Data infrastructures for AI in medical imaging: a report on the experiences of five EU projects

Artificial intelligence (AI) is transforming the field of medical imaging and has the potential to bring medicine from the era of 'sick-care' to the era of healthcare and prevention. The development of AI requires access to large, complete, and harmonized real-world datasets, representative of the population, and disease diversity. However, to date, efforts are fragmented, based on single-institution, size-limited, and annotation-limited datasets. Available public datasets (e.g., The Cancer Imaging Archive, TCIA, USA) are limited in scope, making model generalizability really difficult. In this direction, five European Union projects are currently working on the development of big data infrastructures that will enable European, ethically and General Data Protection Regulation-compliant, quality-controlled, cancer-related, medical imaging platforms, in which both large-scale data and AI algorithms will coexist. The vision is to create sustainable AI cloud-based platforms for the development, implementation, verification, and validation of trustable, usable, and reliable AI models for addressing specific unmet needs regarding cancer care provision. In this paper, we present an overview of the development efforts highlighting challenges and approaches selected providing valuable feedback to future attempts in the area.Key points• Artificial intelligence models for health imaging require access to large amounts of harmonized imaging data and metadata.• Main infrastructures adopted either collect centrally anonymized data or enable access to pseudonymized distributed data.• Developing a common data model for storing all relevant information is a challenge.• Trust of data providers in data sharing initiatives is essential.• An online European Union meta-tool-repository is a necessity minimizing effort duplication for the various projects in the area.

Where should epistaxis education be focused? A comparative study between the public and healthcare workers on knowledge of first aid management methods of epistaxis

BACKGROUND

Epistaxis can be life-threatening. Simple first aid management can stem bleeding. This study compared knowledge of first aid management methods of epistaxis between the general public and healthcare workers.

METHOD

A cross-sectional study of 100 healthcare workers and 103 adult members of the public was conducted at a large London teaching hospital. Respondents completed a survey assessing knowledge on nasal pinching site, head tilt and appropriate adjunct treatment use for first aid management of epistaxis.

RESULTS

Twenty-four per cent and 68 per cent of healthcare workers compared with 25.2 per cent and 37.9 per cent of the public answered correctly on nasal pinching position and head tilt position, respectively, with a statistical difference for head tilt position. Two per cent, 2 per cent and 24 per cent of healthcare workers mentioned ice use on the nose, ice use in the mouth or ice use but not site, respectively, compared with 0 per cent, 0 per cent and 4.9 per cent of the public, with a statistical difference for ice without site.

CONCLUSION

Healthcare workers and the public lack knowledge on first aid management of epistaxis. Improved education on first aid management is required, targeting healthcare workers and the public.

CHAIMELEON Project: Creation of a Pan-European Repository of Health Imaging Data for the Development of AI-Powered Cancer Management Tools

The CHAIMELEON project aims to set up a pan-European repository of health imaging data, tools and methodologies, with the ambition to set a standard and provide resources for future AI experimentation for cancer management. The project is a 4 year long, EU-funded project tackling some of the most ambitious research in the fields of biomedical imaging, artificial intelligence and cancer treatment, addressing the four types of cancer that currently have the highest prevalence worldwide: lung, breast, prostate and colorectal. To allow this, clinical partners and external collaborators will populate the repository with multimodality (MR, CT, PET/CT) imaging and related clinical data. Subsequently, AI developers will enable a multimodal analytical data engine facilitating the interpretation, extraction and exploitation of the information stored at the repository. The development and implementation of AI-powered pipelines will enable advancement towards automating data deidentification, curation, annotation, integrity securing and image harmonization. By the end of the project, the usability and performance of the repository as a tool fostering AI experimentation will be technically validated, including a validation subphase by world-class European AI developers, participating in Open Challenges to the AI Community. Upon successful validation of the repository, a set of selected AI tools will undergo early in-silico validation in observational clinical studies coordinated by leading experts in the partner hospitals. Tool performance will be assessed, including external independent validation on hallmark clinical decisions in response to some of the currently most important clinical end points in cancer. The project brings together a consortium of 18 European partners including hospitals, universities, R&D centers and private research companies, constituting an ecosystem of infrastructures, biobanks, AI/in-silico experimentation and cloud computing technologies in oncology.

