ABSTRACT
In recent years healthcare is undergoing significant changes due to technological innovations, with Artificial Intelligence (AI) being a key trend. Particularly in radiodiagnostics, according to studies, AI has the potential to enhance accuracy and efficiency. We focus on AI's role in diagnosing pulmonary lesions, which could indicate lung cancer, based on chest X-rays. Despite lower sensitivity in comparison to other methods like chest CT, due to its routine use, X-rays often provide the first detection of lung lesions. We present our deep learning-based solution aimed at improving lung lesion detection, especially during early-stage of illness. We then share results from our previous studies validating this model in two different clinical settings: a general hospital with low prevalence findings and a specialized oncology center. Based on a quantitative comparison with the conclusions of radiologists of different levels of experience, our model achieves high sensitivity, but lower specificity than comparing radiologists. In the context of clinical requirements and AI-assisted diagnostics, the experience and clinical reasoning of the doctor play a crucial role, therefore we currently lean more towards models with higher sensitivity over specificity. Even unlikely suspicions are presented to the doctor. Based on these results, it can be expected that in the future artificial intelligence will play a key role in the field of radiology as a supporting tool for evaluating specialists. To achieve this, it is necessary to solve not only technical but also medical and regulatory aspects. It is crucial to have access to quality and reliable information not only about the benefits but also about the limitations of machine learning and AI in medicine.
Subject(s)
Artificial Intelligence , Lung Neoplasms , Radiography, Thoracic , Humans , Lung Neoplasms/diagnostic imaging , Czech Republic , Retrospective Studies , Sensitivity and Specificity , Early Detection of Cancer/methods , Deep LearningABSTRACT
In liver surgery, biliary obstruction can lead to secondary biliary cirrhosis, a life-threatening disease with liver transplantation as the only curative treatment option. Mesenchymal stromal cells (MSC) have been shown to improve liver function in both acute and chronic liver disease models. This study evaluated the effect of allogenic MSC transplantation in a large animal model of repeated biliary obstruction followed by partial hepatectomy. MSC transplantation supported the growth of regenerated liver tissue after 14 days (MSC group, n = 10: from 1087 ± 108 (0 h) to 1243 ± 92 mL (14 days); control group, n = 11: from 1080 ± 95 (0 h) to 1100 ± 105 mL (14 days), p = 0.016), with a lower volume fraction of hepatocytes in regenerated liver tissue compared to resected liver tissue (59.5 ± 10.2% vs. 70.2 ± 5.6%, p < 0.05). Volume fraction of connective tissue, blood vessels and bile vessels in regenerated liver tissue, serum levels of liver enzymes (AST, ALT, ALP and GGT) and liver metabolites (albumin, bilirubin, urea and creatinine), as well as plasma levels of IL-6, IL-8, TNF-α and TGF-ß, were not affected by MSC transplantation. In our novel, large animal (pig) model of repeated biliary obstruction followed by partial hepatectomy, MSC transplantation promoted growth of liver tissue without any effect on liver function. This study underscores the importance of translating results between small and large animal models as well as the careful translation of results from animal model into human medicine.
Subject(s)
Cholestasis/complications , Disease Models, Animal , Liver Diseases/therapy , Mesenchymal Stem Cell Transplantation/methods , Animals , Liver Diseases/etiology , Liver Diseases/pathology , Mesenchymal Stem Cells , SwineABSTRACT
AIM: The aim of this study was to determine the proportion of cystic tumors according to preoperative CT (Bosniak III, IV) among surgically treated patients with histologically confirmed papillary renal cell carcinoma (pRCC) and to assess progression rates among patients with and without cystic appearance on imaging. METHODS: A total of 138 patients with pRCC histology surgically treated in the period of January 2007-March 2017 were included. Clinical and radiological characteristics, type of surgery, histopathology results, and follow-up data were recorded and statistically evaluated. RESULTS: Forty-one cases (29.7%) of cystic lesions (10× BIIF, 14× BIII, 17× BIV) were detected by CT. Patients with pRCC1 significantly more frequently presented with cystic appearance on CT (33/78; 42.3%) in comparison to other papillary types (8/60; 13.3%; p = 0.0002). During a median follow-up time of 49.4 months, only 2 patients with cystic lesions progressed after surgery. CONCLUSIONS: Cystic appearance on imaging methods is mainly a characteristic of pRCC1 (42.3%). Cystic morphology on imaging might predict a relatively indolent behavior of all pRCC types. Preoperative scoring systems including tumor growth patterns (cystic vs. solid) are needed for further classification.
