaging, we demonstrate that DeepGlioma can predict the molecular alterations used by the World Health Organization to define the adult-type diffuse glioma taxonomy (IDH mutation, 1p19q co-deletion and ATRX mutation), achieving a mean molecular classification accuracy of 93.3 ± 1.6%. Our results represent how artificial intelligence and optical histology can be used to provide a rapid and scalable adjunct to wet lab methods for the molecular screening of patients with diffuse glioma.

Rapid ultraviolet photoacoustic remote sensing microscopy using voice-coil stage scanning

There is an unmet need for fast virtual histology technologies that exhibit histological realism and can scan large sections of fresh tissue within intraoperative time-frames. Ultraviolet photoacoustic remote sensing microscopy (UV-PARS) is an emerging imaging modality capable of producing virtual histology images that show good concordance to conventional histology stains. However, a UV-PARS scanning system that can perform rapid intraoperative imaging over mm-scale fields-of-view at fine resolution (<500 nm) has yet to be demonstrated. In this work, we present a UV-PARS system which utilizes voice-coil stage scanning to demonstrate finely resolved images for 2×2 mm² areas at 500 nm sampling resolution in 1.33 minutes and coarsely resolved images for 4×4 mm² areas at 900 nm sampling resolution in 2.5 minutes. The results of this work demonstrate the speed and resolution capabilities of the UV-PARS voice-coil system and further develop the potential for UV-PARS microscopy to be employed in a clinical setting.

Refining the Intraoperative Identification of Suspected High-Grade Glioma Using a Surgical Fluorescence Biomarker: GALA BIDD Study Report

BACKGROUND

Improving intraoperative accuracy with a validated surgical biomarker is important because identifying high-grade areas within a glioma will aid neurosurgical decision-making and sampling.

METHODS

We designed a multicentre, prospective surgical cohort study (GALA-BIDD) to validate the presence of visible fluorescence as a pragmatic intraoperative surgical biomarker of suspected high-grade disease within a tumour mass in patients undergoing 5-aminolevulinic acid (5-ALA) fluorescence-guided cytoreductive surgery.

RESULTS

A total of 106 patients with a suspected high-grade glioma or malignant transformation of a low-grade glioma were enrolled. Among the 99 patients who received 5-ALA, 89 patients were eligible to assess the correlation of fluorescence with diagnosis as per protocol. Of these 89, 81 patients had visible fluorescence at surgery, and 8 patients had no fluorescence. A total of 80 out of 81 fluorescent patients were diagnosed as high-grade gliomas on postoperative central review with 1 low-grade glioma case. Among the eight patients given 5-ALA who did not show any visible fluorescence, none were high-grade gliomas, and all were low-grade gliomas. Of the seven patients suspected radiologically of malignant transformation of low-grade gliomas and with visible fluorescence at surgery, six were diagnosed with high-grade gliomas, and one had no tissue collected.

CONCLUSION

In patients where there is clinical suspicion, visible 5-ALA fluorescence has clinical utility as an intraoperative surgical biomarker of high-grade gliomas and can aid surgical decision-making and sampling. Further studies assessing the use of 5-ALA to assess malignant transformation in all diffuse gliomas may be valuable.

Hydrogel Microneedles Extracting Exosomes for Early Detection of Colorectal Cancer

There are several methods for early diagnosis of tumors, such as detecting circulating tumor DNAs, detecting circulating tumor cells, or imaging with tumor-targeting contrast agents. However, these assays are time-consuming and may cause patient discomfort during the biopsy collecting process. Here, we develop a facile method for early diagnosis of tumors by extracting exosomes from interstitial fluid (ISF) using hydrogel microneedles (MNs). The hydrogel MNs expand in the skin to absorb the ISF, and the tumor exosomes contained in the ISF bind with the glypican-1 antibodies inside the hydrogel of MNs. After removing the hydrogel on the MNs, exosomes are separately purified from the ISF to analyze tumor-related biomarkers. Finally, colorectal cancer can be diagnosed by ELISA for the colorectal cancer-induced model mice. This noninvasive hydrogel MN system to obtain the exosome samples would play an important role in early cancer diagnosis.

High-throughput line-illumination Raman microscopy with multislit detection

Raman microscopy is an emerging tool for molecular imaging and analysis of living samples. Use of Raman microscopy in life sciences is, however, still limited because of its slow measurement speed for spectral imaging and analysis. We developed a multiline-illumination Raman microscope to achieve ultrafast Raman spectral imaging. A spectrophotometer equipped with a periodic array of confocal slits detects Raman spectra from a sample irradiated by multiple line illuminations. A comb-like Raman hyperspectral image is formed on a two-dimensional detector in the spectrophotometer, and a hyperspectral Raman image is acquired by scanning the sample with multiline illumination array. By irradiating a sample with 21 simultaneous illumination lines, we achieved high-throughput Raman hyperspectral imaging of mouse brain tissue, acquiring 1108800 spectra in 11.4 min. We also measured mouse kidney and liver tissue as well as conducted label-free live-cell molecular imaging. The ultrafast Raman hyperspectral imaging enabled by the presented technique will expand the possible applications of Raman microscopy in biological and medical fields.

Multiplex Identification of Post-Translational Modifications at Point-of-Care by Deep Learning-Assisted Hydrogel Sensors

Multiplex detection of protein post-translational modifications (PTMs), especially at point-of-care, is of great significance in cancer diagnosis. Herein, we report a machine learning-assisted photonic crystal hydrogel (PCH) sensor for multiplex detection of PTMs. With closely-related PCH sensors microfabricated on a single chip, our design achieved not only rapid screening of PTMs at specific protein sites by using only naked eyes/cellphone, but also the feasibility of real-time monitoring of phosphorylation reactions. By taking advantage of multiplex sensor chips and a neural network algorithm, accurate prediction of PTMs by both their types and concentrations was enabled. This approach was ultimately used to detect and differentiate up/down regulation of different phosphorylation sites within the same protein in live mammalian cells. Our developed method thus holds potential for POC identification of various PTMs in early-stage diagnosis of protein-related diseases.

Stimulated Raman Scattering Microscopy Enables Gleason Scoring of Prostate Core Needle Biopsy by a Convolutional Neural Network

Focal therapy (FT) has been proposed as an approach to eradicate clinically significant prostate cancer while preserving the normal surrounding tissues to minimize treatment-related toxicity. Rapid histology of core needle biopsies is essential to ensure the precise FT for localized lesions and to determine tumor grades. However, it is difficult to achieve both high accuracy and speed with currently available histopathology methods. Here, we demonstrated that stimulated Raman scattering (SRS) microscopy could reveal the largely heterogeneous histologic features of fresh prostatic biopsy tissues in a label-free and near real-time manner. A diagnostic convolutional neural network (CNN) built based on images from 61 patients could classify Gleason patterns of prostate cancer with an accuracy of 85.7%. An additional 22 independent cases introduced as external test dataset validated the CNN performance with 84.4% accuracy. Gleason scores of core needle biopsies from 21 cases were calculated using the deep learning SRS system and showed a 71% diagnostic consistency with grading from three pathologists. This study demonstrates the potential of a deep learning-assisted SRS platform in evaluating the tumor grade of prostate cancer, which could help simplify the diagnostic workflow and provide timely histopathology compatible with FT treatment.