PRIMAGE project: predictive in silico multiscale analytics to support childhood cancer personalised evaluation empowered by imaging biomarkers

PRIMAGE is one of the largest and more ambitious research projects dealing with medical imaging, artificial intelligence and cancer treatment in children. It is a 4-year European Commission-financed project that has 16 European partners in the consortium, including the European Society for Paediatric Oncology, two imaging biobanks, and three prominent European paediatric oncology units. The project is constructed as an observational in silico study involving high-quality anonymised datasets (imaging, clinical, molecular, and genetics) for the training and validation of machine learning and multiscale algorithms. The open cloud-based platform will offer precise clinical assistance for phenotyping (diagnosis), treatment allocation (prediction), and patient endpoints (prognosis), based on the use of imaging biomarkers, tumour growth simulation, advanced visualisation of confidence scores, and machine-learning approaches. The decision support prototype will be constructed and validated on two paediatric cancers: neuroblastoma and diffuse intrinsic pontine glioma. External validation will be performed on data recruited from independent collaborative centres. Final results will be available for the scientific community at the end of the project, and ready for translation to other malignant solid tumours.

A-26 Performance on Design Fluency and Visuoperception Measures is Related to Single and Dual Task Treadmill Gait Parameters in Healthy Adults

Objective

Research suggests a link between gait and cognition. Executive functions have been related to gait speed, however the relation between design fluency and visuoperception and other spatiotemporal gait characteristics that are related to falling is unclear. The objective of the study was to determine whether performance on design fluency and visuoperception tasks is related to spatiotemporal gait parameters during single and dual task treadmill walking in a sample of healthy adults.

Method

Nineteen healthy adults averaging 40 years of age completed cognitive measures of design fluency, visual attention, and visuoperception. They underwent gait analysis while walking on an instrumented treadmill in single task and dual task conditions.

Results

Performance on Spatial Span significantly correlated with single task stride length, r = 0.47, p = 0.043. Performance on Block Design significantly correlated with dual task stride length, r = 0.46, p = 0.049. Performance on Design Fluency significantly correlated with single task stride length variability, r = -0.50, p = 0.030, dual task stride length variability, r = -0.62, p = 0.005, and dual task step width variability, r = -0.56, p = 0.012. Performance on Picture Completion also correlated with dual task step width variability, r = -0.54, p = 0.017.

Conclusions

Design fluency and visuoperception appear related to spatiotemporal gait parameters in healthy adults. Worse cognitive performance was related to greater variability in dual task stride length and step width, gait characteristics associated with falling in aging and neurological populations.

Population-based cohort study of variation in the use of emergency cholecystectomy for benign gallbladder diseases

Background

The aims of this prospective population-based cohort study were to identify the patient and hospital characteristics associated with emergency cholecystectomy, and the influences of these in determining variations between hospitals.

Methods

Data were collected for consecutive patients undergoing cholecystectomy in acute UK and Irish hospitals between 1 March and 1 May 2014. Potential explanatory variables influencing the performance of emergency cholecystectomy were analysed by means of multilevel, multivariable logistic regression modelling using a two-level hierarchical structure with patients (level 1) nested within hospitals (level 2).

Results

Data were collected on 4744 cholecystectomies from 165 hospitals. Increasing age, lower ASA fitness grade, biliary colic, the need for further imaging (magnetic retrograde cholangiopancreatography), endoscopic interventions (endoscopic retrograde cholangiopancreatography) and admission to a non-biliary centre significantly reduced the likelihood of an emergency cholecystectomy being performed. The multilevel model was used to calculate the probability of receiving an emergency cholecystectomy for a woman aged 40 years or over with an ASA grade of I or II and a BMI of at least 25·0 kg/m2, who presented with acute cholecystitis with an ultrasound scan showing a thick-walled gallbladder and a normal common bile duct. The mean predicted probability of receiving an emergency cholecystectomy was 0·52 (95 per cent c.i. 0·45 to 0·57). The predicted probabilities ranged from 0·02 to 0·95 across the 165 hospitals, demonstrating significant variation between hospitals.

Conclusion

Patients with similar characteristics presenting to different hospitals with acute gallbladder pathology do not receive comparable care.

Adherence to guideline-based antibiotic treatment for acute exacerbations of chronic obstructive pulmonary disease in an Australian tertiary hospital

BACKGROUND

Acute exacerbations of chronic obstructive pulmonary disease (AECOPD) are an important cause of acute hospital admissions and incur significant costs, which include antibiotic costs.

AIMS

This study aimed to (i) define antibiotic prescribing practice in patients admitted to a tertiary hospital with AECOPD and compare this with current locally and nationally recognised antibiotic prescribing guidelines and (ii) correlate variations in guideline-concordant antibiotic prescribing with mean length of stay (LOS) and rates of unplanned readmission to hospital.

METHODS

Retrospective case series of 84 consecutive patients with uncomplicated AECOPD who met pre-specified selection criteria.

