2-[C-11]thymidine imaging of malignant brain tumors

Malignant brain tumors pose diagnostic and therapeutic problems. Despite the advent of new brain imaging modalities, including magnetic resonance imaging (MRI) and [F-18]fluorodeoxyglucose (FDG) positron emission tomography (PET), determination of tumor viability and response to treatment is often difficult. Blood-brain barrier disruption can be caused by tumor or nonspecific reactions to treatment, making MRI interpretation ambiguous. The high metabolic background of the normal brain and its regional variability makes it difficult to identify small or less active tumors by FDG imaging of cellular energetics. We have investigated 2-[C-11]thymidine (dThd) and PET to image the rate of brain tumor cellular proliferation. A series of 13 patients underwent closely spaced dThd PET, FDG PET, and MRI procedures, and the image results were compared by standardized visual analysis. The resulting dThd scans were qualitatively different from the other two scans in approximately 50% of the cases, which suggests that dThd provided information distinct from FDG PET and MRI. In two cases, recurrent tumor was more apparent on the dThd study than on FDG; in two other patients, tumor dThd uptake was less than FDG uptake, and these patients had slower tumor progression than the three patients with both high dThd and FDG uptake. To better characterize tumor proliferation, kinetic modeling was applied to dynamic dThd PET uptake data and metabolite-analyzed blood data in a subset of patients. Kinetic analysis was able to remove the confounding influence of [C-11]CO2, the principal labeled metabolite of 2-[C-11]dThd, and to estimate the flux of dThd incorporation into DNA. Sequential, same-day [C-11]CO2 and [C-11]dThd imaging demonstrated the ability of kinetic analysis to model both dThd and CO2 simultaneously. Images of dThd flux obtained using the model along with the mixture analysis method for pixel-by-pixel parametric imaging significantly enhanced the contrast of tumor compared with normal brain. Comparison of model estimates of dThd transport versus dThd flux was able to discern increased dThd uptake simply on the basis of blood-brain barrier disruption retention on the basis of increased cellular proliferation. This preliminary study demonstrates the potential for imaging brain tumor cellular proliferation to provide unique information for guiding patient treatment.

Use of ridge regression for improved estimation of kinetic constants from PET data

The estimation of parameters in radio-tracer models from positron emission tomography (PET) data by nonlinear least squares (NLS) often leads to results with unacceptable mean square error (ME) characteristics. The introduction of constraints on parameters has the potential to address this problem. We examine a ridge-regression technique that augments the standard NLS criterion by the addition of a term which penalizes estimates which deviate from physiologically reasonable values. A variation on a plug-in methodology of Hoerl et al. [7] is examined for data-dependent selection of the degree of reliance to place on the penalizing term. A simulation study is carried out to evaluate the performance of this approach in the context of estimation of kinetic constants in the three-compartment model used to analyze data from PET studies with fluoro-deoxyglucose (FDG). Results show that over a range of realistic noise levels, the ridge-regression procedure can be expected to reduce the root ME of parameter estimates by 60%. This result is not found to be substantially dependent on the precise formulation of the penalty function used. Thus, the use of ridge regression for estimation of kinetic parameters in PET studies is considered to be a promising tool.

Intrapartum risk factors for newborn encephalopathy: the Western Australian case-control study

OBJECTIVE

To identify intrapartum predictors of newborn encephalopathy in term infants.

DESIGN

Population based, unmatched case-control study.

SETTING

Metropolitan area of Western Australia, June 1993 to September 1995.

SUBJECTS

All 164 term infants with moderate or severe newborn encephalopathy; 400 randomly selected controls.

MAIN OUTCOME MEASURES

Adjusted odds ratio estimates.

