Machine learning predictive performance evaluation of conventional and fuzzy radiomics in clinical cancer imaging cohorts.

BACKGROUND: Hybrid imaging became an instrumental part of medical imaging, particularly cancer imaging processes in clinical routine. To date, several radiomic and machine learning studies investigated the feasibility of in vivo tumor characterization with variable outcomes. This study aims to investigate the effect of recently proposed fuzzy radiomics and compare its predictive performance to conventional radiomics in cancer imaging cohorts. In addition, lesion vs. lesion+surrounding fuzzy and conventional radiomic analysis was conducted. METHODS: Previously published 11C Methionine (MET) positron emission tomography (PET) glioma, 18F-FDG PET/computed tomography (CT) lung, and 68GA-PSMA-11 PET/magneto-resonance imaging (MRI) prostate cancer retrospective cohorts were included in the analysis to predict their respective clinical endpoints. Four delineation methods including manually defined reference binary (Ref-B), its smoothed, fuzzified version (Ref-F), as well as extended binary (Ext-B) and its fuzzified version (Ext-F) were incorporated to extract imaging biomarker standardization initiative (IBSI)-conform radiomic features from each cohort. Machine learning for the four delineation approaches was performed utilizing a Monte Carlo cross-validation scheme to estimate the predictive performance of the four delineation methods. RESULTS: Reference fuzzy (Ref-F) delineation outperformed its binary delineation (Ref-B) counterpart in all cohorts within a volume range of 938-354987 mm3 with relative cross-validation area under the receiver operator characteristics curve (AUC) of +4.7-10.4. Compared to Ref-B, the highest AUC performance difference was observed by the Ref-F delineation in the glioma cohort (Ref-F: 0.74 vs. Ref-B: 0.70) and in the prostate cohort by Ref-F and Ext-F (Ref-F: 0.84, Ext-F: 0.86 vs. Ref-B: 0.80). In addition, fuzzy radiomics decreased feature redundancy by approx. 20%. CONCLUSIONS: Fuzzy radiomics has the potential to increase predictive performance particularly in small lesion sizes compared to conventional binary radiomics in PET. We hypothesize that this effect is due to the ability of fuzzy radiomics to model partial volume effects and delineation uncertainties at small lesion boundaries. In addition, we consider that the lower redundancy of fuzzy radiomic features supports the identification of imaging biomarkers in future studies. Future studies shall consider systematically analyzing lesions and their surroundings with fuzzy and binary radiomics.

Clinical data classification with noisy intermediate scale quantum computers.

Quantum machine learning has experienced significant progress in both software and hardware development in the recent years and has emerged as an applicable area of near-term quantum computers. In this work, we investigate the feasibility of utilizing quantum machine learning (QML) on real clinical datasets. We propose two QML algorithms for data classification on IBM quantum hardware: a quantum distance classifier (qDS) and a simplified quantum-kernel support vector machine (sqKSVM). We utilize these different methods using the linear time quantum data encoding technique ([Formula: see text]) for embedding classical data into quantum states and estimating the inner product on the 15-qubit IBMQ Melbourne quantum computer. We match the predictive performance of our QML approaches with prior QML methods and with their classical counterpart algorithms for three open-access clinical datasets. Our results imply that the qDS in small sample and feature count datasets outperforms kernel-based methods. In contrast, quantum kernel approaches outperform qDS in high sample and feature count datasets. We demonstrate that the [Formula: see text] encoding increases predictive performance with up to + 2% area under the receiver operator characteristics curve across all quantum machine learning approaches, thus, making it ideal for machine learning tasks executed in Noisy Intermediate Scale Quantum computers.

Supervised machine learning enables non-invasive lesion characterization in primary prostate cancer with [68Ga]Ga-PSMA-11 PET/MRI.

PURPOSE: Risk classification of primary prostate cancer in clinical routine is mainly based on prostate-specific antigen (PSA) levels, Gleason scores from biopsy samples, and tumor-nodes-metastasis (TNM) staging. This study aimed to investigate the diagnostic performance of positron emission tomography/magnetic resonance imaging (PET/MRI) in vivo models for predicting low-vs-high lesion risk (LH) as well as biochemical recurrence (BCR) and overall patient risk (OPR) with machine learning. METHODS: Fifty-two patients who underwent multi-parametric dual-tracer [18F]FMC and [68Ga]Ga-PSMA-11 PET/MRI as well as radical prostatectomy between 2014 and 2015 were included as part of a single-center pilot to a randomized prospective trial (NCT02659527). Radiomics in combination with ensemble machine learning was applied including the [68Ga]Ga-PSMA-11 PET, the apparent diffusion coefficient, and the transverse relaxation time-weighted MRI scans of each patient to establish a low-vs-high risk lesion prediction model (MLH). Furthermore, MBCR and MOPR predictive model schemes were built by combining MLH, PSA, and clinical stage values of patients. Performance evaluation of the established models was performed with 1000-fold Monte Carlo (MC) cross-validation. Results were additionally compared to conventional [68Ga]Ga-PSMA-11 standardized uptake value (SUV) analyses. RESULTS: The area under the receiver operator characteristic curve (AUC) of the MLH model (0.86) was higher than the AUC of the [68Ga]Ga-PSMA-11 SUVmax analysis (0.80). MC cross-validation revealed 89% and 91% accuracies with 0.90 and 0.94 AUCs for the MBCR and MOPR models respectively, while standard routine analysis based on PSA, biopsy Gleason score, and TNM staging resulted in 69% and 70% accuracies to predict BCR and OPR respectively. CONCLUSION: Our results demonstrate the potential to enhance risk classification in primary prostate cancer patients built on PET/MRI radiomics and machine learning without biopsy sampling.

Glutamate, glutamine and glutamine synthetase in the neonatal rat brain following hypoxia.