RESULTS

Seventy-two of 84 participants (85.7%) received guideline-discordant antibiotics, of whom the majority (76%) received intravenous antibiotics. Mean LOS was significantly lower among patients receiving guideline-concordant therapy compared with those receiving guideline-discordant therapy (mean 1.6 days vs 3.7 days; P = 0.002). There was no significant difference between groups in rates of readmission. Estimated excess costs per patient associated with guideline-discordant therapy equalled $2642 which, if eliminated, would save approximately $300 000 per annum.

CONCLUSION

In a tertiary hospital, Australian guidelines for treating patients with an AECOPD were rarely followed. The use of guideline-discordant therapy resulted in longer hospital stays and incurred greater costs. Studies are required to determine the reasons behind such discordant practice and to develop initiatives to improve antibiotic prescribing.

Outcome of patients in laparoscopic training courses compared to standard patients

BACKGROUND & AIM

Current Laparoscopic simulators have limited usefulness and patients have been used for training since the dawn of surgery. NUGITS (Northumbrian Upper Gastro Intestinal Team of Surgeons) Laparoscopic Skills courses utilise hands-on experience with simulators moving to live operating on volunteer patients. It is vital to know that the volunteer patient is not disadvantaged by greater surgical risk.

METHODS

This was a case-controlled prospective comparison of patients undergoing both Laparoscopic Cholecystectomy (LC) [n=51] and Laparoscopic Inguinal Hernia (LIH) [n=62] during NUGITS training courses. They are compared with a matched (age, sex and ASA grade) control group LC (n=51) and LIH (n=62) operated on by consultants. The outcome measures were surgical peri-and post-operative complications, post-operative hospital stay, readmission and early recurrence of inguinal hernia (<6 months).

RESULTS

In the LC cohort, there was no significant difference in the length of hospital stay (p=0.07) or readmission (p=0.16) in both the groups. The mean operating time was higher in the trainee compared to the control group (p=0.001). There was no difference in the post-operative morbidity or mortality in either group. In LIH cohort, the mean operating time was higher in the trainee compared with the control group. There was no significant difference in post-operative complications (p>0.05) and early post-operative recurrence of hernia (p>0.05).

CONCLUSION

The post-operative outcomes of patients undergoing laparoscopic surgery during laparoscopic training courses are similar to consultant-operated patients. Thus, it is acceptable and safe to encourage patients to volunteer for laparoscopic training courses.

Rhinosinusitis, symptomatology & absence of polyposis in children with primary ciliary dyskinesia

INTRODUCTION

Primary Ciliary Dyskinesia (PCD) describes a group of inherited disorders which result in functional ciliary defects leading to mucous stasis. Clinical manifestations include otitis media with effusion and chronic rhinosinusitis. Nasal polyposis has previously been thought to be linked to PCD, and current theories of 'polypogenesis' suggest that early and severe polyp formation could be expected among sufferers of this condition.

METHODS

Cross-sectional observational review of all children attending the multi-disciplinary clinic at a national tertiary-referral centre for PCD across a 3-month period. Careful examination was undertaken, and the SNOT-20 questionnaire administered.

RESULTS

Thirty patients were included. No nasal polyps were found, despite children clearly suffering rhinosinusitis and being debilitated by their symptoms. The rhinologically orientated questions of the SNOT-20 produced the most positive responses; however some other questions were found not to be useful in a paediatric population.

CONCLUSIONS

Nasal polyps do not occur in children with PCD, despite the presence of rhinosinusitis. Given that many current theories of polyp pathogenesis hinge on prolongation of proinflammatory stimuli, further investigations are needed into why this should not occur in the situation of chronic mucous stasis which is the hallmark of PCD.

Emotional intelligence and stress coping in dental undergraduates--a qualitative study

OBJECTIVE

To explore how dental undergraduates with different levels of emotional intelligence (EI) cope with stress.

DESIGN

Qualitative unstructured depth interviews.

SETTING

A dental teaching hospital in the UK, 2002.

SUBJECTS AND METHODS

Subjects selected from the undergraduate population of a 5-year dental degree course. A questionnaire survey was carried out to determine the EI scores of the subjects. In each year of study, subjects were divided into low and high EI groups at the median score. From each EI group in each year, one male and one female subject were recruited.

DATA COLLECTION

Unstructured face-to-face interviews.

DATA ANALYSIS

Transcribing, sifting, indexing and charting data according to key themes.

RESULTS

10 males and 10 females with low and high EI, representing all 5 years of study were interviewed. The experience of stress, expressed in emotional terms, ranged from anger and frustration to hatred. Four sets of coping strategies, adopted at varying degrees according to EI, were identified. High EI students were more likely to adopt reflection and appraisal, social and interpersonal, and organisation and time-management skills. Low EI students were more likely to engage in health-damaging behaviours.

