CLINICAL METHODS IN THE CYTOLOGIC DIAGNOSIS OF TUMOR
A. REYNOLDS CRANE, M.D., F.A.C.P. * CANCER in most of its forms is a curable disease, and yet it is the second most frequent cause of death in the United States, being exceeded only by diseases of the cardiovascular system. Cancer, when it is definitely recognizable on clinical grounds alone, is seldom in a curable phase. When the clinical diagnosis is equivocal, it is frequently early and curable. The opportunity of reducing the fatality rate thus lies in the hands of both the specialist and the general practitioner through the early recognition of the condition and the prompt institution of proper therapy. To accomplish early recognition the clinician must use all the methods at his disposal. Of these a most important one is the microscopic examination of material from a suspected lesion. The more one becomes familiar with malignant disease the more one realizes that the one certain means of distinguishing between a malignant tumor and a benign growth or inflammatory process lies in .the microscopic study of the lesion. It is no longer good medicine or even excusable to "watch a lesion" over a period of months until the diagnosis becomes definite on clinical grounds. When this is done the observer has usually seen the progression of the lesion from a curable to an incurable stage, and the opportunity of saving the patient has been lost. Material from every accessible suspicious lesion should be submitted to a competent pathologist as soon as possible. This will allow the institution of proper treatment at an early stage when malignant tumor is found and will prevent the needless overtreatment of nonmalignant conditions. The purpose of this communication is not to evaluate various methods, this having already been done in numerous articles, but rather to present the details of various methods that may be used in establishing the diagnosis of tumor. This paper is particularly addressed to the practitioner who does not have a laboratory and pathologist at his side, and who can still do well by his patient by using proper methods for obtaining specimens to be sent elsewhere for diagnosis. Numerous methods of obtaining material for microscopic study are now available. Each of these is applicable to greatest advantage dependent upon the location, size and distribution of the lesion in question. Material may be obtained by biopsy, by From the Benjamin Franklin Clinic of the Pennsylvania Hospital, Philadelphia.
* Director of the Ayer Clinical Laboratory of the Pennsylvania Hospital and the
Benj amin Franklin Clinic; Associate Professor of Pathology, Graduate School of Medicine of the University of Pennsylvania; Assistant Professor of Pathology, Jefferson Medical College; Consulting Pathologist, Valley Forge General Hospital. 1547
1548
A. REYNOLDS CRANE
smear of secretions, by use of sediment of fluids and by use of the principle of adhesiveness of tissue particles to gelfoam. It must be remembered that the diagnosis is rendered by the pathologist only on the basis of the material submitted for study. Thus only a positive diagnosis serves to establish the nature of the lesion when small samples are submitted, as with a fractional biopsy, secretion smears or gelfoam preparations. A negative report under these conditions does not rule out malignant tumor, and study of further specimens may be advisable, dependent upon the nature of the clinical lesion. BIOPSY
Biopsy is the most certain method of establishing a diagnosis. All biopsies should be taken with a sharp knife and not the cautery knife, as the latter coagulates and dehydrates tissue, destroying cell form and structure. The danger of the spreading of a tumor through biopsy has been overemphasized; there is no clinical evidence to support the occurrence of distant metastases after such procedures,': 2 with the .possible exception of lesions suspected of being malignant melanoma, which should always be widely excised locally. The benefits of accurate histologic diagnosis far outweigh the hypothetical risks involved. Infection and hemorrhage, also possible complications, can be readily controlled by use of the antibiotics and hemostatic measures (cautery, gelfoam, topical thrombin). Biopsy may be (1) an excision biopsy in which the entire lesion is removed, (2) a fractional biopsy in which a portion of a lesion is excised, or (3) a needle biopsy in which a core of tissue is removed through a large bore needle. Excision Biopsy. This is preferable for small lesions in accessible sites, such as skin and subcutaneous tissues. The margin should extend at least 0.5 cm. beyond the lesion at all points, the actual distance being tempered by the size and location of the lesion. Adequate depth of the excision is as important as adequate margins-a fact which is too often forgotten. Fractional Biopsy. This procedure constitutes removal of a portion of a lesion for study either by incision of a subcutaneous mass or excision of a portion of an exposed lesion. A knife, cutting curette or biting forceps may be used. Lesions of the cervix and uterus, vagina and bowel, as well as large lesions of the skin, mucous membrane and subcutaneous tissues not amenable to complete excision, are best so treated. Needle Biopsy. This procedure is the removal of a core of tissue (not simply aspiration of cells) from a mass through a large bore needle under suction applied from a syringe. The most pertinent practical application of this technic is in determining the presence of hepatic tumors," either primary or metastatic, and in lesions of the prostate.' It may also be ap-
