NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Bast RC Jr, Kufe DW, Pollock RE, et al., editors. Holland-Frei Cancer Medicine. 5th edition. Hamilton (ON): BC Decker; 2000.

Cover of Holland-Frei Cancer Medicine

Holland-Frei Cancer Medicine. 5th edition.

Show details

Chapter 30Principles of Imaging


The technologic revolution that is being fueled by the development of increasingly powerful computers and rapid telecommunications is currently affecting all branches of medicine, but none more than diagnostic imaging. Malignant tumors often alter the normal spatial relationships in tissues, and radiologic imaging is critical not only for diagnosing cancer but also for staging tumors and monitoring patients after they have received treatment. In the future, it is expected that technologic advances related to imaging may assist clinicians in evaluating functional parameters in tumors and in assessing the effects of treatment.1–4 Among the imaging studies undergoing rapid development are magnetic resonance (MR) spectroscopy and brain activation analyses, rapid computed tomograph (CT) volume acquisitions, positron emission tomography (PET) with new metabolic agents, and MR blood flow and diffusion studies.

It may be appropriate to review briefly some of the technical similarities, as well as salient differences, between three cross-sectional imaging techniques that are used frequently to study cancer patients: CT, MRI, and PET. These techniques offer computerized image reconstructions of two-dimensional sections of the body in different viewing plans. Three-dimensional volume or surface images are also possible, as well as “see-through” or projectional images. To perform CT, an external radiation source and a detector are required on opposite sides of the body, as in other radiographic imaging techniques. On the other hand, MRI uses powerful magnetic fields and radiofrequency waves to create images of the head or body without the need for ionizing radiation (see below). PET entails the injection of trace amounts of short-lived radionuclides that have been produced in a cyclotron and that concentrate in tumors and various organs. The small amounts of radiation emitted by these radionuclides can be picked up by external radiation detectors, positioned on opposite sides of the body, and then used to create planar or volume images. The radionuclides first emit subatomic particles called positrons. Each positron quickly combines with an electron inside the body to produce two x-ray photons, which leave the body in opposite directions and can be detected. The radiation doeses delivered to tissues in PET are less than those in ordinary radiographic studies.

With CT, the tissue densities depicted on cross-sectional images have the same significance as the densities produced on ordinary x-ray radiographs. These densities correspond to the relative amounts of diagnostic radiation absorbed by each tissue that is depected in the image, but CT is much more sensitive to minor differences in tissue absorption (ie., exhibits greater tissue contrast) than ordinary radiography. In addition, CT can “peel away” superimposed layers of tissue that may obscure detail on ordinary radiographs of the head, chest, abdomen, and extremities, revealing the anatomy of a single layer or section of tissue. An external ionizing radiation source is required, and the x-ray does to tissues are similar to those delivered during ordinary chest, abdomen, bone and skull radiography.

With MRI, a powerful unidirectional magnetic field is used to orient or polarize hydrogen atoms within tissue in the direction of the field. Short pulses of radiowaves are then sent into the body at a frequency that is resonant with the polarized hydrogen atoms. The polarized hydrogen atoms are deflected momentarily from their axes by these radiowave pulses, and subsequently, they emit radiowaves at their resonant frequency. These emissions can be picked up by an external radiowave detector. With a powerful computer, the radiowave emission patterns from the resonant hydrogen atoms can then be used to synthesize a three-dimensional volume image (or multiple adjacent planar images ) of the specific region of the head or body that is under study. MR images, therefore, represent a computer-generated map of the hydrogen atom radiowave emitters in a single body region. The MR emissions from these hydrogen atoms are referred to as echoes. The TE or “echo delay time” for a given MR images is the split-second delay that occurs between the excitation of hydrogen atoms in the tissues by the external pulsed radiowaves and the detection of radiowave “echoes” from the same atoms by an external detector. The interval between each successive radiowave pulse emitted by the MRI machine (which, like the echo delay time, can be selected by the MRI machine operator) is called the TR, or “repetition time.” Depending on the TE and TR settings (measured in milliseconds [ms]) used for each clinical MRI examination, the operator can produce an image that is characterized as a T1–weighted image or a T2-weighted image. With the most commonly used MRI technique, known as spin echo imaging, T1-weighted images can be produced by using relatively short TE and TR settings (e.g., a TE of 40 ms and a TR of 200 ms), whereas T2-weighted images require longer settings (e.g., a TE of 120 ms and a TR of 2,000 ms). Tumors may appear relatively dark on T1-weighted MR images, but they often appear bright on T2-weighted images. Anatomic detail is shown more clearly on T1-weighted images, but tumors (and the edema and reactive tissue that may surround them) often stand out in better contrast to adjacent normal tissues on T2-weighted studies.

