258

are confusing because of a wide variety of image patterns. Radionuclide scans can be evaluated with and without dynamic flow studies, and CT studies with and without contrast. Both examinations can be performed at any time from the onset of symptoms through the recovery period. Thus, few comparative data are available. Moreover, some of the CT data on accuracy are biased because a large proportion of the examinations (42 per cent in one series") were eliminated as "technically suboptimal."

When results from two large studies on ischemic disease were combined, the sensitivity of CT was 72 per cent (42 of 58 patients) for vascular occlusive disease, and that for radionuclides, 83 per cent (48 of 58 patients). 20,21 Most differences in detectability for the two modalities occurred in the period of one to four weeks after the onset of symptoms," but data on the time course for appearance of positive studies are conflicting. Some data suggest that CT is positive 54 per cent of the time by two days and 90 per cent by four days," and other data indicate that up until one week, only 50 per cent of patients with vascular accidents have positive CT scans."

44

Infection. Detection of cerebral abscesses is usually accomplished by radionuclide or CT studies or by angiography. Old data suggest that radionuclide studies are considerably more sensitive than angiography in detection averages of 100 per cent and less than 80 per cent, respectively, from three series. Among the few cases reported thus far, CT and radionuclide studies have shown about the same sensitivity: 56 of 57 patients with CT and 24 of 27 patients for radionuclides.4,49-51 CT is more specific than radionuclide studies and angiography, particularly in differentiating cerebritis (a nonsurgical lesion) from abscess. It is a precise method of localizing cerebral abscesses. 46,50,52 Both CT and radionuclides have been used to follow the course of therapy. CT has been labeled "invaluable" in assessment of response to therapy." Although some data relating early diagnosis with CT to mortality and morbidity of brain abscess indicate little change, other studies suggest a lower mortality and more satisfactory therapeutic outcomes. 50,52

CT may provide a unique contribution relative to other diagnostic modalities in the identification and localization of epidural abscesses" and in the differentiation of many small abscesses from one large one. Hydrocephalus. CT has been successfully applied to evaluation of obstructive internal" and communicating hydrocephalus."-" It has demonstrated the large size and symmetry of the ventricles in normal-pressure hydrocephalus, "," the site and character of any obstruction that may be present, the response to shunting and the persistence of enlargement after shunting in some patients irrespective of the clinical response." The method is safe and reliable for children and is particularly useful in assessment of complications of shunt therapy and in serial examinations." In seven of 20 patients with hydrocephalus, CT showed cortical atrophy not evident on pneumog

Feb. 2, 1978

raphy." Otherwise, CT furnishes few factual data beyond those available from pneumographic examination, but the information is provided more rapidly and with appreciably less risk.

Demyelinating diseases. In leukoencephalopathy, areas of diminished density resembling mild cerebral infarction have been identified. 58-41 Autopsy in one patient demonstrated that the involved area on CT corresponded to the demyelinated tissue in the brain. In multiple sclerosis, CT revealed areas of low attenuation (representing periventricular plaques) and brain atrophy in 36 per cent, brain atrophy alone in 45 per cent and normal scans in 18 per cent of patients." Optic-nerve plaques have also been observed." Thus, changes in demyelinating disease are evident on the CT scan, whereas cerebral angiography, pneumoencephalography, and isotope scanning are rarely help

ful.

Spinal cord. Although the spinal cord has been visible on some CT scans, Sheedy and his co-workers" had problems in defining the cord in 17 patients. Intrathecal introduction of metrizamide improved the information yield and gave better resolution of the thoracic and cervical-cord outline in 10 patients." This technic may be especially valuable in evaluation of disease of the central canal. The single-motion, short-time scanners apparently improve resolution of the spinal cord."

In 18 patients with known or suspected syringomyelia, CT reportedly identified cord cavitation in 11," but others have been unable to duplicate these results." In a small number of patients with spinal arteriovenous malformations, CT correlated well with angiography and improved the appraisal of the extraspinal components."

Eye, ear, nose and throat disease. The CT scan has been extremely useful in depicting the anatomy of the orbit, particularly in exophthalmos," mass lesions in the retrobulbar area and lesions involving the optic nerve and extraocular muscles. The bony structures of the petrous bone, the paranasal sinuses and the nasopharynx can be delineated with a high degree of precision, thereby defining the volume of masses, the effect on adjacent structures and the degree of bone destruction.