Subject(s)
Carcinoma, Renal Cell/diagnostic imaging , Kidney Diseases, Cystic/diagnostic imaging , Kidney Neoplasms/diagnostic imaging , Adult , Aged , Aged, 80 and over , Carcinoma, Renal Cell/pathology , Contrast Media , Disease Progression , Female , Humans , Kaplan-Meier Estimate , Kidney Diseases, Cystic/pathology , Kidney Neoplasms/pathology , Male , Middle Aged , Neoplasm Recurrence, Local , Preoperative Period , Retrospective Studies , Tomography, X-Ray ComputedABSTRACT
: The authors present 2 cases of a subcutaneous biphasic synovial sarcoma with marked apocrine differentiation that potentially may be confused with cutaneous epithelial neoplasms, including malignant apocrine mixed tumor or metaplastic carcinoma with an apocrine glandular component. Microscopically, both neoplasms had a biphasic architecture with the epithelial and spindle cell components. The epithelial component was prominent and consisted of simple glands with round lumina and complex glandular structures with intraluminal bridges forming cribriform areas. The glands were lined by cuboidal to columnar cells with eosinophilic or clear cytoplasm manifesting apical apocrine-like and intraluminal eosinophilic secretions. The spindle cell component was less prominent and was composed of relatively uniform or slightly atypical spindle sells surrounding and merging focally with the glandular structures. Immunohistochemically, the tumor cells in both components were positive for vimentin, AE1/AE3, CK7, and epithelial membrane antigen. Desmin, smooth muscle actin, muscle-specific actin, CD34, and S-100 protein were all negative. SYT-SSX1 gene fusion using fluorescence in situ hybridization and RT-PCR methods was detected in both cases.
Subject(s)
Sarcoma, Synovial/pathology , Soft Tissue Neoplasms/pathology , Apocrine Glands/pathology , Biomarkers, Tumor/analysis , Cell Differentiation , Female , Humans , In Situ Hybridization, Fluorescence , Male , Middle Aged , Oncogene Proteins, Fusion/genetics , Reverse Transcriptase Polymerase Chain Reaction , Sarcoma, Synovial/genetics , Soft Tissue Neoplasms/geneticsABSTRACT
PURPOSE: The new generation of 3TMRI has improved spatial and time resolutions, which are favourable in imaging of the renal vasculature. In this study, we have compared the imaging findings of the renal blood vessels using 3TMRI and CT with intraoperative assessment of the renal vasculature as gold standard. METHODS: This prospective study was approved by the local ethical committee. Between 4/2011 and 12/2011, 80 patients with renal tumours underwent 3TMRA (angiography) (Magnetom SKYRA 3T, Siemens). Twenty of the patients were also examined with CT AG. The results of the CTA- and MRA-imaging studies were correlated with the intraoperative assessment of the renal vessels. RESULTS: Seventy patients (87.5 %) had a detailed intraoperative assessment of the renal vessels. The sensitivities for CTA and MRA were 88.2 and 88.6 %, respectively. All discrepancies between imaging studies and intraoperative findings were due to inability to identify small polar vessels. The results of MRA were concordant with CTA in 85.0 % of cases. The (three) discrepancies between MRI and CT were due to failure of MRI in identifying small polar vessels. CONCLUSIONS: (1) 3TMRA gives detailed information about the renal vasculature including its topographical anatomy. (2) With MRI, small aberrant vessels are more frequently missed than with CTA. (3) CTA remains the gold standard. However, MRA may be used for planning of laparoscopic operations. (4) The quality of the 3D reconstruction is highly depending on the skills of the radiologist.
Subject(s)
Kidney Neoplasms/blood supply , Kidney Neoplasms/surgery , Magnetic Resonance Angiography/methods , Preoperative Care/methods , Humans , Imaging, Three-Dimensional , Kidney/blood supply , Kidney/diagnostic imaging , Kidney/pathology , Kidney Neoplasms/pathology , Laparoscopy/methods , Nephrectomy/methods , Prospective Studies , Retrospective Studies , Sensitivity and Specificity , Tomography, X-Ray ComputedABSTRACT
Angiogenesis in healthy tissue and within malignant tumors differs on many levels, which may partly be explained by vascular mimicry formation resulting in altered contrast material or different radiopharmaceuticals distributions. Failed remodulation results in changes in the molecular exchange through the capillary wall and those consequences affect the behavior of contrast agents and radiopharmaceuticals. One of the most indicative signs of malignant tissue is the increased permeability and the faster molecular exchange that occurs between the extracellular and intravascular spaces. Dynamic imaging can help to assess the changed microenvironment. The fast-distribution of molecules reflects newly developed conditions in blood-flow redistribution inside a tumor and within the affected organ during the early stages of tumor formation. Tumor development, as well as aggressiveness, can be assessed based on the change to the vascular bed development, the level of molecular exchange within the tissue, and/or indicative distribution within the organ. The study of the vascular network organization and its impact on the distribution of molecules is important to our understanding of the image pattern in several imaging methods, which in turn influences our interpretation of the findings. A hybrid imaging approach (including PET/MRI) allows the quantification of vascularization and/or its pathophysiological impressions in structural and metabolic images. It might optimize the evaluation of the pretreatment imaging, as well as help assess the effect of therapy targeting neovascularization; antiVEGF drugs and embolization-based therapies, for example.
Subject(s)
Neoplasms , Radiopharmaceuticals , Humans , Neovascularization, Pathologic/diagnostic imaging , Neovascularization, Pathologic/pathology , Neoplasms/diagnostic imaging , Magnetic Resonance Imaging/methods , Contrast Media , Perfusion , Tumor MicroenvironmentABSTRACT
BACKGROUND/AIM: To test the correlation of 68Ga-PSMA-11 uptake and the expression of PSMA (prostatic specific membrane antigen) with the Gleason score, apparent diffusion coefficient (ADC) and pharmacokinetic parameters obtained from dynamic contrast agent-enhanced MRI/PET. PATIENTS AND METHODS: Forty newly diagnosed, therapy naïve patients with prostatic carcinoma (PC) (mean age of 56.7, range=34-79), who were referred for 68Ga-PSMA-11-PET/MRI for primary staging and had undergone radical prostatectomy (RAPE) were included in this prospective study. Their blood samples were tested for serum levels of prostate-specific antigen (PSA) and proPSA. The patients' prostates were evaluated using whole-mount sections, which helped determine the extent and grade of the tumor; tests were performed to determine immunohistochemical PSMA expression. RESULTS: A correlation between PSMA expression and the accumulation of 68Ga-PSMA-11 was found using the Spearman correlation coefficient (p=0.0011). A stronger correlation was found between Gleason patterns 3 or 4 and PSMA expression (p=0.06). Furthermore, the correlation of Gleason score with the overall 68Ga-PSMA-11 accumulation within the tumor or non-tumor tissue was found to be significant (p=0.0157). A significant relation was found only with the Kep elimination rate constant, which was stronger in Gleason pattern 4 than in Gleason pattern 3. A weaker correlation was found between the accumulation of 68Ga-PSMA-11 and Ktrans in Gleason pattern 4: the most significant relation being between ADCmin and Gleason pattern 3 and 4 (p=0.0074). The total size of the tumor correlated with levels of proPSA (p<0.0001), and its extra prostatic extension correlated with levels of proPSA (p<0.0001). CONCLUSION: 68Ga-PSMA-11 correlates well with the expression of PSMA. Gleason pattern 3 and 4 had a higher correlation with 68Ga-PSMA-11 levels than did Gleason pattern 5. Either no correlation, or a weak correlation, was established with pharmacokinetics.
