The importance of portal circulation in the delivery of drugs and nutrients to colorectal hepatic metastases is controversial. Using 13N (nitrogen 13) amino acids and ammonia with dynamic gamma camera imaging, we demonstrate, for the first time in human beings, a quantitative advantage of hepatic artery compared with portal vein infusion. Eleven patients were studied by hepatic artery injection, five patients were studied by portal vein injection, and two patients had injections through both routes. Data collected from the liver for 10 minutes after rapid bolus injection of 13N L-glutamate, L-glutamine, or ammonia were compared with 99mTc (technetium) macroaggregated albumin (MAA) images produced after injection through the hepatic artery or portal vein at the same session. Tumor regions defined from 99mTc sulfur colloid scans were compared with nearby liver areas of similar thickness. For the 13N compounds, the area-normalized count rate at first pass maximum (Qmax) and the tissue extraction efficiency were computed. The tumor/liver Qmax ratios for MAA and 13N compounds were highly correlated. Both tumor and liver extracted more than 70% of the nitrogenous compounds. The tumor/liver Qmax ratios reflect the relative delivery of injected tracer per unit volume of tissue. After hepatic artery injection the Qmax ratio was 1.03 +/- 0.33 (mean +/- SD), significantly exceeding the Qmax ratio of 0.50 +/- 0.34 after portal vein injection (P less than 0.003). Therefore, more than twice as much of a nutrient substrate is delivered per volume of tumor relative to liver by the hepatic artery as by the portal vein; the high extraction efficiency demonstrates that the hepatic artery flow is nutritive; and the delivery of substance in solution (such as nutrients or drugs) to tumor and liver tissue correlates with the distribution of colloids such as macroaggregated albumin after hepatic arterial and portal venous injection.

By contract with the National Cancer Institute, the accuracy of diagnostic techniques was assessed in 184 patients suspected of having pancreas cancer. Of 138 patients who were operated upon, 89 were found to have pancreas duct cancer, 30 had cancer of a different site of origin in the head of the pancreas region and in 19 there was no evidence of cancer at operation. All of the 46 patients who were not operated upon, 13 proven to have cancer and 33 patients discharged as free of cancer, were followed in our clinic. The majority of our patients presented with signs and symptoms of biliary obstruction. Computerized transaxial tomography (CTT) gave a "correct" diagnosis in 31 of 33 patients (94%) with proven cancer, there were 2 patients with a false negative report and a false positive diagnosis occurred in 8 of 20 patients (40%) without cancer. Celiac angiography (CA) gave a correct diagnosis in 78 of 94 patients (83%) with cancer, a false negative in 17%, and a false positive in 32%. 76Selenomethionine pancreas scan correctly diagnosed 27 of 36 patients (75%) with cancer, gave a false negative in 25% and a false positive in 31%. Ultrasonography gave a correct diagnosis in 18 of 27 patients with cancer (67%), a false negative in 33% and a false positive in 28%. Endoscopic retrograde cholangiopancreatography diagnosed correctly 8 of 11 cases (73%) of cancer, there were false negative diagnoses in 3 cases (27%) and false positives in 3 of 14 patients (21%). Duodenal aspiration techniques gave a very low percentage of correct diagnoses. Chronic pancreatitis most commonly gave rise to a false positive diagnosis. Serum alkaline phosphatase was elevated in 82% of patients, gave 18% false negatives and 33% false positives. Carcinoembryonic antigen (CEA) was elevated (greater than 2.5 ng/ml) in most of the pancreas cancer patients but also in patients with other cancers and with non-cancerous diseases. In our hands, CTT, CA, alkaline phosphatase, 75Se-methionine and ultrasonography, in descending order, have given the highest percentage of correct diagnoses but false positive and false negative diagnoses prevented any single test from being conclusive.

A GREAT many studies have been carried out on the kinetics of iodine in man and a number of models have been proposed (1-9). The most common model has been a three-compartmental one in which iodine in the plasma is taken up by the thyroid and released back into the circulation as thyroid hormone. This in turn gets degraded and the released iodide recycles again. Excretion pathways from plasma iodide into urine and from plasma hormone into feces are also included.During the past ten years we have carried out over 100 kinetic studies on about 50 individuals in various states of thyroid function and found that the three-compartment model is inadequate to explain the data. In a previous publication (7), some of the computational problems encountered in connection with the data analysis were discussed and preliminary attempts to evolve a consistent model were presented. Further modifications of the model were introduced more recently. This paper is intended to summarize the present state of the model, the arguments for it and its shortcomings.The individuals studied included normals, hyperthyroids, euthyroids, hypothyroids and patients with thyroid cancers. Most of the individuals were studied at least twice, usually before and after various treatments. Treatments involved radioiodine therapy, hypophysectomy, and the administration of acetylated TSH.Several studies are chosen to show the spectrum of kinetic responses encountered (Fig. 1). These show both qualitative and quantitative differences in urine, thyroid and plasma responses for various perturbed states. Although it was possible to propose simple models to explain most of the studies individually, the models so derived were quite different from each other qualitatively. Since a "universal" model was sought to fit all the studies, the various features required in fitting the studies individually were integrated into a single model. The complexity of the proposed model evolved after efforts to derive a compatible, less complex model failed.Kinetic studies. The kinetic studies ranged in duration from about five to 30 days. Data were collected roughly in accordance with the schedule outlined below, although not all measurements were made in every study. All studies started with an iv injection of labeled iodide.The schedule of measurements was as follows: