Advances in medical imaging for cancer diagnosis and treatment

Advances in Medical Imaging for Cancer Diagnosis and Treatment Henry N. Wagner Jr, MD, and Peter S. Conti, MD, PhD

Over the last several decades, significant “new eyes” have been developed that improve the diagnosis, treatment, planning, and monitoring of human cancer: computer tomography (CT), magnetic resonance imaging (MRI) and spectroscopy (MRS), single photon emission computed tomography (SPECT), and positron emission tomography (PET). Innovative advances in both morphologic and functional imaging have led to a dramatic improvement in our ability to diagnose and monitor human cancer. Frequently, anatomic detail can be demonstrated in ways that exceed views at surgery, and functional biochemical imaging is being used to show the metabolic activity and receptor status of normal and pathologic states. In vivo functional and biochemical studies differentiate normal from neoplastic or nonviable tissue, and make it possible to measure progression or regression of the disease. Because physiologic changes often precede morphologic findings in many disease processes, the use of in vivo biochemical probes can demonstrate disease before anatomic abnormalities become evident. Gross changes in anatomy are no longer adequate endpoints for therapy protocols. Today, using physiologic imaging, we can evaluate the response to treatment within hours of administration of therapy. Adjuvant metabolic tumor imaging studies provide complimentary information to morphologic evaluation of human cancers that will ultimately lead to better patient care. Cancer 67:1121-1128,1991.

HE HISTORY OF IMAGING of cancer can be summa- T rized as the elaboration of ever more perfect eyes to view a disease in which there is always something more to be seen. Today, images of cancer go far beyond simple detection of disease. The new views of cancer make it possible to characterize the disease biochemically, phys- iologically, and structurally. The images are the basis of diagnosis, prognosis, and the planning and monitoring of treatment.

What is the greatest impact of the new images? Cancer is no longer viewed as a structure, but a process. Structure and function characterize cancer and all living matter. What we call structures are slow processes of long dura-

Presented at the American Cancer Society National Conference on Advances in Cancer Imaging, New York, New York, January 24-26, 1990.

From the Divisions of Nuclear Medicine and Radiation Health Sci- ences, Johns Hopkins Medical Institutions, Baltimore, Maryland.

Supported in part by NIH grants NS-15080 and CA 32845. Address for reprints: Henry N. Wagner, Jr., MD, Division of Nuclear

Accepted for publication June 5 , 1990. Medicine, Johns Hopkins Medical Institute, Baltimore, MD.

tion. What we call functions are fast processes of short duration. Structure and function are two views of a unitary process: the phenomenon of cancer.

The search for the causes of cancer continues and is being expanded to the study of the normal biologic mech- anisms that prevent the development of cancer. While we still do not know the cause of cancer, we are now learning when the normal mechanisms that keep us all from getting cancer have failed. For more than a century, cancer has been characterized by the staining properties of cancer cells removed from the body. Today, cancer can be char- acterized at the molecular and cellular levels in living hu- man beings.

The new advances must not lead to neglect of the old. The greatest success story of imaging in modern oncology is a relatively old technique. New and greatly improved mammography has saved countless lives by making di- agnosis possible at a stage when cure can be accomplished. Mammography is widely accepted, and the only question today is at what age routine mammography should be performed. The scientific evidence of screening mam- mography in women older than 50 years is clear-cut, and a screening procedure (two views) should be carried out

1121

1122 CANCER February 15 Supplement 199 1 Vol. 67

every year. Its value in younger women is highly likely, and it should be recommended every 1 or 2 years for asymptomatic women 40 to 49 years of age. Improve- ments in equipment design have reduced the radiation dose to the breast to between 0.1 and 0.4 cGy per view.

At times we hear that new diagnostic techniques are continually being added in the practice of medicine, but established techniques are never discarded. The fallacy of this belief is indicated by the fact that pneumoencepha- lography and ventriculography are today of historic in- terest only. Even the relatively new technique of x-ray computed tomography (CT) is gradually being replaced by magnetic resonance imaging (MRI) for many purposes. Twenty years from now, MRI and nuclear medicine may be the most widely used imaging techniques in the care of patients with cancer: MRI to identify structure and nuclear medicine or magnetic resonance spectroscopy (MRS) to identify biochemical and cellular abnormalities. MRI can delineate the precise locations of chemical and cellular abnormalities, whereas nuclear medicine tech- niques, such as positron emission tomography (PET) and single photon emission computed tomography (SPECT), together with nuclear MRS, provide new views of cancer as a biochemical and cellular process, rather than simply as a structural aberration. MRI opens up a whole new approach to characterization of the physical properties of cancer, including bioenergetics. MRS, PET, and SPECT extend biochemistry from the test tube to the living human body.

Tomography dominates all imaging techniques. We have moved beyond projection of three-dimensional in- formation onto a single plane. We now view the body from all angles, and compute transaxial, sagittal, and co- ronal images of structure, function, and biochemistry. Tomographic images provide a three-dimensional per- spective of cancer and its relationship to normal organs and functions. The new views of anatomy and gross pa- thology provided by CT and MRI have revolutionized surgery and interventional radiology. The new views of in vivo chemistry provided by PET, SPECT, and MRS will revolutionize pharmacology, improve the selection of the type of therapy, and provide an effective way to plan and monitor treatment. Characterization of cancer on the basis of regional biochemistry and histopathology may provide a new basis of classification of the disease and a new approach to the care of patients, because cancer now can be viewed as a disease of molecules and cells. What we need are new clinical investigations to determine which imaging techniques should be used in the context of a given patient's problems. We need to determine when we need to measure the bioenergetics of cancer, membrane and intracellular receptor concentrations, or enzyme concentrations, and relate these measurements to histo-

pathologic examination and the planning and monitoring of treatment.

In the first half of the last century, a major breakthrough occurred in cancer research when it became possible to characterize cancer on the basis of histopathology and relate the findings to clinical symptoms, signs, and prog- nosis. As students we were taught to relate case histories and physical findings to postmortem findings, initially to the gross pathology at the autopsy table and subsequently at the laboratory bench looking through a microscope at stained tissue. Microscopic examination of biopsy tissue remains the best way to characterize cancer and the best guide to prognosis and treatment, but it is limited by the need to remove the tissue from the body and the fact that the tissue is no longer living. Modern medical imaging makes it possible to apply the principles of histopathology to living patients.

Enormous advances have been made over the past cen- tury by the application of radiographs to the care of pa- tients with cancer. Radiographic examination made it possible to visualize cancer or the alterations of normal structures produced by cancer. The new imaging tech- niques, MRI, MRS, PET, and its close relative, SPECT, have been major advances in the care of patients with cancer.

Magnetic Resonance Imaging and Spectroscopy

The first step in the care of patients with cancer is de- tecting a lesion. Then we need to determine whether it is benign or malignant, the degree of malignancy, and the type of cancer. We need to predict what is going to happen to the patient and try to do something about it. The new imaging techniques, including MRI, do not replace sur- gical histopathology of human biopsy specimens but ex- tend greatly our understanding of the disease in a given patient.

MRI is of greatest value because it produces high res- olution, high quality images of practically every organ of the body, based chiefly on differences in the water content of tissues. Lesions are visualized by displacement of nor- mal structures or in outline because of differences in pro- ton relaxation times. Up until recently cancer has had no specific NMR signals not shared by other diseased or nor- mal tissues.

The brain is ideal for study by MRI because it remains motionless during the time required to obtain the im- age~ . ' -~ Most brain lesions have altered T 1 and T2 signal intensities compared with adjacent normal tissue, which are used to produce high contrast images of benign and malignant tumors, infarction, hemorrhage, and demye- hat ing diseases. Although brain tumors can be readily detected by MRI, determination of the size or extent of

No. 4 CANCER DIAGNOSIS AND TREATMENT * Wagner and Conti 1123

the lesions is difficult because of the infiltrating nature of many of the lesions and the presence of surrounding edema. Enhancement of the lesions with gadolinium has helped to solve some of these

In lung cancer MRI has not been as productive and helpful as in brain tumors because of the motion of tho- racic structures and the weak proton density and relaxa- tion time signals from the lungs.’ In the diagnosis of breast cancer progress with MRI has been slow and is not likely to replace mammography. Recent evidence raises the possibility that short T 1 inversion recovery sequences may permit differentiation of benign from malignant lesions.” In the abdomen lesions of the liver have longer relaxation times and can be visualized readily.” Imaging of pan- creatic tumors by MRI is complicated by the nearby pres- ence of bowel. The digestive tract in general has been difficult to image by MRI, partly because of bowel motion. In the genitourinary tract, MRI imaging is effective in delineation of normal organs because they are usually surrounded by fat.’2313 Cysts and tumors can readily be seen. Pelvic tumors, both benign and malignant, are char- acterized by long relaxation times. In the skeletal system lesions can be easily seen,I4-I6 and especially important advances have been made in the study of the spinal cord. 17,18

In summary, MRI does not reveal cancer directly but indirectly because of its ability to portray normal struc- tures that are often distorted by cancer. MRI is an exquisite form of anatomic imaging that has the particular advan- tage of not involving ionizing radiation. We still need to rely on histopathology because malignant and benign tis- sues cannot be differentiated by measurements of NMR proton relaxation times. Although higher tissue water content results in altered T1 and T2 relaxation times in some malignant tumors, similar findings may OCCUJ in other conditions, such as edema, and other nonmalignant states. There is evidence that the degree of malignancy may be reflected in biochemical changes, such as glucose, amino acid, or nucleoside metabolism, all measurable by PET. 19-21

Nuclear magnetic resonance (NMR) began as a tech- nique in chemistry. Nuclear magnetic resonance spec- troscopy (NMRS) with ‘H and 31P and PET make it pos- sible to examine the chemistry of the living body in situ. Spectroscopic images of phosphorus-containing com- pounds in the brain have been obtained, but the technique continues to suffer from poor sensitivity and spatial res- olution.22 NMR spectra of molecules containing naturally occuring phosphorus-3 1 or administered fluorine- I9 or carbon-13 make it possible to measure molecular con- centrations of important compounds, such as ATP, in- organic phosphorus or phosphocreatine, and chemical re- action rates in different organs and lesions. Thus, the bio-

chemistry and bioenergetics of cancer can now be examined with MRS but only in relatively large anatomic regions and requiring long observation times.

I n Vivo Biochemistry of Cancer

After World War 11, biomedical science was revolu- tionized by the introduction of carbon-14 and tritium as radioactive tracers. Even today, most biologic research is based on their continued use. These two radionuclides could not be used for in sifu studies of living human beings because their beta emissions could not penetrate the body and be detected by imaging devices. PET is based onthe use of carbon- 1 I or fluorine- 18. SPECT is based on iodine-123, technetium 99m, and indium-1 1 1. These nuclides emit photons that can be detected and imaged and extend the study of chemistry to the intact human body. By means of mathematical models and appropriate measurements, the results can be expressed in absolute units of meters, kilograms, and seconds (the MKS system), or more simply in relative terms, such as the percentage of the administered dose of a radiotracer accumulated by a tumor. These nuclear medicine techniques define nor- mal regional chemistry and provide a means of detecting deviations from the normal by statistical analysis. Thus, the technology extends beyond portrayal of images of structure by providing functional and biochemical images. Disease can be defined in terms of variation from normal regional chemistry according to established statistical cri- teria. Interpretation can be objective and subjective. At least in theory, when we can measure the rate of a chemical process in a region of the body, there arises the possibility of the existence of at least two diseases, one in which the rate of the chemical process is abnormally slow and an- other in which the process is abnormally fast.

In general, nuclear medicine studies are of three types: regional blood flow, substrate metabolism, and infor- mation transfer via chemical “recognition sites,” including receptors and enzymes. Nuclear medicine provides in- formation about in vivo regional chemistry with a sensi- tivity and specificity comparable to that obtained by ra- dioimmunoassay in studies of body fluids and can often detect abnormalities before structural changes have oc- curred.

In cancer PET makes it possible to measure the rate of utilization of substrates, such as sugars and amino acids, that supply energy to tumors, or nucleotides that reflect DNA metabolism, or provide pharmacokinetic and phar- macodynamic data concerning radiolabeled cancer che- motherapeutic agents. 19-2’923-32 A n example of the detec- tion of a metastatic melandma by its metabolism of glu- cose is shown in Figure 1, which illustrates a gadolinium- enhanced intracranial mass lesion on MRI with a corre-

1 I24 CANCER February 15 Supplement I99 1 Vol. 67

sponding fluorine- 18 fluorodeoxyglucose PET image in the same transaxial slice.

In clinical practice nuclear medicine techniques can be used to assess the effectiveness of surgery, radiation ther- apy, and chemotherapy, and can document the extent of tumors and progression or regression in response to dif- ferent forms of treatment. Such data permit modifications of the treatment plan sooner than can be determined by the clinical response of the patients or changes in the size of the lesions. Thus, treatment need no longer be based solely on clinicaI response, gross morphology of the le- sions, and histopathologic examination of biopsies. Bio- chemical characterization of tumors with labeled tracers is becoming a new method for classifying tumors and for planning and monitoring their treatment.*'

To develop suitable tracers for these purposes, meta- bolic processes must be identified that occur in a greater or lesser extent in neoplasms relative to the surrounding tissues or organs. 19320 An important characteristic of neo- plastic tissue is its increased rate of cell division. In general, accumulation of thymidine into neoplasms is increased in the presence of increased DNA synthesis. Amino acid transport across tumor cell membranes has also been found to differentiate many malignant from nonmalignant tumors. Not only membrane transport but also protein synthesis can be examined if suitable mathematical mod- eling is used in data analysis. The accumulation of fluo- rodeoxyglucose (FDG) can be used to measure regional glucose utilization. Many malignant tumors have accel- erated glycolysis compared with surrounding tissues.

