Researchers at Karolinska Institutet have developed a method based on artificial intelligence for histopathological diagnosis and grading of prostate cancer. The AI-system has the potential to solve one of the bottlenecks in today’s prostate cancer histopathology by providing more accurate diagnosis and better treatment decisions. The study shows that the AI-system is as good at identifying and grading prostate cancer as world-leading uro-pathologists.
“Our results show that it is possible to train an AI-system to detect and grade prostate cancer on the same level as leading experts,” says Martin Eklund, associate professor at the Department of Medical Epidemiology and Biostatistics at Karolinska Institutet who led the study. “This has the potential to significantly reduce the workload of uro-pathologists and allow them to focus on the most difficult cases.”
A problem in today’s prostate pathology is that there is a certain degree of subjectivity in the assessments of the biopsies. Different pathologists can reach different conclusions even though they are studying the same samples. This leads to a clinical problem where the doctors must pick treatment based on ambiguous information. In this context, the researchers see significant potential to use the AI-technology to increase the reproducibility of the pathological assessments.
To train and test the AI system, the researchers digitized more than 8000 biopsies taken from some 1200 Swedish men in the ages of 50–69 to high-resolution images using digital pathology scanners. About 6,600 of the samples were used to train the AI system to spot the difference between biopsies with or without cancer. The remaining samples, and additional sets of samples collected from other labs, were used to test the AI system. Its results were also compared against the assessments of 23 world-leading uro-pathologists. The study was conducted in collaboration with researchers at Tampere University in Finland.
The findings showed that the AI-system was almost near-perfect in determining whether a sample contained cancer or not, as well as in estimating the length of the cancer tumour in the biopsy. When it comes to determining the severity of the prostate cancer, the so-called Gleason score, the AI system was on par with the international experts.
“When it comes to grading the severity of the prostate cancer, the AI is in the same range as international experts, which is very impressive, and when it comes to diagnostics, to determine whether or not it is cancer, the AI is simply outstanding,” says Lars Egevad, professor in pathology at Karolinska Institutet and co-author of the study.
The initial findings are promising but more validation is needed before the AI system may be rolled out broadly in clinical practice, according to the researchers. That is why a multicenter study spanning nine European countries is currently underway with completion slated by the end of 2020. That study, which is financed by EIT Health, aims to train the AI-system to recognize cancer in biopsies taken from different laboratories, with different types of digital scanners and with very rare growth patterns. In addition, a randomized study starting in 2020 will examine how the AI-model may be implemented in Sweden’s health care system.
“AI-based evaluation of prostate cancer biopsies could revolutionize future health care,” says Henrik Grönberg, professor in cancer epidemiology at Karolinska Institutet and head of the Prostate Cancer Center at St Göran Hospital in Stockholm. “It has the potential to improve the diagnostic quality, and thereby secure a more equitable care at a lower cost.” Karolinska Institute

