I N AN EARLIER paper 2 it was indicated that fluorescein appeared to be

Transcription

1 THE CLINICAL USE OF FLUORESCEIN NEUROSURGERY THE LOCALIZATION OF BRAIN TUMORS* GEORGE E. MOORE, M.D.,~ WILLIAM T. PEYTON, M.D.,~ LYLE A. FRENCH, M.D., ~ A~D WALTER W. WALKER, M.D. w University of Minnesota Medical School, Minneapolis, Minnesota (Received for publication March 5, 1948) I N AN EARLIER paper 2 it was indicated that fluorescein appeared to be helpful in the detection of malignancy, particularly in the recognition of brain tumors. This report concerns the clinical use of fluorescein as an aid in the localization and recognition of intracranial neoplasms. Fluorescein has been used for this purpose during the past year in this clinic and has proved to be of definite clinical value. Although brain tumors can usually be localized by neurological examination and ventriculographic study, this is not always true. Cerebral edema, which is so frequently present in association with subcortical tumors, often renders precise preoperative determination of the size and location of tumor impossible. Unless some evidence of the tumor appears in or through the dura at the time of operation, there is doubt concerning the exact site at which the dura should be opened, for one should confine the dural opening to the smallest defect consistent with adequate exposure. In this clinic, as well as in many others, it has been customary to locate subcortical tumors with a brain needle before the dura is opened widely. In these cases an opening is made just large enough to pass freely a brain needle into the most probable site of the tumor. A needle biopsy is obtained by creating a vacuum in the syringe attached to the needle. This biopsy material may be sectioned to determine the presence and even the type of tumor. This procedure, however, is time-consuming and the presence of edema or necrosis often makes the sectioning and mounting of the minute pieces difficult. But with the method described in this report the surgeon or his assistant can determine the presence of tumor at once. The mere presence of tumor is the essential information desired, for regardless of the type, the tumor should be removed so that an internal decompression is obtained to make recovery from the operation more certain and more rapid. Since neoplastic tissue as well as edematous tissue adjacent to a neoplasm can be recognized by applying the fluorescein technique, the value of this procedure can readily be appreciated. IN * Supported by grants from the National Cancer Institute and the United States Public Health Service, the Malignant Disease Fund, and the Flora L. Rosenblatt Fund for Cancer Research. t Senior Research Fellow, United States Public Health Service; Department of Surgery. ~/Division of Neurosurgery. w Department of Pathology. 39~

2 CLINICAL USE OF FLUORESCEIN IN NEUROSURGERY 393 TECHNIQUE At present, the following technique is being used in this clinic. Immediately before or upon the completion of ventriculography, 5 cc. of a ~0 per cent solution of sodium fluorescein (1 gm.) is injected very slowly intravenously. With this high concentration the patient may become nauseated and vomit if the dye is injected too rapidly, but the 20 per cent solution is still being used because the small containers in which it is bottled are conveniently sterilized, and when injected more slowly the frequency of disagreeable reactions encountered is low. Another satisfactory method is to add the fluoresccin to the first 500 cc. of intravenous fluid that is started immediately preoperatively. Children require relatively greater amounts of dye than adults. A convenient volume containing the equivalent of.85 to.5 gin. of the dye gives satisfactory results. It is most important that the dye be given early enough so that an interval of at least 1 hour or more intervenes between the time of injection and the examination of the tissue under suspicion. If the tissue is examined too soon after injecting the dye, the differential fluorescence of normal and neoplastic tissue is insufficient for diagnosis. On the other hand, if the interval is greater than 5 hours, the degree of fluorescence is diminished. The maximum tissue fluorescence is attained in about ~ hours. At the time of operation, the material obtained by needle biopsy is collected on gauze squares and examined under ultraviolet light. A CH-4 Mercury Vapor Lamp with a Wood's filter has been found to be convenient and efficient for this purpose. Since it is not necessary to darken the room in order to view the difference in fluorescence, the examination can be carried out in the operating room. CLINICAL EXPERIENCE In 46 patients operated upon because of a clinical diagnosis of a possible brain tumor, fluorescein has been injected intravenously before operation and 5~ needle biopsies have been obtained and examined under ultraviolet light. In addition, the gross specimen was examined and the presence or absence of fluorescence recorded. The pathological examination of individual needle biopsies was done by one of us (W. W. Walker) and the final diagnoses are taken from the records of Dr. A. B. Baker, Division of Neuropathology. The results appear in Table 1. Of the 46 patients subjected to the fluorescein technique the presence or absence of tumor tissue was correctly determined in 44 instances. In one case (#31), the correct diagnosis was suggested but not unequivocally. In one instance (Case #15), a needle biopsy appeared to fluoresce even though no neoplastic cells were found by examination of multiple frozen sections. No reason for this discrepancy can be given. In another instance, (Case #17), six pieces of tumor tissue (angioblastoma) did not fluoresce when examined under the ultraviolet lamp. It is possible that in this case, the excessive amount of clotted blood present obscured such fluorescence as might have been present. It is noteworthy that the fluid aspirated from cysts associated with brain tumors have invariably fluoresced a brilliant yellow. The specificity of this phenomenon is not known since no cysts have been explored that were not associated with tumor tissue. Fluid obtained from the ventricles does fluoresce faintly, but can easily be distinguished from cyst fluid. After removal of infiltrating brain tumors, the residual cavity can be examined directly under the ultraviolet lamp. In several cases, additional small pieces of tumor tissue which had escaped notice when viewed with

