Novel rapid intraoperative qualitative tumor detection by a residual convolutional neural network using label-free stimulated Raman scattering microscopy

Image Quality Assessment and Reliability Analysis of Artificial Intelligence-Based Tumor Classification of Stimulated Raman Histology of Tumor Biobank Samples

BACKGROUND

Stimulated Raman histology (SRH) is a label-free optical imaging method for rapid intraoperative analysis of fresh tissue samples. Analysis of SRH images using Convolutional Neural Networks (CNN) has shown promising results for predicting the main histopathological classes of neurooncological tumors. Due to the relatively low number of rare tumor representations in CNN training datasets, a valid prediction of rarer entities remains limited. To develop new reliable analysis tools, larger datasets and greater tumor variety are crucial. One way to accomplish this is through research biobanks storing frozen tumor tissue samples. However, there is currently no data available regarding the pertinency of previously frozen tissue samples for SRH analysis. The aim of this study was to assess image quality and perform a comparative reliability analysis of artificial intelligence-based tumor classification using SRH in fresh and frozen tissue samples.

METHODS

In a monocentric prospective study, tissue samples from 25 patients undergoing brain tumor resection were obtained. SRH was acquired in fresh and defrosted samples of the same specimen after varying storage durations at -80 °C. Image quality was rated by an experienced neuropathologist, and prediction of histopathological diagnosis was performed using two established CNNs.

RESULTS

The image quality of SRH in fresh and defrosted tissue samples was high, with a mean image quality score of 1.96 (range 1-5) for both groups. CNN analysis showed high internal consistency for histo-(Cα 0.95) and molecular (Cα 0.83) pathological tumor classification. The results were confirmed using a dataset with samples from the local tumor biobank (Cα 0.91 and 0.53).

CONCLUSIONS

Our results showed that SRH appears comparably reliable in fresh and frozen tissue samples, enabling the integration of tumor biobank specimens to potentially improve the diagnostic range and reliability of CNN prediction tools.

Accuracy of Raman spectroscopy in discriminating normal brain tissue from brain tumor: A systematic review and meta-analysis

IMPORTANCE

There are several methods of intraoperative tumor border identification, but none of them is perfect. There is a need of a new tool.

OBJECTIVE

Raman spectroscopy, being a noninvasive, requiring no tissue preparation, quick technique of substance structure identification, is a potential tool for intraoperative identification of brain tumor. This meta-analysis aimed to assess the accuracy of Raman spectroscopy in differentiation of normal brain tissue from brain tumor.

DATA SOURCES

PubMed, Google Scholar, Scopus and Web of Science databases were searched until October 1, 2024.

STUDY SELECTION

All English-language articles reporting efficacy and accuracy of Raman spectroscopy for brain tumor differentiation were analyzed, sufficient data to construct 2x2 table was extracted.

EXCLUSION CRITERIA

studies using data from national databases; reviews, conference abstracts, case studies, letters to the editor; studies with irrelevant or not sufficient data; not human tissue used in the experiment. 6112 records were found; after exclusion, the suitability of 64 full-text articles was evaluated. 18 studies were reviewed and included into the meta-analysis.

DATA EXTRACTION AND SYNTHESIS

The meta-analysis was performed in accordance with PRISMA guidelines and recommendations. Methodological quality was assessed according to the QUADAS-2 guidelines. Data were extracted by multiple observers and any discrepancies were resolved by discussion and consensus. Data were pooled using a random-effects model.

MAIN OUTCOME(S) AND MEASURE(S)

The primary outcome was pooled sensitivity, specificity and diagnostic odds ratio (DOR) for Raman spectroscopy.

RESULTS

The manuscript presents 18 studies which were used to calculate pooled values. The pooled sensitivity, specificity and pooled diagnostic odds ratio (DOR) of RS for discriminating glioma and normal brain tissues were 0,965, 0,738 and 61,305 respectively. For GBM the results were 0,948, 0,506 and 78,420 respectively. For meningioma pooled values were 0,896, 0,913, and 149,59. For metastases pooled values were 0,946, 0,862 and 133,90 respectively.