Concarlo Holdings has received a US patent for IpY, a novel therapeutic peptide that addresses drug-resistant breast cancer by targeting a unique cellular pathway — p27Kip1. The patent is the latest step in Concarlo’s journey to commercialize revolutionary medicines for metastatic breast cancer.
Concarlo has also announced that a new provisional patent application has been filed for modified versions of the therapeutic peptide that are believed to exhibit enhanced bioavailability. Concarlo is a Brooklyn, New York-based biotechnology innovator dedicated to developing sophisticated, targeted therapies and diagnostics in the oncology space. The IpY technology is the first to address the high incidence of drug-refractory disease that develops with currently available CDK4 inhibitor (CDK4i) treatments. Such a solution has the potential to drastically increase overall survival of breast cancer patients.
The recent introduction of CDK4i drugs, a class of medicines that directly targets the CDK4/6 pathway implicated in many malignancies, has had a significant impact on the way in which metastatic breast cancer is managed. However, such therapeutics are associated with patients transitioning to a treatment-resistant form of the condition, despite initial extended periods of remission. Backed by more than 20 years of research and development expertise, Concarlo has developed IpY and a companion diagnostic, ApY, to effectively overcome the issue of CDK4i resistance and roll out a more targeted treatment approach for optimized patient outcomes.
“Despite the clinical efficacy of CDK4 inhibitors, we’re seeing that primary or secondary resistance to therapy is presenting a significant challenge to overall survival,” said Dr. Dominique Bridon, Chief Development Officer at Concarlo. “With the IpY technology and its unique mechanism of action, we’re effectively targeting CDK4 while simultaneously inhibiting another target — CDK2 — which has been found to be a key molecular player in the development of drug resistance. In doing so, we are the first company to successfully address the CDK4i resistance issue to provide long-term durable tumour arrest. Combined with its highly specific targeting and low toxicity profile, the positive impact of this drug on the breast cancer treatment landscape is hard to understate.”
Concarlo was formed in 2016 and is supported by a team of internationally renowned experts forming its Scientific Advisory Board. To date, the company has raised more than $3.1 million to support the development, improvement, and commercialization of its IpY and ApY technologies to bring a precision medicine approach to breast cancer management. The newly issued patent for IpY and the provisional patent application for modified versions of the peptide are the first key milestones in Concarlo’s plan to build an extensive patent estate to maintain market exclusivity for its clinically relevant therapeutics.

National Institutes of Health (NIH) scientists have developed an ultrasensitive new test to detect abnormal forms of the protein tau associated with uncommon types of neurodegenerative diseases called tauopathies.  This advance gives them hope of using cerebrospinal fluid, or CSF – an accessible patient sample – to diagnose these and perhaps other, more common neurological diseases, such as Alzheimer’s disease.

Scientists have linked the abnormal deposition of tau in the brain to at least 25 different neurodegenerative diseases. However, to accurately diagnose these diseases, brain tissue often must be analysed after the patient has died. For their study, the researchers used the same test concept they developed when using postmortem brain tissue samples to detect the abnormal tau types associated with Pick disease, Alzheimer’s disease and chronic traumatic encephalopathy (CTE). They adapted the test to use CSF for the detection of abnormal tau of progressive supranuclear palsy (PSP), corticobasal degeneration (CBD), and other less common tauopathies.

They detected abnormal tau in CSF from both living and deceased patients. In one case, the test led to a corrected diagnosis in a patient who had died from CBD, but who was initially diagnosed with PSP. The new test is called 4R RT-QuIC – which stands for 4-repeat tau protein amplified in a real-time, quaking-induced conversion process.

The researchers plan to continue evaluating the clinical performance of 4R RT-QuIC by analysing larger sets of CSF samples. One focus will be to compare test results from tauopathy patients who agree to provide CSF samples both before and after death. The scientists hope this type of evaluation will help them better understand how abnormal tau in CSF evolves during brain disease.

NIHwww.niaid.nih.gov/news-events/nih-scientists-develop-test-uncommon-brain-diseases