RESULTS

The birth prevalence of moderate or severe newborn encephalopathy was 3.8/1000 term live births. The neonatal fatality was 9.1%. Maternal pyrexia (odds ratio 3.82), a persistent occipitoposterior position (4.29), and an acute intrapartum event (4.44) were all risk factors for newborn encephalopathy. More case infants than control infants were induced (41.5% and 30.5%, respectively) and fewer case infants were delivered by caesarean section without labour (3.7% and 14.5%, respectively). Operative vaginal delivery (2.34) and emergency caesarean section (2.17) were both associated with an increased risk. There was an inverse relation between elective caesarean section (0.17) and newborn encephalopathy. After application of a set of consensus criteria for elective caesarean section only three (7%) eligible case mothers compared with 33 (65%) eligible control mothers were sectioned electively. Of all the case infants, 113 (69%) had only antepartum risk factors for newborn encephalopathy identified; 39 (24%) had antepartum and intrapartum factors; eight (5%) had only intrapartum factors; and four (2%) had no recognised antepartum or intrapartum factors.

CONCLUSIONS

The causes of newborn encephalopathy are heterogeneous and many relate to the antepartum period. Elective caesarean section has an inverse association with newborn encephalopathy. Intrapartum hypoxia alone accounts for only a small proportion of newborn encephalopathy. These results question the view that most risk factors for newborn encephalopathy lie in the intrapartum period.

Antepartum risk factors for newborn encephalopathy: the Western Australian case-control study

OBJECTIVE

To ascertain antepartum predictors of newborn encephalopathy in term infants.

DESIGN

Population based, unmatched case-control study.

SETTING

Metropolitan area of Western Australia, June 1993 to September 1995.

SUBJECTS

All 164 term infants with moderate or severe newborn encephalopathy; 400 randomly selected controls.

MAIN OUTCOME MEASURES

Adjusted odds ratio estimates.

RESULTS

The birth prevalence of moderate or severe newborn encephalopathy was 3.8/1000 term live births. The neonatal fatality was 9.1%. The risk of newborn encephalopathy increased with increasing maternal age and decreased with increasing parity. There was an increased risk associated with having a mother who was unemployed (odds ratio 3.60), an unskilled manual worker (3.84), or a housewife (2.48). Other risk factors from before conception were not having private health insurance (3.46), a family history of seizures (2.55), a family history of neurological disease (2.73), and infertility treatment (4.43). Risk factors during pregnancy were maternal thyroid disease (9.7), severe pre-eclampsia (6.30), moderate or severe bleeding (3.57), a clinically diagnosed viral illness (2.97), not having drunk alcohol (2.91); and placenta described at delivery as abnormal (2.07). Factors related to the baby were birth weight adjusted for gestational age between the third and ninth centile (4.37) or below the third centile (38.23). The risk relation with gestational age was J shaped with 38 and 39 weeks having the lowest risk.

CONCLUSIONS

The causes of newborn encephalopathy are heterogeneous and many of the causal pathways start before birth.

Clinical and immunologic effects of a PRIMATIZED anti-CD4 monoclonal antibody in active rheumatoid arthritis: results of a phase I, single dose, dose escalating trial

OBJECTIVE

The goal of this single infusion, dose escalation study was to evaluate the safety of the PRIMATIZED anti-CD4 monoclonal antibody (Mab), IDEC-CE9.1, in patients with rheumatoid arthritis (RA).

METHODS

Twenty-five patients received single infusions of IDEC-CE9.1 in dose escalation form (0.03 to 4 mg/kg). Cohorts consisted of 3 patients each with seropositive RA. Following treatment, patients were monitored for 2 weeks before initiation of treatment of the next cohort. Peripheral blood samples were taken during and after treatment to measure immune function. Flow cytometry of peripheral blood mononuclear cells and in vitro proliferative responses to antigens and recall antigens were assessed pre and post-treatment. Cell surface markers CD3, CD4 (OKT4 and Leu 3a), CD8, CD20, CD25, CD45Ro, CD45Ra and DR were analyzed, and proliferation to mitogens and recall antigens was measured.

RESULTS

No infusion related adverse events were noted and other drug related adverse events were mild. Reduction in peripheral CD4 T cell number was brief (3 to 7 days) and not associated with infection. CD4 cell surface antigen downmodulation was observed postinfusion. Suppression of CD25 expression was associated with a positive clinical response. In vitro proliferative responses to mitogens and antigen were inhibited for up to one month with no association to positive clinical response.