CONCLUSIONS AND RELEVANCE

Raman spectroscopy has a potential to serve as a tool for differentiation of brain tumor from normal brain tissue. Not only could it be helpful in distinguishing malignant lesion from benign with high sensitivity and specificity, but also indicate type of tumor. There is a need for more studies examining the accuracy of spectroscopy in differentiating brain tumors from healthy tissues, especially in vivo and in differentiation of brain tumor subtypes.

Artificial Intelligence-Assisted Stimulated Raman Histology: New Frontiers in Vibrational Tissue Imaging

Frozen section biopsy, introduced in the early 1900s, still remains the gold standard methodology for rapid histologic evaluations. Although a valuable tool, it is labor-, time-, and cost-intensive. Other challenges include visual and diagnostic variability, which may complicate interpretation and potentially compromise the quality of clinical decisions. Raman spectroscopy, with its high specificity and non-invasive nature, can be an effective tool for dependable and quick histopathology. The most promising modality in this context is stimulated Raman histology (SRH), a label-free, non-linear optical process which generates conventional H&E-like images in short time frames. SRH overcomes limitations of conventional Raman scattering by leveraging the qualities of stimulated Raman scattering (SRS), wherein the energy gets transferred from a high-power pump beam to a probe beam, resulting in high-energy, high-intensity scattering. SRH's high resolution and non-requirement of preprocessing steps make it particularly suitable when it comes to intrasurgical histology. Combining SRH with artificial intelligence (AI) can lead to greater precision and less reliance on manual interpretation, potentially easing the burden of the overburdened global histopathology workforce. We review the recent applications and advances in SRH and how it is tapping into AI to evolve as a revolutionary tool for rapid histologic analysis.

Streamlined Intraoperative Brain Tumor Classification and Molecular Subtyping in Stereotactic Biopsies Using Stimulated Raman Histology and Deep Learning

PURPOSE

Recent artificial intelligence algorithms aided intraoperative decision-making via stimulated Raman histology (SRH) during craniotomy. This study assesses deep learning algorithms for rapid intraoperative diagnosis from SRH images in small stereotactic-guided brain biopsies. It defines a minimum tissue sample size threshold to ensure diagnostic accuracy.

EXPERIMENTAL DESIGN

A prospective single-center study examined 121 SRH images from 84 patients with unclear intracranial lesions undergoing stereotactic brain biopsy. Unprocessed, label-free samples were imaged using a portable fiber laser Raman scattering microscope. Three deep learning models were tested to (i) identify tumorous/nontumorous tissue as qualitative biopsy control; (ii) subclassify into high-grade glioma (central nervous system World Health Organization grade 4), diffuse low-grade glioma (central nervous system World Health Organization grades 2-3), metastases, lymphoma, or gliosis; and (iii) molecularly subtype IDH and 1p/19q statuses of adult-type diffuse gliomas. Model predictions were evaluated against frozen section analysis and final neuropathologic diagnoses.

RESULTS

The first model identified tumorous/nontumorous tissue with 91.7% accuracy. Sample size on slides impacted accuracy in brain tumor subclassification (81.6%, κ = 0.72 frozen section; 73.9%, κ = 0.61 second model), with SRH images being smaller than hematoxylin and eosin images (4.1 ± 2.5 mm2 vs. 16.7 ± 8.2 mm2, P < 0.001). SRH images with more than 140 high-quality patches and a mean squeezed sample of 5.26 mm2 yielded 89.5% accuracy in subclassification and 93.9% in molecular subtyping of adult-type diffuse gliomas.

CONCLUSIONS

Artificial intelligence-based SRH image analysis is non-inferior to frozen section analysis in detecting and subclassifying brain tumors during small stereotactic-guided biopsies once a critical squeezed sample size is reached. Beyond frozen section analysis, it enables valid molecular glioma subtyping, allowing faster treatment decisions in the future; however, refinement is needed for long-term application.