CONCLUSIONS

Future research needs to establish whether the enhancement of EI in dental students would lead to improved stress-coping, and better physical and psychological health.

Membrane phospholipids and high-energy metabolites in childhood ataxia with CNS hypomyelination

BACKGROUND

Childhood ataxia with CNS hypomyelination (CACH) is a leukodystrophy with extreme rarefaction of white matter caused by mutations in one of the five subunits of the translation initiation factor 2B (eIF2B).

METHODS

Seven children with this disease and nine age-matched control subjects were studied with proton-decoupled phosphorus magnetic resonance (MR) spectroscopy.

RESULTS

In patients with CACH, cerebral concentrations of high-energy phosphate metabolites were abnormal. Of the metabolites involved in biosynthesis and catabolism of membrane phospholipids, glycerophosphorylethanolamine was reduced (0.24 +/- 0.18 mmol/kg brain vs 0.44 +/- 0.14; p < 0.02), and phosphorylethanolamine was increased (2.32 +/- 0.53 vs 1.53 +/- 0.22; p < 0.01), whereas the choline-containing phosphorylated metabolites were unchanged. Nucleoside triphosphate (NTP) was reduced (2.44 +/- 0.34 mmol/kg brain tissue vs 3.09 +/- 0.58; p < 0.01), phosphocreatine was elevated (4.11 +/- 0.63 vs 3.27 +/- 0.33; p < 0.01), and inorganic phosphate was reduced (0.77 +/- 0.32 vs 1.06 +/- 0.26; p < 0.05). Intracellular pH was elevated in patients (7.03 +/- 0.04 vs 6.99 +/- 0.02; p < 0.02).

CONCLUSIONS

The authors found an altered energy state of the residual cell population investigated. Together with previously identified replacement of white matter by CSF, the present findings raise the possibility that the genetic defect in eIF2B may result in impairment of myelin membrane synthesis or myelin membrane transport in the in vivo CACH brain. Ethanolamine metabolites constitute the plasmalogens, and the present findings may include a defect in plasmalogen metabolism.

T cell adhesion and cytolysis of pancreatic cancer cells: a role for E-cadherin in immunotherapy?

Pancreatic cancer is an aggressive and potent disease, which is largely resistant to conventional forms of treatment. However, the discovery of antigens associated with pancreatic cancer cells has recently suggested the possibility that immunotherapy might become a specific and effective therapeutic option. T cells within many epithelia, including those of the pancreas, are known to express the alphaEbeta7-integrin adhesion molecule, CD103. The only characterised ligand for CD103 is E-cadherin, an epithelial adhesion molecule which exhibits reduced expression in pancreatic cancer. In our study, CD103 was found to be expressed only by activated T cells following exposure to tumour necrosis factor beta 1, a factor produced by many cancer cells. Significantly, the expression of this integrin was restricted mainly to class I major histocompatibility complex-restricted CD8+ T cells. The human pancreatic cancer cell line Panc-1 was transfected with human E-cadherin in order to generate E-cadherin negative (wild type) and positive (transfected) sub-lines. Using a sensitive flow cytometric adhesion assay it was found that the expression of both CD103 (on T cells) and E-cadherin (on cancer cells) was essential for efficient adhesion of activated T cells to pancreatic cancer cells. This adhesion process was inhibited by the addition of antibodies specific for CD103, thereby demonstrating the importance of the CD103-->E-cadherin interaction for T-cell adhesion. Using a 51Cr-release cytotoxicity assay it was found that CD103 expressing T cells lysed E-cadherin expressing Panc-1 target cells following T cell receptor stimulation; addition of antibodies specific for CD103 significantly reduced this lysis. Furthermore, absence of either CD103 from the T cells or E-cadherin expression from the cancer cells resulted in a significant reduction in cancer cell lysis. Therefore, potentially antigenic pancreatic cancer cells could evade a local anti-cancer immune response in vivo as a consequence of their loss of E-cadherin expression; this phenotypic change may also favour metastasis by reducing homotypic adhesion between adjacent cancer cells. We conclude that effective immunotherapy is likely to require upregulation of E-cadherin expression by pancreatic cancer cells or the development of cytotoxic immune cells that are less dependent on this adhesion molecule for efficient effecter function.