CONCLUSION

IDEC-CE9.1 appears to have a benign safety profile and may modulate immune function rather than deplete CD4+ T cells.

Glucose metabolism in human malignant gliomas measured quantitatively with PET, 1-[C-11]glucose and FDG: analysis of the FDG lumped constant

Calculation of the glucose metabolic rate (MRGlc) in brain with PET and 2-[18F]fluoro-2-deoxy-D-glucose (FDG) requires knowing the rate of uptake of FDG relative to glucose from plasma into metabolite pools in the tissue. The proportionality factor for this is the FDG lumped constant (LC[FDG]), the ratio of the volumes of distribution of FDG and glucose multiplied by the hexokinase phosphorylation ratio for the two hexoses, Km(Glc) x Vm(FDG)/Km(FDG) x Vm(Glc) x MRGlc equals the FDG metabolic rate (MRFDG) divided by the LC(FDG), i.e., MRGlc = MRFDG/LC(FDG) and LC(FDG) = MRFDG/MRGlc. This investigation tested the hypothesis that LC(FDG) is significantly higher in gliomas than it is in brain uninvolved with tumor.

METHODS

We imaged 40 patients with malignant gliomas with 1-[11C]glucose followed by FDG. The metabolic rates MRGlc and MRFDG were estimated for glioma and contralateral brain regions of interest by an optimization program based on three-compartment, four-rate constant models for the two hexoses.

RESULTS

The LC(FDG), estimated as MRFDG/MRGlc, in gliomas was 1.40 +/- 0.46 (mean +/- s.d.; range = 0.72-3.10), whereas in non-tumor-bearing contralateral brain, it was 0.86 +/- 0.14 (range = 0.61-1.21) (p < 0.001, glioma versus contralateral brain).

CONCLUSION

These data strongly suggest that the glioma LC(FDG) exceeds that of contralateral brain, that quantitation of the glioma MRGlc with FDG requires knowing the LC(FDG) specific for the glioma and that the LC(FDG) of normal brain is higher than previously reported estimates of about 0.50. 2-Fluoro-2-deoxy-D-glucose/PET studies in which glioma glucose metabolism is calculated by the autoradiographic approach with normal brain rate constants and LC(FDG) will overestimate glioma MRGlc, to the extent that the glioma LC(FDG) exceeds the normal brain LC(FDG). "Hot spots" visualized in FDG/PET studies of gliomas represent regions where MRGlc, LC(FDG) or their product is higher in glioma than it is in uninvolved brain tissue.

Control of blood glucose in a group of diabetic scuba divers

A preliminary study to examine the hypothesis that the ability of well-controlled (defined as no hypoglycemic episodes within the last 12 mo., HbAlc < 9.0%, and none of the long-term complications of diabetes type I) diabetic scuba divers to control their serum glucose and dive without becoming hypoglycemic during a simulated dive to 27 meters of seawater in controlled environment is impaired. An open, controlled, crossover study compared blood glucose levels, hematocrits, and hematologic cell counts in a group of eight type I diabetic scuba divers to those from eight age- and sex-matched, normoglycemic control scuba divers. Each diver did one simulated dive and one control exercise on the surface on 2 consecutive days. The simulated dive was done to depth of 375 kPa in a hyperbaric chamber, the control exercise was done at ambient pressure. The order of the dive and the control exercise was randomized. No statistically significant differences were observed between serum glucose levels in the diabetic divers measured during the simulated dive to 375 kPa vs. the serum glucose levels in the diabetic divers measured during the control exercise at the same time points. All divers with type I diabetes remained free of symptoms and signs of hypoglycemia throughout the course of the trial, and no diabetic subject had a serum glucose less than 4 mmol/liter before the end of the trial. As the sample size was small, larger studies including subject with type II diabetes will be necessary to extend these results to the diabetic diving population at large. The authors conclude that, contrary to advice issued by most diving agencies to scuba divers, it may be safe to allow well-controlled subjects with type I diabetes with no long-term complications to undertake scuba diving, and that high partial pressures of oxygen do not seem to lower serum glucose levels significantly in the diabetic diver during the dive.