Subject(s)
Carcinoma , Prostatic Neoplasms , Male , Humans , Middle Aged , Prostate/pathology , Positron Emission Tomography Computed Tomography , Neoplasm Grading , Prospective Studies , Oligopeptides , Edetic Acid , Gallium Radioisotopes , Prostatic Neoplasms/diagnostic imaging , Prostatic Neoplasms/surgery , Prostatic Neoplasms/metabolism , Positron-Emission Tomography , Magnetic Resonance ImagingABSTRACT
BACKGROUND/AIMS: To find out whether the total tumor mass and post-ablation necrosis volume influence the disease-free survival of patients following radiofrequency ablation (RFA). METHODOLOGY: Fifty nine patients with RFA of primary and secondary tumors were evaluated retrospectively in a four year period. Total liver mass, post-ablation necrosis volume and their ratio were evaluated using computed tomography examination in the relationship with the risk of insufficient tumor ablation and the disease-free patients survival. RESULTS: A complete ablation was performed in 51 patients, non-ablation in 8 (13.6%) patients. Tumor, necrosis volume were 19.2±19.5, 58.7±44.7mL, respectively. The tumor and necrosis mass ratio was 0.39±0.45. The tumor or necroses mass volume or the tumor/necroses mass ratio had no effect on the patients progression-free survival. Patients with a necrosis volume <25mL had a 10-times higher risk of insufficient ablation (OR=9.9; 95% CI=1.9-51.5; p<0.002) and patients with the tumor/necrosis mass ratio >0.4 had a 8-times higher risk of insufficient ablation (OR=7.9; 95% CI=1.4-44.6; p<0.01). CONCLUSIONS: Necrosis volume after RFA and tumor/necrosis mass ratio are the important factors for insufficient ablation but do not have any influence on the patients progression-free survival.
Subject(s)
Carcinoma, Hepatocellular/pathology , Catheter Ablation , Colorectal Neoplasms/pathology , Liver Neoplasms/pathology , Liver/pathology , Aged , Carcinoma, Hepatocellular/surgery , Disease-Free Survival , Female , Humans , Kaplan-Meier Estimate , Liver Neoplasms/secondary , Liver Neoplasms/surgery , Male , Middle Aged , Necrosis/diagnostic imaging , Necrosis/pathology , Retrospective Studies , Tomography, X-Ray ComputedABSTRACT
BACKGROUND/AIMS: Portal vein embolization (PVE) extends the resecability of liver tumours.The issue of PVE is an insufficient growth of the liver parenchyma or a tumour progression in some patients. We evaluated the effect of the volume and the number of liver tumours on the effect of PVE. METHODOLOGY: PVE was performed in 40 patients with liver tumours due to an insufficient future remnant liver volume. The number and the volume of the tumours were evaluated and compared with the final PVE effect. RESULTS: In patients without any increase of the liver volume after PVE (n=3) the number and the volume of the tumours before PVE were 2.7±2.1 and 2205.1±2432.7mm3, respectively. In patients with sufficient growth of the liver (n=22) it was 3.8±2.2 (NS) and 1164.9±1392.1mm3 (NS), respectively. In patients with tumour progression (n=11) it was 5.6±2.2 and 6971.4±5189.5mm3, respectively (p<0.04 and p<0.005, respectively). Four patients were treated by radiofrequency ablation only due to worsening of their health state. Patients with >4 foci (OR 4.7) and a tumour volume >400mm3 (OR=13.0) had a higher probability of cancer progression or insufficient growth of the liver tissue. Patients with <6 foci and a tumour volume <3100mm3 had an 87.5% probability of a successful liver hypertrophy after PVE. CONCLUSIONS: The tumour number and volume were crucial for progression of a malignant disease and growth of the liver parenchyma after PVE.
Subject(s)
Embolization, Therapeutic , Liver Neoplasms/therapy , Liver Regeneration , Neoadjuvant Therapy , Neoplasms, Multiple Primary/therapy , Portal Vein , Adult , Aged , Cone-Beam Computed Tomography , Disease Progression , Embolization, Therapeutic/adverse effects , Female , Hepatectomy , Humans , Liver Neoplasms/blood supply , Liver Neoplasms/diagnostic imaging , Liver Neoplasms/secondary , Liver Neoplasms/surgery , Logistic Models , Male , Middle Aged , Neoadjuvant Therapy/adverse effects , Neoplasms, Multiple Primary/blood supply , Neoplasms, Multiple Primary/diagnostic imaging , Neoplasms, Multiple Primary/pathology , Neoplasms, Multiple Primary/surgery , Odds Ratio , Predictive Value of Tests , ROC Curve , Time Factors , Treatment Outcome , Tumor BurdenABSTRACT
BACKGROUND/AIMS: Portal vein ligation (PVL) could multiply the future liver remnant volume (FLRV). Tumor necrosis factor- alpha (TNF-α) is a pleiotropic cytokine that is connected with initial phase of liver regeneration. The aim of this basic pilot study was to accelerate regeneration of liver parenchyma after PVL. The experimental porcine model was developed to be as much compatible as possible with portal vein embolization (PVE) in human medicine. METHODOLOGY: After ligation of portal branches of caudate and right lateral and right medial liver lobes recombinant porcine TNF-α (TNF-α group) or physiological solution (control group) were applied into non-occluded portal vein branches. The biochemical and immunoanalytical parameters were assessed. The compensatory hypertrophy was evaluated by periodic ultrasonography. The histological examination of liver was performed. RESULTS: The acceleration of growth of hypertrophic liver lobes was maximal at the 7th postoperative day in comparison with the control group (p<0.05); nevertheless this stimulating effect was lost at the end of experiment. The important differences in biochemical or histological studied parametres between study groups were not proved. CONCLUSIONS: The achieved acceleration of growth of hypertrophic liver lobes after application of TNF-α confirms the role of studied cytokine in priming of liver regeneration.
Subject(s)
Hepatocytes/drug effects , Liver Regeneration/drug effects , Liver/blood supply , Liver/drug effects , Portal Vein/surgery , Tumor Necrosis Factor-alpha/pharmacology , Animals , Animals, Newborn , Biomarkers/blood , Cell Proliferation/drug effects , Disease Models, Animal , Hepatocytes/pathology , Ligation , Liver/diagnostic imaging , Liver/metabolism , Liver/pathology , Recombinant Proteins/pharmacology , Swine , Time Factors , Tumor Necrosis Factor-alpha/blood , UltrasonographyABSTRACT
BACKGROUND/AIMS: TGF-ß1 is a pleiotropic cytokine that is over expressed in terminal phase of liver regeneration. METHODOLOGY: Twenty-four hours after partial portal vein ligation monoclonal antibody against TGF-ß1 (TGF-ß1 group, 7 piglets) or physiological solution (control group, 9 piglets) were applied into the central venous catheter. The biochemical parameters (bilirubin, urea, creatinine, alkaline phosphatase, gamma- glutamyl transferase, cholinesterase, aspartate aminotransferase, alanine aminotransferase and albumin) were assessed. The compensatory hypertrophy of the non-occluded liver lobes was evaluated by periodic ultrasonography during the next fourteen days and by histological examination. RESULTS: The acceleration of growth of the hypertrophic liver lobes was maximal between 3rd and 7th postoperative days in comparison with the control group (p<0.05). No important differences in the biochemical or studied histological parameters were proved. CONCLUSIONS: The present study describes a new usage of monoclonal antibody against TGF-ß1 in large animal experimental model of partial portal vein ligation.
Subject(s)
Antibodies, Monoclonal/pharmacology , Cell Proliferation/drug effects , Liver Regeneration/drug effects , Liver/drug effects , Portal Vein/surgery , Transforming Growth Factor beta1/antagonists & inhibitors , Animals , Animals, Newborn , Biomarkers/metabolism , Hypertrophy , Ligation , Liver/blood supply , Liver/metabolism , Liver/pathology , Models, Animal , Swine , Time Factors , Transforming Growth Factor beta1/metabolismABSTRACT
Tissue hypoxia becomes the driver of the micro-environmental transformation of the malignant tumorous tissue in majority of solid tumors. Even more, during the development of solid tumors, a large amount of nutrients is consumed due to rapid proliferation of tumor cells. The way of use of nutrients could be assessed using the probes during positron emission tomography (PET), the most frequent is the use of 18F-fluorodeoxyglucose (18F-FDG) demonstrating the turn-over of glucose exploitation in a manner of Warburg phenomenon. Even tumor cells can undergo metabolic reprogramming by changing the expression of glycolysis-related proteins, development of hypoxic condition increases glucose uptake to promote their growth. The hypoxic conditions are stimulation also the development of drug and/or radiation resistance, thus, the identification of the hypoxic tissue could play an important role in therapy planning. The process that enables use of hypoxia specific PET radiopharmaceutical starts when it enters cell by passive diffusion. Intracellularly, a model substance - 18F-fluoromisonidazol (18F-FMISO) is reduced by nitroreductase enzymes to become trapped in cells within reduced tissue oxygen partial pressure. The identification of the hypoxic tissue aids to plan the target to radiotherapy in squamous cell carcinomas of head and neck or cervical carcinoma, when 18F-FMISO-PET/CT is used as a planning procedure for a target volume, while 18F-FDG-PET/MRI is being the local and whole body staging procedure. Hypoxia imaging helps to identify the high-grade component of an astrocytic tumor before stereotactic biopsy.