Palliative stenting of malignant large bowel obstruction

OBJECTIVE: Untreated malignant large bowel obstruction is rapidly fatal. Short-term palliation of symptoms can be achieved by formation of a stoma in those patients for whom resection surgery is inappropriate. In the final months of life, a stoma represents a significant burden for both patients and carers. Palliative endoluminal stenting may therefore be an attractive alternative option for this poor prognosis group. In this paper, we examine our experience of palliative endoluminal colonic stenting. PATIENTS: Twenty patients, 11 males and 9 females of median age 81 years were referred for stenting. All had left sided colonic cancers. Ten patients had confirmed metastases on presentation, four had fixed rectal cancers and the remainder had severe comorbidity limiting surgical options. Stents were placed endoscopically using a radiologically controlled 'stent over wire' technique. RESULTS: Stenting successfully relieved the obstruction in 18 of the 20 patients attempted. In one patient the stricture could not be negotiated and the procedure was abandoned. Eleven patients have died of their disease, their median duration of palliation was 50 days (3-152 days). The rest of the patients continue in follow-up and have had 80 days median palliation (14-257 days). One stent-related complication has been observed in a patient who suffered anal pain due to fracture and migration of part of a stent into the low rectum. This complication occurred after 250 days and the distal stent fragment was removed with further symptom relief. CONCLUSION: Carefully selected patients benefit from colonic endoluminal stenting with relief of obstructive symptoms. They may be spared the potential problems associated with palliative stoma formation.

Preoperative portal vein embolisation for primary and metastatic liver tumours: volume effects, efficacy, complications and short-term outcome

BACKGROUND

The aim of portal vein embolisation is to induce hyperplasia of normal tissue when resection of a cancerous portion of the liver is contraindicated only by the volume of liver that would remain following operation.

METHODS

Eight patients with inoperable liver tumours (3 women and 5 men, median age 69.5 years, 3 colorectal hepatic metastasts, 2 choloangiocarcinomas and 3 hepatocellular cancers) were selected for portal vein embolisation. Selected portal branches were occluded with microparticles and coils. Liver volumes were determined by magnetic resonance imaging (MRI) before embolisation and again before operation.

RESULTS

Embolisation was successfully performed in all 8 patients, 7 by the percutaneous-transhepatic route, while one patient required open cannulation of a mesenteric vein. Management was altered in 6 patients who proceded to 'curative' resection; projected remaining liver volumes increased (Wilcoxon's matched pairs test p=0.02) from a median of 361 cc to a median of 550 cc; two patients had disease progression such that operation was no longer indicated. In one patient a misplaced coil unintentionally occluded a portal branch to normal liver.

CONCLUSIONS

Portal vein embolisation produced appreciable hyperplasia of the normal liver and extended the option of 'curative' operation to 6 out of the 8 cases attempted. Complications can occur. The long-term results following operation are unknown.

Regulation of T-cell apoptosis: a mixed lymphocyte reaction model

Despite the capacity for antigen-specific activation and rapid clonal expansion, homeostatic mechanisms ensure that the mature immune system contains a relatively stable number of T cells. In recent years, it has become apparent that this stability is a consequence of apoptotic death of most of the specific T cells generated during an immune response. Clearly this process must be tightly regulated in order to retain sufficient T-cell progeny to mediate an effective response, whilst allowing the rapid deletion of these cells at the end of the response to prevent lymphadenopathy and cross-reactive autoimmunity. In this study, the factors that regulate the sensitivity of T cells to apoptosis were investigated in vitro after the induction of primary T-cell activation within a mixed lymphocyte reaction (MLR). It was found that activated T cells rapidly acquire the expression of both Fas and Fas ligand (FasL) on their surface and contain high levels of the precursor form of the pro-apoptotic enzyme, caspase 8 (FLICE). However, these T cells were resistant for up to 5 days to apoptosis following the stimulation of Fas; a maximal apoptotic response was observed after 7 days. This time point coincided with a marked reduction in expression of the FLICE inhibitory protein (FLIP) and maximal activity of caspase 8. At time points beyond day 7, the number of viable cells in the MLR decreased further despite a reduction in the expression of FasL. However, the expression of interleukin-2 (IL-2) at these late time points was low, resulting in a decrease in expression of the anti-apoptotic protein Bcl-2. This can produce apoptosis by allowing leakage of cytochrome-c from mitochondria resulting in direct activation of the caspase cascade. In this study, it is shown that T cells are resistant to apoptosis for the first 5 days after activation as a consequence of insensitivity of the Fas pathway and the presence of intracellular Bcl-2. After between 5 and 7 days, the cells become sensitive to Fas-mediated apoptosis while retaining Bcl-2 expression. At later time points, Fas ligation is reduced but the cells respond to a decreased availability of IL-2 by reducing Bcl-2 expression; this encourages further apoptosis by allowing the direct activation of caspase enzymes.