Co-expression of MDR-associated markers, including P-170, MRP and LRP and cytoskeletal proteins, in three resistant variants of the human ovarian carcinoma cell line, OAW42

Variants of the human ovarian carcinoma cell line, OAW42, exhibiting low-level intrinsic resistance (OAW42-SR) and drug-induced higher-level resistance (OAW42-A1 & OAW42-A), were studied along with a sensitive clonal population (OAW42-S) which was isolated from OAW42-SR. Expression of the MDR-associated protein P-170, the more recently discovered LRP (lung resistance-related protein) and MRP (multidrug resistance-associated protein), topoisomerase II alpha and beta, GST pi and the cytoskeletal proteins, cytokeratin 8 and vimentin, were studied (using immunocytochemistry and Western blotting techniques) in conjunction with drug (doxorubicin) accumulation and subcellular distribution. Expression of mRNA for P-170, MRP, topoisomerase 11 alpha and beta and GST pi was studied using RT-PCR (reverse transcriptase polymerase chain reaction). Results indicate differential co-expression of four MDR-associated parameters (P-170, MRP, LRP and reduced topoisomerase II alpha and beta) in the OAW42-SR and OAW42-A1 variants, whereas resistance in the OAW42-A variant appeared to be mainly P-170 mediated. Comparable amounts of MRP and greater amounts of LRP were detected in the OAW42-S cells compared to the OAW42-SR variant (which showed increased resistance compared to the OAW42-S cells), but all cell lines expressed similar low-level amounts of MRP mRNA (by RT-PCR). GST pi levels did not differ markedly between variants. Increased levels of the cytoskeletal proteins were observed with increasing levels of resistance. The relative resistance of the variants, OAW42-SR and OAW42-A1, compared with OAW42-S was seen to change during increased serial passaging of the cells. There was greater drug accumulation by the sensitive OAW42-S cell line compared with that of the resistant variants, particularly the most highly resistant OAW42-A cells. Both verapamil and cyclosporin A effectively restored the accumulation defects seen in the resistant variants, cyclosporin A being the more effective of the two. Sub-cellular location of drug was predominantly in the nucleus with maximum levels seen in the sensitive OAW42-S variant and minimum levels in the most resistant OAW42-A clone.

Feasibility of imaging pentose cycle glucose metabolism in gliomas with PET: studies in rat brain tumor models

The feasibility of imaging pentose cycle (PC) glucose utilization in human gliomas with PET was explored in two rat glioma models by means of glucose radiolabeled in either the carbon-1 (C-1) or carbon-6 (C-6) position.

METHODS

In vitro, monolayers of T-36B-10 glioma, tissue slices of intracerebral glioma grafts or slices of normal brain were fed [1-14C]glucose or [6-14C]glucose, and the generated [14C]CO2 was trapped to quantitate the ratio of [14C]CO2 from 14C-1 versus 14C-6. In vivo, rats bearing grafts of either T-36B-10 or T-C6 rat gliomas at six subcutaneous sites received simultaneous intravenous injections of either [1-11C]glucose and [6-14C]glucose, or [1-14C]glucose and [6-11C]glucose. Tumors were excised between 5 and 55 min postinjection to quantify tracer uptake while arterial plasma was collected to derive time-activity input curves.