Exposing 7-day-old rat pups to hypoxia, 8% oxygen/92% nitrogen, for 3 h alters glutamate (GLU), glutamine and glutamine synthetase (GS) activity in the striatum, frontal cortex and hippocampus. Immediately following the hypoxic insult there is a rapid transient elevation of GLU followed by a fall and then recovery to control values within 6 h. Glutamine content initially decreased after the termination of the insult, rose thereafter and approached control values within 6 h. GS activity was depressed after hypoxia and gradually returned to normal levels within 6 h. GS mRNA was increased in the three brain regions studied after hypoxia and returned to control values within 24 h. These results suggest that hypoxia alters GLU metabolism in the immature brain.

Neonatal hypoxia: early neurotransmitter responses and the consequences of treatment with GM1 ganglioside.

Brain neurotransmitter content and uptake activity were assessed in the brains of 7-day-old rats 15 min after exposure to hypoxia (8% O2-92% N2) for 3 hr. Glutamate, dopamine and choline uptake were decreased in the striatum, hippocampus and frontal cortex of the hypoxic animals. Moreover, the content of glutamate, dopamine and serotonin as well as the acidic metabolites of the two biogenic amines increased in the same tissues. Acetylcholine content was decreased in all three brain regions as well. Treating the animals with GM1 ganglioside before the insult prevented all neurochemical changes in the hypoxic neonatal brain. GM1 also prevented an hypoxia-induced decrease in phorbol ester binding. Finally, GM1 ganglioside reduced the mortality rate resulting from the hypoxic insult. Our results along with those in the literature suggest that GM1 might be useful for combating the pathology associated with perinatal hypoxia.

Modulation of dopamine metabolism in the retina via dopamine D2 receptors.

When rats are placed in a lighted environment from the dark retinal DOPAC increases. There is no significant change of retinal dopamine (DA) under either lighting condition. Blockade of aromatic L-amino acid decarboxylase results in a more rapid accumulation of DOPA in the retina of animals in the light than in the dark implying that DA synthesis and metabolism are more rapid in the light than in the dark. Retinal DOPAC increases in the dark and in the light when rats are treated with the DA D2 antagonists sulpiride and spiperone. Treatment with the D2 agonist, quinpirole, lowers the content of DA in the retina of rats kept in the dark or exposed to light. D1 receptor drugs induce only limited changes in DA metabolism. We conclude that D2 receptors play a principal role for modulating DA synthesis and metabolism in the rat retina.

Aromatic L-amino acid decarboxylase is modulated by D1 dopamine receptors in rat retina.

Aromatic L-amino acid decarboxylase (AAAD) activity of rat retina increases when animals are placed in a lighted environment from the dark. The increase of activity can be inhibited by administering the selective dopamine D1 receptor agonist SKF 38393, but not the selective D2 agonist quinpirole, or apomorphine. Conversely, in the dark, enzyme activity can be enhanced by administering the selective D1 antagonist SCH 23390 or haloperidol, but not the selective D2 antagonist (-)-sulpiride. Furthermore, in animals exposed to room light for 3 h, the D1 agonist SKF 38393 reduced retinal AAAD activity, and this effect was prevented by prior administration of SCH 23390. In contrast, quinpirole had little or no effect when administered to animals in the light. Kinetic analysis indicated that the apparent Vmax for the enzyme increases with little change in the apparent Km for the substrate 3,4-dihydroxyphenylalanine or the cofactor pyridoxal-5'-phosphate. We suggest that dopamine released in the dark tonically occupies D1 receptors and suppresses AAAD activity. When the room light is turned on, D1 receptors are vacated and selective D1 agonists can either prevent the rise of AAAD or reverse light-enhanced AAAD activity.

Modulation of retinal aromatic L-amino acid decarboxylase via alpha 2 adrenoceptors.

Aromatic L-amino acid decarboxylase (AAAD) activity of the rat retina increases when animals are placed in a lighted environment from the dark. The rise of activity can be inhibited by administering alpha 2 adrenoceptor agonists. In the dark, the enzyme activity can be made to increase by administering alpha 2 adrenoceptor antagonist drugs. Kinetic analysis indicates that the maximum velocity of the enzyme increases with little change of the Km for the substrate L-3,4-dihydroxyphenylalanine or the cofactor pyridoxal-5'-phosphate. The rise of activity in the light and in the dark after alpha 2 antagonists can be blocked by administering cycloheximide, suggesting that protein synthesis is needed for the response. We speculate that epinephrine released in the dark from a subpopulation of retinal amacrine cells onto alpha 2 receptors suppresses AAAD activity that is associated with dopaminergic amacrines.

Aromatic L-amino acid decarboxylase activity of the rat retina is modulated in vivo by environmental light.

Aromatic L-amino acid decarboxylase (AAAD) activity of rat retina is low in animals placed in the dark. When the room lights are turned on, activity rises for almost 3 h and reaches values that are about twice the values found in the dark. A study of the kinetics of the enzyme revealed that the apparent Km values for L-3,4-dihydroxyphenylalanine and pyridoxal-5'-phosphate were unchanged in light- and dark-exposed animals, whereas the Vmax increased in the light. Treating the animals with cycloheximide before exposure to light prevented the increase of enzyme activity. Immunotitration with antibodies to AAAD suggested that more enzyme molecules are present in the light than in the dark. When the room lights are turned off AAAD activity drops rapidly at first and then more slowly, suggesting that at least two processes are responsible for the fall of enzyme activity. Exposure to short periods of dark followed by light results in a rapid increase of AAAD activity. Mixing homogenates from light- and dark-exposed rats results in activity values that are less than expected, suggesting the presence of an endogenous inhibitor(s). These studies demonstrate that AAAD activity is modulated in vivo.