3 394 G.E. MOORE, W.T. PEYTON, L.A. FRENCH AND W. W. WALKER TABLE 1. Compar~on ~fluoresce~ew~hh~tolog~al di~nos~ ~ ~a~tumors Case Presence of Diagnosis from No. Initials Specimen Fluorescence Histological Section Evaluation 1 L.M. Needle Biopsy Nolle No Tumor Correct Needle Biopsy Present Astroblastoma Correct Tumor Present Astroblastoma Correct M.M. 3 J.M. Needle Biopsy Present Astrocytoma Correct Tumor Present Astrocytoma Correct 4 H.H. Needle Biopsy Present Astrocytoma Correct Cyst Fluid Present Correct Tumor Present Astrocytoma Correct 5 H.M. Tumor Present Meningioma Correct 6 C.D. Tumor Present Meningioma Correct 7 A.C. Needle Biopsy Present Metastatic Correct Tumor Present Metastatic Correct 8 J.H. Needle Biopsy Present Astroblastoma Correct Tumor Present Astroblastoma Correct 9 J.G. Tumor Present Acoustic Neuroma Correct 10 L.P. Needle Biopsy None Tuberculoma Correct 11 M.K. Tumor Present Meningioma Correct 1~ F.R. Tumor Present Glioblastoma Correct 13 D.T. Needle Biopsy Present Meningioma Correct Tumor Present Meningioma Correct 14 S.R. Tumor None Cholesteatoma Correct lb H.K. Needle Biopsy Present No Tumor Error Needle Biopsy Present Astroblastoma Correct Tumor Present Astroblastoma Correct 16 S.S. Tumor Present Chromophobe Adenoma Correct 17 I.K. 3 Needle Biopsies None Angioblastoma Error Tumor Nolle Angioblastoma Error 18 G.H. Needle Biopsy Present Astrocytoma Correct Tumor Present Astrocytoma Correct 19 F.J. ~0 J.S. Tumor Present Ependymoma Correct ~1 T.H. ~ R.K. ~3 L.M. ~4 W.N. Needle Biopsy Present Anglioblastoma Correct Tumor Present Angioblastoma Correct ~5 L.H. Tumor Present Astrocytoma Correct ~6 P.K. Needle Biopsy Negative No Tumor Correct ~7 G.B. Tumor Present Astrocytoma Correct Cyst Fluid Present Correct 28 H.P. Tumor None No Tumor Correct Tumor Present Astrocytoma & Cyst Correct ~9 O.K. Tumor Present Chromophobe Adenoma Correct