Certainly regarding diagnosis and prognosis

For the patient who has a neurologic symptom (headache, for example) and whose major concern is brain tumor, the physician serves his most important function either by excluding neoplasm or by defining an organic cause of the symptom and introducing ap

Vol. 298 No. 5

propriate management. In the former situation small gains in sensitivity and specificity may be considerably more important to the patient and his physician than simple numerical calculations would suggest.

For example, assuming that CT is much more sensitive (0.95 instead of 0.80) and more specific (0.98 instead of 0.75)* than radionuclide studies, use of these tests on a patient population in which 20 per cent are diseased increases the certainty that a patient with a normal result will be actually disease free from only 0.94 to 0.99." Whether this change results in either greater peace of mind in the patient or physician or a reduction in the number of unnecessary additional tests is not known. In practice, the quantitation of such an effect is difficult. It cannot be plotted against days, weeks or months; the patient and physician alike convey its importance only in the most qualitative and subjective terms. Thus, one of the important immediate applications of new diagnostic technologies is not susceptible to persuasive analytic approaches.

Mortality and morbidity of the diagnostic evaluation

Most hospitals have observed a marked drop in the number of pneumoencephalograms (22 to 80 per cent) as well as a smaller drop in the number of cerebral angiograms (20 to 65 per cent). 84-87 On average, the mortality of pneumoencephalography is 0.25 per cent, and that of carotid angiography less than per cent." Both have a variety of associated nonfatal morbid complications (including meningeal reactions, hemorrhages and arterial emboli). A recent co-operative study indicated that over 10 per cent of patients having cerebral angiography for transient ischemic attacks have one or more complications." Thus, if the number of pneumoencephalograms and angiograms actually decreases as a result of CT, and if the procedures performed are less elaborate, the morbidity and mortality of the diagnostic evaluation of neurologic patients should also decrease. No large-scale study has as yet provided adequate data to support these assumptions, however.

Changes in therapy

According to two recent studies, CT has resulted in changes in therapy. In one study of patients with a variety of neurologic complaints, CT led to change in therapy 20 per cent of the time: about 15 per cent had a complete change in their treatment strategy, and 5 per cent had a modification of an existing plan of treatment. In another study focusing on a particular class of patient rather than on a wide variety of diseases and symptoms, Ambrose and his co-workers' found a 94 per cent reduction in craniotomies in patients having CT after acute head trauma. Theoretically, CT should also influence the choice of therapy in some patients with stroke and massive intracranial

*For illustrative purposes these numbers have been deliberately chosen to represent extreme positions that overestimate the sensitivity and specificity of CT and underestimate the sensitivity and specificity of radionuclide studies. They thus show dramatically that when most of a population is free of disease, changes in sensitivity and specificity of the order indicated here have little effect on these certainty calculations.

259

bleeding, but such an effect has not yet been documented.

Long-term survival

Changes in long-term survival are appropriate end points for evaluating cerebral CT in only a few situations because of the marked dependence of survival on treatment rather than on diagnosis. For example, the five-year survival of patients with brain tumors has changed only slightly over the past 15 years (from 24 to 29 per cent) despite earlier detection of disease with radionuclide scanning, more accurate surgical technic due to arteriographic guidance, continuous advances in surgical technic and postsurgical care and introduction of supervoltage radiation therapy." At Johns Hopkins Hospital, the introduction of radionuclide brain scans and a rapid rise in the number of cerebral angiograms in the early 1960's did not improve survival in patients with tumors." It is equally unlikely that CT of the head will statistically improve survival of patients with brain tumors in the next decade without marked changes in therapeutic efficacy. Financial considerations

Cost calculations related to CT in the head have been considered from two reference points. The first one evaluates changes in hospital admissions and length of stay and the concomitant drop in use of invasive procedures. The second one compares the average and marginal costs of diagnostic evaluation with and without CT for patients with similar symptoms.

Actual cost savings have not yet been clearly documented although theoretical calculations suggest that outpatient evaluation by CT may obviate hospitalization for many patients, that hospital stays may be reduced as a result of averted invasive procedures, and that specialized intensive care may be decreased as a result of more correct diagnosis (e.g., untreatable intracranial bleeding). The Toronto General Hospital has estimated that averted hospital admissions as a result of outpatient neurologic evaluations have saved over 10 bed days per patient and inpatient stays have been shortened by 1.6 days per patient." These theoretical estimates have not been confirmed in that institution, however. At the Frenchay Hospital in England the average stay for neurologic patients decreased from 17 days in 1974 to 13 days in 1976. Although this decrease started before the introduction of CT, it appears to have been accelerated by it." About 400 patients who would ordinarily have been admitted were worked up as outpatients in 1975-1976. Although a saving in hospital beds has not been realized because all beds are still full, the previously lengthy waiting list for new surgical patients has disappeared.