~
CLINICAL METHODS IN CYTOLOGIC DIAGNOSIS OF TUMOR
1549
plied to other masses, such as breast and bone lesions, where cortical bone is destroyed and not amenable to incisional fractional biopsy. In cases of splenomegaly needle biopsy is of limited practical value, actually being applicable for diagnostic purposes only in cases of lipid disturbances, such as Gaucher's disease. The material obtained is not of sufficient amount as a rule to make the diagnosis possible in such conditions as leukemia, Banti's syndrome and hemolytic jaundice; furthermore,
( ~========~
~t::::::=:::::=========:==::n:>
~~~~~~~
~t:===============2!!!5===~ Fig. 443. Diagram showing approaches for needle biopsy of liver and prostate. The Vim-Silverman needle is also shown. The parts from above are stylet, trocar, cutting needle, trocar and cutting needle assembled.
studies of peripheral blood, bone marrow and liver function, as well as various other clinical pathologic procedures, are more pertinent in establishing a diagnosis. The disadvantage of the method lies in the fact that the specimen is a random one and may not be representative of the lesion. Again only a positive report is of significance. TECHNIC OF NEEDLE BIOPSY. Materials. A local anesthetic agent (lor 2 per cent procaine), a 2 cc. syringe and a 22 gauge needle for injection of the anesthetic, a large 20 cc. syringe for applying suction and an aspirat-
1550
A. REYNOLDS CRANE
ing needle are necessary. A standard 16 or 18 gauge needle 5 to 12 cm. in length may be used, but specially designed needles for biopsy purposes are preferable. Of these the Vim-Silverman needle has proved generally successful. It consists of an outer trocar, a stylet to fit the outer needle and keep it from becoming plugged with tissue, and an inner longer split cutting needle which fits through the outer needle and which actually obtains the specimen (Fig. 443). Procedure. The superficial tissues about the lesion or organ to be biopsied are infiltrated with procaine. A small slit (0.1 to 0.2 cm. in length) is made with a scalpel in the covering surface to allow easy introduction of the needle, especially through skin surfaces. The outer needle with the stylet in place is inserted through the covering structures to just penetrate the lesion to be aspirated. The stylet is removed and the inner aspirating needle is inserted through the outer needle and into the lesion. The depth of penetration is dependent upon the size of the lesion. The outer needle is then pushed with a slight rotary motion into the lesion over the inner cutting needle. Both are then rotated about 360 degrees to cut the end of the plug and both are simultaneously and rapidly withdrawn. If a simple 16 gauge needle is used, constant negative pressure must be maintained with the 20 cc. syringe. While satisfactory specimens may be so obtained, the frequency of failure to obtain an actual core of tissue is enough to make the use of special biopsy needles desirable. Special details are necessary in regard to needle biopsies of the liver and prostate. The field should be sterilized and sterile drapes and gloves used as well as aseptic technic. In either case a mild sedative is necessary only for apprehensive patients, local infiltration by 1 per cent procaine at the site of the needle penetration of the skin and a slit incision (0.2 cm.) of the skin serving to remove the most discomforting aspects of the procedures insofar as the patient is concerned. For hepatic biopsy the transthoracic approach is preferred if the liver is not enlarged. The liver is entered through the lower border of the eighth or ninth interspace (dependent upon liver dullness to percussion) in the right midaxillary line (Fig. 443). The patient is instructed to exhale completely and hold his breath in expiration. A subcostal approach may also be used by entering the liver 2 cm. to the right of the xiphoid and below the costal margin. This approach is not advisable in the presence of ascites because of the danger of overriding of the liver by the transverse colon. The outer needle with stylet is inserted through the capsule, the stylet is removed, the inner needle passed through the outer into the liver, the outer needle passed over the inner, both are rotated 360 degrees and withdrawn rapidly. If the liver is found enlarged upon palpation, any other specific point desired may be chosen for direct entry. Because of the danger of hemorrhage, especially in the presence of jaundice, the