Fast intraoperative detection of primary CNS lymphoma and differentiation from common CNS tumors using stimulated Raman histology and deep learning

Accurate intraoperative diagnosis is crucial for differentiating between primary CNS lymphoma (PCNSL) and other CNS entities, guiding surgical decision-making, but represents significant challenges due to overlapping histomorphological features, time constraints, and differing treatment strategies. We combined stimulated Raman histology (SRH) with deep learning to address this challenge. We imaged unprocessed, label-free tissue samples intraoperatively using a portable Raman scattering microscope, generating virtual H&E-like images within less than three minutes. We developed a deep learning pipeline called RapidLymphoma based on a self-supervised learning strategy to (1) detect PCNSL, (2) differentiate from other CNS entities, and (3) test the diagnostic performance in a prospective international multicenter cohort and two additional independent test cohorts. We trained on 54,000 SRH patch images sourced from surgical resections and stereotactic-guided biopsies, including various CNS tumor/non-tumor lesions. Training and test data were collected from four tertiary international medical centers. The final histopathological diagnosis served as ground-truth. In the prospective test cohort of PCNSL and non-PCNSL entities (n=160), RapidLymphoma achieved an overall balanced accuracy of 97.81% ±0.91, non-inferior to frozen section analysis in detecting PCNSL (100% vs. 78.94%). The additional test cohorts (n=420, n=59) reached balanced accuracy rates of 95.44% ±0.74 and 95.57% ±2.47 in differentiating IDH-wildtype diffuse gliomas and various brain metastasis from PCNSL. Visual heatmaps revealed RapidLymphoma's capabilities to detect class-specific histomorphological key features. RapidLymphoma is valid and reliable in detecting PCNSL and differentiating from other CNS entities within three minutes, as well as visual feedback in an intraoperative setting. This leads to fast clinical decision-making and further treatment strategy planning.

Practical Application of Deep Learning in Diagnostic Neuropathology-Reimagining a Histological Asset in the Era of Precision Medicine

In the past few decades, neuropathology has experienced several paradigm shifts with the introduction of new technologies. Deep learning, a rapidly progressing subfield of machine learning, seems to be the next innovation to alter the diagnostic workflow. In this review, we will explore the recent changes in the field of neuropathology and how this has led to an increased focus on molecular features in diagnosis and prognosis. Then, we will examine the work carried out to train deep learning models for various diagnostic tasks in neuropathology, as well as the machine learning frameworks they used. Focus will be given to both the challenges and successes highlighted therein, as well as what these trends may tell us about future roadblocks in the widespread adoption of this new technology. Finally, we will touch on recent trends in deep learning, as applied to digital pathology more generally, and what this may tell us about the future of deep learning applications in neuropathology.

Virtual Staining of Nonfixed Tissue Histology

Surgical pathology workflow involves multiple labor-intensive steps, such as tissue removal, fixation, embedding, sectioning, staining, and microscopic examination. This process is time-consuming and costly and requires skilled technicians. In certain clinical scenarios, such as intraoperative consultations, there is a need for faster histologic evaluation to provide real-time surgical guidance. Currently, frozen section techniques involving hematoxylin and eosin (H&E) staining are used for intraoperative pathology consultations. However, these techniques have limitations, including a turnaround time of 20 to 30 minutes, staining artifacts, and potential tissue loss, negatively impacting accurate diagnosis. To address these challenges, researchers are exploring alternative optical imaging modalities for rapid microscopic tissue imaging. These modalities differ in optical characteristics, tissue preparation requirements, imaging equipment, and output image quality and format. Some of these imaging methods have been combined with computational algorithms to generate H&E-like images, which could greatly facilitate their adoption by pathologists. Here, we provide a comprehensive, organ-specific review of the latest advancements in emerging imaging modalities applied to nonfixed human tissue. We focused on studies that generated H&E-like images evaluated by pathologists. By presenting up-to-date research progress and clinical utility, this review serves as a valuable resource for scholars and clinicians, covering some of the major technical developments in this rapidly evolving field. It also offers insights into the potential benefits and drawbacks of alternative imaging modalities and their implications for improving patient care.