SIGNIFICANCE

A platform combining stimulated Raman scattering microscopy and a convolutional neural network provides rapid histopathology and automated Gleason scoring on fresh prostate core needle biopsies without complex tissue processing.

A deep learning model for detection of leukocytes under various interference factors

The accurate detection of leukocytes is the basis for the diagnosis of blood system diseases. However, diagnosing leukocyte disorders by doctors is time-consuming and requires extensive experience. Automated detection methods with high accuracy can improve detection efficiency and provide recommendations to inexperienced doctors. Current methods and instruments either fail to automate the identification process fully or have low performance and need suitable leukocyte data sets for further study. To improve the current status, we need to develop more intelligent strategies. This paper investigates fulfilling high-performance automatic detection for leukocytes using a deep learning-based method. We established a new dataset more suitable for leukocyte detection, containing 6273 images (8595 leukocytes) and considering nine common clinical interference factors. Based on the dataset, the performance evaluation of six mainstream detection models is carried out, and a more robust ensemble model is proposed. The mean of average precision (mAP) @IoU = 0.50:0.95 and mean of average recall (mAR)@IoU = 0.50:0.95 of the ensemble model on the test set are 0.853 and 0.922, respectively. The detection performance of poor-quality images is robust. For the first time, it is found that the ensemble model yields an accuracy of 98.84% for detecting incomplete leukocytes. In addition, we also compared the test results of different models and found multiple identical false detections of the models, then provided correct suggestions for the clinic.

Label-free intraoperative histology of bone tissue via deep-learning-assisted ultraviolet photoacoustic microscopy

Obtaining frozen sections of bone tissue for intraoperative examination is challenging. To identify the bony edge of resection, orthopaedic oncologists therefore rely on pre-operative X-ray computed tomography or magnetic resonance imaging. However, these techniques do not allow for accurate diagnosis or for intraoperative confirmation of the tumour margins, and in bony sarcomas, they can lead to bone margins up to 10-fold wider (1,000-fold volumetrically) than necessary. Here, we show that real-time three-dimensional contour-scanning of tissue via ultraviolet photoacoustic microscopy in reflection mode can be used to intraoperatively evaluate undecalcified and decalcified thick bone specimens, without the need for tissue sectioning. We validate the technique with gold-standard haematoxylin-and-eosin histology images acquired via a traditional optical microscope, and also show that an unsupervised generative adversarial network can virtually stain the ultraviolet-photoacoustic-microscopy images, allowing pathologists to readily identify cancerous features. Label-free and slide-free histology via ultraviolet photoacoustic microscopy may allow for rapid diagnoses of bone-tissue pathologies and aid the intraoperative determination of tumour margins.

Advances in the intraoperative delineation of malignant glioma margin

Surgery plays a critical role in the treatment of malignant glioma. However, due to the infiltrative growth and brain shift, it is difficult for neurosurgeons to distinguish malignant glioma margins with the naked eye and with preoperative examinations. Therefore, several technologies were developed to determine precise tumor margins intraoperatively. Here, we introduced four intraoperative technologies to delineate malignant glioma margin, namely, magnetic resonance imaging, fluorescence-guided surgery, Raman histology, and mass spectrometry. By tracing their detecting principles and developments, we reviewed their advantages and disadvantages respectively and imagined future trends.

Polysulfide Serves as a Hallmark of Desmoplastic Reaction to Differentially Diagnose Ductal Carcinoma In Situ and Invasive Breast Cancer by SERS Imaging

Pathological examination of formalin-fixed paraffin-embedded (FFPE) needle-biopsied samples by certified pathologists represents the gold standard for differential diagnosis between ductal carcinoma in situ (DCIS) and invasive breast cancers (IBC), while information of marker metabolites in the samples is lost in the samples. Infrared laser-scanning large-area surface-enhanced Raman spectroscopy (SERS) equipped with gold-nanoparticle-based SERS substrate enables us to visualize metabolites in fresh-frozen needle-biopsied samples with spatial matching between SERS and HE staining images with pathological annotations. DCIS (n = 14) and IBC (n = 32) samples generated many different SERS peaks in finger-print regions of SERS spectra among pathologically annotated lesions including cancer cell nests and the surrounding stroma. The results showed that SERS peaks in IBC stroma exhibit significantly increased polysulfide that coincides with decreased hypotaurine as compared with DCIS, suggesting that alterations of these redox metabolites account for fingerprints of desmoplastic reactions to distinguish IBC from DCIS. Furthermore, the application of supervised machine learning to the stroma-specific multiple SERS signals enables us to support automated differential diagnosis with high accuracy. The results suggest that SERS-derived biochemical fingerprints derived from redox metabolites account for a hallmark of desmoplastic reaction of IBC that is absent in DCIS, and thus, they serve as a useful method for precision diagnosis in breast cancer.

Raman spectroscopy: A prospective intraoperative visualization technique for gliomas

The infiltrative growth and malignant biological behavior of glioma make it one of the most challenging malignant tumors in the brain, and how to maximize the extent of resection (EOR) while minimizing the impact on normal brain tissue is the pursuit of neurosurgeons. The current intraoperative visualization assistance techniques applied in clinical practice suffer from low specificity, slow detection speed and low accuracy, while Raman spectroscopy (RS) is a novel spectroscopy technique gradually developed and applied to clinical practice in recent years, which has the advantages of being non-destructive, rapid and accurate at the same time, allowing excellent intraoperative identification of gliomas. In the present work, the latest research on Raman spectroscopy in glioma is summarized to explore the prospect of Raman spectroscopy in glioma surgery.

Rapid, label-free histopathological diagnosis of liver cancer based on Raman spectroscopy and deep learning

Biopsy is the recommended standard for pathological diagnosis of liver carcinoma. However, this method usually requires sectioning and staining, and well-trained pathologists to interpret tissue images. Here, we utilize Raman spectroscopy to study human hepatic tissue samples, developing and validating a workflow for in vitro and intraoperative pathological diagnosis of liver cancer. We distinguish carcinoma tissues from adjacent non-tumour tissues in a rapid, non-disruptive, and label-free manner by using Raman spectroscopy combined with deep learning, which is validated by tissue metabolomics. This technique allows for detailed pathological identification of the cancer tissues, including subtype, differentiation grade, and tumour stage. 2D/3D Raman images of unprocessed human tissue slices with submicrometric resolution are also acquired based on visualization of molecular composition, which could assist in tumour boundary recognition and clinicopathologic diagnosis. Lastly, the potential for a portable handheld Raman system is illustrated during surgery for real-time intraoperative human liver cancer diagnosis.