It is also worthy of note that (1) MR images can be generated easily in any plane (not just the axial plane, as with ordinary CT images); (2) the contrast between soft tissues of different types (e.g., tumor and adjacent muscle) is better with MRI than with CT; and (3) one can discern flowing blood in vessels on MRI without necessarily injecting contrast material into the bloodstream (parenthetically, Doppler ultrasound [US] methods can also demonstrate flow in tumor vessels noninvasively and may have future applications in clinical cancer diagnosis). Among the limitations of MRI, small amounts of calcium cannot be detected easily with MRI, whereas CT is quite sensitive to calcium deposits in tissues. Most MR scans also take longer to obtain than CT scans; they may require several minutes rather than a few seconds to perform. Body motion, therefore, remains a potential problem with some MRI studies. The physical confines of MRI machines are usually quire restrictive for patients, and up to 10 to 15% of patients may experience claustrophobic reactions in the “tunnel” of the machine that preclude their undergoing the studies. Because of limited physical access to the patient in position in the scanner, very ill patients who are on life support systems are also difficult to study with most MRI machines. These limitations in MRI equipment may be less of a problem in the future, as faster MRI studies become feasible and specialized open field units with easy access to patients become available for certain clinical applications.

It is also possible to do semiquantitative spectroscopic analyses of metabolites within living tissues using MRI, and MR images can be used to pinpoint the regions of interest for spectroscopic analyses. MR spectroscopy might be used in the future to improve diagnostic specificity or to assess early responses of tumors to therapy. These ideas are still undergoing investigation.

PET, like MR spectroscopy, may also be capable of providing unique information on the metabolism of human tumors, including early changes that may result from treatment. The short-lived positron-emitting radionuclides that are used in PET can serve as “tags” on certain metabolites (e.g., radioactive fluorine-labeled glucose analogues). The metabolites are injected intravenously prior to scanning, and the rate of accumulation of the metabolites is then determined to assess tumor metabolism and the alterations caused by treatment. Unsuspected deposits of metastatic tumor can also be detected, in some instances, with PET because of their metabolic activity. With its ability to sample the metabolism of targeted volumes of tissue in situ, PET, therefore, may offer a powerful new tool for studying tumors in the laboratory as well as in selected clinical situations.

The technologic advances that have led to transmission (CT) and positron emission (PET) computed tomography, as well as to MR imaging or spectroscopy and modern US methods, have also led to fundamental innovations in our ability to visualize tumors and to assess their metabolism. Although the traditional radiologic imaging techniques that are still used throughout the world (e.g., radiography of the chest, skeleton, and abdomen; gastrointestinal (GI) barium studies; radionuclide scintigraphy of bone) may demonstrate tumors directly (e.g., show tumor nodules in the lung) or indirectly (e.g., demonstrate widening of the duodenal loop on a GI series by a pancreatic mass), cross-sectional imaging techniques, often with contrast materials, are capable of producing direct images of tumors anywhere in the head or body. In addition to showing the internal structure or “texture” of an individual tumor, these techniques can accurately delineate the tumor’s margins and demonstrate its effects on adjacent structures. Some tumors may become more visible (“enhance”) after vascular contrast materials are infused during CT or MRI, or when certain MRI sequences are used (e.g., T2-weighted images). The ability to image deep tumors directly as well as to show their effects on surrounding organs in vivo and the corresponding ability of cross-sectional imaging techniques to strip away overlying structures that may obscure tumor masses constitute fundamental advances for managing cancer patients that have come about in the last two decades with these imaging techniques.5–7

