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.

Cherenkov Luminescence in Tumor Diagnosis and Treatment: A Review

Malignant tumors rank as a leading cause of death worldwide. Accurate diagnosis and advanced treatment options are crucial to win battle against tumors. In recent years, Cherenkov luminescence (CL) has shown its technical advantages and clinical transformation potential in many important fields, particularly in tumor diagnosis and treatment, such as tumor detection in vivo, surgical navigation, radiotherapy, photodynamic therapy, and the evaluation of therapeutic effect. In this review, we summarize the advances in CL for tumor diagnosis and treatment. We first describe the physical principles of CL and discuss the imaging techniques used in tumor diagnosis, including CL imaging, CL endoscope, and CL tomography. Then we present a broad overview of the current status of surgical resection, radiotherapy, photodynamic therapy, and tumor microenvironment monitoring using CL. Finally, we shed light on the challenges and possible solutions for tumor diagnosis and therapy using CL.

Prospective testing of clinical Cerenkov luminescence imaging against standard-of-care nuclear imaging for tumour location

In oncology, the feasibility of Cerenkov luminescence imaging (CLI) has been assessed by imaging superficial lymph nodes in a few patients undergoing diagnostic ¹⁸F-fluoro-2-deoxyglucose (¹⁸F-FDG) positron emission tomography/computed tomography (PET/CT). However, the weak luminescence signal requires the removal of ambient light. Here we report the development of a clinical CLI fiberscope with a lightproof enclosure, and the clinical testing of the setup using five different radiotracers. In an observational prospective trial (ClinicalTrials.gov identifier NCT03484884 ) involving 96 patients with existing or suspected tumours, scheduled for routine clinical FDG PET or ¹³¹I therapy, the level of agreement of CLI with standard-of-care imaging (PET or planar single-photon emission CT) for tumour location was 'acceptable' or higher (≥3 in the 1-5 Likert scale) for 90% of the patients. CLI correlated with the concentration of radioactive activity, and captured therapeutically relevant information from patients undergoing targeted radiotherapy or receiving the alpha emitter ²²³Ra, which cannot be feasibly imaged clinically. CLI could supplement radiological scans, especially when scanner capacity is limited.

Serum lncRNA-ANRIL and SOX9 expression levels in glioma patients and their relationship with poor prognosis

Background

lncRNA-CDKN2B antisense RNA 1 (ANRIL) and SRY-box transcription factor 9 (SOX9) has abnormal expression in many tumors including glioma, but the underlying molecular mechanism is unclear. This study set out to investigate the serum lncRNA-ANRIL and SOX9 levels in glioma patients and their effects on prognosis.

Methods

We enrolled 142 glioma patients admitted to our hospital from May 2014 to May 2016 into the research group (RG) and 120 healthy subjects receiving concurrent physical examinations into the control group (CG). Fasting peripheral blood (4 mL each) was sampled from subjects from the two groups. Using the quantitative real-time polymerase chain reaction (qRT-PCR), lncRNA-ANRIL and SOX9 were measured to explore their values in the early diagnosis of glioma. Patients from RG were followed up for 3 years to analyze the influence of lncRNA-ANRIL and SOX9 on patient prognosis. We purchased glioma cell lines U251 and U87 and grouped them according to the transfection of different plasmids. We conducted CCK8 assay to test cell proliferation, Transwell assay to test cell invasion, the flow cytometry to test cell apoptosis, and Western Blot assay to measure bcl-2 and bax protein levels.