Advanced intraoperative MRI in pediatric brain tumor surgery

Introduction: In the pediatric brain tumor surgery setting, intraoperative MRI (ioMRI) provides "real-time" imaging, allowing for evaluation of the extent of resection and detection of complications. The use of advanced MRI sequences could potentially provide additional physiological information that may aid in the preservation of healthy brain regions. This review aims to determine the added value of advanced imaging in ioMRI for pediatric brain tumor surgery compared to conventional imaging. Methods: Our systematic literature search identified relevant articles on PubMed using keywords associated with pediatrics, ioMRI, and brain tumors. The literature search was extended using the snowball technique to gather more information on advanced MRI techniques, their technical background, their use in adult ioMRI, and their use in routine pediatric brain tumor care. Results: The available literature was sparse and demonstrated that advanced sequences were used to reconstruct fibers to prevent damage to important structures, provide information on relative cerebral blood flow or abnormal metabolites, or to indicate the onset of hemorrhage or ischemic infarcts. The explorative literature search revealed developments within each advanced MRI field, such as multi-shell diffusion MRI, arterial spin labeling, and amide-proton transfer-weighted imaging, that have been studied in adult ioMRI but have not yet been applied in pediatrics. These techniques could have the potential to provide more accurate fiber tractography, information on intraoperative cerebral perfusion, and to match gadolinium-based T1w images without using a contrast agent. Conclusion: The potential added value of advanced MRI in the intraoperative setting for pediatric brain tumors is to prevent damage to important structures, to provide additional physiological or metabolic information, or to indicate the onset of postoperative changes. Current developments within various advanced ioMRI sequences are promising with regard to providing in-depth tissue information.

Extent of use of artificial intelligence & machine learning protocols in cancer diagnosis: A scoping review

Background & objectives

Artificial intelligence (AI) and machine learning (ML) have shown promising results in cancer diagnosis in validation tests involving retrospective patient databases. This study was aimed to explore the extent of actual use of AI/ML protocols for diagnosing cancer in prospective settings.

Methods

PubMed was searched for studies reporting usage of AI/ML protocols for cancer diagnosis in prospective (clinical trial/real world) setting with the AI/ML diagnosis aiding clinical decision-making, from inception till May 17, 2021. Data pertaining to the cancer, patients and the AI/ML protocol were extracted. Comparison of AI/ML protocol diagnosis with human diagnosis was recorded. Through a post hoc analysis, data from studies describing validation of various AI/ML protocols were extracted.

Results

Only 18/960 initial hits (1.88%) utilized AI/ML protocols for diagnostic decision-making. Most protocols used artificial neural network and deep learning. AI/ML protocols were utilized for cancer screening, pre-operative diagnosis and staging and intra-operative diagnosis of surgical specimens. The reference standard for 17/18 studies was histology. AI/ML protocols were used to diagnose cancers of the colorectum, skin, uterine cervix, oral cavity, ovaries, prostate, lungs and brain. AI/ML protocols were found to improve human diagnosis, and had either similar or better performance than the human diagnosis, especially made by the less experienced clinician. Validation of AI/ML protocols was described by 223 studies of which only four studies were from India. Also there was a huge variation in the number of items used for validation.

Interpretation & conclusions

The findings of this review suggest that a meaningful translation from the validation of AI/ML protocols to their actual usage in cancer diagnosis is lacking. Development of regulatory framework specific for AI/ML usage in healthcare is essential.

Neurooncology: 2023 update

This article presents some of the author's neuropathological highlights in the field on neuro-oncology research encountered in 2022. Major advances were made in the development of more precise, faster, easier, less invasive and unbiased diagnostic tools ranging from immunohistochemical prediction of 1p/19q loss in diffuse glioma, methylation analyses in CSF samples, molecular profiling for CNS lymphoma, proteomic analyses of recurrent glioblastoma, integrated molecular diagnostics for better stratification in meningioma, intraoperative profiling making use of Raman effect or methylation analysis, to finally, the assessment of histological slides by means of machine learning for the prediction of molecular tumor features. In addition, as the discovery of a new tumor entity may also be a highlight for the neuropathology community, the newly described high-grade glioma with pleomorphic and pseudopapillary features (HPAP) has been selected for this article. Regarding new innovative treatment approaches, a drug screening platform for brain metastasis is presented. Although diagnostic speed and precision is steadily increasing, clinical prognosis for patients with malignant tumors affecting the nervous system remains largely unchanged over the last decade, therefore future neuro-oncological research focus should be put on how the amazing developments presented in this article can be more sustainably applied to positively impact patient prognosis.

Deep learning-based high-speed, large-field, and high-resolution multiphoton imaging

Multiphoton microscopy is a formidable tool for the pathological analysis of tumors. The physical limitations of imaging systems and the low efficiencies inherent in nonlinear processes have prevented the simultaneous achievement of high imaging speed and high resolution. We demonstrate a self-alignment dual-attention-guided residual-in-residual generative adversarial network trained with various multiphoton images. The network enhances image contrast and spatial resolution, suppresses noise, and scanning fringe artifacts, and eliminates the mutual exclusion between field of view, image quality, and imaging speed. The network may be integrated into commercial microscopes for large-scale, high-resolution, and low photobleaching studies of tumor environments.

Histological diagnosis of unprocessed breast core-needle biopsy via stimulated Raman scattering microscopy and multi-instance learning

Core-needle biopsy (CNB) plays a vital role in the initial diagnosis of breast cancer. However, the complex tissue processing and global shortage of pathologists have hindered traditional histopathology from timely diagnosis on fresh biopsies. In this work, we developed a full digital platform by integrating label-free stimulated Raman scattering (SRS) microscopy with weakly-supervised learning for rapid and automated cancer diagnosis on un-labelled breast CNB. Methods: We first compared the results of SRS imaging with standard hematoxylin and eosin (H&E) staining on adjacent frozen tissue sections. Then fresh unprocessed biopsy tissues were imaged by SRS to reveal diagnostic histoarchitectures. Next, weakly-supervised learning, i.e., the multi-instance learning (MIL) model was conducted to evaluate the ability to differentiate between benign and malignant cases, and compared with the performance of supervised learning model. Finally, gradient-weighted class activation mapping (Grad-CAM) and semantic segmentation were performed to spatially resolve benign/malignant areas with high efficiency. Results: We verified the ability of SRS in revealing essential histological hallmarks of breast cancer in both thin frozen sections and fresh unprocessed biopsy, generating histoarchitectures well correlated with H&E staining. Moreover, we demonstrated that weakly-supervised MIL model could achieve superior classification performance to supervised learnings, reaching diagnostic accuracy of 95% on 61 biopsy specimens. Furthermore, Grad-CAM allowed the trained MIL model to visualize the histological heterogeneity within the CNB. Conclusion: Our results indicate that MIL-assisted SRS microscopy provides rapid and accurate diagnosis on histologically heterogeneous breast CNB, and could potentially help the subsequent management of patients.

Machine learning can aid in prediction of IDH mutation from H&E-stained histology slides in infiltrating gliomas

While Machine Learning (ML) models have been increasingly applied to a range of histopathology tasks, there has been little emphasis on characterizing these models and contrasting them with human experts. We present a detailed empirical analysis comparing expert neuropathologists and ML models at predicting IDH mutation status in H&E-stained histology slides of infiltrating gliomas, both independently and synergistically. We find that errors made by neuropathologists and ML models trained using the TCGA dataset are distinct, representing modest agreement between predictions (human-vs.-human κ = 0.656; human-vs.-ML model κ = 0.598). While no ML model surpassed human performance on an independent institutional test dataset (human AUC = 0.901, max ML AUC = 0.881), a hybrid model aggregating human and ML predictions demonstrates predictive performance comparable to the consensus of two expert neuropathologists (hybrid classifier AUC = 0.921 vs. two-neuropathologist consensus AUC = 0.920). We also show that models trained at different levels of magnification exhibit different types of errors, supporting the value of aggregation across spatial scales in the ML approach. Finally, we present a detailed interpretation of our multi-scale ML ensemble model which reveals that predictions are driven by human-identifiable features at the patch-level.

Raman Spectroscopy on Brain Disorders: Transition from Fundamental Research to Clinical Applications

Brain disorders such as brain tumors and neurodegenerative diseases (NDs) are accompanied by chemical alterations in the tissues. Early diagnosis of these diseases will provide key benefits for patients and opportunities for preventive treatments. To detect these sophisticated diseases, various imaging modalities have been developed such as computed tomography (CT), magnetic resonance imaging (MRI), and positron emission tomography (PET). However, they provide inadequate molecule-specific information. In comparison, Raman spectroscopy (RS) is an analytical tool that provides rich information about molecular fingerprints. It is also inexpensive and rapid compared to CT, MRI, and PET. While intrinsic RS suffers from low yield, in recent years, through the adoption of Raman enhancement technologies and advanced data analysis approaches, RS has undergone significant advancements in its ability to probe biological tissues, including the brain. This review discusses recent clinical and biomedical applications of RS and related techniques applicable to brain tumors and NDs.

Deep Learning with a Dataset Created Using Kanno Saitama Macro, a Self-Made Automatic Foveal Avascular Zone Extraction Program

The extraction of the foveal avascular zone (FAZ) from optical coherence tomography angiography (OCTA) images has been used in many studies in recent years due to its association with various ophthalmic diseases. In this study, we investigated the utility of a dataset for deep learning created using Kanno Saitama Macro (KSM), a program that automatically extracts the FAZ using swept-source OCTA. The test data included 40 eyes of 20 healthy volunteers. For training and validation, we used 257 eyes from 257 patients. The FAZ of the retinal surface image was extracted using KSM, and a dataset for FAZ extraction was created. Based on that dataset, we conducted a training test using a typical U-Net. Two examiners manually extracted the FAZ of the test data, and the results were used as gold standards to compare the Jaccard coefficients between examiners, and between each examiner and the U-Net. The Jaccard coefficient was 0.931 between examiner 1 and examiner 2, 0.951 between examiner 1 and the U-Net, and 0.933 between examiner 2 and the U-Net. The Jaccard coefficients were significantly better between examiner 1 and the U-Net than between examiner 1 and examiner 2 (p < 0.001). These data indicated that the dataset generated by KSM was as good as, if not better than, the agreement between examiners using the manual method. KSM may contribute to reducing the burden of annotation in deep learning.

Innovation in Neurosurgery: Lessons Learned, Obstacles, and Potential Funding Sources

Innovation is central to neurosurgery and has dramatically increased over the last twenty years. Although the specialty innovates as a whole, only 3-4.7% of practicing neurosurgeons hold patents. Various roadblocks to innovation impede this process such as lack of understanding, increasing regulatory complexity, and lack of funding. Newly emerging technologies allow us to understand how to innovate and how to learn from other medical specialties. By further understanding the process of innovation, and the funding that supports it, Neurosurgery can continue to hold innovation as one of its's central tenets.

Deep learning-based age estimation from chest X-rays indicates cardiovascular prognosis

BACKGROUND

In recent years, there has been considerable research on the use of artificial intelligence to estimate age and disease status from medical images. However, age estimation from chest X-ray (CXR) images has not been well studied and the clinical significance of estimated age has not been fully determined.

METHODS

To address this, we trained a deep neural network (DNN) model using more than 100,000 CXRs to estimate the patients' age solely from CXRs. We applied our DNN to CXRs of 1562 consecutive hospitalized heart failure patients, and 3586 patients admitted to the intensive care unit with cardiovascular disease.

RESULTS

The DNN's estimated age (X-ray age) showed a strong significant correlation with chronological age on the hold-out test data and independent test data. Elevated X-ray age is associated with worse clinical outcomes (heart failure readmission and all-cause death) for heart failure. Additionally, elevated X-ray age was associated with a worse prognosis in 3586 patients admitted to the intensive care unit with cardiovascular disease.

CONCLUSIONS

Our results suggest that X-ray age can serve as a useful indicator of cardiovascular abnormalities, which will help clinicians to predict, prevent and manage cardiovascular diseases.

Advances, technological innovations, and future prospects in stereotactic brain biopsies

Stereotactic brain biopsy is one of the most frequently performed brain surgeries. This review aimed to expose the latest cutting-edge and updated technologies and innovations available to neurosurgeons to safely perform stereotactic brain biopsy by minimizing the risks of complications and ensuring that the procedure is successful, leading to a histological diagnosis. We also examined methods for improving preoperative, intraoperative, and postoperative workflows. We performed a comprehensive state-of-the-art literature review. Intraoperative histology, fluorescence, and imaging techniques appear as smart tools to improve the diagnostic yield of biopsy. Constant innovations such as optical methods and augmented reality are also being made to increase patient safety. Robotics and integrated imaging techniques provide an enhanced intraoperative workflow. Patients' management algorithms based on early discharge after biopsy optimize the patient's personal experience and make the most efficient possible use of the available hospital resources. Many new trends are emerging, constantly improving patient care and safety, as well as surgical workflow. A parameter that must be considered is the cost-effectiveness of these devices and the possibility of using them on a daily basis. The decision to implement a new instrument in the surgical workflow should also be dependent on the number of procedures per year, the existing stereotactic equipment, and the experience of each center. Research on patients' postbiopsy management is another mandatory approach to enhance the safety profile of stereotactic brain biopsy and patient satisfaction, as well as to reduce healthcare costs.