by Dr Ann-Christin Niehoff
Multimodal imaging by mass spectrometry offers a spatially resolved analysis of tissue sections as an additional tool in clinical research. Here, matrix-assisted laser desorption/ionization mass spectrometry for molecular imaging and laser ablation inductively coupled plasma mass spectrometry for elemental imaging are used to tackle two drug applications.
Mass spectrometry imaging
In recent years, mass spectrometry imaging (MSI) has gained more and more interest in the field of clinical, biological and pharmaceutical research. In contrast to hyphenated chromatographic techniques (e.g. LC-MS or GC-MS), MSI provides spatially resolved information while maintaining high sensitivity. With today’s techniques, high spatial resolution down to the low micrometre range can be achieved and is therefore a good combination with existing clinical imaging approaches from pathology.
Multimodal imaging describes the combination of different imaging techniques, as none by itself is a gold standard to answer all questions. Since MSI works with tissue sections, it can be combined easily with various microscopy applications, providing an additional input to clinical histology. Although protocols for different kinds of tissue sections exist, the preference here is to work on cryosections rather than formalin-fixed paraffin-embedded sections; this helps to avoid wash out of analytes from the tissue during the fixation and embedding steps.
Different MSI techniques can be used to focus on molecular or elemental imaging. In this article, the focus will be on matrixassisted laser desorption ionization mass spectrometry (MALDIMS) for molecular and laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) for elemental imaging.
MALDI-MSI
MALDI-MS is the most frequently used imaging technique in molecular MS. The analysis requires coating of the tissue section with a matrix, typically a small organic compound, and performing soft ionization of desorbed molecules by a pulsed laser. Ionization efficiency is highly dependent on molecular structure, matrices and laser wavelength. By scanning over the sample, a full mass spectrum is generated for each pixel.
Using the tandem MS (MS/MS) mode, fragmentation studies can provide information on molecular structures. Matrix preparation is one of the critical aspects and a potential disadvantage of MALDI-MS, Microscopy and Imaging in the Clinical Lab June 2020 21 | Figure1. Multimodal imaging of myocardial infarction Microscopic images of the two parallel sections (a & b) with the area of myocardial infarct (marked with a black line), quantified distribution of gadolinium determined with LA-ICP-MS (c) at 15|μm spot size, distribution of the ligand from Gadofluorine P (d) at 40|μm spot size, as well as the structure of the ligand and the theoretical spectrum (cyan bars) and the measured spectra (black line) with MALDI-MS (e). as it may influence limits of detection and spatial resolution due to analyte extraction of the sample by the matrix. Different instruments for matrix preparation are therefore commercially available to improve homogenous distribution and reproducibility. Owing to matrix effects in molecular MS, quantification is challenging, but is possible to achieve for single analytes via internal standards or standard addition with matrix matched standards.
Here, matrix was sublimated using the iMLayer (Shimadzu). MALDI-MS experiments were performed with the iMScope TRIO (Shimadzu), equipped with a fluorescence microscope, atmospheric pressure MALDI-source and an ion trap/time-of-flight (IT-TOF) mass analyser. IMAGEREVEAL MS (Shimadzu) was used for data analysis.
LA-ICP-MSI
In the field of elemental MSI, LA-ICP-MS provides major, minor and trace elemental information on surfaces and tissue sections. A laser is scanned over the sample and the ablated material is transported by a carrier gas into the ICP source, where the particles are atomized and ionized. To obtain spatially resolved images, transient signals of the respective analyte are required.
As mass analyser, quadrupoles are most frequently used. Although less matrix dependent than MALDI-MS, a fundamental aspect of recent research is method development for reliable quantification strategies, mainly via matrix matched standards. The major disadvantages of LA-ICP-MS are its destructive nature with loss of molecular information.
In this study, experiments were performed with the LSX-G2+ laser ablation system (Teledyne Cetac Technologies) coupled to the quadrupole based ICPMS-2030 (Shimadzu).
Complementary bioimaging of Gadofluorine P in myocardial infarction in mice
Magnetic resonance imaging is a widely used imaging technique in daily clinical practice. To enhance contrast during this examination, several different contrast agents are available. While most gadoliniumbased contrast agents (GBCAs) distribute systemically, some targetspecific GBCAs are under investigation as well. Gadofluorine P is one of these target-specific contrast agents and shows high affinity towards the collagen-rich extracellular matrix which is secreted in the event of myocardial infarction (MI) [1].