4 CLINICAL USE OF FLUORESCEIN IN NEUROSURGERY 595 Table 1 (continued) Case Presence of Diagnosis from No. Initials Specimen Fluorescence Histological Section Evaluation 80 E.A. Needle Biopsy Present Angioblastoma Correct Tumor Present Angioblastoma Correct 81 C.R. Needle Biopsy Negative? No Tumor Doubtful Needle Biopsy Present (slight) Meningioma Doubtful Tumor Meningioma Correct 8~ J.C. Tumor Present Glioblastoma Correct 83 E.J. Needle Biopsy None Clotted Blood Needle Biopsy Slight Edema Correct Needle Biopsy Present Astrocytoma Correct Tumor Present Astrocytoma Correct 34 E.S. Needle Biopsy Present Ependymoblastoma Correct Tumor Present Ependymoblastoma Correct S5 T.B. Needle Biopsy Present Meningioma Correct Needle Biopsy Present Meningioma Correct Tumor Present Meningioma Correct 36 P.K. Needle Biopsy Present Metastatic Carcinoma Correct Tumor Present Metastatic Carcinoma Correct 37 A.W. Tumor Present Meningioma Correct 38 H.L. Needle Biopsy NoDe No Tumor Correct Tumor Present Glioblastoma Correct 89 J. ti. Tumor Present Ependymoblastoma Correct 40 J.B. Tumor Present Meningioma Correct 41 W.W. Tumor Present Medulloblastoma Correct 4~ L.W. Tumor Present Acoustic Neuroma Correct 43 A.B. Tumor Present Meningioma Correct 44 U.G. Tumor Present Meningi3ma Correct 45 J.S. Tumor Present Metastatic Tumor Correct 46 A.B. Cyst Fluid Present Glioblastoma Correct Tumor Present Glioblastoma Correct ordinary illumination, were detected by their fluorescence. These pieces were removed and proved to be tumor by microscopic examination of paraffin sections. By this method, an additional check on the thoroughness of the operative procedure is obtained. The usefulness of this survey of the operative defect is limited by the tendency of even a small amount of hemorrhage to obscure fluorescence. This use of ultraviolet lamp in the operating room, however, can be done safely only if a non-explosive anesthetic is being used, although no accidents have occurred in the two years

5 396 G. E. MOORE, W. T. PEYTON, L. A. FRENCH AND W. W. WALKER these lamps have been employed. At this clinic, a combination of Baird's solution* and nitrous oxide is used for the majority of intracranial operations. As previously mentioned, the surgeons themselves can examine the suspected tissue. A brilliant yellow-green fluorescence characterizes tumor tissue whereas the normal brain, which retains appreciably less dye, appears white. Edematous tissue surrounding the tumor does fluoresce, but to a lesser degree so that it can be readily separated both from normal brain and the tumor itself. It is well to warn the patient or his relatives that his skin temporarily will have a yellow color. Generally, the dye disappears from the skin in 18 to ~4 hours. It has been noted that patients with severe liver disease do not excrete the dye as rapidly as other patients, and therefore the yellow tinge of their skin persists for several days. As can be noted in Table 1, both meningeal and glial tumors have been recognized with equal facility, even though the degree of fluorescence was usually greater in the gliomas. In ~ eases of metastatic tumor to the brain subjected to the fluoreseein technique (Cases #36 and #45), there was a definite, brilliant fluorescence. Unfortunately, the primary lesions were unavailable for comparative studies. DISCUSSION We have no explanation at the present time for the increased concentration of fluorescein found in malignant tissue. If vaseularity were the chief factor concerned, it might be reasonable to assume that the greatest degree of fluorescence would be observed immediately after injecting the dye, but this does not occur. Since an interval of several hours must intervene before the differential fluorescence of normal and tumor tissue is obtained, the possibility of a difference in the retention of the dye must be entertained. Since fluorescein is an acid-chromagen dye, the so-called "blood brain barrier" may play an important role in the selectivity of the dye for tumor tissue. This could be based only on the postulation that the blood vessels of tumor tissue differ physiologically from those of normal brain. No brain abscesses suitable for testing with fluorescein have been encountered since this study was begun, therefore it can only be suggested that acute inflammatory tissue will probably take up a large amount of dye, but may not retain it as long as tumor tissue. In this regard it is noteworthy that Sorsby ~ and his associates have reported successful staining of the retina by amphoteric dyes only in animals whose retinae had first been damaged experimentally. Attempts to apply vital staining to brain surgery were not uniformly successful, but it was concluded that: "It was probable that pathological structures within the * A solution of pentathol and curare containing 23 milligrams of sodium pentathol and 5 units of d-tubocurarine chloride.