Impact of preprocessing methods on the Raman spectra of brain tissue

Delineating cancer tissue while leaving functional tissue intact is crucial in brain tumor resection. Despite several available aids, surgeons are limited by preoperative or subjective tools. Raman spectroscopy is a label-free optical technique with promising indications for tumor tissue identification. To allow direct comparisons between measurements preprocessing of the Raman signal is required. There are many recognized methods for preprocessing Raman spectra; however, there is no universal standard. In this paper, six different preprocessing methods were tested on Raman spectra (n > 900) from fresh brain tissue samples (n = 34). The sample cohort included both primary brain tumors, such as adult-type diffuse gliomas and meningiomas, as well as metastases of breast cancer. Each tissue sample was classified according to the CNS WHO 2021 guidelines. The six methods include both direct and iterative polynomial fitting, mathematical morphology, signal derivative, commercial software, and a neural network. Data exploration was performed using principal component analysis, t-distributed stochastic neighbor embedding, and k-means clustering. For each of the six methods, the parameter combination that explained the most variance in the data, i.e., resulting in the highest Gap-statistic, was chosen and compared to the other five methods. Depending on the preprocessing method, the resulting clusters varied in number, size, and associated spectral features. The detected features were associated with hemoglobin, neuroglobin, carotenoid, water, and protoporphyrin, as well as proteins and lipids. However, the spectral features seen in the Raman spectra could not be unambiguously assigned to tissue labels, regardless of preprocessing method. We have illustrated that depending on the chosen preprocessing method, the spectral appearance of Raman features from brain tumor tissue can change. Therefore, we argue both for caution in comparing spectral features from different Raman studies, as well as the importance of transparency of methodology and implementation of the preprocessing. As discussed in this study, Raman spectroscopy for in vivo guidance in neurosurgery requires fast and adaptive preprocessing. On this basis, a pre-trained neural network appears to be a promising approach for the operating room.

Early repeat resection for residual glioblastoma: decision-making among an international cohort of neurosurgeons

OBJECTIVE

The importance of extent of resection (EOR) in glioblastoma (GBM) has been thoroughly demonstrated. However, few studies have explored the practices and benefits of early repeat resection (ERR) when residual tumor deemed resectable is unintentionally left after an initial resection, and the survival benefit of ERR is still unknown. Herein, the authors aimed to internationally survey current practices regarding ERR and to analyze differences based on geographic location and practice setting.

METHODS

The authors distributed a survey to the American Association of Neurological Surgeons and Congress of Neurological Surgeons Tumor Section, Society of British Neurological Surgeons, European Association of Neurosurgical Society, and Latin American Federation of Neurosurgical Societies. Neurosurgeons responded to questions about their training, practice setting, and current ERR practices. They also reported the EOR threshold below which they would pursue ERR and their likelihood of performing ERR using a Likert scale of 1-5 (5 being the most likely) in two sets of 5 cases, the first set for a patient's initial hospitalization and the second for a referred patient who had undergone resection elsewhere. The resection likelihood index for each respondent was calculated as the mean Likert score across all cases.

RESULTS

Overall, 180 neurosurgeons from 25 countries responded to the survey. Neurosurgeons performed ERRs very rarely in their practices (< 1% of all GBM cases), with an EOR threshold of 80.2% (75%-95%). When presented with 10 cases, the case context (initial hospitalization vs referred patient) did not significantly change the surgeon ERR likelihood, although ERR likelihood did vary significantly on the basis of tumor location (p < 0.0001). Latin American neurosurgeons were more likely to pursue ERR in the provided cases. Neurosurgeons were more likely to pursue ERR when the tumor was MGMT methylated versus unmethylated, with a resection likelihood index of 3.78 and 3.21, respectively (p = 0.004); however, there was no significant difference between IDH mutant and IDH wild-type tumors.

CONCLUSIONS

Results of this survey reveal current practices regarding ERR, but they also demonstrate the variability in how neurosurgeons approach ERR. Standardized guidelines based on future studies incorporating tumor molecular characteristics are needed to guide neurosurgeons in their decision-making on this complicated issue.

Toward digital histopathological assessment in surgery for central nervous system tumors using stimulated Raman histology

OBJECTIVE

Intraoperative neuropathological assessment with conventional frozen sections supports the neurosurgeon in optimizing the surgical strategy. However, preparation and review of frozen sections can take as long as 45 minutes. Stimulated Raman histology (SRH) was introduced as a novel technique to provide rapid high-resolution digital images of unprocessed tissue samples directly in the operating room that are comparable to conventional histopathological images. Additionally, SRH images are simultaneously and easily accessible for neuropathological judgment. Recently, the first study showed promising results regarding the accuracy and feasibility of SRH compared with conventional histopathology. Thus, the aim of this study was to compare SRH with conventional H&E images and frozen sections in a large cohort of patients with different suspected central nervous system (CNS) tumors.

METHODS

The authors included patients who underwent resection or stereotactic biopsy of suspected CNS neoplasm, including brain and spinal tumors. Intraoperatively, tissue samples were safely collected and SRH analysis was performed directly in the operating room. To enable optimal comparison of SRH with H&E images and frozen sections, the authors created a digital databank that included images obtained with all 3 imaging modalities. Subsequently, 2 neuropathologists investigated the diagnostic accuracy, tumor cellularity, and presence of diagnostic histopathological characteristics (score 0 [not present] through 3 [excellent]) determined with SRH images and compared these data to those of H&E images and frozen sections, if available.

RESULTS

In total, 94 patients with various suspected CNS tumors were included, and the application of SRH directly in the operating room was feasible in all cases. The diagnostic accuracy based on SRH images was 99% when compared with the final histopathological diagnosis based on H&E images. Additionally, the same histopathological diagnosis was established in all SRH images (100%) when compared with that of the corresponding frozen sections. Moreover, the authors found a statistically significant correlation in tumor cellularity between SRH images and corresponding H&E images (p < 0.0005 and R = 0.867, Pearson correlation coefficient). Finally, excellent (score 3) or good (2) accordance between diagnostic histopathological characteristics and H&E images was present in 95% of cases.

CONCLUSIONS

The results of this retrospective analysis demonstrate the near-perfect diagnostic accuracy and capability of visualizing relevant histopathological characteristics with SRH compared with conventional H&E staining and frozen sections. Therefore, digital SRH histopathology seems especially useful for rapid intraoperative investigation to confirm the presence of diagnostic tumor tissue and the precise tumor entity, as well as to rapidly analyze multiple tissue biopsies from the suspected tumor margin. A real-time analysis comparing SRH images and conventional histological images at the time of surgery should be performed as the next step in future studies.

The diagnostic accuracy of intraoperative differentiation and delineation techniques in brain tumours

Brain tumour identification and delineation in a timeframe of seconds would significantly guide and support surgical decisions. Here, treatment is often complicated by the infiltration of gliomas in the surrounding brain parenchyma. Accurate delineation of the invasive margins is essential to increase the extent of resection and to avoid postoperative neurological deficits. Currently, histopathological annotation of brain biopsies and genetic phenotyping still define the first line treatment, where results become only available after surgery. Furthermore, adjuvant techniques to improve intraoperative visualisation of the tumour tissue have been developed and validated. In this review, we focused on the sensitivity and specificity of conventional techniques to characterise the tumour type and margin, specifically fluorescent-guided surgery, neuronavigation and intraoperative imaging as well as on more experimental techniques such as mass spectrometry-based diagnostics, Raman spectrometry and hyperspectral imaging. Based on our findings, all investigated methods had their advantages and limitations, guiding researchers towards the combined use of intraoperative imaging techniques. This can lead to an improved outcome in terms of extent of tumour resection and progression free survival while preserving neurological outcome of the patients.