Even with the availability of these new diagnostic imaging techniques, it is important to emphasize that the radiologist cannot make tissue diagnoses; relevant clinical information is always needed to make the best use of our diagnostic images. Furthermore, US, CT, and MRI have not replaced standard radiologic techniques, which have been at the core of the diagnostic armamentarium of clinicians for almost a century. The newer imaging methods are powerful complements to these techniques. Cross-sectional imaging methods have made cancer diagnosis, tumor staging, and patient follow-up more accurate (particularly with regard to diagnostic sensitivity), more rapid, and less invasive than ever before. Despite their relatively high cost, particularly for CT and MRI, the impact of these techniques on the net costs of cancer care may not be self-evident. A good argument can be made that cross-sectional techniques decrease the overall costs of patient care in several important ways.8 First, planar imaging methods have lessened the need for invasive diagnostic techniques, such as angiography and standard myelography, and they have eliminated painful and hazardous studies like pneumoencephalography. They have also expedited the diagnosis of cancer (as well as our efforts to rule it out), and they may obviate unnecessary surgery in patients who are shown to have metastatic or locally extensive disease. The newer imaging methods are also invaluable for directing percutaneous aspirations or needle biopsies and for planning open biopsies and surgical resections, when necessary, and they have lessened the requirement for hospitalization by facilitating diagnosis and tumor staging on an outpatient basis. For these reasons and many others, modern diagnostic imaging techniques have not only improved care and decreased patient suffering, they may also have reduced overall medical expenditure and loss of income to patients by simplifying diagnostic procedures, shortening the time to diagnosis, decreasing the need for hospitalization, and helping to tailor therapeutic approaches to individual patient needs.

In contrast to modern radiology’s impact on diagnosis, staging, and patient follow-up, only one imaging technique has had a significant impact on screening asymptomatic individuals for cancer: low-dose mammography. Breast cancer mortality in a defined population of women can be reduced by up to one-third through regular screening mammography in accordance with nationally published guidelines.9,10

Physicians are also cautioned that, although cross-sectional imaging techniques may have increased sensitivity for detecting tumor masses and delineating their extent, the specificity of these techniques in diagnosing cancer may not have improved to the same degree.11 Both the clinician and the diagnostic radiologist should exercise restraint in interpreting positive radiologic findings in patients with established or suspected cancer diagnoses, particularly findings that suggest metastases or extensive local disease for the first time. Many conditions besides metastases can present as lung nodules on CT, or as “hot spots” on skeletal radionuclide scintigraphy. As always, careful correlation of the abnormalities seen on an imaging study with a patient’s clinical, laboratory, and other imaging findings is essential. Comparing serial images that have been obtained over days, weeks, or months will improve diagnostic specificity when certain kinds of abnormalities have been noted: tumor masses in the lungs cannot be expected to double in size over a few days; conversely, primary tumor masses or metastatic lesions ordinarily do not remain unchanged for many months or years. Comparing a new study with a baseline imaging study obtained 6 months a year previously can sometimes be the critical step in reaching a correct diagnosis. Correlating findings on different types of imaging studies from the same patient (e.g., comparing focal radionuclide bone scan abnormalities with radiographs of precisely the same areas in patients who have suspected breast cancer metastases) may also be extremely helpful.