Results

ANRIL and SOX9 were evidently higher in RG than in CG (P<0.01). The receiver operating characteristic (ROC) curve revealed that the diagnostic sensitivity of ANRIL combined with SOX9 for glioma was 81.62%, and the specificity was 90.83% (P<0.01). ANRIL and SOX9 were closely related to tumor grade, tumor diameter, distant metastasis, and family history of glioma (P<0.01). In total, 135 patients were successfully followed up (95.07%). Patients with high levels of ANRIL and SOX9 had a markedly poorer prognosis than those with low levels (P<0.05). ANRIL and SOX9 were markedly higher in glioma cell lines (U251 and U87) than in normal brain cells (P<0.01). The proliferation and invasion of U251 cells were notably reduced after the transfection of ANRIL and SOX9 inhibitory sequences (P<0.01), but the apoptosis was notably increased (P<0.01). Bcl-2 expression was markedly increased in lncRNA-ANRIL-inhibitor and SOX9-inhibitor (P<0.01), while bax expression was markedly reduced in lncRNA-ANRIL-inhibitor and SOX9-inhibitor (P<0.01).

Conclusion

lncRNA-ANRIL and SOX9 levels were higher in glioma patients than in healthy people. High-lncRNA-ANRIL and SOX9 levels were strongly associated with unfavorable prognosis of patients. The testing of biological behaviors revealed that lncRNA-ANRIL and SOX9 worked as tumor-promoting genes in glioma.

Synthesis and in vitro preliminary evaluation of prostate-specific membrane antigen targeted upconversion nanoparticles as a first step towards radio/fluorescence-guided surgery of prostate cancer

Over the last decade, upconversion nanoparticles (UCNP) have been widely investigated in nanomedicine due to their high potential as imaging agents in the near-infrared (NIR) optical window of biological tissues. Here, we successfully develop active targeted UCNP as potential probes for dual NIR-NIR fluorescence and radioactive-guided surgery of prostate-specific membrane antigen (PSMA)(+) prostate cancers. We designed a one-pot thermolysis synthesis method to obtain oleic acid-coated spherical NaYF4:Yb,Tm@NaYF4 core/shell UCNP with narrow particle size distribution (30.0 ± 0.1 nm, as estimated by SAXS analysis) and efficient upconversion luminescence. Polyethylene glycol (PEG) ligands bearing different anchoring groups (phosphate, bis- and tetra-phosphonate-based) were synthesized and used to hydrophilize the UCNP. DLS studies led to the selection of a tetra-phosphonate PEG(2000) ligand affording water-dispersible UCNP with sustained colloidal stability in several aqueous media. PSMA-targeting ligands (i.e., glutamate-urea-lysine derivatives called KuEs) and fluorescent or radiolabelled prosthetic groups were grafted onto the UCNP surface by strain-promoted azide-alkyne cycloaddition (SPAAC). These UCNP, coated with 10 or 100% surface density of KuE ligands, did not induce cytotoxicity over 24 h incubation in LNCaP-Luc or PC3-Luc prostate cancer cell lines or in human fibroblasts for any of the concentrations evaluated. Competitive binding assays and flow cytometry demonstrated the excellent affinity of UCNP@KuE for PSMA-positive LNCaP-Luc cells compared with non-targeted UCNP@CO2H. Furthermore, the binding of UCNP@KuE to prostate tumour cells was positively correlated with the surface density of PSMA-targeting ligands and maintained after 125I-radiolabelling. Finally, a preliminary biodistribution study in LNCaP-Luc-bearing mice demonstrated the radiochemical stability of non-targeted [125I]UCNP paving the way for future in vivo assessments.

Radiopharmaceutical and Eu3+ doped gadolinium oxide nanoparticles mediated triple-excited fluorescence imaging and image-guided surgery