Immune selection in neoplasia: towards a microevolutionary model of cancer development

The dual properties of genetic instability and clonal expansion allow the development of a tumour to occur in a microevolutionary fashion. A broad range of pressures are exerted upon a tumour during neoplastic development. Such pressures are responsible for the selection of adaptations which provide a growth or survival advantage to the tumour. The nature of such selective pressures is implied in the phenotype of tumours that have undergone selection. We have reviewed a range of immunologically relevant adaptations that are frequently exhibited by common tumours. Many of these have the potential to function as mechanisms of immune response evasion by the tumour. Thus, such adaptations provide evidence for both the existence of immune surveillance, and the concept of immune selection in neoplastic development. This line of reasoning is supported by experimental evidence from murine models of immune involvement in neoplastic development. The process of immune selection has serious implications for the development of clinical immunotherapeutic strategies and our understanding of current in vivo models of tumour immunotherapy.

Extending the indications for curative liver resection by portal vein embolization

AIMS: The aim of ipsilateral portal vein embolization is to induce hypertrophy of normal tissue when resection of a cancerous portion of the liver is contraindicated only by the volume of liver that would remain following surgery. This study reports its use in primary and metastatic liver tumours. METHODS: Eight patients with inoperable liver tumours (three women and five men of median age 68. 5 years; three colorectal hepatic metastases, two cholangiocarcinomas and three hepatocellular cancers) were selected for portal vein embolization. Selected portal branches were occluded distally with microbeads and proximally with coils. Liver volumes were determined by magnetic resonance imaging before embolization and again before surgery, 6-8 weeks later. RESULTS: Embolization was performed successfully in seven patients by the percutaneous-transhepatic route; one further patient required an open cannulation of the inferior mesenteric vein. Management was altered in six patients, who proceeded to 'curative' surgery. The projected remaining (predominantly left lobe) liver volumes increased significantly from a median of 350 to 550 ml (P < 0.05, Wilcoxon matched pairs test). Two patients had disease progression such that surgery was no longer indicated. One patient, whose disease progressed, had the left portal branch occluded unintentionally by a misplaced coil that was successfully retrieved, although the left portal branch remained occluded. CONCLUSIONS: Portal vein embolization produced significant hypertrophy of the normal liver and extended the option of 'curative' surgery to six of the eight patients in whom it was attempted. It appears to be equally effective for primary and metastatic liver tumours in selected patients.

Selection of metastatic tumour phenotypes by host immune systems

Metastasis and the processes underlying this phenomenon make epithelial cancers so malignant. Local control of cancers by surgery is sometimes possible but locoregional and distant recurrence commonly lead to the failure of treatment with ensuing morbidity and mortality. Tumour cells express a range of new antigens during growth and there are opportunities for the host immune system to interact with these antigens. This immune interaction eliminates the tumour or allows selection of phenotypic variants. Cell phenotypes selected by an incomplete immune response resemble the cell type commonly associated with metastases. Thus we propose that the host immune system may be responsible for selection of this phenotype and progression of the disease.

Evidence that metastasis is less common in cirrhotic than normal liver: a systematic review of post-mortem case-control studies

BACKGROUND

It has been hypothesized that the cirrhotic liver is afforded protection against metastasis. The evidence has been examined and the plausibility of such a phenomenon is reviewed.

METHODS

A systematic literature review was conducted with analysis of combined data from post-mortem case-control studies.

RESULTS

Overall, the crude rate of metastasis to normal liver was 37.3 per cent, while the rate to cirrhotic liver was 23.7 per cent. The Mantel-Haenszel (MH) fixed-effects estimate of the odds ratio was 0. 47 (95 per cent confidence interval (c.i.) 0.41-0.53; chi2 = 136, 11 d.f., P < 0.001). The DerSimonian-Laird (DL) random-effects estimation of the odds ratio was 0.42 (95 per cent c.i. 0.31-0.58; chi2 = 28, 1 d.f., P < 0.001). For tumours arising within the distribution of the portal vein, the crude rate of metastasis to normal liver was 47.6 per cent, whereas the rate to cirrhotic liver was 29.8 per cent. The MH estimate of the odds ratio was 0.45 (95 per cent c.i. 0.37-0.54; chi2 = 68.2, 5 d.f., P < 0.001). The DL pooled odds ratio was 0.44 (95 per cent c.i. 0.28-0.70; chi2 = 12.3, 1 d.f., P < 0.001). The MH and DL pooled estimates of the odds ratio were similar for groups of patients from the East (Japan) and the West (Europe and the USA).