RESULTS

In vitro, the C-1/C-6 ratio for CO2 production from T-36B-10 monolayers was 8.8 +/- 0.4 (s.d.), in glioma slices it was 6.1 +/- 2.1 and in normal brain slices it was 1.1 +/- 0.7. PC metabolism in T-36B-10 was 1.8% +/- 0.5 of total glucose utilization. In vivo, tumor radioactivity levels normalized by plasma isotopic glucose levels showed that retained C-1 relative to C-6 radiolabeled glucose was significantly lower in both gliomas, 4.9% lower in T-36B-10 (p < 0.01) and 4.7% lower in T-C6 (p < 0.01). In an additional group of rats bearing T-36B-10 gliomas and exposed to 10 Gy of 137Cs irradiation 4 hr before isotope injection, the C-1 level was 5.6% lower than that for C-6 (p < 0.05). These results were analyzed with a model of glucose metabolism that simultaneously optimized parameters for C-1 and C-6 glucose kinetics by simulating the C-1 and C-6 tumor time-activity curves. The rate constant for loss of radiolabeled carbon from the tumors, k4, was higher for C-1 than for C-6 in all groups of rats (19% higher for T-36B-10 unirradiated, 32% for T-36B-10 irradiated and 32% for T-C6 unirradiated).

CONCLUSION

Mathematical modeling, Monte Carlo simulations and construction of receiver-operator-characteristic curves show that if human gliomas have a similar fractional use of the PC, it should be measurable with PET using sequential studies with [1-11C]glucose and [6-11C]glucose.

Effects of pressure on whole blood glucose measurements using the Bayer Glucometer 4 blood glucose meter

The effect of pressure was investigated on the readings of whole blood glucose obtained from the Bayer Glucometer 4 blood glucose meter which uses the hexokinase enzymatic reaction. Sixteen subjects (eight normal and eight insulin-dependent diabetics) were exercised in a hyperbaric chamber at a depth of 3.7 atm abs. Venous blood samples were monitored at regular intervals for whole blood glucose concentration as measured by a Glucometer 4 inside the chamber. The blood samples were immediately placed in an airlock and taken to 1 atm abs, where whole blood glucose concentrations were measured using an identical instrument. The remaining blood was then analyzed in duplicate for serum glucose concentration using standard laboratory methods. The results show a significant difference between whole blood glucose concentrations measured at pressure and those measured at atmospheric pressure. Significant differences are also observed between whole blood glucose concentrations measured under pressure and serum blood glucose concentrations measured at atmospheric pressure.

Fibrin glue: an alternative method of wound closure in glaucoma surgery

PURPOSE AND METHODS

A commercially available tissue glue has been used to effect conjunctival wound closure after trabeculectomy. In four cases it was used in conjunction with sutures and in two cases alone to achieve watertight closure of the conjunctiva. A small transient leak was noted postoperatively in one case, and no case of flat anterior chamber occurred.

RESULTS AND CONCLUSION

Intraocular pressure was controlled in all cases 3 months postoperatively. Tissue glue can be an effective method of achieving conjunctival wound closure in glaucoma surgery.

Plasma atrial natriuretic peptide is elevated in patients with hypertrophic cardiomyopathy

STUDY OBJECTIVES

To determine if plasma levels of atrial natriuretic peptide are elevated in patients with hypertrophic cardiomyopathy and to determine the relationship of atrial natriuretic peptide to symptoms and echocardiographic indices of left ventricular structure and diastolic function in these patients.

DESIGN

A prospective study in which atrial natriuretic peptide was measured in peripheral venous plasma in 14 patients (age 44 +/- 14 years) with hypertrophic cardiomyopathy and 17 healthy controls. Echocardiography was performed in all cases and 30 controls to examine indices of left heart structure and function. All patients underwent clinical evaluation.

RESULTS

The concentration of atrial natriuretic peptide was significantly higher in patients with hypertrophic cardiomyopathy than controls, (17.86 +/- 8.72 vs. 6.22 +/- 3.26 pmol/l, P = 0.0001). Diastolic dysfunction was observed in 11 of 14 patients with hypertrophic cardiomyopathy. No correlation was demonstrated between atrial natriuretic peptide levels and the degree of diastolic dysfunction, septal or free wall thickness, left atrial size, degree of mitral regurgitation or New York Heart Association functional class.

CONCLUSIONS

Plasma levels of atrial natriuretic peptide are elevated in patients with hypertrophic cardiomyopathy but do not correlate with symptoms or echocardiographically-derived indices of left ventricular structure or diastolic function.