FIGS. 1 A AND I B. (A) Transaxial T I-weighted image with gadolinium enhancement demonstrating large mass lesion right frontal temporal re- gion. Pathologic diagnosis was metastatic malignant melanoma. (B) Transaxial PET image with F- I8 FDG at same level with marked elevated activity in tumor region.

In addition to measuring blood flow to tumors, blood volume, amino acid and glucose utilization, and DNA synthesis, PET and SPECT can be used to measure the number and affinity of hormone receptors that charac- terize certain tumors. Estrogen receptors are increased in many breast tumors in both the primary and metastatic

Dopamine receptors are often increased in pi- tuitary a d e n o m a ~ . ~ ~ Finally, metabolic substrates, such as putrescine, can also be used to reflect increased spermine and so permidine synthesis rates may be useful for the characterization of some tumor^.^'

Let us consider, for example, the case of a brain tumor. Measurement of the rate of fluorine-18 FDG uptake is helpful in determining the degree of malignancy of the tumor. DiChiro et al. at the National Institutes of Health first reported in 1982 the rough correlation of the rate of glucose utilization with the histologic grading of brain tumor malignancy and a good predictor of the patient's life expe~tancy.*~,~' If the tumor has a low rate of metab- olism and is in a particularly dangerous location from the standpoint of surgical therapy, the lesion may be followed with periodic reassessment for signs of increasing metab- olism of the lesion. Assessment of tumor metabolism can be based on the absolute glucose metabolism in the tumor or the relative metabolism of the tumor compared with a corresponding region on the opposite side of the brain or compared with global glucose metabolism. Both visual and objective criteria have been used to make these as- sessments. A major problem in absolute quantification is the variability that occurs not only between persons but

No. 4 CANCER DIAGNOSIS AND TREATMENT * Wagner and Conti 1125

also at different times in the same person. The range may be as great as 20% or 30% of the mean and is probably due to systemic biochemical processes rather than simply technical factors. After treatment, measurement of the metabolic activity of the tumor relative to normal brain helps to discriminate between persistence or recurrence of the tumor and damage to normal brain tissue, such as that resulting from radiation necrosis.39 Outside the cen- tral nervous system studies of tumor recurrence are being carried out with fluorine-18 FDG in colorectal tumors.24

Another substrate, (1 1 -C)-methionine, is useful for de- lineating the boundaries of brain tumors, providing in- formation of value in the planning and performance of brain surgery, by permitting differentiation of the metab- olizing brain tumor from simple disruption of the blood- brain b a r ~ i e r . ~ ~ . ~ ~ Carbon- 1 1 and nitrogen- 13 labeled amino acids have great potential for imaging both pe- ripheral and central nervous system neoplasms. As early as 1979, nitrogen-1 3-labeled amino acids were used to assess the treatment response of patients with osteogenic ~arcoma.~’

In the United States, there were 70 PET scanners and 43 cyclotrons which had been installed or ordered by the end of 1989. The clinical application of PET in the care of patients with cancer is therefore certain to increase in the 1990s.

Molecular Recognition Sites

Since the classical work of Claude Bernard, the body has been viewed as made up of cells bathed in a sea of extracellular fluid. Communication among cells is based to a large degree on the existence of “recognition sites,” which identify the sources of energy to be incorporated into the cells, or “messages” from other cells, which de- termine the subsequent activity of the receptor cells.

How the human body recognizes foreign substances remains one of the most intriguing and important ques- tions in biomedical science. The process seems to be a general phenomenon in which billions of specific mole- cules, some free in the blood and others attached to mobile cells, circulate until they encounter a stereospecific fit with another free or cell-bound molecule. After the encounter of the “key” with the “lock,” the molecule binds to the stereospecific recognition site, whether the site is an en- zyme, receptor, or antigen. A basic concern is how we are able to tolerate our own molecules. Since 1950 it has been known that each of the trillion lymphocytes in the human body has specific receptors that identify specific antigens and then respond by proliferation, a process called the “clonal selection theory.” These new cells then attack and eliminate the antigen from the body. The lymphocyte re- ceptors include B-cells, which recognize unaltered anti-

gens, and T-cells, which bind via connecting proteins (MHC proteins) to smaller polypeptide chunks of the broken down original antigen. The T-cells include cyto- toxic cells and helper cells. The cytotoxic lymphocytes release lymphokines, including the interleukins, and in- terferons. Not only tumor antigens but also damaged nor- mal cells are recognized and dealt with in this way. Tol- erance of self is believed to occur as a result of early ex- posure of immature lymphocytes that die after they encounter the antigens of normal cells. In essence, the lymphocytes that recognize “self” are filtered out early in life. It is now possible to label these B- and T-lymphocytes with antibodies or antibody fragments, labeled with tech- netium-99m. To date, technetium-99m-labeled anti- granulocyte antibodies have been used to visualize the normal distribution of granulocytes and their increased concentrations in infections.