In this application, mice underwent injection of Gadofluorine P solution as contrast agents 6|weeks after an induced MI. Afterwards the mice were sacrificed and the hearts were removed for cryosections preparation. By multimodal imaging, LA-ICP-MS was used to generate quantified elemental imaging of gadolinium, while MALDI-MS validated the findings (Fig. 1) and could further provide information for phospholipids and heme b distribution (data not shown).
Figure 1 shows the microscopic images (a & b) of the two thin sections analysed. With LA-ICP-MS (c), a homogeneous distribution of the gadolinium in the healthy myocardium with an average concentration of about 50|μg/g was detected. The infarct region contained two times higher gadolinium concentrations of about 110|μg/g with maximum values up to 370|μg/g.
Higher gadolinium concentrations could also be found in the ventricle due to the intravenous administration of the contrast agent. These distributions could be verified with MALDI-MS imaging (d).
In this experiment, only the protonated ligand of Gadofluorine P rather than the intact complex could be detected (e). The main peak (m/z 1168.39) was used to create the image, which showed good correlation to the gadolinium distribution determined with LA-ICP-MS. The highest intensities of the molecular probe were found in MI and ventricle regions, whereas healthy myocardium showed low and homogenous intensities.
Multimodal imaging of photosensitizers in 3D tumour cell models
Photodynamic therapy offers an alternative cancer treatment. A photosensitive compound (photosensitizer; PS) is administered and the tumour is subsequently irradiated. The activation of the PS leads to the formation of a reactive oxygen species and subsequently to cell apoptosis. One main challenge in the development of PS is the hydrophobic character of the compounds, which hinders tissue penetration. Additionally, the orally administered compound needs to pass through the mucus layer in the gastrointestinal tract. Thus, the determination of the penetration depth of these compounds is of great interest.
The use of 3D tumour spheroids enables in vitro drug screening, while simulating the tumour environment better than 2D cell cultures. The photosensitizer 5,10,15,20-tetrakis (3-hydroxy-phenyl)-porphyrin (mTHPP) and its palladiumtagged analogue mTHPP-Pd were analysed in this study. Here, multimodal imaging is used to visualize the penetration depth of mTHPP and the lipid distribution in 3D tumour spheroid by MALDI-MS (5|μm spot size) as well as to quantify the drug by LA-ICP-MS (7|μm spot size) [2,3].
The MALDI-MS and LA-ICP-MS images of a tumour spheroid treated with mTHPP or mTHPP-Pd are shown in Figure 2. In the microscopic image, an almost spherical tissue section with a diameter of approx. 550|μm can be seen. The distribution map of mTHPP shows a ring-shaped distribution, which can be precisely correlated with the outer cell layer of the tumour spheroid. The PS is distributed homogeneously inside the outer layer and not around the spheroid, although it does not penetrate deeper into the tissue.
Nevertheless, the MALDI-MS experiments revealed that the PS can be detected as an intact molecule without substantial decomposition during the sample preparation. The LA-ICP-MS results for a spheroid incubated with mTHPP-Pd show the same distribution as the mTHPP detected by MALDI-MS. Since the metal-tagged PS is needed for ICP-MS analysis, only spheroids treated with this compound could be investigated. Conversely, this modification of the molecule could no longer be detected using MALDI-MS. Owing to the loading with palladium, the preferred protonation sites of the molecule are unavailable, impairing the detection.
However, before LA-ICP-MS experiments, MALDI-MS can be used to identify phospholipids as shown in Figure 3. Palladium concentrations up to 10|μg/g with an average of 1.9|μg/g were detected (Fig. 3b). This represents an enrichment of PS by a factor of 3 (average) up to 18 (highest concentration) compared to the incubation concentration. The phospholipids PC(34:1), PC(34:0) and PC(30:0) could be detected and show different distributions coherent with the different metabolic zones in a tumour spheroid.
Conclusion
In conclusion, the two applications shown provide an example of how to add MSI to clinical research. Multimodal imaging has successfully been performed to address drug penetration and enrichment in different kinds of tissue based on the combination of elemental imaging and molecular imaging by LA-ICP-MS and MALDI-MS.
The author
Ann-Christin Niehoff PhD
European Innovation Center, Shimadzu Europa GmbH,
47259 Duisburg, Germany