Subject(s)
Carcinoma, Squamous Cell , Head and Neck Neoplasms , Female , Fluorodeoxyglucose F18 , Humans , Hypoxia , Magnetic Resonance Imaging , Positron Emission Tomography Computed Tomography , Positron-Emission Tomography/methods , Radiopharmaceuticals , Tomography, X-Ray Computed , Tumor MicroenvironmentABSTRACT
OBJECTIVES: The first objective of our study was to determine the radiation exposure received by patients during tin-filtrated ultra-low-dose computed tomography (TFULDCT) of sacroiliac joints and to compare those to conventional X-ray doses. For comparison, we added a cohort examined by low-dose CT (LDCT) without tin filtration. The second objective was to compare the results of TFULDCT and X-ray in the detection of sacroiliitis. METHODS: Our retrospective study covered 45 patients, who were examined for suspected axial spondyloarthritis (AxSpA). The first group underwent TFULDCT as well as conventional radiography (CR); the second group underwent LDCT only without tin filtration. Effective doses of TFULDCT, LDCT and CR were calculated by an experienced medical physicist. TFULDCT and CR were independently evaluated by three investigators, who decided on the presence or absence of rheumatoid inflammatory bone changes. The results were statistically evaluated. RESULTS: In our cohort, the median effective dose for TFULDCT was 0.11 mSv, range (0.06-0.40 mSv), for LDCT 0.5 mSv (0.29-0.89 mSv), and for CR 0.25 mSv (0.06-1.87 mSv). We proved that TFULDCT produces a significantly lower percentage of uncertain results (23.3%; 95% CI: 11.3-41.6%) than CR (66.7%; 95% CI: 48.3-81.1%). CONCLUSIONS: Tin filtration helps to reduce CT radiation exposure to values lower than those resulting from CR. TFULDCT offers better overall diagnostic performance than CR. Our results prove that TFULDCT can replace CR in the diagnosis of sacroiliitis in the radiographical stage of AxSpA.
Subject(s)
Sacroiliitis , Tin , Humans , Radiation Dosage , Radiography , Retrospective Studies , Sacroiliitis/diagnostic imaging , Tomography, X-Ray Computed/methods , X-RaysABSTRACT
PURPOSE: The aims of this work were the retrospective analysis of a cohort of patients with acute mediastinitis treated at the authors' worksite over a 15-year period and the identification of factors that significantly affect the outcomes of the therapy. METHODS: During the period 2006-2020, 80 patients with acute mediastinitis were treated. Within the cohort, the following were observed: the causes and the type of acute mediastinitis, length of anamnesis, comorbidities, diagnostic methods, time from the diagnosis to surgery, types and number of surgical procedures, results of microbiological tests, complications, and outcomes of the treatment. RESULTS: The most common type of acute mediastinitis was descending mediastinitis (48.75%). A total of 116 surgical procedures were performed. Ten patients in the cohort died (12.5%). Patients older than 60 years were at a 6.8 times higher risk of death. Patients with more than two comorbidities were at a 14.3 times higher risk of death. The presence of yeasts in the culture material increased the risk of death by 4.4 times. CONCLUSION: Early diagnosis, removal of the cause of mediastinitis, sufficient mediastinal debridement, and multiple drainage thereof with the possibility of continual postoperative lavage are essential for the successful treatment of acute mediastinitis.
Subject(s)
Mediastinitis , Acute Disease , Drainage/methods , Humans , Mediastinitis/diagnosis , Mediastinitis/etiology , Mediastinitis/surgery , Prognosis , Retrospective Studies , Treatment OutcomeABSTRACT
Liver volumetry is an important tool in clinical practice. The calculation of liver volume is primarily based on Computed Tomography. Unfortunately, automatic segmentation algorithms based on handcrafted features tend to leak segmented objects into surrounding tissues like the heart or the spleen. Currently, convolutional neural networks are widely used in various applications of computer vision including image segmentation, while providing very promising results. In our work, we utilize robustly segmentable structures like the spine, body surface, and sagittal plane. They are used as key points for position estimation inside the body. The signed distance fields derived from these structures are calculated and used as an additional channel on the input of our convolutional neural network, to be more specific U-Net, which is widely used in medical image segmentation tasks. Our work shows that this additional position information improves the results of the segmentation. We test our approach in two experiments on two public datasets of Computed Tomography images. To evaluate the results, we use the Accuracy, the Hausdorff distance, and the Dice coefficient. Code is publicly available at: https://gitlab.com/hachaf/liver-segmentation.git.
ABSTRACT
BACKGROUND/AIM: This study determined whether computed tomography (CT) is an appropriate means by which to differentiate non-invasive and minimally invasive forms of pulmonary adenocarcinoma from the invasive variant. PATIENTS AND METHODS: A total of 64 patients (38 men and 26 women, aged 42-76, mean age 64), who underwent surgery for pulmonary adenocarcinoma and a chest CT no less than 1 month before surgery, were included in the study. Lesions exhibiting ground glass opacity or ground glass opacity with a solid component of 5 mm or smaller, were defined as minimally invasive or non-invasive adenocarcinomas. CT findings were correlated with histopathological examination. RESULTS: Distinguishing minimally invasive and non-invasive adenocarcinoma from invasive adenocarcinoma using CT was achieved with a sensitivity of 77.7%, a specificity of 97.8%, a positive predictive value of 93.3%, and a negative predictive value of 91.8%. CONCLUSION: CT can be useful in assessing the degree of invasiveness of pulmonary adenocarcinoma and is a potential tool for the individualization of treatment.
Subject(s)
Adenocarcinoma of Lung/pathology , Adenocarcinoma of Lung/surgery , Lung Neoplasms/pathology , Lung Neoplasms/surgery , Radiographic Image Interpretation, Computer-Assisted/methods , Adenocarcinoma of Lung/diagnostic imaging , Adult , Aged , Diagnosis, Differential , Female , Humans , Lung Neoplasms/diagnostic imaging , Male , Middle Aged , Precision Medicine , Sensitivity and Specificity , Tomography, X-Ray ComputedABSTRACT
OBJECTIVES: to assess the feasibility of CT with an integrated photon-counting-detector system (PC-CT) in the body imaging of clinical patients. METHODS: 120 examinations using photon counting detector CT were evaluated in six groups: 1/ a standard-dose lung, 2/ low-dose lung, 3/ ultra-high resolution (UHR) lung, 4/ standard-dose abdominal, 5/ lower-dose abdominal, 6/ UHR abdominal CTA. All CT examinations were performed on a single-source prototype device equipped with a photon counting detector covering a 50 cm scan field of view. Standard dose examinations were performed with the use of detector element size of 0.4 mm, ultra-high-resolution examinations with the detector element size of 0.2 mm, respectively. The stability of the system during imaging was tested. The diagnostic quality of the acquired images was assessed based on the imaging of key structures and the noise level in five-point scale, the effective dose equivalent, dose length product and noise level, and also relation to body mass index and body surface area were compared with three similar groups of CT images made with energy integrating high end scanner. The parameters were evaluated using Wilcoxon test for independent samples, the independence was tested using Kruskal-Wallis test. RESULTS: When PC-CT images radiation dose is compared with the similar imaging using energy integrating CT, the PC-CT shows lower dose in ultra-high resolution mode, the dose is significantly lower (p < 0.0001), the standard dose examinations were performed with the comparable radiation doses. PC-CT exhibited the significantly higher ratio between parenchyma signal and background noise both in lung and in abdominal imaging (p < 0.0001). CONCLUSIONS: PC-CT showed imaging stability and excellent diagnostic quality at dose values that are comparable or better to the quality of energy integrating CT, the better signal and improved resolution is most important advantage of photon counting detector CT over energy integrating detector CT.