Beta-carotene inhibits atherosclerosis in hypercholesterolemic rabbits

Oxidatively damaged LDL may be of central importance in atherogenesis. Epidemiological evidence suggests that high dietary intakes of beta-carotene and vitamin E decreases the risk for atherosclerotic vascular disease, raising the possibility that lipid-soluble antioxidants slow vascular disease by protecting LDL from oxidation. To test this hypothesis, we fed male New Zealand White rabbits a high-cholesterol diet or the same diet supplemented with either 1% probucol, 0.01% vitamin E, 0.01% all-trans beta-carotene, or 0.01% 9-cis beta-carotene; then we assessed both the susceptibility of LDL to oxidation ex vivo and the extent of aortic atherosclerosis. As in earlier studies, probucol protected LDL from oxidation and inhibited lesion formation. In contrast, vitamin E modestly inhibited LDL oxidation but did not prevent atherosclerosis. While beta-carotene had no effect on LDL oxidation ex vivo, the all-trans isomer inhibited lesion formation to the same degree as probucol. Moreover, all-trans beta-carotene was undetectable in LDL isolated from rabbits fed the compound, although tissue levels of retinyl palmitate were increased. The effect of all-trans beta-carotene on atherogenesis can thus be separated from the resistance of LDL to oxidation, indicating that other mechanisms may account for the ability of this compound to prevent vascular disease. Our results suggest that metabolites derived from all-trans beta-carotene inhibit atherosclerosis in hypercholesterolemic rabbits, possibly via stereospecific interactions with retinoic acid receptors in the artery wall.

Pattern electroretinogram and peripheral colour contrast thresholds in ocular hypertension and glaucoma: comparison and correlation of results

AIMS

Both pattern electroretinogram and peripheral colour contrast thresholds have been shown to be abnormal in glaucoma and ocular hypertension. This study evaluates each of these tests as tools for the early diagnosis of glaucoma, compares and contrasts the results, and examines the relation between the two tests in a large cohort of ocular hypertensive patients.

METHODS

Transient and steady state pattern electroretinograms and peripheral colour contrast thresholds were performed in 45 normal, 37 glaucomatous, and 206 ocular hypertensive eyes. The results were analysed using receiver operating characteristic curves, together with evaluation of sensitivity and specificity of the tests. The relation between the two tests was examined by direct statistical correlation of the results.

RESULTS

All tests showed high sensitivity and specificity for discriminating between normal and glaucomatous eyes. However, there was a significant difference between the two tests for the number of ocular hypertensives considered as abnormal. Forty per cent of ocular hypertensives had abnormal pattern electroretinogram compared with 30% with abnormal peripheral colour vision. Peripheral colour contrast thresholds showed a significant correlation with both transient and steady state pattern electroretinogram.

CONCLUSION

Both of these tests have been shown to be promising new tools for the early detection of glaucoma but the number of ocular hypertensive patients showing abnormal results is rather higher than expected considering the natural history of the condition. Sensitivity in ocular hypertension may be increased by using a combination of both tests. The significant correlation between these psychophysical and electrophysiological tests is discussed.

Metabolic images from dynamic positron emission tomography studies

A dynamic sequence of positron emission tomography (PET) images gives rise to the possibility of creating images of in vivo tissue metabolism. For this reason PET is potentially a valuable instrument in the study of human biology and medicine. The analysis of dynamic PET data to produce metabolic images is a challenging problem from a statistical point of view. For example, a typical data set arising in the study of cerebral glucose utilization has on the order of 30 time-binned images per cross-sectional slice of tissue under examination, each of dimension 128 x 128 pixels. Metabolic imaging requires that the time series at each pixel, known as the time activity curve (TAC), be analysed to produce an estimate of local metabolism. This paper describes a mixture analysis approach to the construction of such metabolic images. In the approach the TAC at a given pixel is expressed as a weighted sum of sub-TACs corresponding to homogeneous tissues represented at the pixel. Estimates of tissue metabolism at the pixel are then constructed as a weighted sum of the metabolism associated with the individual sub-TACs. The procedure is illustrated by application to a [F-18]-labelled deoxyglucose study in a patient with a brain tumour. The ability to map simultaneously a range of parameters related to the transport and biochemical transformation of the radio-tracer, demonstrates the potential power of dynamic PET.