E-mail: acn@shimadzu.eu

Researchers at Rady Children’s Institute for Genomic Medicine (RCIGM) have utilized automated machine-learning and clinical natural language processing (CNLP) to diagnose rare genetic diseases in record time. This new method is speeding answers to physicians caring for infants in intensive care and opening the door to increased use of genome sequencing as a first-line diagnostic test for babies with cryptic conditions.

“Some people call this artificial intelligence, we call it augmented intelligence,” said Stephen Kingsmore, MD, DSc, President and CEO of RCIGM. “Patient care will always begin and end with the doctor. By harnessing the power of technology, we can quickly and accurately determine the root cause of genetic diseases. We rapidly provide this critical information to intensive care physicians so they can focus on personalizing care for babies who are struggling to survive.”

The workflow and research were led by the RCIGM team in collaboration with leading technology and data-science developers —Alexion, Clinithink, Diploid, Fabric Genomics and Illumina.

Dr. Kingsmore’s team has pioneered a rapid Whole Genome Sequencing process to deliver genetic test results to neonatal and paediatric intensive care (NICU/PICU) physicians to guide medical intervention. RCIGM is the research arm of Rady Children’s Hospital-San Diego.

By reducing the need for labour-intensive manual analysis of genomic data, the supervised automated pipeline provided significant time-savings. In February 2018, the same team achieved the Guinness World Record for fastest diagnosis through whole genome sequencing. Of the automated runs, the fastest times – averaging 19 hours – were achieved using augmented intelligence.

“This is truly pioneering work by the RCIGM team—saving the lives of very sick newborn babies by using AI to rapidly and accurately analyse their whole genome sequence “ says Eric Topol, MD, Professor of Molecular Medicine at Scripps Research and author of the new book Deep Medicine.

RCIGM has optimized and integrated several time-saving technologies into a rapid Whole Genome Sequencing (rWGS) process to screen a child’s entire genetic makeup for thousands of genetic anomalies from a blood sample.

Key components in the rWGS pipeline come from Illumina, the global leader in DNA sequencing, including Nextera DNA Flex library preparation, whole genome sequencing via the NovaSeq 6000 and the S1 flow cell format. Speed and accuracy are enhanced by Illumina’s DRAGEN (Dynamic Read Analysis for GENomics) Bio-IT Platform.

Other pipeline elements include Clinithink’s clinical natural language processing platform CliX ENRICH that quickly combs through a patient’s electronic medical record to automatically extract comprehensive patient phenotype information.

Another core element of the machine learning system is MOON by Diploid. The platform automates genome interpretation using AI to automatically filter and rank likely pathogenic variants. Deep phenotype integration, based on natural language processing of the medical literature, is one of the key features driving this automated interpretation. MOON takes five minutes to suggest the causal mutation out of the 4.5 million variants in a whole genome.

In addition, Alexion’s rare disease and data science expertise enabled the translation of clinical information into a computable format for guided variant interpretation.

As part of this study, the genetic sequencing data was fed into automated computational platforms under the supervision of researchers. For comparison and verification, clinical medical geneticists on the team used Fabric Genomics’ AI-based clinical decision support software, OPAL (now called Fabric Enterprise)—to confirm the output of the automated pipeline. Fabric software is part of RCIGM’s standard analysis and interpretation workflow.

The study titled “Diagnosis of genetic diseases in seriously ill children by rapid whole-genome sequencing and automated phenotyping and interpretation,” found that automated, retrospective diagnoses concurred with expert manual interpretation (97 percent recall, 99 percent precision in 95 children with 97 genetic diseases).

Researchers concluded that genome sequen-cing with automated phenotyping and interpretation—in a median 20:10 hours—may spur use in intensive care units, thereby enabling timely and precise medical care. “Using machine-learning platforms doesn’t replace human experts. Instead it augments their capabilities,” said Michelle Clark, PhD, statistical scientist at RCIGM and the first author of the study. “By informing timely targeted treatments, rapid genome sequencing can improve the outcomes of seriously ill children with genetic diseases.”
Rady Children’s Institutewww.radygenomics.org/category/news/pr/