CONCLUSION

The post-mortem evidence reviewed suggests that the likelihood of metastasis to the cirrhotic liver is lower than that to normal liver. The degree of protection for tumours arising from within the distribution of the portal vein is neither greater nor less than it is overall. Eastern and Western populations appear to have a similar degree of risk reduction. The differences noted were significant on testing in the meta-analysis, but confounding bias accounting for these differences has not been excluded.

In vivo magnetic resonance spectroscopy of human fetal neural transplants

To better define the survival and cellular composition of human fetal neurotransplants in vivo, we performed quantitative 1H MRS to determine the concentration of the neuronal amino acid [N-acetylaspartate] within MRI-visible grafts. In all, 71 grafts in 38 patients [24 Parkinson's disease (PD), 14 Huntington's disease (HD)] were examined, as well as 24 untreated PD and HD patients and 13 age-matched normal controls. MRI appearances of edema were present in three out of 71 grafts, the remainder being consistent with histologically identified viable neural transplant tissue. N-acetylaspartate (NAA), creatine, choline, myoinositol and glutamine plus glutamate (Glx) were identified in all post-transplant putamens, with abnormal metabolites, lactate and/or lipid detectable in only three patients. Of 71 grafts, 19 occupied more than 60% of the MRS-examined volume (VOI) (mean 84.2 +/- 3%; range 61-100%). In those, [NAA] was 8.50 +/- 0.99 mM in eight PD spectra and 6.59 +/- 0.81 mM in 11 HD spectra, and was not significantly different from controls. In contrast, transplanted fetal neurones contain less than 0.4 mM of the neuronal amino acid NAA. This suggests that established fetal neurotransplants in the human putamen of both PD and HD patients are populated by adult neurones, axons and dendrites.

Changes in esophageal function after vertical banded gastroplasty as demonstrated by esophageal scintigraphy

BACKGROUND

The effects of surgery for morbid obesity on the function of the upper gastrointestinal (GI) tract are of interest to bariatric surgeons. This study was undertaken to determine any changes in esophageal function, following vertical banded gastroplasty (VBG) in morbidly obese patients, as detected by esophageal scintigraphy.

METHODS

Ten consecutive morbidly obese patients (six female and four male) underwent preoperative esophageal scintigraphy and upper GI endoscopy. These investigations were repeated 12 months after VBG to coincide with expected appreciable weight reduction. The results were tabulated together with body mass indices, crude weights and percentage excess weight lost.

RESULTS

Before VBG one patient gave a history of mild heartburn, one had mild dyspepsia and the remaining eight patients had no GI symptoms. No patient had a hiatus hernia or endoscopic evidence of reflux esophagitis. Preoperatively all patients had abnormal scintiscans. The abnormalities were esophageal retention (all) and intraesophageal reflux (five out of 10 patients). Gastroesophageal reflux was not identified in any patient. Postoperatively scintiscans were normal or improved in six out of 10 patients and unchanged in four out of 10 patients. In three patients the scans were normal and three showed overall improvement in esophageal function, although in one of these latter patients gastroesophageal reflux was observed.

CONCLUSIONS

In this series of morbidly obese patients, esophageal function as assessed by scintigraphy was abnormal. Following VBG it improved in six out of 10 patients and was unchanged in four out of 10. However, in one patient, who had shown an overall improvement in esophageal function, gastroesophageal reflux was demonstrated when it had not been seen preoperatively. This was asymptomatic. Thus, adverse changes in esophageal function after VBG were uncommon.

Quantitation of presynaptic cardiac sympathetic function with carbon-11-meta-hydroxyephedrine

The purpose of this study was to validate an axially distributed blood-tissue exchange model for the quantitation of cardiac presynaptic sympathetic nervous system function that could be applied to PET images. The model accounts for heterogeneity in myocardial blood flow, differences in transport rates of 11C-meta-hydroxyephedrine (mHED) across the capillary endothelium and/or neuronal membranes, the virtual volumes of distribution in the interstitial space and neuron and retention of mHED in the neuronal vesicles.

METHODS

Multiple indicator outflow dilution and residue detection methods were used to measure the kinetics of radiolabeled intravascular space and interstitial space markers and 11C-mHED in isolated perfused rat heart at baseline and during norepinephrine neuronal transporter blockade with desipramine (DMI). The outflow dilution and residue detection data were modeled with a multiple pathway, four-region, axially distributed model of blood-tissue exchange describing flow in the capillary and exchange between regions using permeability-surface area products with units of clearance of milliliters per minute per gram. Meta-hydroxyephedrine may enter the nerve terminal via membrane transport, where it may be sequestered by first-order unidirectional uptake within vesicles. Release of mHED from the vesicles is modeled via exchange with the interstitial space.