CLINICAL METHODS IN CYTOLOGIC DIAGNOSIS OF TUMOR
1551
prothrombin time should be checked, and if prolonged should be restored to 70 per cent of normal by administration of vitamin K (vitamin K 1 or K 1 oxide-o.5 to 1.0 gm. intravenously") before biopsy is attempted. For prostatic lesions the approach is through the perineum (Fig. 443). One finger is placed in the rectum and guides the outer needle until it enters the prostate. The finger then is hooked above the prostate and the cutting needle guided into the gland, with subsequent steps as in a liver biopsy. Specimen. The specimen must be placed in a fixative if it is to be shipped for examination. The specimen from- needle biopsy is removed from the aspirating needle and preferably placed onto a small square (1 to 2 em.) of absorbent toweling or filter paper to which it will adhere. This obviates handling of the specimen and crushing with destruction of histologic detail, as the material can be moved by handling the paper rather than the tissue. The fixative used should be that preferred by the pathologist who is to study the section. In general 10 per cent formalin is most widely used; the volume of fixative should be ten times the volume of tissue. It is not adequate to use any old acid or alcohol solution that happens to be in the office. Rubbing alcohol and hydrochloric acid, while usually handy, are not good or even satisfactory fixing agents. Materials from excision or fractional biopsy should also be placed in fixatives if they are to be transported for study. In general all fixatives penetrate only a short distance quickly, 0.2 cm. being a good standard penetration. Therefore it is pointless to insert large masses into a fixative. When a specimen is greater than 0.5 cm. in thickness it should be cut. If the specimen is large, blocks 1 to 2 em. square and 0.3 to 0.4 cm. thick should be cut out and put in the fixative along with the bulk so that portions at least will show good histologic detail. If the procedure is done in a hospital the handling of the specimen should be governed by the procedure established by the pathologist. Specimens must not be placed in water or bottles of saline, as this destroys cell structure. SECRETIONS
Any tumor located in a site where the cells may exfoliate is amenable to study by spreading these secretions upon a slide and observing microscopically the individual cell form and structure. The widest use of this principle has been in lesions of the endometrium and cervix through studies of 'vaginal and cervical smears. It has also been successfully applied to sputum, bronchial washings, gastric and duodenal washings, urinary sediments and prostatic secretions. It is extremely important for the clinician to use meticulous care in obtaining and preparing the specimen for staining and study by the laboratory staff. Interpretation of variants of normal cells and their differentiation from malignant cells
1552
A. REYNOLDS CRANE
require a broad cytologic experience, so that the actual interpretation of the specimens cannot as yet be made an office procedure for use by the inexperienced. The value of secretions has been well established by many investigators. Because false positive diagnoses occur even in the best of hands, however, a positive report is not acceptable as final proof of the nature of the lesion, but rather indicates the necessity for immediate and further study utilizing other methods. The study of vaginal smears has its great-
Fig. 444. X-ray film of lung fields showing lesion in left upper lung field.
est value as a screening measure in the detection of unsuspected carcinoma that is without clinical signs or symptoms.": 7, 8 As such it is a particularly useful supplementary tool in the hands of the practitioner who would make a thorough examination of his patients. Application of the method to gastric," duodenal'? and urogenital sediments" will allow the corroboration of a suspected clinical diagnosis by a method short of maj or surgery. Perhaps the most useful application of the smear technic is in pulmonary lesions where either the study of bronchial aspirations or sputum examinations, or both, will produce more positive diagnoses than bron-
CLINICAL METHODS IN CYTOLOGIC DIAGNOSIS OF TUMOR
1553
choscopy and biopsy, because of the inaccessibility of many early lesions. The following illustrative case is a good example of the value of study of bronchial aspirations in detecting small lesions which cannot be visualized or biopsied. J. B. W., a 48 year old male Negro, was admitted with a history of evening chills, fever and anorexia and weakness for eight weeks, having been previously well. Physical examination showed only a questionable decrease in resonance
Fig. 445. Bronchial lesion in the division to the anterior segment of the upper division of the left upper lobe producing the x-ray lesion shown in Fig. 444.
over the anterior chest in the region of second and third ribs in the mid clavicular line; there .were no rales or changes in breath sounds. The remainder of the physical examination was entirely normal. Laboratory determinations were all normal. X-ray showed a density in the left upper lobe (Fig. 444) believed to be carcinoma Or tuberculosis. Repeated attempts to demonstrate tubercle bacilli were unsuccessful. A bronchoscopic examination was done with the report: "The right main bronchus was negative. Careful examination was made of the left upper lobe bronchus and it was even possible to see the subdivision of the lingular portion of the left upper lobe. The mucous membrane was smooth and no
1554
A. REYNOLDS CRANE
tumors or granulations were seen. Only a small amount of secretion was present, which was sent to the laboratory for study." These bronchial secretions revealed tumor cells. A pneumonectomy was done. The specimen showed a fine granulation of the upper lobe. bronchus starting 2 cm. from its origin and extending for a distance of 1.5 cm. The bronchial wall here was thickened to 0.3 cm. This small tumor is shown in Figure 445. The diagnosis was bronchogenic carcinoma.
Uterine Lesions. Smears may betaken either from the posterior fornix or directly from the endocervix. The smears must be taken before pelvic examination is done or on a subsequent visit. Manipulation, insertion of lubricants, use of wet instruments and douching all destroy and distort cytological detail. Since the diagnosis depends on fine changes in cellular and nuclear configuration, and upon thecharacter of nuclear chromatin, the cytologist must have well preserved' cells and carefully taken specimens for study. While material which accumulates on the posterior tongue of a speculum may be used, it usually contains many degenerated cell forms and also represents a marked dilution of actualcervical secretion. The method using a glass pipette originally described by Papanicolaou is preferable. Materials. An 8 inch slightly curved glass pipette which has a narrowed distal opening is used. Such pipettes can now be obtained through any surgical supply house. A 1 or 2 ounce rubber bulb which will tightly fit the pipette and furnish strong suction is required. We prefer glass slides with etched ends. These can readily be inscribed with either lead . pencil or India ink, and both inscriptions will survive through fixing until the slides reach the laboratory for permanent marking. A glass jar containing equal parts of 95 per cent alcohol and ether is needed. We prefer to use a Coplin staining jar with a screw top, which will take a number of slides and can, if necessary, be readily transported. Procedure. The pipette, with bulb attached and compressed, is inserted through the labia and deep into the vagina, without the use of lubricant, until the tip lies in the posterior fornix. As pressure on the bulb is released the tip of the pipette is moved gently from side to side and is withdrawn when the bulb is fully expanded, and the material is treated immediately as noted below. After the vaginal smear has thus been secured, a speculum is inserted, preferably without the use of a lubricant. The cervical os is exposed and material scraped from the edges of the mucosquamous junction with either an Ayre spatula or a tongue blade (Fig. 446). Specimen. The aspirate from the fornix is blown upon a glass slide as soon as it is removed from the vagina. The secretion is spread thinly with a rotary motion. It is most essential that the smear be thin so that individual cells can be clearly studied. When the secretion has been spread, the slide is placed immediately in the alcohol-ether solution. The cervical secretions should be thinly spread in the same manner and immediately
CLINICAL METHODS IN CYTOLOGIC DIAGNOSIS OF TUMOR
1555
placed in the fixative. If the slides can be delivered to the examining laboratory by messenger it is best to leave them in the fixative. If the slides must be sent by mail for examination, they must be allowed to fix for not less than two hours, after which they may be allowed to dry and may be shipped to the examining laboratory in the usual cardboard containers. Pulmonary Lesions. Pulmonary lesions may be studied either through examination of sputum or of material aspirated directly from a bronchus.