Cerenkov luminescence imaging (CLI) is a novel optical imaging technique that has been applied in clinic using various radionuclides and radiopharmaceuticals. However, clinical application of CLI has been limited by weak optical signal and restricted tissue penetration depth. Various fluorescent probes have been combined with radiopharmaceuticals for improved imaging performances. However, as most of these probes only interact with Cerenkov luminescence (CL), the low photon fluence of CL greatly restricted it's interaction with fluorescent probes for in vivo imaging. Therefore, it is important to develop probes that can effectively convert energy beyond CL such as β and γ to the low energy optical signals. In this study, a Eu3+ doped gadolinium oxide (Gd2O3:Eu) was synthesized and combined with radiopharmaceuticals to achieve a red-shifted optical spectrum with less tissue scattering and enhanced optical signal intensity in this study. The interaction between Gd2O3:Eu and radiopharmaceutical were investigated using 18F-fluorodeoxyglucose (18F-FDG). The ex vivo optical signal intensity of the mixture of Gd2O3:Eu and 18F-FDG reached 369 times as high as that of CLI using 18F-FDG alone. To achieve improved biocompatibility, the Gd2O3:Eu nanoparticles were then modified with polyvinyl alcohol (PVA), and the resulted nanoprobe PVA modified Gd2O3:Eu (Gd2O3:Eu@PVA) was applied in intraoperative tumor imaging. Compared with 18F-FDG alone, intraoperative administration of Gd2O3:Eu@PVA and 18F-FDG combination achieved a much higher tumor-to-normal tissue ratio (TNR, 10.24 ± 2.24 vs. 1.87 ± 0.73, P = 0.0030). The use of Gd2O3:Eu@PVA and 18F-FDG also assisted intraoperative detection of tumors that were omitted by preoperative positron emission tomography (PET) imaging. Further experiment of image-guided surgery demonstrated feasibility of image-guided tumor resection using Gd2O3:Eu@PVA and 18F-FDG. In summary, Gd2O3:Eu can achieve significantly optimized imaging property when combined with 18F-FDG in intraoperative tumor imaging and image-guided tumor resection surgery. It is expected that the development of the Gd2O3:Eu nanoparticle will promote investigation and application of novel nanoparticles that can interact with radiopharmaceuticals for improved imaging properties. This work highlighted the impact of the nanoprobe that can be excited by radiopharmaceuticals emitting CL, β, and γ radiation for precisely imaging of tumor and intraoperatively guide tumor resection.

An Optimized Integrin α6-Targeted Magnetic Resonance Probe for Molecular Imaging of Hepatocellular Carcinoma in Mice

Introduction

Integrin α6 is an attractive diagnostic biomarker for molecular imaging of hepatocellular carcinoma (HCC) as it has an extremely high positive rate (approximately 94%) in clinical early-stage HCC. In this study, based on our previously identified integrin α6-targeted peptide, we developed an optimized integrin α6-targeted magnetic resonance (MR) probe dubbed DOTA(Gd)-ANADYWR for MR imaging of HCC in mice.

Materials and Methods

The longitudinal (R₁) relaxivity of DOTA(Gd)-ANADYWR was measured on a 3.0 T MR system . The specific tumor enhancement of the agent was investigated in four distinct mouse models, including subcutaneous, orthotopic, genetically engineered and chemically induced HCC mice.

Results

The R₁ relaxivity value of DOTA(Gd)-ANADYWR is 5.11 mM⁻¹s⁻¹ at 3.0 T, which is similar to that of the nonspecific clinical agent Gadoteridol. DOTA(Gd)-ANADYWR generated superior enhanced MR signal in HCC lesions and provided complementary enhancement MR signals to the clinically available hepatobiliary MR contrast agent gadoxetate disodium (Gd-EOB-DTPA). Importantly, DOTA(Gd)-ANADYWR could efficiently visualize small HCC lesion (approximately 1 mm) which was hardly detected by the clinical Gd-EOB-DTPA.

Conclusion

These findings suggest the potential application of this integrin α6-targeted MR probe for the detection of HCC, particularly for small HCC.

Real-time intraoperative glioma diagnosis using fluorescence imaging and deep convolutional neural networks

Purpose

Surgery is the predominant treatment modality of human glioma but suffers difficulty on clearly identifying tumor boundaries in clinic. Conventional practice involves neurosurgeon’s visual evaluation and intraoperative histological examination of dissected tissues using frozen section, which is time-consuming and complex. The aim of this study was to develop fluorescent imaging coupled with artificial intelligence technique to quickly and accurately determine glioma in real-time during surgery.