A deep learning model designed for Raman spectroscopy with a novel hyperparameter optimization method

Raman spectroscopy is a spectroscopic technique typically used to determine vibrational modes of molecules and is commonly used in chemistry to provide a structural fingerprint by which molecules can be identified. With the help of deep learning, Raman spectroscopy can be analyzed more efficiently and thus provide more accurate molecular information. However, no general neural network is designed for one-dimensional Raman spectral data so far. Furthermore, different combinations of hyperparameters of neural networks lead to results with significant differences, so the optimization of hyperparameters is a crucial issue in deep learning modeling. In this work, we propose a deep learning model designed for Raman spectral data and a hyperparameter optimization method to achieve its best performance, i.e., a method based on the simulated annealing algorithm to optimize the hyperparameters of the model. The proposed model and optimization method have been fully validated in a glioma Raman spectroscopy dataset. Compared with other published methods including linear regression, support vector regression, long short-term memory, VGG and ResNet, the mean squared error is reduced by 0.1557 while the coefficient determination is increased by 0.1195 on average.

Updates in intraoperative strategies for enhancing intra-axial brain tumor control

To ensure excellent postoperative clinical outcomes while preserving critical neurologic function, neurosurgeons who manage patients with intra-axial brain tumors can use intraoperative technologies and tools to achieve maximal safe resection. Neurosurgical oncology revolves around safe and optimal extent of resection, which further dictates subsequent treatment regimens and patient outcomes. Various methods can be adapted for treating both primary and secondary intra-axial brain lesions. We present a review of recent advances and published research centered on different innovative tools and techniques, including fluorescence-guided surgery, new methods of drug delivery, and minimally invasive procedural options.

Novel intraoperative strategies for enhancing tumor control: Future directions

Maximal safe surgical resection plays a key role in the care of patients with gliomas. A range of technologies have been developed to aid surgeons in distinguishing tumor from normal tissue, with the goal of increasing tumor resection and limiting postoperative neurological deficits. Technologies that are currently being investigated to aid in improving tumor control include intraoperative imaging modalities, fluorescent tumor makers, intraoperative cell and molecular profiling of tumors, improved microscopic imaging, intraoperative mapping, augmented and virtual reality, intraoperative drug and radiation delivery, and ablative technologies. In this review, we summarize the aforementioned advancements in neurosurgical oncology and implications for improving patient outcomes.

Stimulated Raman Histology for Rapid Intra-Operative Diagnosis of Sinonasal and Skull Base Tumors

OBJECTIVE

Intra-operative stimulated Raman histology (SRH) is a novel technology that uses laser spectroscopy and color-matching algorithms to create images similar to the formalin-fixed paraffin-embedded (FFPE) section. We aim to assess the accuracy of SRH in a novel range of sinonasal and skull base tumors.

METHODS

Select patients undergoing sinonasal and skull base surgery using the Invenio Imaging™ Nio™ Laser Imaging SRH system between June 2020 and September 2021 were assessed. The SRH images were reviewed for pathologic features similar to frozen section (FS) and FFPE. Time taken for results and diagnostic concordance was assessed.

RESULTS

Sixty-seven SRH images from 7 tumor types in 12 patients were assessed. Pathologies included squamous cell carcinoma, rhabdomyosarcoma, inverted papilloma, adenoid cystic carcinoma, SMARCB1-deficient sinonasal carcinoma, mucosal melanoma, metastatic colonic adenocarcinoma, and meningioma. Tumor was identified in 100% of lesional specimens, with characteristic diagnostic features readily appreciable on SRH. Median time for diagnosis was significantly faster for SRH (4.3 min) versus FS (44.5 min; p = <.0001). Where SRH sample site matched precisely to FS (n = 32/67, 47.8%), the same diagnosis was confirmed in 93.8%. Sensitivity, specificity, precision, and overall accuracy of SRH were 93.3%, 94.1%, 93.8%, and 93.3%, respectively. Near-perfect concordance was seen between SRH and FS (Cohen's kappa [κ] = 0.89).

CONCLUSION

Stimulated Raman histology can rapidly produce images similar to FFPE H&E in sinonasal and skull base tumors. This technology has the potential to act as an adjunct or alternative to standard FS.

LEVEL OF EVIDENCE

4 Laryngoscope, 132:2142-2147, 2022.

Evaluation of an Integrated Spectroscopy and Classification Platform for Point-of-Care Core Needle Biopsy Assessment: Performance Characteristics from Ex Vivo Renal Mass Biopsies

PURPOSE

To evaluate a transmission optical spectroscopy instrument for rapid ex vivo assessment of core needle cancer biopsies (CNBs) at the point of care.

MATERIALS AND METHODS

CNBs from surgically resected renal tumors and nontumor regions were scanned on their sampling trays with a custom spectroscopy instrument. After extracting principal spectral components, machine learning was used to train logistic regression, support vector machines, and random decision forest (RF) classifiers on 80% of randomized and stratified data. The algorithms were evaluated on the remaining 20% of the data set held out during training. Binary classification (tumor/nontumor) was performed based on a decision threshold. Multinomial classification was also performed to differentiate between the subtypes of renal cell carcinoma (RCC) and account for potential confounding effects from fat, blood, and necrotic tissue. Classifiers were compared based on sensitivity, specificity, and positive predictive value (PPV) relative to a histopathologic standard.

RESULTS

A total of 545 CNBs from 102 patients were analyzed, yielding 5,583 spectra after outlier exclusion. At the individual spectra level, the best performing algorithm was RF with sensitivities of 96% and 92% and specificities of 90% and 89%, for the binary and multiclass analyses, respectively. At the full CNB level, RF algorithm also showed the highest sensitivity and specificity (93% and 91%, respectively). For RCC subtypes, the highest sensitivity and PPV were attained for clear cell (93.5%) and chromophobe (98.2%) subtypes, respectively.

CONCLUSIONS

Ex vivo spectroscopy imaging paired with machine learning can accurately characterize renal mass CNB at the time of tissue acquisition.

Investigation of the Spatial Generation of Stimulated Raman Scattering Using Computer Simulation and Experimentation

Stimulated Raman scattering is a phenomenon with potential use in providing real-time molecular information in three-dimensions (3D) of a sample using imaging. For precise imaging, the knowledge about the spatial generation of stimulated Raman scattering is essential. To investigate the spatial behavior in an idealized case, computer simulations and experiments were performed. For the computer simulations, diffraction theory was used for the beam propagation complemented with nonlinear phase modulation describing the interaction between the light and matter. For the experiments, a volume of ethanol was illuminated by an expanded light beam and a plane inside the volume was imaged in transmission. For generating stimulated Raman scattering, a pump beam was focused into this volume and led to a beam dump after passing the volume. The pulse duration of the two beams were 6 ns and the pump beam energy ranged from 1 to 27 mJ. The effect of increasing pump power on the spatial distribution of the Raman gain and the spatial growth of the signal at different interaction lengths between the beam and the sample was investigated. The spatial width of the region where the stimulated Raman scattering signal was generated for experiments and simulation was 0.21 and 0.09 mm, respectively. The experimental and simulation results showed that most of the stimulated Raman scattering is generated close to the pump beam focus and the maximum peak of the Stokes intensity spatially comes shortly after the peak of the pump intensity.