An example of the use of PET to assess the presence of receptors in tumors is breast cancer. Fluorine-1 8 estra- diol accumulation as determined by PET makes it possible to tailor the treatment of a specific patient on the basis of the number of estrogen receptor^.^^-^^ A tumor con- taining estrogen receptors is more likely to be treated suc- cessfully with estrogen receptor blocking drugs, such as tamoxifen, than cancers that do not contain estrogen re- ceptors. Radioactive tracers that bind to estrogen receptors make it possible to assess the status of the primary breast cancer and regional metastatic deposits. This is an excel- lent example of the new biochemical approach to the characterization of disease, an approach directly related to prognosis and therapy. Histopathology alone is no longer the only required criterion for diagnosis, prognosis, and therapy.

Another example is the study of receptors on pituitary tumors.36 Using the dopamine receptor binding agent (1 1- C)-N-methylspiperone, it has been possible to classify pi- tuitary adenomas according to whether they possess do- pamine receptors. If the tumors contain such receptors, they can be treated chemically rather than surgically, i.e., by administering the dopamine receptor agonist brorno- cryptine.

Immunoscintigraph y

Hellstrom and Hellstrom4’ found melanoma antibodies in plasma and proposed that certain antigens might be specific for tumors. Mice could be immunized with hu- man melanomas by hybridization techniques to yield large quantities of antibodies. The goal was to use these mono- clonal antibodies as diagnostic markers or carriers of ther- apeutic doses of radionuclides or tumoricidal chemo- therapeutic agents.

They stated: “We are enthusiastic about the future of

1126 CANCER February 15 Supplement 199 1 Vol. 61

nuclear imaging with monoclonal antibodies (fragments), since it can not only demonstrate a ‘lump’-which may be detected even better by other methods-but can show whether this lump expresses a known tumor marker, and therefore represents a neoplasm . . . Several technical problems have to be worked out, however, including the need to develop labeling techniques for isotopes with bet- ter imaging characteristics than I- 13 1 and ways to decrease the degree of antibody uptake by normal tissues, including the liver.”41

The basic premise of immunoscintigraphy is that while tumor-associated antigens are present on normal tissues, the concentration on the tumor cells is much greater, which makes possible their detection by SPECT. In some cases antigens may be present in malignant cells at several hundred times the concentration in normal cells.4’

The major application of immunoscintigraphy to date has been making the initial surgical treatment of colon cancer, ovarian cancer, and melanoma more effective. These diseases are successfully treated by surgery only in the early stages, before the disease has spread to other organs. In addition, the sites of recurrence can be identified and the decision made with respect to whether repeat sur- gery and/or radiotherapy is indicated. Perhaps the single most important problem is the immune response to the diagnostic agents, which may limit the number of ex- aminations in a given patient. Monoclonal antibodies with human components are now being developed to reduce antigenicity.

There have been great improvements in the surgical treatment of colon cancer, in which sphincters can be spared and the postoperative quality of life greatly im- proved. Persistence and recurrence of tumor remain a problem. Preoperative staging is of great importance, and modern imaging makes it possible to delineate lymph nodes by means of technetium-99m sulfur colloid injec- tions or lymph node metasases by means of monoclonal antibodies (SPECT) or tumor substrate metabolism (PET). Careful follow-up can then reveal recurrences at an early stage in which it is still worthwhile to remove the recur- rences by a second operation. Imaging is used in associ- ation with measurement of serum carcinoembryonic an- tigen (CEA) levels. The resectability rate for recurrent dis- ease is about 12% to 15%, and the number of resections of liver metastases is i nc rea~ ing .~~

In malignant melanoma, monoclonal antibodies play a role in clinical staging, therapeutic decisions, and prog- n ~ s i s . ~ ~ Particularly in the detection of axillary nodes, im- munoscintigraphy has demonstrated a sensitivity of greater than 80% and a specificity of The finding of axillary metastases clarifies prognosis and can be an indication for therapeutic lymphadenectomy. The exis- tence of this possibility decreases the need for prophylactic

lymphadenectomy because a randomized study has dem- onstrated no differences in the survival rate of patients with prophylactic versus therapeutic lymph node resec- t i ~ n . ~ ~ - ~ ’ Metastatic nodes in remote regions can also be d e t e ~ t e d . ~ ~ - ~ ’ The particular value of immunoscintigraphy relative to clinical and other imaging examinations is the improved specificity. The 5-year survival of patients who have had lymph node resection is about 33% to 35%.51

In ovarian cancer, metastases are often difficult to detect by CT or other anatomic imaging methods. Thus, second or third operations are often carried out. Furthermore, pelvic masses detected by these imaging techniques may or may not represent metastases. For these reasons and because the metabolic response to treatment can be as- sessed directly, biochemical imaging methods have been employed. Another approach is to use monoclonal anti- bodies to increase the specificity of diagnosis by identifying high concentrations of tumor-associated antigens. Fibrotic deposits are seen after surgery and must be differentiated from viable tumor.52 The method has a role in staging of disease and in detecting local tumor in patients with el- evated serum markers of ovarian cancer. The primary role is in assessing the effects of chemotherapy, demon- strating residual or recurrent disease, and distinguishing postoperative fibrosis from viable tumor. In such cases the need for “second-look” operations is reduced.

The Future

Max Planck once said: “A new scientific truth is not usually presented in a way that convinces its opponents . . . rather the opponents gradually die off, and a rising generation becomes familiar with the truth from the start.” Ready acceptance is the exception rather than the rule. Histopathology remains the cornerstone of cancer diag- nosis and treatment, but in vivo chemistry is an idea whose time has come.

The two new eyes for imaging cancer are MRI and PET. The former is widely accepted today: the latter is likely to achieve wide recognition in the future. The value of PET is to make it possible to characterize cancer itself rather than by its effect on normal structure and function.

The availability of carbon-14 and tritium after World War I1 revolutionized biochemistry. The development of PET and the use of carbon- 1 1 and fluorine- 1 8, along with MRS with fluorine- 19, carbon- 13, and phosphorus-3 1, have extended biochemistry from the study of body fluids to the study of the regional biochemistry of cancer and the body’s defenses against cancer.

Physicist Karl K. Darrow has said: “One of the things which distinguishes ours from all earlier generations is that we have seen our atoms.” The human body is made up of molecules. A fundamental unit in biology today is

No. 4 CANCER DIAGNOSIS AND TREATMENT - Wagner and Conti 1127

not only the cell but also the molecule. Most physicians, basic scientists, and the public view cancer as something foreign that has invaded the body. A virus, chemical or photons of ionizing radiation, is believed to enter the body and cause cancer. If cancer is viewed as a foreign invasion, the treatment is to cut it out, destroy it by radiation, or poison it with chemicals. The postmodern view is that all of us are constantly in danger of developing cancer, but normal control mechanisms keep it from happening. Thus, cancer is not a capricious invasion of the body, but a failure of normal control processes, processes that can now be examined and quantified in normal persons and identified as deficient in patients with cancer. If the cause is a failure of normal control processes, the treatment is to restore these deficient control mechanisms. For ex- ample, growth factors bind to cellular membrane receptors and activate the growth stimulatory process via small reg- ulatory molecules, called second messengers. Eventually, the message to proliferate is signaled to the nucleus, which then divides. Then growth inhibitory signals prevent cel- lular proliferation from proceeding indefinitely. Eluci- dation of the biochemistry of these signal-transducing functions can be extended to living human beings via PET, SPECT, and MRS. Such studies of the biochemistry of cancer can be used not only to detect cancer but also to plan its treatment and monitor its effectiveness. The molecular approach changes the rules of the established game. As Whitehead stated: “A science that hesitates to forget its founders is lost.” It is unwise to reach a turning point and not turn.

REFERENCES

1 . Holland BA, Haas DK, Norman D et a/. MRI of normal brain maturation. AJNR 1986; 7:201-208.

2. Brant-Zawadzki M, Kelly W. Brain tumors. In: Brant-Zawadzki M, Norman D, eds. Magnetic Resonance Imaging of the Central Nervous System. New York: Raven Press, 1987; 151-186.

3. Atlas SW, Grossman RI, Gomori JM, et a/. Hemorrhagic intra- cranial malignant neoplasms: Spin-echo MR imaging. Radiology 1987; 164:7 1-77.

4. Sze G , Krol G, Olsen WL et a/. Hemorrhagic neoplasms: MR mimics of occult vascular malformations. AJNR 1987; 8:795-802.

5. Brasch RC. Methods of contrast enhancement for NMR imaging and potential applications. Radiology 1983; 147:78 1-788.

6. Stark DD, Weissleder R, Elizondo G eta/. Superparamagnetic iron oxide: Clinical application as a contrast agent for magnetic resonance imaging. Rudidogy 1988; 168:297-30 1.