Subject(s)
Photons , Tomography, X-Ray Computed , Humans , Lung , Phantoms, ImagingABSTRACT
AIM: The aim of the study was to evaluate the yields of 2-deoxy-2-[18F]-fluoro-D-glucose positron-emission tomography/magnetic resonance imaging (18F-FDG-PET/MRI) and 18F-FDG-PET/computed tomography (CT) for the detection of metastatic involvement of locoregional neck nodes in patients with head and neck malignancy. PATIENTS AND METHODS: A total of 90 patients (66 men and 24 women) met the inclusion criteria. Preoperative staging was performed: 53 examinations using PET/CT and 37 using PET/MRI. RESULTS: For PET/MRI, the study demonstrated a sensitivity of 89%, specificity of 64%, positive predictive value of 85%, and negative predictive value of 70%; PET/CT had a sensitivity of 95%, specificity 47%, positive predictive value of 82%, and negative predictive value of 78%. CONCLUSION: Both methods have a high yield in N-clinical staging with statistically insignificant differences. We assume PET/MRI to be the first-choice method in organ-targeted examinations, for pediatric patients and repeat examinations. In cases of MRI contraindication, PET/CT can be used with no impact on the quality of care.
Subject(s)
Fluorodeoxyglucose F18 , Positron Emission Tomography Computed Tomography , Child , Female , Humans , Lymph Nodes/diagnostic imaging , Lymphatic Metastasis/diagnostic imaging , Magnetic Resonance Imaging , Male , Neoplasm Staging , Positron-Emission Tomography , Radiopharmaceuticals , Sensitivity and Specificity , Tomography, X-Ray ComputedABSTRACT
BACKGROUND: In patients with colorectal liver metastases, the possibility for radical liver resection can be limited by oxaliplatin-induced sinusoidal obstruction syndrome (SOS). This study investigates the potential of mesenchymal stem cells (MSC) to improve the outcome of liver resections in pigs with SOS. MATERIALS AND METHODS: SOS was induced in all animals (n=20) on day 0. Animals in the experimental group (n=8) received allogeneic MSC on day 7. Liver resection was performed in all animals on day 14 and the animals were observed until day 28. Ultrasound volumetry, biochemical analysis and histological examination of liver parenchyma was performed during the follow-up period. RESULTS: Six animals from the control group died prematurely, while all animals survived in the experimental group. According to histology, biochemical analysis and ultrasound volumetry, there were no significant differences between the groups documenting the effect of MSC. CONCLUSION: Single dose allogeneic MSC administration improved survival of animals with SOS undergoing partial liver resection. Further experiments with different timing of liver resection and MSC administration should be performed to investigate the effect of MSC in more detail.
Subject(s)
Hepatectomy , Hepatic Veno-Occlusive Disease/pathology , Hepatic Veno-Occlusive Disease/therapy , Mesenchymal Stem Cell Transplantation , Mesenchymal Stem Cells , Animals , Biomarkers , Colorectal Neoplasms/pathology , Combined Modality Therapy , Disease Models, Animal , Female , Hepatectomy/methods , Hepatic Veno-Occlusive Disease/etiology , Immunohistochemistry , Immunophenotyping , Liver Neoplasms/complications , Liver Neoplasms/secondary , Male , Mesenchymal Stem Cells/cytology , Swine , Treatment OutcomeABSTRACT
The purpose of this study was to assess virtual unenhanced brain computed tomography (CT) images obtained by dual-energy CT angiography (CTA) for the detection of intracranial bleeding. In total, 25 patients were included in the study (average age 53.2 years, range 25-75 years, 14 male, 11 female), all with intracranial bleeding on unenhanced brain CT and who underwent additional CTA performed on a dual-source CT in a dual-energy acquisition mode. The two X-ray tubes were operated at 140 and 80 kV, respectively. Data were analyzed using dual-energy evaluation software. Virtual unenhanced images were calculated by removing the relative iodine content from each voxel. The virtual unenhanced images were evaluated by a radiologist blinded to the findings of the conventional images related to the presence of intracranial bleeding. The image quality and contrast-to-noise ratio (CNR) between bleeding and brain tissue were assessed. The virtual image quality was found to be sufficient in 96%. The agreement in detection of intracranial bleeding on virtual and conventional unenhanced images reached 96% in per-lesion analysis and 100% in per-patient analysis. The averaged CNR reached 2.63 in virtual unenhanced images and 3.27 in conventional. Virtual unenhanced images are sufficient for the detection of intracranial bleeding.