Imaging radiotracer model parameters in PET: a mixture analysis approach

Two methodologies for fitting radiotracer models on a pixel-wise basis to PET data are considered. The first method does parameter optimization for each pixel considered as a separate region of interest. The second method also does pixel-wise analysis but incorporates an additive mixture representation to account for heterogeneity effects induced by instrumental and biological blurring. Several numerical and statistical techniques including cluster analysis, constrained nonlinear optimization, subsampling, and spatial filtering are used to implement the methods. A computer simulation experiment, modeling a standard F-18 deoxyglucose (FDG) imaging protocol using the UW-PET scanner, is conducted to evaluate the statistical performance of the parametric images obtained by the two methods. The results obtained by mixture analysis are found to have substantially improved mean square error performance characteristics. The total computation time for mixture analysis is on the order of 0.7 s/pixel on a 16 MIPS workstation. This results in a total computation time of about 1 h per slice for a typical FDG brain study.

Data-dependent bandwidth selection for emission computed tomography reconstruction

At present, the choice of bandwidth in emission computed tomography (ECT) reconstruction is done by subjective means. The authors develop an automated objective selection technique for linear reconstruction algorithms such as filtered backprojection. The approach is based on the method of unbiased risk estimation. A set of 2-D validation studies using computer simulated and physical phantom data from the Hoffman et al. (1990) brain phantom are carried out. These 2-D studies incorporate measured corrections for object attenuation and lack of uniformity in detector sensitivity. It is found that the unbiased risk approach works very well. Over a range of count rates and brain slice source distributions, the root mean square (RMS) error of the fully automated reconstruction, with the data-dependent choice of bandwidth, is around 5% greater than the RMS error for the reconstruction with an ideal choice of the bandwidth.

Reducing negativity artifacts in emission tomography: post-processing filtered backprojection solutions

The problem of negative artifacts in emission tomography reconstructions computed by filtered backprojection (FBP) is of practical concern particularly in low count studies. Statistical reconstruction methods based on maximum likelihood (ML) are automatically constrained to be non-negative but their excessive computational overhead (orders of magnitude greater than FBP) has limited their use in operational settings. Motivated by the statistical character of the negativity artifact, the authors develop a simple post-processing technique that iteratively adjusts negative values by cancellation with positive values in a surrounding local neighborhood. The compute time of this approach is roughly equivalent to 2 applications of FBP. The approach was evaluated by numerical simulation in 1- and 2-dimensional (2D) settings. In 2D, the source distributions included the Hoffman, the Shepp and Vardi, and a digitized version of the Jaszczak cold spheres phantoms. The authors' studies compared smoothed versions of FBP, the post-processed FBP, and ML implemented by the expectation-maximization algorithm. The root mean square (RMS) error between the true and estimated source distribution was used to evaluate performance; in 2D, additional region-of-interest-based measures of reconstruction accuracy were also employed. In making comparisons between the different methods, the amount of smoothing applied to each reconstruction method was adapted to minimize the RMS error-this was found to be critical.

Amaurosis fugax in young people

Nine young adults (median age 19.5 years) who suffered from amaurosis fugax (AF) are described. The attacks of AF were short in duration and preceded by premonitory symptoms in five cases and by a migrainous headache in two. In five patients the visual loss progressed in a lacunar pattern unlike the 'curtain' pattern characteristic of AF in older patients. Investigation revealed no evidence of an embolic or atheromatous aetiology. In two cases a minor abnormality was found on echocardiography. We conclude that AF in young adults has a different clinical pattern and may have a different aetiology, possibly migrainous, compared with that seen in older patients. The pattern of visual loss in some of the cases suggests that the choroidal circulation rather than the retinal circulation is primarily affected.

The pattern electroretinogram in glaucoma and ocular hypertension

Thirty one eyes with established glaucoma, 61 high risk ocular hypertensive (OHTs) eyes, 66 medium risk OHT eyes, 58 low risk OHT eyes, and 47 control eyes have been followed for up to 2 years by clinical examination and pattern electroretinography (PERG). The study was 'masked' so electrophysiological and clinical data were kept separate. Criteria have been devised which enable PERG measurements to distinguish all established glaucomatous eyes from all normal controls; these criteria demonstrate abnormalities in some OHT eyes, particularly those at high risk. The PERG abnormality is greatest in eyes with established glaucoma in which the intraocular pressure has been lowered by treatment. The PERG becomes smaller as the degree of clinical abnormality increases. Test-retest variability of the PERG is sufficiently low to ensure that most of those first described as abnormal continue to be so.

Macular colour contrast sensitivity in ocular hypertension and glaucoma: evidence for two types of defect

Colour contrast sensitivity (CCS) of a large cohort of glaucomatous patients, ocular hypertensive patients (OH), and normal persons was measured at six-month intervals during a two-year period. The OHs were graded into high, medium, and low risk groups. 69% of glaucomatous patients and 32% of all OHs had CCS thresholds greater than the mean plus 2 SDs of the controls. Satisfactory specificity and sensitivity could not be obtained by adjusting the criterion of threshold. In abnormal eyes, progressive small increases of threshold occurred during the study, but glaucomatous eyes with normal thresholds on the first visit retained normal thresholds in the subsequent visits. Although our system is very sensitive and precise, the proportion of abnormalities detected is no greater than with other techniques. In some glaucomatous patients there is a true preservation of colour vision which does not merely reflect the limitations of the test employed.

CEA and CA 15-3 in primary and recurrent breast cancer

To evaluate CA 15-3, a new breast cancer associated antigen, and to compare it with carcinoembryonic antigen (CEA), all patients presenting with breast cancer had preoperative and serial (3-monthly) postoperative levels measured. Of 124 patients with primary breast cancer, 23% had an elevated CA 15-3 (greater than 25 units/ml) while 11% had an elevated CEA (greater than 5 ng/ml) (p = not significant). Neither marker was an indicator of spread to regional lymph nodes in primary breast cancer. In 45 recurrences of breast cancer, CA 15-3 was elevated at the time of first recurrence in 58% while CEA was elevated in 47% (p = not significant). Of 17 patients with locoregional recurrence alone, none had a CA 15-3 above 40 units/ml while 11 of 12 with synchronous locoregional and distant recurrence had a CA 15-3 level greater than 40 units/ml (chi 2: 21.36, p less than 0.0001). This study shows that CA 15-3, like CEA, is of little clinical value in primary breast cancer. CA 15-3, however, is an accurate indicator (overall accuracy, 97%) of synchronous distant metastases in patients with locoregional recurrence from breast cancer. This information has important implications for further investigation and management of such patients.

Deconvolution of episodic hormone data: an analysis of the role of season on the onset of puberty in cows

A new approach to the analysis of episodic hormone data is described. The method involves a stochastic model in which measured blood hormone concentration is represented as a convolution of individual pulses, each of which is thought of as the response to a burst of neural activity. Individual pulses are not constrained to occur in a fixed regular pattern in time. The methodology takes a series of blood hormone measurements and produces a spike train of pulse peak times together with a set of pulse shape parameters. This decomposition motivates some fresh approaches to the analysis of hormone data. For a given number of pulses the model is fit by minimizing a residual sum of squares criterion. This is a difficult combinatorial optimization problem. A randomized local adjustment algorithm is developed. Generalized cross-validation is used to select the number of pulses. The technique seems to produce reliable results on simulated data sets. The methodology is used to study some data concerned with the role of season of birth on the onset of puberty in bovine females. The analysis raises some interesting questions related to the maturation of the pituitary and hypothalamus.