In all cases, these reductions are most meaningful when they ultimately reduce the total number of beds required to handle the patient-care load generated within a particular community. Otherwise, the fixed costs of maintaining beds will critically erode any potential savings. Moreover, these cost calculations cannot ignore any additional costs resulting from the performance of more CT examinations than the studies

1

260

that CT replaces. An analysis of special neurodiagnostic studies at the Peter Bent Brigham and Beth Israel hospitals demonstrated a small decrease in angiograms and a substantial decrease in pneumograms, but a marked increase in the total number of neurodiagnostic procedures since the introduction of CT (Table 2). During the same period there was no parallel increase in the number of neurology and neurosurgery inpatients nor of outpatient-clinic neurologic evaluations. These data suggest that because of its noninvasive character, CT is being applied with more liberal indications than invasive neurodiagnostic procedures. After the introduction of CT at George Washington University, a similar jump in neurodiagnostic examinations from 1834 (1973) to 5538 (1975) occurred," but with a larger patient population served in 1975 than in 1973. This increase was accompanied by a rise in charges from $267,650 (1973) to $1,132,475 (1975).“

Theoretically, the 94 per cent reduction in craniotomies for head trauma (which Ambrose et al." found when CT was used) should produce formidable savings by reducing the number of costly surgical procedures. Such savings would only be realized fully, however, if the number of operating-room personnel, supplies and overhead were reduced, a major difficulty in fixed-cost installations. Otherwise, the general effect of fewer operations is an increase in unit cost.

Data on average and marginal costs of diagnostic evaluation with and without CT are not available Their estimation requires a group of patients who present in the same manner, are studied by both CT and conventional neuroradiologic studies, and are then followed to obtain proof of the presence or absence of disease as well as unbiased analysis of their image data.

REFERENCES

1. Fineberg HV, Parker GS, Pearlman LA: CT scanners: distribution and planning status in the United States. N Engl J Med 297:216-218, 1977

2. Banta HD, Sanes JR, Gomer PA, et al: The Computed Tomography (CT or CAT) Scanner and Its Implications for Health Policy. Washington, DC, Office of Technology Assessment, 1976

Feb. 2, 1978

3. Bogue T, Wolfe SM: CAT Scanners: Is fancier technology worth a bil lion dollars of health consumers' money? Washington, DC, Health Research Group, 1976

4. Creditor MC, Garrett JB: The information base for diffusion of technology: computed tomography scanning. N Engl J Med 297 49-52, 1977 5. Brooks RA, Di Chiro G: Theory of image reconstruction in computed tomography. Radiology 117:561-572, 1975

6. Idem. Principles of computer assisted tomography (CAT) in radiographic and radioisotopic imaging. Phys Med Biol 21:689-732, 1976 7. Johns HE: New methods of imaging in diagnostic radiology. Br J Ra diol 49:745-764, 1976

8. Ambrose J, Gooding MR, Griver J, et al: A quantitative study of the EMI values obtained for normal brain cerebral infarction and certain tumours. Br J Radiol 49:827-830, 1976

9. Adams DF. Hessel SJ, Judy PF, et al: Differing attenuation coeffi cients of normal and infarcted myocardium. Science 192.467-469, 1976 10. Idem: Computed tomography of the normal and infarcted myocardium. Am J Roentgenol 126.786-791, 1976

11. Powell WJ Jr, Wittenberg J, Maturi RA, et al: Detection of edema associated with myocardial ischemia by computerized tomography in 150lated, arrested canine hearts. Circulation 55.99-108, 1977

12. Evens RG, Alfidi RJ, Haaga JR, et al: Body computed tomography: a clinically important and efficacious radiologic procedure. Radiology 123:239-240, 1977

13. McNeil BJ, Adelstein SJ: Determining the value of diagnostic and screening tests. J Nucl Med 17:439-448, 1976

14. Wildavsky A: Doing better and feeling worse: the political pathology of health policy. Daedalus 106(1):105-124, 1977

15. Pendergrass HP, McKusick KA, New PFJ, et al: Relative efficacy of radionuclide imaging and computed tomography of the brain. Radiology 116:363-366, 1975

