Breast Tumor Characterization Using [18F]FDG-PET/CT Imaging Combined with Data Preprocessing and Radiomics

Simple Summary

Breast cancer is the second most common diagnosed malignancy in women worldwide. In this study, we examine the feasibility of breast tumor characterization based on [¹⁸F]FDG-PET/CT images using machine learning (ML) approaches in combination with data-preprocessing techniques. ML prediction models for breast cancer detection and the identification of breast cancer receptor status, proliferation rate, and molecular subtypes were established and evaluated. Furthermore, the importance of most repeatable features was investigated. Results displayed high performance of malignant/benign tumor differentiation and triple negative tumor subtype ML models. We observed high repeatability of radiomic features for both high performing predictive models.

Abstract

Background: This study investigated the performance of ensemble learning holomic models for the detection of breast cancer, receptor status, proliferation rate, and molecular subtypes from [¹⁸F]FDG-PET/CT images with and without incorporating data pre-processing algorithms. Additionally, machine learning (ML) models were compared with conventional data analysis using standard uptake value lesion classification. Methods: A cohort of 170 patients with 173 breast cancer tumors (132 malignant, 38 benign) was examined with [¹⁸F]FDG-PET/CT. Breast tumors were segmented and radiomic features were extracted following the imaging biomarker standardization initiative (IBSI) guidelines combined with optimized feature extraction. Ensemble learning including five supervised ML algorithms was utilized in a 100-fold Monte Carlo (MC) cross-validation scheme. Data pre-processing methods were incorporated prior to machine learning, including outlier and borderline noisy sample detection, feature selection, and class imbalance correction. Feature importance in each model was assessed by calculating feature occurrence by the R-squared method across MC folds. Results: Cross validation demonstrated high performance of the cancer detection model (80% sensitivity, 78% specificity, 80% accuracy, 0.81 area under the curve (AUC)), and of the triple negative tumor identification model (85% sensitivity, 78% specificity, 82% accuracy, 0.82 AUC). The individual receptor status and luminal A/B subtype models yielded low performance (0.46–0.68 AUC). SUV_max model yielded 0.76 AUC in cancer detection and 0.70 AUC in predicting triple negative subtype. Conclusions: Predictive models based on [¹⁸F]FDG-PET/CT images in combination with advanced data pre-processing steps aid in breast cancer diagnosis and in ML-based prediction of the aggressive triple negative breast cancer subtype.

Assessment of the kidney function parameters split function, mean transit time, and outflow efficiency using dynamic FDG-PET/MRI in healthy subjects

Background

Traditionally, isotope nephrography is considered as the method of choice to assess kidney function parameters in nuclear medicine. We propose a novel approach to determine the split function (SF), mean transit time (MTT), and outflow efficiency (OE) with 2-deoxy-2-[18F]fluoro-D-glucose (FDG) dynamic positron emission tomography (PET).

Materials and methods

Healthy adult subjects underwent dynamic simultaneous FDG-PET and magnetic resonance imaging (PET/MRI). Time-activity curves (TACs) of total kidneys, renal cortices, and the aorta were prospectively obtained from dynamic PET series. MRI images were used for anatomical correlation. The same individuals were subjected to dynamic renal Technetium-99 m-mercaptoacetyltriglycine (MAG3) scintigraphy and TACs of kidneys; the perirenal background and the left ventricle were determined. SF was calculated on the basis of integrals over the TACs, MTT was determined from renal retention functions after deconvolution analysis, and OE was determined from MTT. Values obtained from PET series were compared with scintigraphic parameters, which served as the reference.

Results

Twenty-four subjects underwent both examinations. Total kidney SF, MTT, and OE as estimated by dynamic PET/MRI correlated to their reference values by r = 0.75, r = 0.74 and r = 0.81, respectively, with significant difference (p < 0.0001) between the means of MTT and OE. No correlations were found for cortex FDG values.

Conclusions

The study proofs the concept that SF, MTT, and OE can be estimated with dynamic FDG PET/MRI scans in healthy kidneys. This has advantages for patients receiving a routine PET/MRI scan, as kidney parameters can be estimated simultaneously to functional and morphological imaging with high accuracy.