7. Felix R, Schorner W, Lamado M ef a/. Brain tumors: MR imaging with gadolinium-DTPA. Radiology 1985; 156:681-688.

8. Graif M, Bydder GM, Steiner RE, Niendorf HP, Thomas DGT, Young IR. Contrast enhanced MRI of malignant brain tumors. A.JNR

9. Von Schulthess GK, McMardo K, Tscholakoff D, deGeer G, Gamsu G , Higgins CB. Mediastinal masses: MR imaging. Radiology 1986; 158: 289-296.

10. Zobel BB, Tella S, Patrizio G et a/. MR STIR imaging versus spin-echo imaging of the breast: lesion characterization. Radiology 1989; I73:354.

1986; 61855-862.

I I . Freeny PC, Marks WM. Patterns of contrast enhancement of be- nign and malignant hepatic neoplasms during bolus dynamic and delayed CT. Radiology 1986; 160:613-618.

12. Heiken JP, Lee JKT. MR imaging ofthe pelvis. Radiology 1988;

13. Hricak H, Dooms GC, Jeffrey RB et a/. Prostatic carcinoma: Staging by clinical assessment, CT and MR imaging. Radiology 1987;

14. Totty WG, Murphy WA, Lee JKT. Soft-tissue tumors: MR im- aging. Radiology 1986; 160:135-141.

15. Pettersson H, Gillespy T 111, Hamplin DJ et al. Primary mus- culoskeletal tumors: Examination with MR imaging compared with conventional modalities. Radiology 1987; 164237-241.

16. Vanel D, Lacombe MJ, Couanet D, Kalita C, Spielmann M, Genin J. Musculoskeletal tumors: Follow-up with the MR imaging after treat- ment with surgery and radiation therapy. Radiokogy 1987; 164:243-245.

17. DiChiro G, Doppman JL, Dwyer AJ et a/. Tumors and arterio- venous malformations of the spinal cord: Assessment using MR. Ra- diology 1985; 156:689-697.

18. Williams AL, Haughton UM, Pojunas KW, Daniels DL, Kilgoe DP. Differentiation of intramedullary neoplasms and cysts by MR. AJNR 1987; 8:527-532.

19. Conti PS, Bading JR, Wong DF et a/. Positron emitting radiotra- cers: Potential aids in the diagnosis and management of neoplasms. Tzi- mor Diagnostik Therapie 1988; 9: 175.

20. Schmall B, Conti PS, Kelinert EL, Whitmore WF Jr. Metabolic characterization of rat prostatic adenocarcinoma in vivo with radiolabeled alpha-aminoisobutyric acid, amonicyclopentanecarboxylic acid, 2-deoxy- D-glucose and thymidine (Abstr). Biochemistry 1983; 22:36.

21. Brownell GL, Kariento AL, Swartz M et a/. PET in oncology: The Massachusetts General Hospital experience. Semin NuclMed 1985; 15:201-209.

22. Weiner MW. MR spectroscopic imaging approaches clinical realm. Diagn Imag Clin Med 1989; I1:48-52.

23. DiChiro G, Dela Paz RL, Brooks RA et a/. Glucose utilization of cerebral gliomas measurement by ( 18F) fluorodeoxyglucose and pos- itron emission tomography. Neurology 1982; 32: 1323- 1329.

24. Strauss LG, Clorius JH, Schlag P et al. Recurrence of colorectal tumors: PET evaluation. Radiology 1989; 170:329.

25. Gelbard AS, Benua RS, Laughlin JS et a/. Quantitative scanning of osteogenic sarcoma with nitrogen-I 3 labeled L-glutamate. J NuclMcJd

26. Conti PS, Sordillo PP, Schmall B et a/. Tumor imaging with car- bon-I 1 labeled alpha-aminoisobutyric acid (AIB) in a patient with ad- vanced malignant melanoma. Eur J Nucl Med 1986; 12:353-356.

27. Schmall B, Conti PS, Bigler RE et a/. Imaging studies of patients with malignant fibrous histiocytoma using C- 1 1 alpha-aminoisobutyric acid (AIB). Clin NuclMed 1987; 1:22-26.

28. Bergstrom M, Ericson K, Hagenfeldt L et a/. PET study of me- thionine accumulation of glioma and normal brain: Competition with branchidl chain amino acids. J Comput Assist Tomogr 1987; 11:208.

29. Martiat P, Ferrant A, Labar D et a/. In vivo measurement of carbon-I 1 thymidine uptake in non-Hodglun’s lymphoma using positron emission tomography. J Nucl Med 1988; 29: 1633.

30. Conti PS, Grossman SA, Wilson AA eta/. Brain tumor imaging with C-l I labeled thymidine and PET. Radiology 1990; 177:234.