6 FIG. 1. ~IG. ~. PLATE I. Depicting the appearance of a recurrent brain tumor (~39 ependymoblastoma) removed at operation, under ordinary illumination (Fig. 1) and when viewed with ultraviolet light (Fig. ~). On the left in eachifigure are dishes containing normal (top) and tumor (bottom) tissue. These differences in appearance are illustrative of the results commonly found with the fluorescein technique. (Photographs by Mr. W. Holmquist of the Medical Photography Laboratory.)

7 CLINICAL USE OF FLUORESCEIN IN NEUROSURGERY 397 central nervous system containing connective tissue or marked degeneration will show contrast staining". 6 Since the initiation of the present study, a very interesting and pertinent paper was encountered while scanning the literature concerned with cerebral vascular disorders. Broman 1 postulated that the permeability-regulating function of the brain vessels did not cease with the death of the animal and attempted a similar study of human brains removed soon after death. He reported that with a perfusion of 0.~ per cent trypan blue, and subsequent washing out of the residual dye in the vessels, the normal brain remained unstained. On the contrary, areas of non-necrotic tumor tissue, abscesses, edema, and demyelination (multiple sclerosis) showed diffuse staining. He concluded, "... tumors of the brain, and their vessels seem to lack the special permeability function of the blood-brain-barrier... " Broman's attempted clinical use of iodine-containing compounds resembling trypan blue to enhance roentgen visualization of the damaged areas was unsuccessful. An attempt has been made to quantitate by chemical fluorometric methods the amount of dye present in both tumor and normal tissue after varying time intervals. However, it was found after chemical extraction that different samples of the same tissue gave variable fluorometric measurements. This might be due either to a differential absorption of the dye by protein from the normal and tumor tissues, and/or a more rapid quenching of fluorescence by the tumor extract. With the help of Dr. Gerald Boyack of the Department of Chemistry and Dr. Wallace Armstrong of the Department of Physiologieal Chemistry, radioactive iodine (131) has been incorporated into diiodofluoreseein. It was hoped to utilize the gamma radiation in an attempt to localize brain tumors clinically, 3 and to measure quantitatively the relative amounts of dye in tumor and adjacent "normal" brain tissue removed at operation. 4 In the instance of the first patient who received radioactive dye and subsequently had tissue removed at operation, the activity of the tumor sample was ~.8r times greater than that of adjacent brain tissue. In all probability this does not represent accurately the true ratio of dye concentration in tumor and normal brain, since usually the brain tissue immediately surrounding the tumor is edematous and edematous brain tissue is known to retain a greater amount of dye than otherwise normal brain. Several other iodine-containing fluorescein dyes have been synthesized. The least toxic and most radiopaque dye so far has been designated as sodium tetra-iodo-phthalic fluorescein. Preliminary attempts to localize brain tumors by roentgenography have not been uniformly encouraging. Other related radiopaque and radioactive dyes are undergoing experimental trials at the present time.

8 398 G. E. MOORE, W. T. PEYTON, L. A. FRENCH AND W. W. WALKER SUMMARY 1. Fluorescein has proved useful in determining the presence or absence of tumor tissue in needle biopsies of brain tissue. ~. Fluorescein technique to confirm the complete removal of infiltrating gliomas is suggested. 3. Present and future studies of radioactive and radiopaque fluoresceinlike dyes are outlined. REFERENCES 1. BROMAN, T. Supravital analysis of disorders in cerebral vascular permeability in man. Acta reed. Scand., 1944, 118: MOORE, G.E. Fluorescein as an agent in the differentiation of normal and malignant tissues. Science, 1947, 106: MOORE, G. E. Use of radioactive diiodofluorescein in the diagnosis and localization of brain tumors. Science, 1948, 107: MOORE, G.E. (In preparation) 5. SORSBY, A. Vital staining of the retina: preliminary clinical note. Brit. J. Ophthal., 1939, 23: ~0-~4. 6. SOaSBV, A., WnmnT, A. D., and ELKELES, A. Vital staining in brain surgery. A preliminary note. Proc. roy. Soc. Med., 194~, 36: 137.