Multiple haploids, triploids, and tetraploids found in modern-day "living fossil" Ginkgo biloba

Ginkgo biloba, the last extant representative of a lineage of Mesozoic gymnosperms, is one of the few seed plants with an exceptionally long (~300 Myr) evolutionary history free of genome-wide duplications (polyploidy). Despite this genome conservatism, we have recently found a viable spontaneous tetraploid Ginkgo sapling during routine screening of several plants, demonstrating that natural polyploidy is possible in Ginkgo. Here we provide a much wider flow cytometry survey of ploidy in some European Ginkgo collections, and own seedlings (>2200 individuals and ~200 cultivars). We found a surprisingly high level of ploidy variation in modern-day Ginkgo and documented altogether 13 haploid, 3 triploid, and 10 tetraploid Ginkgo plants or cultivars, most of them being morphologically distinct from common diploids. Haploids frequently produced polyploid (dihaploid) buds or branches. Tetraploids showed some genome size variation. The surveyed plants provide a unique resource for future Ginkgo research and breeding, and they might be used to accelerate the modern diversification of this nearly extinct plant lineage.

Assessing the kidney function parameters glomerular filtration rate and effective renal plasma flow with dynamic FDG-PET/MRI in healthy subjects

Background

A method was developed to assess the kidney parameters glomerular filtration rate (GFR) and effective renal plasma flow (ERPF) from 2-deoxy-2-[¹⁸F]fluoro-d-glucose (FDG) concentration behavior in kidneys, measured with positron emission tomography (PET) scans.

Twenty-four healthy adult subjects prospectively underwent dynamic simultaneous PET/magnetic resonance imaging (MRI) examination. Time activity curves (TACs) were obtained from the dynamic PET series, with the guidance of MR information. Patlak analysis was performed to determine the GFR, and based on integrals, ERPF was calculated. Results were compared to intra-individually obtained reference values determined from venous blood samples.

Results

Total kidney GFR and ERPF as estimated by dynamic PET/MRI were highly correlated to their reference values (r = 0.88/p < 0.0001 and r = 0.82/p < 0.0001, respectively) with no significant difference between their means.

Conclusions

The study is a proof of concept that GFR and ERPF can be assessed with dynamic FDG PET/MRI scans in healthy kidneys. This has advantages for patients getting a routine scan, where additional examinations for kidney function estimation could be avoided. Further studies are required for transferring this PET/MRI method to PET/CT applications.

Electronic supplementary material

The online version of this article (10.1186/s13550-018-0389-1) contains supplementary material, which is available to authorized users.

Recent Borexino results and perspectives of the SOX measurement

Borexino is a liquid scintillator detector sited underground in the Laboratori Nazionali del Gran Sasso (Italy). Its physics program, until the end of this year, is focussed on the study of solar neutrinos, in particular from the Beryllium, pp, pep and CNO fusion reactions. Knowing the reaction chains in the sun provides insights towards physics disciplines such as astrophysics (star physics, star formation, etc.), astroparticle and particle physics. Phase II started in 2011 and its aim is to improve the phase I results, in particular the measurements of the neutrino fluxes from the pep and CNO processes. By the end of this year, data taking from the sun will be over and a new project is scheduled to launch: Short distance Oscillation with boreXino (SOX), which uses a Cerium source for neutrinos (100÷150 kCi of activity) and aims to confirm or rule out the presence of sterile neutrinos. This particle is hypothesised to justify the reactor, Gallium and LSND anomalies found and can reject extensions to the standard model. The work presented is a summary of the solar neutrino results achieved so far, which lead not only to a precise study of the processes in the sun, but also to more Standard Model oriented measurements (such as the stability of the charge, i.e. the life time of the electron). Furthermore, the perspectives of the SOX program are discussed showing the experiment sensitivity to a fourth neutrino state covering almost entirely 3σ of the preferred region of the anomalous neutrino experiments, and additional applications of the detector such as the study of geo-neutrinos.

Impact of the Fukushima accident on tritium, radiocarbon and radiocesium levels in seawater of the western North Pacific Ocean: A comparison with pre-Fukushima situation

Tritium, radiocarbon and radiocesium concentrations in water column samples in coastal waters offshore Fukushima and in the western North Pacific Ocean collected in 2011-2012 during the Ka'imikai-o-Kanaloa (KoK) cruise are compared with other published results. The highest levels in surface seawater were observed for ¹³⁴Cs and ¹³⁷Cs in seawater samples collected offshore Fukushima (up to 1.1 Bq L^-1), which represent an increase by about three orders of magnitude when compared with the pre-Fukushima concentration. Tritium levels were much lower (up to 0.15 Bq L^-1), representing an increase by about a factor of 6. The impact on the radiocarbon distribution was measurable, but the observed levels were only by about 9% above the global fallout background. The ¹³⁷Cs (and similarly ¹³⁴Cs) inventory in the water column of the investigated western North Pacific region was (2.7 ± 0.4) PBq, while for ³H it was only (0.3 ± 0.2) PBq. Direct releases of highly contaminated water from the damaged Fukushima NPP, as well as dry and wet depositions of these radionuclides over the western North Pacific considerably changed their distribution patterns in seawater. Presently we can distinguish Fukushima labeled waters from global fallout background thanks to short-lived ¹³⁴Cs. However, in the long-term perspective when ¹³⁴Cs will decay, new distribution patterns of ³H, ¹⁴C and ¹³⁷Cs in the Pacific Ocean should be established for future oceanographic and climate change studies in the Pacific Ocean.