16. New PFJ, Scott WR, Schnur JA, et al: Computerized axial tomogra phy with the EMI scanner. Radiology 110.109-123, 1974

17. Ambrose J, Gooding MR, Richardson AE: An assessment of the accuracy of computerized transverse axial scanning (EMI scanner) in the di agnosis of intracranial tumour: a review of 366 patients. Brain 93.569582, 1975

18. Reich NE, Zelch JV, Alfidi RJ, et al: Computed tomography in the detection of juxtasellar lesions. Radiology 113:333-335, 1976

19. Niendorf HP, Bull U, Kazner E, et al: The value of serial scintigraphy with Te pertechnetate compared with axial computed tomography in the diagnosis of cerebral tumours. Fortschr Roentgenstr 126.259306, 1977

20. Alderson PO, Mikhael M, Coleman RE, et al: Optimal utilization of computerized cranial tomography and radionuclide brain imaging Neurology (Minneap) 26:803-807, 1976

21. Christie JH, Mori H, Go RT, et al: Computed tomography and radionuclide studies in the diagnosis of intracranial disease. Am J Roentge nol 127:171-174, 1976

22. Gado M, Huete 1, Mikhael M: Computerized tomography of infratentorial tumors. Semin Roentgenol 12:109-120, 1977

23. Mori H, Lu CH, Chiu LC, et al: Reliability of computed tomography: correlation with neuropathologic findings. Am J Roentgenol 128:795798, 1977

24. Naidich TP, Lin JP, Leeds NE, et al: Computed tomography in the diagnosis of extra-axial posterior fossa masses. Radiology 120.333-339,

1976

25. Naidich TP, Pinto RS, Kushner MJ, et al. Evaluation of sellar and parasellar masses by computed tomography. Radiology 120.91-99, 1976 26. Chernak ES, Rodriguez-Antunez A, Jelden GL, et al: The use of com. puted tomography for radiation therapy treatment planning. Radiclogy 117:613-614, 1975

27. Zimmerman RA, Bilaniuk LT, Grundy G, et al: Computed tomo graphic localization for radiotherapy of cerebral tumors. Radiology 119:230-231, 1976

28. Schnabel K, Guillaume R, Hermann HJ, et al: Computed tomography of the whole body as a means of radiation planning. Strahlentherapie 153:51-56, 1977

29. Lee KF, Suh JH, Mazziotta JC, et al: The value of computed tomogra phy in the radiotherapeutic management of juxtasellar tumors Comput Tomogr 1:111-119, 1977

30. Wortzman G: Computerized transaxial tomography its role in the post-operative tumor case. Can J Neurol Sci 3-51-58. 1976

31. Norman D, Enzmann DR, Levin VA, et al. Computed tomography L the evaluation of malignant glioma before and after therapy. Radiology 121.85-88, 1976

32. Brismar J, Roberson GH, Davis KR Radiation necrosis of the brain. neuroradiological considerations with computed tomography. Neuroradiology 12:109-113, 1976

33. Evens RG, Rujanavech N, Mikhael MA: Utilization, reliability, and core effectiveness of cranial computed tomography in evaluating pseudotumor cerebri. Am J Roentgenol 129:263-265, 1977

Vol. 298 No. 5

- ABRAMS AND MCNEIL

261

34. Dublin AB, French BN, Rennick JM: Computed tomography in head trauma. Radiology 122:365-369, 1977

35. Ambrose J, Gooding MR. Uttley D: E.M.I. scan in the management of bead injuries. Lancet 1:847-848, 1976

36. Radiology of the Skull and Brain: Angiography. Specific Disease Processes. Edited by TH Newton, DG Potts. St Louis, CV Mosby, 1974

37. Messina AV, Chernik NL: Computed tomography: the "resolving" intracerebral hemorrhage. Radiology 118.609-613, 1975

38. Alavi A, Uzzell BP, Kuhl DE, et al: Radionuclide and computerized tomography scans in evaluation of head injured patients. J Nucl Med 18:614, 1977

39. Liliequist B, Lindqvist M, Valdimarsson E: Computed tomography and subarachnoid hemorrhage. Neuroradiology 14:21-26, 1977

40. Scott WR, New PFJ, Davis KR, et al: Computerized axial tomography of intracerebral and intraventricular hemorrhage. Radiology 112:73-80, 1974

41. Steiner L, Bergvall U, Zwetnow N: Quantitative estimation of intracerebral and intraventricular hematoma by computer tomography. Acta Radiol [Suppl) (Stockh) 346:143-154, 1975

42. Walshe TM. Hier DB, Davis KR: The diagnosis of hypertensive intracerebral hemorrhage: the contribution of computed tomography. Comput Tomogr 1.63-69, 1977

43. Walshe TM, Davis KR, Fisher CM: Thalamic hemorrhage: a computed tomographic-clinical correlation. Neurology (Minneap) 27:217-222, 1977

44. Davis KR, Taveras JM, New PFJ, et al: Cerebral infarction diagnosis by computerized tomography: analysis and evaluation of findings. Am J Roentgenol 124:643-660, 1975

45. Steinhoff H, Aulich A, Gnumme TH, et al: Diagnosis and followun studies in cerebral infarcts, Proceedings of the International Symposium on Computer Assisted Tomography of Nontumoral Diseases of the Brain, Spinal Cord and Eye. Bethesda, Maryland, October 11-15, 1976, PP 86-87

46. Nielsen H. Gyldensted C: Computed tomography in the diagnosis of cerebral abscess. Neuroradiology 12:207-217, 1977

47. Jordan CE, James AE Jr. Hodges FJ III: Comparison of the cerebral angiogram and the brain radionuclide image in brain abscess. Radiology 104:327-331, 1972

48. Crocker EF, McLaughlin AF, Morris JG, et al: Technetium brain scanning in the diagnosis and management of cerebral abscess. Am J Med 56:192-201, 1974

49. Lout T, El Gammal T, Dasilva R, et al. Evaluation of brain and epidural abscesses by computed tomography. Radiology 122:371-376, 1977 30. New PFJ, Davis KR, Ballantine HT Jr. Computed tomography in cerebral abscess. Radiology 121.641-646, 1976

51. Clavena LE, du Boulay GH, Moseley IF: Intracranial infections: investigation by computerized axial tomography. Neuroradiology 12:5971, 1976

52. Joubert MJ, Stephanov S: Computerized tomography and surgical treatment in intracranial suppuration: report of 30 consecutive unselected cases of brain abscess and subdural empyema. J Neurosurg 47:73-78, 1977

53. Shaw MDM, Russell JA Value of computed tomography in the diag. nosis of intracranial abscess. J Neurol Neurosurg Psychiatry 40:214220, 1977

54. Garfield J: Management of supratentorial abscess: a reveiw of 200 cases. Br Med J 27-11, 1969

55. Naidich TP, Epstein F, Lin JP, et al: Evaluation of pediatric hydrocephalus by computed tomography. Radiology 119:337-345, 1976

56. Wiggli U, Müller-Brand J. Müller HR, et al: Contribution of computerized axial tomography to the diagnosis of malresorptive hydrocephalus. Rev Neurol 132 405-414, 1976

57. Jacobs L, Kinkel W. Computerized axial transverse tomography in normal pressure hydrocephalus. Neurology (Minneap) 26.501-507, 1976 58. Conomy JP, Weinstein MA, Agamanolis D, et al: Computed tomography in progressive multifocal leukoencephalopathy. Am J Roentgenol 127.663-665, 1976

39. Bosch EP, Cancilla PA, Cornell SH: Computerized tomography in progressive multifocal leukoencephalopathy. Arch Neurol 33-216, 1976 60. Carroll BA, Lane B, Norman D, et al: Diagnosis of progressive multifocal leukoencephalopathy by computed tomography. Radiology 122:137-141, 1977

61. Bjorgen JE, Gold LHA Computed tomographic appearance of methotrexate-induced necrotizing leukoencephalopathy. Radiology 122:377378, 1977

62. Gyldensted C: Computer tomography of the cerebrum in multiple sclerosis. Neuroradiology 12.33-42, 1976

63. Cala LA, Mastaglia FL: Computerised axial tomography in multiple sclerosis. Lancet 1:659, 1976

64. Enzmann DR, Hayward RW, Norman D, et al: Cranial computed tomographic scan appearance of Sturge-Weber disease: unusual presentation. Radiology 122:721-724, 1977

65. Martin GI, Kaiserman D, Liegler D, et al: Computer-assisted cranial tomography in early diagnosis of tuberous sclerosis. JAMA 235:23232324, 1976

66. Murphy JV, D'Souza JB, Haughton VM: CT scans and tuberous sclerosis. JAMA 236:1115, 1976

67. Becker MH, McCarthy JG, Chase N, et al: Computerized axial tomography of craniofacial malformations: a preliminary report. Am J Dis Child 130:17-20, 1976

68. Di Chiro G, Axelbaum SP, Schellinger D, et al: Computerized axial tomography in syringomyelia. N Engl J Med 292:13-16, 1975

69. Sheedy PF, Stephens DH, Hattery RR, et al: Computed tomography of the body: initial clinical trial with the EMI prototype. Am J Roentgenol 127.23-51, 1976

70. Di Chiro G, Schellinger D: Computed tomography of spinal cord after lumbar intrathecal introduction of metrizamide (computer-assisted myelography). Radiology 120:101-104, 1976

71. Hilal SK, Shepp LA, Stein JA, et al: High resolution CT scanning of the neck. Presented at the International Symposium and Course on Computed Tomography. Miami Beach, Florida, April 3-8, 1977

72. Hammerschlag SB, Wolpert SM, Carter BL: Computed tomography of the spinal canal. Radiology 121 361-367, 1976

73. Di Chiro G, Doppman JL, Wener L: Computed tomography of spinal cord arteriovenous malformations. Radiology 123-351-354, 1977

74. Enzmann D, Marshall WH Jr, Rosenthal AR, et al: Computed tomography in Graves' ophthalmopathy. Radiology 118.615-620, 1976 75. Hilal SK, Trokel SL: Computerized tomography of the orbit using thin sections. Semin Roentgenol 12:137-147, 1977

76. Pasquini U, Salvolini U, Cabanis EA: Computer assisted tomography observations in one hundred cases of exophthalmos. Proceedings of the International Symposium on Computer Assisted Tomography in Nontumoral Diseases of the Brain, Spinal Cord and Eye, Bethesda, Maryland, October 11-15, 1976, pp 11-15

77. Cabanis EA, Salvolini U: Shape and size of optic nerves in patients with papilledema. Proceedings of the International Symposium on Computer Assisted Tomography in Nontumoral Diseases of the Brain, Spinal Cord and Eye, Bethesda, Maryland, October 11-15, 1976, pp 11-15 78. Chu FC: Radiological methods of studying the orbit. Comput Tomogr 1:45-50, 1977

79. Bergström K: Computer tomography of the orbits. Acta Radiol [Suppl] (Stockh) 346:155-160, 1975

80. Baker HL Jr, Kearns TP, Campbell JK, et al: Computerized transaxial tomography in neuro-ophthalmology. Am J Ophthalmol 78:285-294,

1974

81. Sanders MD: Computer-assisted tomography (EMI scan) in orbital and neuro-ophthalmologic diagnosis. Trans Pac Coast Otoophthalmol Soc 56:17-42, 1975

82. Wortzman G, Holgate RC: Computerized tomography (CT) in otolaryngology. Laryngoscope 86:1552-1562, 1976

83. McNeil BJ, Kecler E, Adelstein SJ. Primer on certain elements of medical decision making. N Engl J Med 293 211-215, 1975

84. Thomson JLG: Cost-effectiveness of an EMI brain scanner. A review of a 2 year experience. Health Trends 9:16-19, 1977

85. Jonsson E, Marké L-A: CAT scanners: the Swedish experience. Health Care Manage Rev 2(2):37-53, 1977

86. Baker HL Jr: The impact of computed tomography on neuroradiologic practice. Radiology 116:637-640, 1975

87. Isherwood I, Rutherford RA, Pullen BR: Computerized tomography of brain. Br Med J 2:746-747, 1975

88. Abbott KH, Gay JR, Goodall RJ: Clinical complications of cerebral angiography. J Neurosurg 9:258-274, 1952

89. Swanson PD, Calanchini PR, Dyken ML, et al: A cooperative study of hospital frequency and character of transient ischemic attacks. II. Performance of angiography among six centers. JAMA 237:2202-2206,

1977

90. Fineberg HV, Bauman R, Sosman M: Computerized cranial tomography: some effects on diagnostic and therapeutic plans. JAMA 238:224227, 1977