Radiomics in neuro-oncological clinical trials

The development of clinical trials has led to substantial improvements in the prevention and treatment of many diseases, including brain cancer. Advances in medicine, such as improved surgical techniques, the development of new drugs and devices, the use of statistical methods in research, and the development of codes of ethics, have considerably influenced the way clinical trials are conducted today. In addition, methods from the broad field of artificial intelligence, such as radiomics, have the potential to considerably affect clinical trials and clinical practice in the future. Radiomics is a method to extract undiscovered features from routinely acquired imaging data that can neither be captured by means of human perception nor conventional image analysis. In patients with brain cancer, radiomics has shown its potential for the non-invasive identification of prognostic biomarkers, automated response assessment, and differentiation between treatment-related changes from tumour progression. Despite promising results, radiomics is not yet established in routine clinical practice nor in clinical trials. In this Viewpoint, the European Organization for Research and Treatment of Cancer Brain Tumour Group summarises the current status of radiomics, discusses its potential and limitations, envisions its future role in clinical trials in neuro-oncology, and provides guidance on how to address the challenges in radiomics.

Artificial intelligence for multimodal data integration in oncology

In oncology, the patient state is characterized by a whole spectrum of modalities, ranging from radiology, histology, and genomics to electronic health records. Current artificial intelligence (AI) models operate mainly in the realm of a single modality, neglecting the broader clinical context, which inevitably diminishes their potential. Integration of different data modalities provides opportunities to increase robustness and accuracy of diagnostic and prognostic models, bringing AI closer to clinical practice. AI models are also capable of discovering novel patterns within and across modalities suitable for explaining differences in patient outcomes or treatment resistance. The insights gleaned from such models can guide exploration studies and contribute to the discovery of novel biomarkers and therapeutic targets. To support these advances, here we present a synopsis of AI methods and strategies for multimodal data fusion and association discovery. We outline approaches for AI interpretability and directions for AI-driven exploration through multimodal data interconnections. We examine challenges in clinical adoption and discuss emerging solutions.

Innovations in the Diagnosis and Surgical Management of Low-Grade Gliomas

Low-grade gliomas are a broad category of tumors that can manifest at different stages of life. As a group, their prognosis has historically been considered to be favorable, and surgery is a mainstay of treatment. Advances in the molecular characterization of individual lesions has led to newer classification systems, a better understanding of the biological behavior of different neoplasms, and the identification of previously unrecognized entities. New prospective genetic and molecular data will help delineate better treatment paradigms and will continue to change the taxonomy of central nervous system tumors in the coming years. Advances in the field of radiomics will help predict the molecular profile of a particular tumor through noninvasive testing. Similarly, more precise methods of intraoperative tumor tissue analysis will aid surgical planning. Improved surgical outcomes propelled by novel surgical techniques and intraoperative adjuncts and emerging forms of medical treatment in the field of immunotherapy have enriched the management of these lesions. We review the contemporary management and innovations in the treatment of low-grade gliomas.

Challenges in the Diagnosis and Management of Low-Grade Gliomas

Low-grade gliomas are clinically challenging entities. Patients with these tumors tend to be relatively young at presentation, and lesions are often incidental findings or are identified because the patient presents with a seizure. Rapidly emerging and evolving molecular classifications of gliomas have influenced treatment paradigms. Importantly, low-grade gliomas can be classified on the basis of IDH mutation status, whereby low-grade astrocytomas harbor the IDH mutation, while oligodendrogliomas are defined by both IDH mutant status and 1p/19q co-deletion. Given the importance of molecular classification for diagnosis, treatment planning, and prognostication, tissue samples are necessary for proper management. Literature supports improved overall survival and outcomes with increased extent of resection for low-grade glioma. Awake craniotomies and resection of insular low-grade gliomas both have been demonstrated as safe and improve outcomes for patients with lesions located in eloquent areas. Given the younger age at diagnosis of these lesions compared with higher-grade gliomas, fertility, fertility preservation, and potential malignant transformation should be discussed with patients of childbearing age.

Intraoperative Glioma Grading Using Neural Architecture Search and Multi-Modal Imaging

Glioma grading during surgery can help clinical treatment planning and prognosis, but intraoperative pathological examination of frozen sections is limited by the long processing time and complex procedures. Near-infrared fluorescence imaging provides chances for fast and accurate real-time diagnosis. Recently, deep learning techniques have been actively explored for medical image analysis and disease diagnosis. However, issues of near-infrared fluorescence images, including small-scale, noise, and low-resolution, increase the difficulty of training a satisfying network. Multi-modal imaging can provide complementary information to boost model performance, but simultaneously designing a proper network and utilizing the information of multi-modal data is challenging. In this work, we propose a novel neural architecture search method DLS-DARTS to automatically search for network architectures to handle these issues. DLS-DARTS has two learnable stems for multi-modal low-level feature fusion and uses a modified perturbation-based derivation strategy to improve the performance on the area under the curve and accuracy. White light imaging and fluorescence imaging in the first near-infrared window (650-900 nm) and the second near-infrared window (1,000-1,700 nm) are applied to provide multi-modal information on glioma tissues. In the experiments on 1,115 surgical glioma specimens, DLS-DARTS achieved an area under the curve of 0.843 and an accuracy of 0.634, which outperformed manually designed convolutional neural networks including ResNet, PyramidNet, and EfficientNet, and a state-of-the-art neural architecture search method for multi-modal medical image classification. Our study demonstrates that DLS-DARTS has the potential to help neurosurgeons during surgery, showing high prospects in medical image analysis.

Multistrategy Improved Sparrow Search Algorithm Optimized Deep Neural Network for Esophageal Cancer

Deep neural network is a complex pattern recognition network system. It is widely favored by scholars for its strong nonlinear fitting ability. However, training deep neural network models on small datasets typically realizes worse performance than shallow neural network. In this study, a strategy to improve the sparrow search algorithm based on the iterative map, iterative perturbation, and Gaussian mutation is developed. This optimized strategy improved the sparrow search algorithm validated by fourteen benchmark functions, and the algorithm has the best search accuracy and the fastest convergence speed. An algorithm based on the iterative map, iterative perturbation, and Gaussian mutation improved sparrow search algorithm is designed to optimize deep neural networks. The modified sparrow algorithm is exploited to search for the optimal connection weights of deep neural network. This algorithm is implemented for the esophageal cancer dataset along with the other six algorithms. The proposed model is able to achieve 0.92 under all the eight scoring criteria, which is better than the performance of the other six algorithms. Therefore, an optimized deep neural network based on an improved sparrow search algorithm with iterative map, iterative perturbation, and Gaussian mutation is an effective approach to predict the survival rate of esophageal cancer.

Deep-Learning to Predict BRCA Mutation and Survival from Digital H&E Slides of Epithelial Ovarian Cancer

BRCA 1/2 genes mutation status can already determine the therapeutic algorithm of high grade serous ovarian cancer patients. Nevertheless, its assessment is not sufficient to identify all patients with genomic instability, since BRCA 1/2 mutations are only the most well-known mechanisms of homologous recombination deficiency (HR-d) pathway, and patients displaying HR-d behave similarly to BRCA mutated patients. HRd assessment can be challenging and is progressively overcoming BRCA testing not only for prognostic information but more importantly for drugs prescriptions. However, HR testing is not already integrated in clinical practice, it is quite expensive and it is not refundable in many countries. Selecting patients who are more likely to benefit from this assessment (BRCA 1/2 WT patients) at an early stage of the diagnostic process, would allow an optimization of genomic profiling resources. In this study, we sought to explore whether somatic BRCA1/2 genes status can be predicted using computational pathology from standard hematoxylin and eosin histology. In detail, we adopted a publicly available, deep-learning-based weakly supervised method that uses attention-based learning to automatically identify sub regions of high diagnostic value to accurately classify the whole slide (CLAM). The same model was also tested for progression free survival (PFS) prediction. The model was tested on a cohort of 664 (training set: n = 464, testing set: n = 132) ovarian cancer patients, of whom 233 (35.1%) had a somatic BRCA 1/2 mutation. An area under the curve of 0.7 and 0.55 was achieved in the training and testing set respectively. The model was then further refined by manually identifying areas of interest in half of the cases. 198 images were used for training (126/72) and 87 images for validation (55/32). The model reached a zero classification error on the training set, but the performance was 0.59 in terms of validation ROC AUC, with a 0.57 validation accuracy. Finally, when applied to predict PFS, the model achieved an AUC of 0.71, with a negative predictive value of 0.69, and a positive predictive value of 0.75. Based on these analyses, we have planned further steps of development such as proving a reference classification performance, exploring the hyperparameters space for training optimization, eventually tweaking the learning algorithms and the neural networks architecture for better suiting this specific task. These actions may allow the model to improve performances for all the considered outcomes.

Neural Networks-Based On-Site Dermatologic Diagnosis through Hyperspectral Epidermal Images

Cancer originates from the uncontrolled growth of healthy cells into a mass. Chromophores, such as hemoglobin and melanin, characterize skin spectral properties, allowing the classification of lesions into different etiologies. Hyperspectral imaging systems gather skin-reflected and transmitted light into several wavelength ranges of the electromagnetic spectrum, enabling potential skin-lesion differentiation through machine learning algorithms. Challenged by data availability and tiny inter and intra-tumoral variability, here we introduce a pipeline based on deep neural networks to diagnose hyperspectral skin cancer images, targeting a handheld device equipped with a low-power graphical processing unit for routine clinical testing. Enhanced by data augmentation, transfer learning, and hyperparameter tuning, the proposed architectures aim to meet and improve the well-known dermatologist-level detection performances concerning both benign-malignant and multiclass classification tasks, being able to diagnose hyperspectral data considering real-time constraints. Experiments show 87% sensitivity and 88% specificity for benign-malignant classification and specificity above 80% for the multiclass scenario. AUC measurements suggest classification performance improvement above 90% with adequate thresholding. Concerning binary segmentation, we measured skin DICE and IOU higher than 90%. We estimated 1.21 s, at most, consuming 5 Watts to segment the epidermal lesions with the U-Net++ architecture, meeting the imposed time limit. Hence, we can diagnose hyperspectral epidermal data assuming real-time constraints.

Diagnostics and theranostics of central nervous system diseases based on aggregation-induced emission luminogens

Central nervous system (CNS) diseases include Alzheimer's disease (AD), Parkinson's disease (PD), brain tumors, strokes, and other important diseases that are harmful and fatal to human beings. CNS diseases have the characteristics of high fatality rates, difficult diagnosis, and costly treatment. The diagnosis and treatment of CNS diseases by molecular imaging are usually limited by the depth of tissue penetration and the blood-brain barrier (BBB). Therefore, it is still a huge challenge to distinguish between the lesion and the surrounding parenchymal boundary with high sensitivity and specificity. Compared with traditional fluorophores with aggregation-caused quenching effect, luminogens with aggregation-induced emission (AIE) characteristics have strong near-infrared deep penetration, large Stokes shift, excellent biocompatibility, light stability, and desirable BBB permeability. In view of this, developing novel AIE-based materials for diagnostics and theranostics of CNS diseases is promising and of great significance. Herein, we highlight the recent research progress in this field with a special focus on near-infrared imaging and AIE nanorobots for CNS diseases. The design principle of AIE probes is discussed in detail, and the outlook is presented as well.

Next Generation Digital Pathology: Emerging Trends and Measurement Challenges for Molecular Pathology

Digital pathology is revolutionising the analysis of histological features and is becoming more and more widespread in both the clinic and research. Molecular pathology extends the tissue morphology information provided by conventional histopathology by providing spatially resolved molecular information to complement the structural information provided by histopathology. The multidimensional nature of the molecular data poses significant challenge for data processing, mining, and analysis. One of the key challenges faced by new and existing pathology practitioners is how to choose the most suitable molecular pathology technique for a given diagnosis. By providing a comparison of different methods, this narrative review aims to introduce the field of molecular pathology, providing a high-level overview of many different methods. Since each pixel of an image contains a wealth of molecular information, data processing in molecular pathology is more complex. The key data processing steps and variables, and their effect on the data, are also discussed.

PDBL: Improving Histopathological Tissue Classification With Plug-and-Play Pyramidal Deep-Broad Learning

Histopathological tissue classification is a simpler way to achieve semantic segmentation for the whole slide images, which can alleviate the requirement of pixel-level dense annotations. Existing works mostly leverage the popular CNN classification backbones in computer vision to achieve histopathological tissue classification. In this paper, we propose a super lightweight plug-and-play module, named Pyramidal Deep-Broad Learning (PDBL), for any well-trained classification backbone to improve the classification performance without a re-training burden. For each patch, we construct a multi-resolution image pyramid to obtain the pyramidal contextual information. For each level in the pyramid, we extract the multi-scale deep-broad features by our proposed Deep-Broad block (DB-block). We equip PDBL in three popular classification backbones, ShuffLeNetV2, EfficientNetb0, and ResNet50 to evaluate the effectiveness and efficiency of our proposed module on two datasets (Kather Multiclass Dataset and the LC25000 Dataset). Experimental results demonstrate the proposed PDBL can steadily improve the tissue-level classification performance for any CNN backbones, especially for the lightweight models when given a small among of training samples (less than 10%). It greatly saves the computational resources and annotation efforts. The source code is available at: https://github.com/linjiatai/PDBL.