Validation of an automated morphological MRI-based (123)I-FP-CIT SPECT evaluation method

INTRODUCTION

Dopamine transporter imaging with (123)I-FP-CIT single photon emission computed tomography (SPECT) is helpful for the differential diagnosis between Parkinsonian syndrome (PS) and essential tremor (ET). Although visual assessment and time-consuming manual evaluation techniques are readily available, a fully objective and automated dopamine transporter quantification technique is always preferable, at least in research and follow-up investigations. Our aim was to develop a novel automated magnetic resonance imaging (MRI)-based evaluation technique of dopamine transporter SPECT images and to compare its diagnostic accuracy with those of the gold-standard visual grading and manual dopamine transporter binding quantification methods.

METHODS

(123)I-FP-CIT SPECT and MRI sessions were conducted in 33 patients with PS (15 men; mean age: 60.3 ± 9.7 years) and 15 patients with ET (8 men; mean age: 54.7 ± 16.3 years). Striatal dopamine transporter binding was visually classified by 2 independent experts as normal or abnormal grade I, II and III. Caudal and putaminal specific uptake ratios were calculated by both automated MRI-based and manual evaluation techniques.

RESULTS

We found almost perfect agreement (κ = 0.829) between the visual scores by the 2 observers. The automated method showed strong correlation with the visual and manual evaluation techniques and its diagnostic accuracy (sensitivity = 97.0%; specificity = 93.3%) was also comparable to these methods. The automatically determined uptake parameters showed negative correlation with the clinical severity of parkinsonism. Based on ordinal regression modelling, the automated MRI-based method could reliably determine the visual grading scores.

CONCLUSION

The novel MRI-based evaluation of (123)I-FP-CIT SPECT images is useful for the differentiation of PS from ET.

Test of Electric Charge Conservation with Borexino

Borexino is a liquid scintillation detector located deep underground at the Laboratori Nazionali del Gran Sasso (LNGS, Italy). Thanks to the unmatched radio purity of the scintillator, and to the well understood detector response at low energy, a new limit on the stability of the electron for decay into a neutrino and a single monoenergetic photon was obtained. This new bound, τ≥6.6×10^{28}  yr at 90% C.L., is 2 orders of magnitude better than the previous limit.

Lifetimes of (214)Po and (212)Po measured with Counting Test Facility at Gran Sasso National Laboratory

The decays of (214)Po into (210)Pb and of (212)Po into (208)Pb tagged by the previous decays from (214)Bi and (212)Bi have been studied inserting quartz vials inside the Counting Test Facility (CTF) at the underground laboratory in Gran Sasso (LNGS). We find that the mean lifetime of (214)Po is (236.00 ± 0.42(stat) ± 0.15(syst)) μs and that of (212)Po is (425.1 ± 0.9(stat) ± 1.2(syst)) ns. Our results are compatible with previous measurements, have a much better signal to background ratio, and reduce the overall uncertainties.

Growth regulator requirement for in vitro embryogenic cultures of snowdrop (Galanthus nivalis L.) suitable for germplasm preservation

In this study, we report on the production of bulb scale-derived tissue cultures capable of efficient shoot and plant regeneration in three genotypes of snowdrop (Galanthus nivalis L., Amaryllidaceae), a protected ornamental plant. For culture line A, high auxin and low cytokinin concentration is required for callus production and plant regeneration. The type of auxin is of key importance: α-naphthaleneacetic acid (NAA) in combination with indole-3-acetic acid (IAA) at concentrations of 2 mg L-1 or 2-10 mg L-1 NAA with 1 mg L-1 N6-benzyladenine (BA), a cytokinin on full-strength media are required for regeneration. Cultures showing regeneration were embryogenic. When lines B and C were induced and maintained with 2 mg L-1 NAA and 1 mg L-1 BA, they produced mature bulblets with shoots, without roots. Line A produced immature bulblets with shoots under the above culture condition. Amplified Fragment Length Polymorphism (AFLP) analysis showed that (i) genetic differences between line A and its bulb explants were not significant, therefore these tissue cultures are suitable for germplasm preservation, and (ii) different morphogenetic responses of lines A, B and C originated from genetic differences. Culture line A is suitable for field-growing, cultivation and germplasm preservation of G. nivalis and for the production of Amaryllidaceae alkaloids.

Phylogenetic and comparative functional analysis of the cell-separation α-glucanase Agn1p in Schizosaccharomyces

The post-cytokinetic separation of cells in cell-walled organisms involves enzymic processes that degrade a specific layer of the division septum and the region of the mother cell wall that edges the septum. In the fission yeast Schizosaccharomyces pombe, the 1,3-α-glucanase Agn1p, originally identified as a mutanase-like glycoside hydrolase family 71 (GH71) enzyme, dissolves the mother cell wall around the septum edge. Our search in the genomes of completely sequenced fungi identified GH71 hydrolases in Basidiomycota, Taphrinomycotina and Pezizomycotina, but not in Saccharomycotina. The most likely Agn1p orthologues in Pezizomycotina species are not mutanases having mutanase-binding domains, but experimentally non-characterized hypothetical proteins that have no carbohydrate-binding domains. The analysis of the GH71 domains corroborated the phylogenetic relationships of the Schizosaccharomyces species determined by previous studies, but suggested a closer relationship to the Basidiomycota proteins than to the Ascomycota proteins. In the Schizosaccharomyces genus, the Agn1p proteins are structurally conserved: their GH71 domains are flanked by N-terminal secretion signals and C-terminal sequences containing the conserved block YNFNA(Y)/HTG. The inactivation of the agn1(Sj) gene in Schizosaccharomyces japonicus, the only true dimorphic member of the genus, caused a severe cell-separation defect in its yeast phase, but had no effect on the hyphal growth and yeast-to-mycelium transition. It did not affect the mycelium-to-yeast transition either, only delaying the separation of the yeast cells arising from the fragmenting hyphae. The heterologous expression of agn1(Sj) partially rescued the separation defect of the agn1Δ cells of Schizosaccharomyces pombe. The results presented indicate that the fission yeast Agn1p 1,3-α-glucanases of Schizosaccharomyces japonicus and Schizosaccharomyces pombe share conserved functions in the yeast phase.

Monitoring of tritium, 60Co and 137Cs in the vicinity of the warm water outlet of the Paks Nuclear Power Plant, Hungary

Danube water, sediment and various aquatic organisms (snail, mussel, predatory and omnivorous fish) were collected upstream (at a background site) and downstream of the outlet of the warm water channel of Paks Nuclear Power Plant. Gamma emitters, tissue free-water tritium (TFWT) and total organically-bound tritium (T-OBT) measurements were performed. A slight contribution of the power plant to the natural tritium background concentration was measured in water samples from the Danube section downstream of the warm water channel. Sediment samples also contained elevated tritium concentrations, along with a detectable amount of (60)Co. In the case of biota samples, TFWT exhibited only a very slight difference compared to the tritium concentration of the Danube water, however, the OBT was higher than the tritium concentration in the Danube, independent of the origin of the samples. The elevated OBT concentration in the mollusc samples downstream of the warm water channel may be attributed to the excess emission from the nuclear power plant. The whole data set obtained was used for dose rate calculations and will be contributed to the development of the ERICA database.

PET-CT based automated lung nodule detection

An automatic method is presented in order to detect lung nodules in PET-CT studies. Using the foreground and background mean ratio independently in every nodule, we can detect the region of the nodules properly. The size and intensity of the lesions do not affect the result of the algorithm, although size constraints are present in the final classification step. The CT image is also used to classify the found lesions built on lung segmentation. We also deal with those cases when nearby and similar nodules are merged into one by a split-up post-processing step. With our method the time of the localization can be decreased from more than one hour to maximum five minutes. The method had been implemented and validated on real clinical cases in Interview Fusion clinical evaluation software (Mediso). Results indicate that our approach is very effective in detecting lung nodules and can be a valuable aid for physicians working in the daily routine of oncology.

New workflows and algorithms of bone scintigraphy based on SPECT-CT

Gold standard bone scintigraphy workflow contains acquisition of planar anterior and posterior images and if necessary, additional SPECTs as well. Planar acquisitions are time consuming and not enough for accurately locating hotspots. Current paper proposes a novel workflow for fast whole body bone SPECT scintigraphy. We present a novel stitching method to generate a whole body SPECT based on the SPECT projections. Our stitching method is performed on the projection series not on the reconstructed SPECTs, thus stitching artifacts are greatly reduced. Our workflow does not require any anterior-posterior image pairs, since these images are derived from the reconstructed whole body SPECT automatically. Our stitching method has been validated on real clinical data performed by medical physicians. Results show that our method is very effective for whole body SPECT generations leaving no signs of artifacts. Our workflow required overall 16 minutes to acquire a whole body SPECT which is comparable to the 60 minutes acquisition time required for gold standard techniques including planar images and additional SPECT acquisitions.