91. Cutler SJ, Myers MM. Green SB: Trends in survival rates of patients with cancer. N Engl J Med 293.122-124, 1975

92. George RA, Wagner HN: Ten years of brain tumor scanning at Johns Hopkins: 1962-1972, Noninvasive Brain Imaging. Edited by HJ DeBlanc, JA Sorenson. New York, Society of Nuclear Medicine, 1975, PP 3-16

93. Wortzman G, Horgate RC, Morgan PP: Cranial computed tomography: an evaluation of cost effectiveness. Radiology 117.75-77, 1975 94. Knaus WA, Schroeder SA, Davis DO: Impact of new technology: the CT scanner. Med Care 15:533-542, 1977

310

Feb. 9, 1978

Efficacy of Body CT

The evaluation of body CT is more complex than that of head CT because of the need to consider separately so many organ systems and presenting symptom complexes (e.g., patients with unexplained abdominal pain, jaundice or a pelvic mass). Furthermore, there are many effective noninvasive competing diagnostic modalities for studying diseases of the body.

Diagnostic accuracy

Lungs and pleura. CT has considerable theoretical application to the lungs, and some practical impact has already been shown. CT may help distinguish benign from malignant tumors and tumor from inflammation, "." and its attenuation coefficient numbers may ultimately be useful in distinguishing pulmonary edema from pneumonia, pneumonia from tumor, and infarct from pneumonia. These are applications in which its usefulness must be confirmed. It can detect small metastases not visible on conventional radiography and tomography" and may differentiate pulmonary from extrapulmonary densities, pleural loculations and pleural masses." It can also demonstrate occult effusions, distinguish collapsed or consolidated lung from pleural fluid, and define tumors of the pleu." CT allows better evaluation of lateral extentions of hilar abnormalities for biopsy.

ra."

Nonetheless, neither the value nor the accuracy of lung CT has yet been documented.* In one study in 13 of 16 patients with abnormal chest radiographs showing solitary nodules, "CT was no more definitive or correct than plain radiographic or conventional tomographic examinations."""" On the other hand, CT showed additional nodules in 12 of 23 patients when compared with chest x-ray films and whole-lung tomograms in patients with multiple pulmonary nodules or suspected metastases," and in 10 of 29 pa

From the Department of Radiology, Harvard Medical School, Peter Bent Brigham Hospital, and the Sidney Farber Cancer Institute (address reprint requests to Dr. Abrams at the Department of Radiology, Harvard Medical School, 25 Shattuck Street, Boston, MA 02115).

Supported in part by grants (GM 18674 and HL 20895) from the U.S. Public Health Service and by a grant from the Henry A. Kaiser Family Foundation and the James Picker Foundation (Dr. McNeil is the recipient of a research career-development award (K04 GM 00194) from the U.S. Public Health Service).

Presented in part as the annual Walter W. Herbert Memorial Lecture, University of California Medical School, San Francisco, CA, March 10, 1977 (H.L.A.).

*With a few exceptions the literature to date is largely anecdotal, speculative and nonquantitative. For example, "From a theoretical point of view pulmonary densitometric studies may enable detection of some form of dif fuse lung disease (sarcoidosis, pneumoconiosis)."

Although CT has clearly provided additional infor mation in specific cases of mediastinal disease, in a small group of patients with bronchogenic carcinoma or lymphoma and normal mediastinum on conventional chest films, CT failed to define any mediastinal abnormalities."

Heart. CT imaging of isolated hearts has demonstrated that myocardial infarcts have a sufficiently dif ferent attenuation coefficient from normal myocardium to delineate the area of infarction. The attenuation coefficient of normal and infarcted myocardium correlates closely with the measurement of blood flow to the infarcted and ischemic areas and with changes in wall thickness. In acute coronary occlusion, the decreased attenuation coefficient was linearly related to the severity of myocardial edema. With contrast enhancement, areas of ischemia were detectable in vitro, and the increased iodine content of the intraventricular blood pool readily differentiated the cavity, the free wall, the papillary muscles and the interventricular septum. Relatively accurate measurements of wall thickness and cavity volumes were obtained in arrested canine hearts.

9,10

A gated short-time scanner should afford in vivo scanning with CT and provide hemodynamic data such as end-systolic and end-diastolic volumes, stroke volume and ejection fraction. Such data may have particular value in following patients with heart failure before or after operation.

CT scanning has been shown to be useful in detecting pericardial effusion and distinguishing between a fat pad and a pericardial cyst with its higher attenuation coefficient. 191