Methods

Glioma patients (N = 23) were enrolled and injected with indocyanine green for fluorescence image–guided surgery. Tissue samples (N = 1874) were harvested from surgery of these patients, and the second near-infrared window (NIR-II, 1000–1700 nm) fluorescence images were obtained. Deep convolutional neural networks (CNNs) combined with NIR-II fluorescence imaging (named as FL-CNN) were explored to automatically provide pathological diagnosis of glioma in situ in real-time during patient surgery. The pathological examination results were used as the gold standard.

Results

The developed FL-CNN achieved the area under the curve (AUC) of 0.945. Comparing to neurosurgeons’ judgment, with the same level of specificity >80%, FL-CNN achieved a much higher sensitivity (93.8% versus 82.0%, P < 0.001) with zero time overhead. Further experiments demonstrated that FL-CNN corrected >70% of the errors made by neurosurgeons. FL-CNN was also able to rapidly predict grade and Ki-67 level (AUC 0.810 and 0.625) of tumor specimens intraoperatively.

Conclusion

Our study demonstrates that deep CNNs are better at capturing important information from fluorescence images than surgeons’ evaluation during patient surgery. FL-CNN is highly promising to provide pathological diagnosis intraoperatively and assist neurosurgeons to obtain maximum resection safely.

Trial registration

ChiCTR ChiCTR2000029402. Registered 29 January 2020, retrospectively registered

Supplementary Information

The online version contains supplementary material available at 10.1007/s00259-021-05326-y.

NIR-II/NIR-I Fluorescence Molecular Tomography of Heterogeneous Mice Based on Gaussian Weighted Neighborhood Fused Lasso Method

Fluorescence molecular tomography (FMT), which can visualize the distribution of fluorescence biomarkers, has become a novel three-dimensional noninvasive imaging technique for in vivo studies such as tumor detection and lymph node location. However, it remains a challenging problem to achieve satisfactory reconstruction performance of conventional FMT in the first near-infrared window (NIR-I, 700-900nm) because of the severe scattering of NIR-I light. In this study, a promising FMT method for heterogeneous mice was proposed to improve the reconstruction accuracy using the second near-infrared window (NIR-II, 1000-1700nm), where the light scattering significantly reduced compared with NIR-I. The optical properties of NIR-II were analyzed to construct the forward model for NIR-II FMT. Furthermore, to raise the accuracy of solution of the inverse problem, we proposed a novel Gaussian weighted neighborhood fused Lasso (GWNFL) method. Numerical simulation was performed to demonstrate the outperformance of GWNFL compared with other algorithms. Besides, a novel NIR-II/NIR-I dual-modality FMT system was developed to contrast the in vivo reconstruction performance between NIR-II FMT and NIR-I FMT. To compare the reconstruction performance of NIR-II FMT with traditional NIR-I FMT, numerical simulations and in vivo experiments were conducted. Both the simulation and in vivo results showed that NIR-II FMT outperformed NIR-I FMT in terms of location accuracy and spatial overlap index. It is believed that this study could promote the development and biomedical application of NIR-II FMT in the future.

A novel in vivo Cerenkov luminescence image-guided surgery on primary and metastatic colorectal cancer

Intraoperative Cerenkov luminescence imaging (CLI) can effectively improve the performance of tumor surgery. Nevertheless, the existing approaches are still unsatisfying to the clinical demands of open surgery. This study develops a novel intraoperative in vivo CLI approach to investigate the potential and value of Cerenkov luminescence (CL) image-guided surgery. A system characterized with high sensitivity (19.61 kBq mL^{-1 18} F-FDG) and desirable spatial resolution (88.34 μm) is developed. CL image-guided surgery is performed on colorectal cancer (CRC) models of mice and swine. Tumor surgery is guided by the static CL images, and the resection quality is evaluated quantitatively and contrasted with other imaging modalities exemplified by bioluminescence imaging (BLI). The in vivo results demonstrated the effectiveness of the proposed intraoperative CLI approach for removing primary and metastatic CRC. Safety of performing in vivo CL image-guided surgery is verified as well through radiation measurements of related staffs. Overall, the developed intraoperative in vivo CLI approach can efficiently improve the cancer treatment.