RESULTS

After intracoronary injection, mHED transport across the capillary endothelium and in the interstitial space closely followed that of sucrose. Subsequently, mHED was retained in the heart, whereas sucrose washed out rapidly. With DMI the outflow dilution curves more closely resembled those of sucrose. Model parameters reflecting capillary-interstitial kinetics and volumes of distribution were unchanged by DMI, whereas parameters reflecting the neuronal transporter process and volumes of distribution in the nerve terminal and vesicular sequestration were markedly decreased by DMI. Application of the model to a pilot set of canine PET images of mHED suggests the feasibility of this approach.

CONCLUSION

Meta-hydroxyephedrine kinetics in the heart can be quantitated using an axially distributed, blood-tissue exchange model that accounts for heterogeneity of flow, reflects changes in neuronal function and is applicable to PET images.

Ion-exchange column chromatographic method for assaying purine metabolic pathway enzymes

High energy phosphate levels fall rapidly during cardiac ischemia and recover slowly (more than one week) during reperfusion. The slow recovery of ATP may reflect a lack of purine metabolic precursors and/or increased activity of purine catabolic enzymes such as 5'-nucleotidase (5'-NT, EC 3.1.3.5) and adenosine deaminase (ADA, EC 3.5.4.4). The activity of enzymes involved in both the catabolism of ATP precursors (5-NT and ADA) and the restoration of ATP from slow synthetic pathways [adenosine kinase (AK, EC 2.7.1.20), adenine phosphoribosyl transferase (APRT, EC 2.4.2.7) and hypoxanthine phosphoribosyl transferase (HPRT, EC 2.4.2.8)] may directly affect the rate of ATP recovery. Strategies to enhance recovery will depend on the relative activity of these enzymes following ischemia. Their activity in different species and their response to ischemia are not well characterized. Hence, rapid assay methods for these enzymes would facilitate detailed time course studies of their activities in postischemic myocardium. We modified a single ion-exchange column chromatographic method using DEAE-Sephadex to determine the products of incubation of 5'-NT, AK, APRT and HPRT with their respective substrates. The uniformity of the final product measurement procedure for all assays permits the activities of the four enzymes to be rapidly determined in a single tissue sample and facilitates the study of a large number of samples. This technique should also be useful for enzymes of the pyrimidine metabolic pathway.

Assessment of shoulder dimensions and angles of porcelain bonded to metal crown preparations

The metal ceramic crown is the most popular extracoronal restoration in the United Kingdom. These restorations may fail because of fracture or esthetics. A potential cause of failure is the quality and width of the facial shoulder preparation. In this study 24 extracted human teeth were prepared to receive metal ceramic crowns by one of three dentists. Preparations were replicated and scanned in the midfacial plane by a coordinate measuring machine with a noncontact probe. The x, y, and z surface coordinates were recorded. The results indicated a mean (+/-SD) shoulder width value of 0.752 mm (+/-0.174 mm) and a shoulder angle of 108.54 (+/-15.06) degrees. From these data it would appear that there are deficiencies in shoulder preparations, particularly in width. These inadequacies may have implications for longevity of the restoration and periodontal health in a clinical situation.

Impairments on neuropsychologic tests of temporohippocampal and frontohippocampal functions and word fluency in remitting schizophrenia and affective disorders

Experimental neuropsychologic tests were administered to acute patients with schizophrenia and affective psychosis and to normal controls. Patients had remitting illnesses. Tests included memory for recurring digit and block spans (left and right temporohippocampal), digit and block spans (lateralized parietal/frontal), spatial and nonspatial conditional associate learning (frontohippocampal), and oral word fluency to letter-designated categories (frontal) and semantic-designated categories (left-sided). In 81% of schizophrenic patients patterned deficits incompatible with generalized losses of function were disclosed. Patterns were heterogeneous and characterized by (1) the frequency and severity of left temporohippocampal impairment; (2) asymmetric frontohippocampal function such that severity of bilateral impairment was associated with poorer nonspatial learning and superior performance with better nonspatial learning; (3) syndrome relationships predicted by the hemisphere imbalance syndrome model pertaining to positive and negative symptoms and the catatonic syndrome; (4) a generalized deficit independent of temporohippocampal functions; and (5) no relationship between performance and computed tomographic signs or medication. Patients with affective disorders had patterned deficits characterized by bilateral impairments that disclosed a preponderance of deficits in spatial learning and memory; depressives demonstrated impairments in digit span.