CI>--------. '"'\'------""") Fig. 446. Diagram showing aspiration technic for vaginal smear studies, aspirating pipette and bulb, and a tongue blade trimmed for obtaining scrapings of the mucosquamous cervical junction.
Liebow and his co-workers" report greater success with sputa than with bronchial washings; but our own experiences have been more productive with material taken directly from a bronchus after the methods described by Herbut and Clerf;'" and McKay and his associates." Materials. For sputum the only materials necessary are a Petri or other flat dish, glass slides and fixative (95 per cent alcohol and ether in equal parts). For bronchoscopic aspirations a Clerf aspirator with a glass suction trap is used. A special aspirator that will hold a cotton plug" may be used. Procedure. To be fully satisfactory a sputum specimen must be obtained from the lower respiratory tract. This method is described below under "specimen."
1556
A. REYNOLDS CRANE
If bronchial aspirations are to be used it is of utmost importance that the lesion be localized radiologically and that material be obtained from the corresponding bronchus. In the presence of tumor, secretions from an affected bronchus are usually copious, and we have had little difficulty in obtaining adequate specimens. If the bronchus is dry a small amount of saline solution (2 or 3 cc.) may be run into the affected bronchus and then aspirated. Specimen. As with all other methods concerned with' diagnosis on the basis of single cell study, the proper handling of the material obtained is of the utmost importance in obtaining good results. The sediment of the material aspirated into a collecting cup is spread thinly with a rotary motion on a glass slide and immersed immediately in the fixing solution (95 per cent alcohol and ether in equal parts). Fixation should be for at least two hours if the specimen has to be transported in dry containers and then handled as described under vaginal smears. If the aspirator with cotton plug is used, the plug should be removed and the secretions transferred to a slide directly from the plug, using a blotting motion. The slide is then placed directly into the fixative and handled as above. If abundant sediment is obtained this may, in addition to the smears, be transferred to an absorbent paper square and placed in 10 per cent formalin or other fixative preferred by the laboratory to do the interpretation and sent to that laboratory to be treated as a specimen to be cut and stained. For this purpose we have found gelfoam most satisfactory, as will be described later. Sputum must be fresh when used as a specimen source. Several (three) specimens collected on separate days should be submitted for study. The patient is instructed how to raise material from the lower respiratory tract, particular emphasis being placed upon disregarding saliva and material from the nasopharynx. The pulmonary specimen is expectorated into a clean Petri dish. As soon as a satisfactory specimen is produced, flecks preferably blood-tinged or gray-white and resembling tissue, are picked out, placed upon a slide and spread out with an applicator to form a thin film on the slide. The slide is then immediately placed in the fixative and handled as for other specimens. Gastric Lesions. Materials necessary are a Levin tube in which several additional openings have been cut in the side, a 50 cc. syringe for aspiration purposes, saline solution and centrifuge tubes, slides and fixatives. Procedure. The tube is measured and marked so that it will lie well within the stomach. It may be passed either by nose or mouth with the patient lying prone. The stomach is aspirated dry and the material placed in a container. Saline solution 100 cc. (tap water may be used) is then placed in the stomach through the tube and partially washed in and out several times. It is then left in and the patient massaged lat-
CLINICAL METHODS IN CYTOLOGIC DIAGNOSIS OF TUMOR
1557