Finally, much research still needs to be done on the most appropriate applications of diagnostic imaging techniques for cancer management. It is entirely possible that the divergent results reported in different cancer therapy trials might be caused, in part, by inaccurate stratification of patients.7 A heterogeneous group of cancer patients who are improperly stratified in a trial of therapy may have stages of disease different from those reported in another trial. The applications of imaging techniques to oncology practice and, in particular, to clinical research studies are still less than optimal, and some reported patients trials may, in fact be “mixing apples and organges.”12 The more judicious use of radiologic techniques in clinical trial protocols may have much to offer here.

An area for continuing research is the appropriate type of radiologic studies to use and the appropriate intervals between radiologic examinations for monitoring cancer patients following completion of treatment. The clinical questions inherent in this area of concern must be subjected to carefully controlled trials. Whether or not to use imaging examinations at all for the post-treatment surveillance of cancer patients should depend on whether effective palliative or salvage methods (second- or third-line treatments) are currently available. Equally germane to following cancer patients who have received definitive treatment is whether or not a second- or third-line therapy is more effective when a recurrence has been detected early as opposed to late: some treatments for recurrent disease may be just as effective (or just as ineffective) when a recurrence has become manifest through new symptoms or physical findings, rather than through regular surveillance with laboratory studies and/or imaging examinations. It is likely that some diagnostic imaging techniques are currently being used in inappropriate ways to monitor treated patients at treatment centers as well as in community practice, in the absence of reliable data from controlled trials to determine what the actual effects of periodic imaging studies are on patient outcomes.

The basic principles of diagnostic imaging and their current applications to cancer management, as described in this section, are the informed recommendations of several contributing experts. While their recommendations must suffice for now, additional clinical studies and more corroborative data will be needed if recent advances in diagnostic imaging are to be applied optimally to the care of cancer patients.


Carrasquillo J A, Bunn P A, Keenan A M. et al. Radioimmunodetection of cutaneous T-cell lymphoma with111In-labelled T101 monoclonal antibody. N Engl J Med. 1986;315:673. [PubMed: 3489185]
Dodd G D. Advances in cancer diagnosis. Cancer. 1990;65(3 Suppl):595. [PubMed: 2405989]
Larson SM. Positron emission tomography in oncology and allied disease. In: DeVite V, Hillman S, Rosenberg S, editors. Principles and practice of oncology updates. Philadelphia, PA: JB Lippincott; 1989. p. 1.
Schlom J. Innovations in monoclonal antibody tumors targeting: diagnostic and therapeutic implications. JAMA. 1989;261:744. [PubMed: 2911171]
Platt J F, Glazer G M, Gross B H. et al. CT evaluation of mediastinal lymph nodes in lung cancer: influence of lobar site of the primary neoplasm. AJR Am J Roentgenol. 1987;149:683. [PubMed: 3498314]
Siegelman S S, Khouri N F, Leo F P. et al. Solitary pulmonary nodules: CT assessment. Radiology. 1986;160:307. [PubMed: 3726105]
Zerhouni G A, Stitik F P, Siegelman S S. et al. CT of the pulmonary nodule: a cooperative study. Radiology. 1986;160:319. [PubMed: 3726107]
Kuhns LR, Thornbury JR, Tryback D. Decision making in imaging. Chicago, IL: Year Book; 1989.
Cady B. New diagnostic, staging, and therapeutic aspects of early breast cancer. Cancer. 1990;65(3 Suppl):634. [PubMed: 2405990]
Henson D E, Ries L A. Progress in early breast cancer detection. Cancer. 1990;65 (Suppl 9):2155. [PubMed: 2328479]
Steckel R J, Kagan A R. Pitfalls in the diagnosis of metastatic disease or local tumor extension with modern imaging techniques. Invest Radiol. 1990;25:818. [PubMed: 2202693]
Simon R. The importance of prognostic factors in clinical trials. Cancer Treat Rev. 1984;68:185. [PubMed: 6362863]
© 2000, BC Decker Inc.
Bookshelf ID: NBK20847


  • PubReader
  • Print View
  • Cite this Page

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...