31. Ginos JZ, Cooper AJL, Dhawan V et a/. [”N] cisplatin PET to assess pharmacokinetics of intra-arterial versus intravenous chemotherapy for malignant brain tumors. J Nucl Med 1987; 28: 1844.

32. Strauss LG, Dimitrakopoulu A, Clorius JH, Schlag P, Helus F, Maier-Borst W. PET measurements of F-18 fluorouracil uptake in liver metastases and response to chemotherapy. Radiobgy 1989; 173: 152.

33. Mintun MA, Welch MJ, Mathias CJ et a/. Application of 16- alpha-( 18F) fluoro- 17-beta estradiol for the assessment of estrogen re- ceptors in human breast carcinoma. J Nucl Med 1987; 28:300.

34. Zielinski JE, Larner JM, Hoffer PB et a/. Synthesis of 1 1-beta- ethoxy-[ 16 alpha-’*’I] iodoestradiol and its interaction with the estrogen receptor in vivo and in vitro. J Nucl Med 1989; 30:209.

35. Mintun MA, Welch MJ, Siege1 BA et al. Breast cancer: PET im- aging of estrogen receptors. Radiology 1988; 169:45.

16611 1-16.

162~331-336.

1979; 20:782-784.

1128 CANCER Ft.hmary 15 SuppIement I991 Vol. 67

36. Muhr C, Bergstrom M, Mundberg PO 01 a/. Dopamine receptors in pituitary adenomas: PET visualization. J Compuf Assist 7bmogr 1986;

37. Wolkow N, Goldman SS, Flam ES, Cravioto H, Wolf AP, Brodie JD. Labeled putrescine as a probe in brain tumors. Scicwce 1983; 221:

38. Patronas NJ, DiChiro G, Kufta C el a/. Prediction of survival in glioma patients by means of positron emission tomography. JNeirrosurg

39. Patronas NJ, DiChiro G , Brooks RA r( al. ( 1 8F) Fluorodeoxy- glucose and positron emission tomography in evaluation of radiation necrosis of the brain. Radiokogy 1982; 144385-889.

40. O T u a m a LA, Guilarte TR, Douglass KH et ul. Assessment of ["C]-~-methionine transport into the human brain. J Ccwb B/ood F h Metab 1988; 8:341-345.

41. Hellstrom I , Hellstrom KE. Anti-tumor antibodies for therapy. Nucl Med Biol 1989; 16:6 13-6 16.

42. Bucci L, Meucci E. Clinical diagnosis and therapy of colo-rectal cancer: State of the art. Nucl Med Biol 1989: 16:6 11-6 19.

43. Belli F, Santinami M, Baldin1 MT, Testori A, Maio A, Cascinelli N. Clinical status of diagnosis and therapy of malignant melanoma. Niicl Mrd Biol 1989; 16:62 1-624.

44. Buraggi GL, Gasparini M . Immunoscintigraphy of melanoma us- ing radiolaheled antibodies: Pilot, prospective studies and Italian mul- ticenter clinical trials. Nucl Med Biol 1989; 16:629-632.

10: 175- 180.

673-675.

1985; 62x816-822.

45. Veronesi U, Adamus J, Bandiera DC et a/. Inefficiency of im- mediate node dissection in Stage I melanoma of the limbs. N Engl .I Med 1977; 2971627.

46. Veronesi U, Adamus J. Bandiera DC el a/. Stage J melanoma of the limbs: Immediate versu.y delayed node dissection. Tumori 1980; 66: 373.

47. Veronesi U, Adamus J , Bandiera DC el a/. Delayed regional lymph node dissection in Stage 1 melanoma ofthe skin ofthe lower extremities. Cancer 1982; 49:2420.

48. Larson SM. Radiolabeled monoclonal anti-tumor antibodies in diagnosis and therapy. J Nucl Med 1985; 26:538-545.

49. Baraggi GL, Callegaro L, Mariani G eta/. Imaging with 13'1-lahelled monoclonal antibodies to a high weight-melanoma associated antigen: Efficacy of whole Ig and its Fab2 fragments. Cancer Res 1985; 45:3378- 3387.

50. Cascinelli N, Attili A, Belli F el d. Anti-melanoma monoclonal antibody 225-28s: Evaluation of toxicity in man. Timori 1988; 74:35- 40.

5 1. Roses DF, Hams MN, Rigel D el a/. Local and in-transit metas- tases following definitive excision for primary cutaneous malignant mel- anoma. Ann SurK 1983; 1985-69.

52. Granowski M, Britton KE, Mather S et a/. Diagnosis of ovarian cancer with radiolabeled antibodies. Nucl Med Biol 1989; 16:653-657.