[Intraoperative stimulated Raman histology for personalized brain tumor surgery]

BACKGROUND

In brain tumor surgery a personalized surgical approach is crucial to achieve a maximum safe tumor resection. The extent of resection decisively depends on the histological diagnosis. Stimulated Raman histology (SRH), a fiber laser-based optical imaging method, offers the possibility for evaluation of an intraoperative diagnosis in a few minutes.

OBJECTIVE

To provide an overview on the applications of SRH in neurosurgery and transference of the technique to other surgical disciplines.

METHODS

Description of the technique and review of the current literature on SRH.

RESULTS

The SRH technique was successfully used in multiple neuro-oncological tumor entities. Initial pilot projects showed the potential for analysis of extracranial tumors.

CONCLUSION

The use of SRH provides a near real-time diagnosis with high diagnostic accuracy and provides further developmental potential to improve personalized tumor surgery.

Machine Learning in Neurosurgery: Toward Complex Inputs, Actionable Predictions, and Generalizable Translations

Machine learning can predict neurosurgical diagnosis and outcomes, power imaging analysis, and perform robotic navigation and tumor labeling. State-of-the-art models can reconstruct and generate images, predict surgical events from video, and assist in intraoperative decision-making. In this review, we will detail the neurosurgical applications of machine learning, ranging from simple to advanced models, and their potential to transform patient care. As machine learning techniques, outputs, and methods become increasingly complex, their performance is often more impactful yet increasingly difficult to evaluate. We aim to introduce these advancements to the neurosurgical audience while suggesting major potential roadblocks to their safe and effective translation. Unlike the previous generation of machine learning in neurosurgery, the safe translation of recent advancements will be contingent on neurosurgeons' involvement in model development and validation.

Advances in Intraoperative Glioma Tissue Sampling and Infiltration Assessment

Gliomas are infiltrative brain tumors that often involve functional tissue. While maximal safe resection is critical for maximizing survival, this is challenged by the difficult intraoperative discrimination between tumor-infiltrated and normal structures. Surgical expertise is essential for identifying safe margins, and while the intraoperative pathological review of frozen tissue is possible, this is a time-consuming task. Advances in intraoperative stimulation mapping have aided surgeons in identifying functional structures and, as such, has become the gold standard for this purpose. However, intraoperative margin assessment lacks a similar consensus. Nonetheless, recent advances in intraoperative imaging techniques and tissue examination methods have demonstrated promise for the accurate and efficient assessment of tumor infiltration and margin delineation within the operating room, respectively. In this review, we describe these innovative technologies that neurosurgeons should be aware of.

Rapid, label-free classification of glioblastoma differentiation status combining confocal Raman spectroscopy and machine learning

Label-free identification of tumor cells using spectroscopic assays has emerged as a technological innovation with a proven ability for rapid implementation in clinical care. Machine learning facilitates the optimization of processing and interpretation of extensive data, such as various spectroscopy data obtained from surgical samples. The here-described preclinical work investigates the potential of machine learning algorithms combining confocal Raman spectroscopy to distinguish non-differentiated glioblastoma cells and their respective isogenic differentiated phenotype by means of confocal ultra-rapid measurements. For this purpose, we measured and correlated modalities of 1146 intracellular single-point measurements and sustainingly clustered cell components to predict tumor stem cell existence. By further narrowing a few selected peaks, we found indicative evidence that using our computational imaging technology is a powerful approach to detect tumor stem cells in vitro with an accuracy of 91.7% in distinct cell compartments, mainly because of greater lipid content and putative different protein structures. We also demonstrate that the presented technology can overcome intra- and intertumoral cellular heterogeneity of our disease models, verifying the elevated physiological relevance of our applied disease modeling technology despite intracellular noise limitations for future translational evaluation.

Intraoperative microscopic autofluorescence detection and characterization in brain tumors using stimulated Raman histology and two-photon fluorescence

Introduction

The intrinsic autofluorescence of biological tissues interferes with the detection of fluorophores administered for fluorescence guidance, an emerging auxiliary technique in oncological surgery. Yet, autofluorescence of the human brain and its neoplasia is sparsely examined. This study aims to assess autofluorescence of the brain and its neoplasia on a microscopic level by stimulated Raman histology (SRH) combined with two-photon fluorescence.

Methods

With this experimentally established label-free microscopy technique unprocessed tissue can be imaged and analyzed within minutes and the process is easily incorporated in the surgical workflow. In a prospective observational study, we analyzed 397 SRH and corresponding autofluorescence images of 162 samples from 81 consecutive patients that underwent brain tumor surgery. Small tissue samples were squashed on a slide for imaging. SRH and fluorescence images were acquired with a dual wavelength laser (790 nm and 1020 nm) for excitation. In these images tumor and non-tumor regions were identified by a convolutional neural network that reliably differentiates between tumor, healthy brain tissue and low quality SRH images. The identified areas were used to define regions.of- interests (ROIs) and the mean fluorescence intensity was measured.

Results

In healthy brain tissue, we found an increased mean autofluorescence signal in the gray (11.86, SD 2.61, n=29) compared to the white matter (5.99, SD 5.14, n=11, p<0.01) and in the cerebrum (11.83, SD 3.29, n=33) versus the cerebellum (2.82, SD 0.93, n=7, p<0.001), respectively. The signal of carcinoma metastases, meningiomas, gliomas and pituitary adenomas was significantly lower (each p<0.05) compared to the autofluorescence in the cerebrum and dura, and significantly higher (each p<0.05) compared to the cerebellum. Melanoma metastases were found to have a higher fluorescent signal (p<0.01) compared to cerebrum and cerebellum.

Discussion

In conclusion we found that autofluorescence in the brain varies depending on the tissue type and localization and differs significantly among various brain tumors. This needs to be considered for interpreting photon signal during fluorescence-guided brain tumor surgery.

Technical report: surgical preparation of human brain tissue for clinical and basic research

BACKGROUND

The study of the distinct structure and function of the human central nervous system, both in healthy and diseased states, is becoming increasingly significant in the field of neuroscience. Typically, cortical and subcortical tissue is discarded during surgeries for tumors and epilepsy. Yet, there is a strong encouragement to utilize this tissue for clinical and basic research in humans. Here, we describe the technical aspects of the microdissection and immediate handling of viable human cortical access tissue for basic and clinical research, highlighting the measures needed to be taken in the operating room to ensure standardized procedures and optimal experimental results.

METHODS

In multiple rounds of experiments (n = 36), we developed and refined surgical principles for the removal of cortical access tissue. The specimens were immediately immersed in cold carbogenated N-methyl-D-glucamine-based artificial cerebrospinal fluid for electrophysiology and electron microscopy experiments or specialized hibernation medium for organotypic slice cultures.

RESULTS

The surgical principles of brain tissue microdissection were (1) rapid preparation (<1 min), (2) maintenance of the cortical axis, (3) minimization of mechanical trauma to sample, (4) use of pointed scalpel blade, (5) avoidance of cauterization and blunt preparation, (6) constant irrigation, and (7) retrieval of the sample without the use of forceps or suction. After a single round of introduction to these principles, multiple surgeons adopted the technique for samples with a minimal dimension of 5 mm spanning all cortical layers and subcortical white matter. Small samples (5-7 mm) were ideal for acute slice preparation and electrophysiology. No adverse events from sample resection were observed.

CONCLUSION

The microdissection technique of human cortical access tissue is safe and easily adoptable into the routine of neurosurgical procedures. The standardized and reliable surgical extraction of human brain tissue lays the foundation for human-to-human translational research on human brain tissue.

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.

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.