A novel Cerenkov luminescence tomography approach using multilayer fully connected neural network

Cerenkov luminescence tomography (CLT) has been proved as an effective tool for various biomedical applications. Because of the severe scattering of Cerenkov luminescence, the performance of CLT remains unsatisfied. This paper proposed a novel CLT reconstruction approach based on a multilayer fully connected neural network (MFCNN). Monte Carlo simulation data was employed to train the MFCNN, and the complex relationship between the surface signals and the true sources was effectively learned by the network. Both simulation and in vivo experiments were performed to validate the performance of MFCNN CLT, and it was further compared with the typical radiative transfer equation (RTE) based method. The experimental data showed the superiority of MFCNN CLT in terms of accuracy and stability. This promising approach for CLT is expected to improve the performance of optical tomography, and to promote the exploration of machine learning in biomedical applications.

Effectiveness and safety of indocyanine green fluorescence imaging-guided hepatectomy for liver tumors: A systematic review and first meta-analysis

BACKGROUND

This meta-analysis was conducted to evaluate the effectiveness and safety of indocyanine green fluorescence imaging-guided hepatectomy(FIGH) for liver tumors.

METHODS

Clinical studies were retrieved from the PubMed, Embase, Cochrane Library, Medline and Web of Science electronic databases. Primary outcomes included operative time, blood loss, blood transfusion, hospital stay, R0 resection, postoperative complications, postoperative mortality and 1-year recurrence rate. Study-specific effect sizes and their 95% confidence intervals (CIs) were combined to calculate the pooled value using a fixed-effects or random-effects model.

RESULTS

Six studies comprising 587 patients were included. Major operative time (mean difference [MD] = -55.45; 95% CI = -78.85- -32.05), blood loss (MD = 12.99; 95% CI = 12.00-13.97), hospital stay (rate difference [RD] = -12.61; 95% CI = -15.06- -10.17), and postoperative complications (RD = -0.07; 95% CI = -0.12- -0.01) were all less in the FIGH group than in the traditional hepatectomy(TH) group. No differences were found in blood transfusion, R0 resection or 1-year recurrence rate. No perioperative mortality was observed in either group.

CONCLUSION

Based on current evidence, applying indocyanine green fluorescence imaging technology to accurately diagnose and treat liver tumors can effectively reduce operative time, blood loss, hospital stay and postoperative complications.

Radiopharmaceuticals and Fluorescein Sodium Mediated Triple-Modality Molecular Imaging Allows Precise Image-Guided Tumor Surgery

Radical resection is the most effective method for malignant tumor treatments. However, conventional imaging cannot fully satisfy the clinical needs of surgical navigation. This study presents a novel triple-modality positron emission tomography (PET)-Cerenkov radiation energy transfer (CRET)-confocal laser endomicroscopy (CLE) imaging strategy for intraoperative tumor imaging and surgical navigation. Using clinical radiopharmaceuticals and fluorescein sodium (FS), this strategy can accurately detect the tumor and guide the tumor surgery. The FS emission property under Cerenkov radiation excitation is investigated using 2-deoxy-2-18F-fluoroglucose and 11C-choline. Performances of the PET-CRET-CLE imaging and the CRET-CLE image-guided surgery are evaluated on mouse models. The CRET signal at 8 mm depth is stronger than the Cerenkov luminescence at 1 mm depth in phantoms. In vivo experiments indicate that 0.5 mL kg-1 of 10% FS generates the strongest CRET signal, which can be observed immediately after FS injection. A surgical navigation study shows that the tumors are precisely detected and resected using intraoperative CRET-CLE. In summary, a PET-CRET-CLE triple-modality imaging strategy is developed. This strategy can detect the tumors and precisely guide the tumor resection using clinical pharmaceuticals. This triple-modality imaging shows high potential in surgical navigation research and clinical translation.

Endoscopic Cerenkov luminescence imaging and image-guided tumor resection on hepatocellular carcinoma-bearing mouse models

Detecting deep tumors inside living subject is still challenging for Cerenkov luminescence imaging (CLI). In this study, a high-sensitivity endoscopic CLI (ECLI) system was developed with a dual-mode deep cooling approach to improve the imaging sensitivity. System was characterized through a series of ex vivo studies. Furthermore, subcutaneous and orthotropic human hepatocellular carcinoma (HCC) mouse models were established for ECLI guided tumor resection in vivo. The results showed that the ECLI system had spatial resolution (62.5 μm) and imaging sensitivity (6.29 × 10^-2 kBq/μl ¹⁸F-FDG). The in vivo experimental data from the HCC mouse models demonstrated that the system was effective to intraoperatively guide the surgery of deep tumors such as liver cancer. Overall, the developed system exhibits promising potential for the applications of tumor precise resection and novel nanoprobe based optical imaging.

Cerenkov luminescence imaging: physics principles and potential applications in biomedical sciences

Cerenkov luminescence imaging (CLI) is a novel imaging modality to study charged particles with optical methods by detecting the Cerenkov luminescence produced in tissue. This paper first describes the physical processes that govern the production and transport in tissue of Cerenkov luminescence. The detectors used for CLI and their most relevant specifications to optimize the acquisition of the Cerenkov signal are then presented, and CLI is compared with the other optical imaging modalities sharing the same data acquisition and processing methods. Finally, the scientific work related to CLI and the applications for which CLI has been proposed are reviewed. The paper ends with some considerations about further perspectives for this novel imaging modality.

Nanoparticle-mediated radiopharmaceutical-excited fluorescence molecular imaging allows precise image-guided tumor-removal surgery

Fluorescent molecular imaging technique has been actively explored for optical image-guided cancer surgery in pre-clinical and clinical research and has attracted many attentions. However, the efficacy of the fluorescent image-guided cancer surgery can be compromised by the low signal-to-noise ratio caused by the external light excitation. This study presents a novel nanoparticle-mediated radiopharmaceutical-excited fluorescent (REF) image-guided cancer surgery strategy, which employs the internal dual-excitation of europium oxide nanoparticles through both gamma rays and Cerenkov luminescence emitted from radiopharmaceuticals. The performance of the novel image-guided surgery technique was systematically evaluated using subcutaneous breast cancer 4 T1 tumor models, orthotropic and orthotropic-ectopic hepatocellular carcinoma tumor-bearing mice. The results reveal that the novel REF image-guided cancer surgery technique exhibits high performance of detecting invisible ultra-small size tumor (even less than 1 mm) and residual tumor tissue. Our study demonstrates the high potential of the novel image-guided cancer surgery for precise tumor resection.

Cerenkov luminescence imaging (CLI) for image-guided cancer surgery

Cerenkov luminescence imaging (CLI) is a novel molecular optical imaging technique based on the detection of optical Cerenkov photons emitted by positron emission tomography (PET) imaging agents. The ability to use clinically approved tumour-targeted tracers in combination with small-sized imaging equipment makes CLI a particularly interesting technique for image-guided cancer surgery. The past few years have witnessed a rapid increase in proof-of-concept preclinical studies in this field, and several clinical trials are currently underway. This article provides an overview of the basic principles of Cerenkov radiation and outlines the challenges of CLI-guided surgery for clinical use. The preclinical and clinical trial literature is examined including applications focussed on image-guided lymph node detection and Cerenkov luminescence endoscopy, and the ongoing clinical studies and technological developments are highlighted. By intraoperatively guiding the oncosurgeon towards more accurate and complete resections, CLI has the potential to transform current surgical practice, and improve oncological and cosmetic outcomes for patients.