erally, prone, supine and sitting, in an attempt to make the fluid wash the stomach. The stomach is then again aspirated dry and the material placed in a container. The method may also be applied to duodenal contents if care is taken to see that the tube openings are in the duodenum.'" Specimen. Gastric washings must be immediately transferred to slides and fixed to avoid the destructive action of gastric secretions. The samples obtained are first centrifuged at 2000 r.p.m. for twenty minutes. It is best to centrifuge the entirety of each specimen. If large (50 cc.) tubes are available this can be done at once. If not, the specimen can be distributed among smaller tubes, centrifuged, the supernatant discarded, sediment mixed from several tubes, and recentrifuged until the entire sediment is in one tube. The sediment is then placed on glass slides that have been freshly coated with egg albumen, this being necessary as the sediment of gastric secretions will not adhere well to clean slides. The sediment is then thinly spread on the slide with an applicator or wire loop and the slides are placed immediately in the fixative with subsequent handling as previously described. Renal and Bladder Lesions may be studied either by use of whole urine or that selectively obtained from ureteral catheterization. A freshly voided specimen is centrifuged as with other bulk fluid specimens and the sediment is thinly spread upon a slide that has been coated with egg albumen. Immediate fixation and subsequent handling are as with other specimens. Prostatic Lesions. Specimens are obtained by digital prostatic massage. When a drop of secretions appears at the meatus it is collected on a glass slide, spread thinly with either an applicator or a second slide and placed immediately in the fixative with subsequent handling as previously described. Fluids aspirated from pleural or abdominal cavities if properly studied for tumor cells will allow proper treatment of nonmalignant cases and serve to establish the diagnosis in many due to carcinomatosis. The entire fluid must be used as the source of the specimen. To the specimen enough glacial acetic acid is added to make a 2 per cent solution. This will dissolve red blood cells. The entire fluid is then centrifuged and smears of the concentrated sediment made on glass slides with fixation and handling as previously noted. If sediment is abundant this may also be piled upon absorbent paper or picked up with gelfoam, either then being fixed in 10 per cent formalin and submitted for study as a tissue block. GELFOAM
Gladstone'": 16 has shown that tissue particles will adhere well to an absorbing sponge of gelatin (Gelfoam No. 12-Upjohn). If a given ul-
1558
A. REYNOLDS CRANE
cerative lesion can be visualized as in the mouth, cervix, or on a skin surface, it is first wiped clean of necrotic material with a piece of sterile gauze. A piece of gelfoam about 1.5 by 1.5 by 0.5 em. is then vigorously rubbed over the raw surface until it is well soaked with tissue juices and suspended contents. The piece of sponge is then placed in 10 per cent formalin or other fixative and sent for pathological examination as with any tissue fragment. This absorbent property of gelfoam may be used as well for the diagnosis of gastric Iesions'" by tying numerous pieces of the gelfoam to a piece of braided silk string or to a Levin tube. After swallowing such a string of pledgets the patient is manipulated as in doing gastric washing for tumor cells, then the pledgets are retrieved, fixed in formalin and sent for sectioning. We have also found gelfoam to be most useful in absorbing sediment from bronchial secretions, gastric washings and fluids of various types. SUMMARY
1. Clinical procedures for office use in obtaining specimens for .eytologic study by biopsy, smear and gelfoam methods are outlined. 2. Meticulous care by the physician in procuring proper specimens at the bedside will increase the accuracy of diagnosis by these methods. 3. The interpretation of material should not be attempted in the office of the practitioner, as it requires special training and a broad experience with normal and abnormal cell morphology. REFERENCES 1. Ackerman, L. V. and Regato, J. A.: Cancer Diagnosis, Prognosis and Treatment. St. Louis, C. W. Mosby Co., 1947. 2. Greenough, R. B.: Ann. Surge 102: 233-238, 1935. 3. Davis, W. D., Scott, R. W. and Lund, H. Z.: Am. J. M. Sc. 212:449-461,1946. 4. Roth, A. A. and Turkel, H.: J. Urol. 51 :66-68, 1944. 5. Miller, R., Harvey, W. P. and Finch, C. A.: New England J. Med. 242:2]6, 1950. 5. Miller, R., Harvey, W. P. and Finch, C. A.: New England J. Med. 242:211216, 1950. " 6. Meigs, J. V., Graham, R. M., Fremont-Smith, M., Janzen, L. T. and Nelson, C. B.: Surg., Gynec. & Obst. 81 :337-345, 1945. 7. Jones, C. A. Neustaedetor, T. and MacKenzie, L. L.: Am. J. Obst. & Gynec. 49 :159-168, 1945. 8. Gates, 0., MacMillan, J. C. and Middleton, M.: Cancer 2 :838-844, 1949. 9. Fremont-Smith, M., Graham, R. M. and Meigs, J. V.: J.A.M.A. 138:469-474, 1948. 10. Lemon, H. M. and Byrnes, W. W.: J.A.M.A. 141:254-257,1949. 11. Herbut, P. A. and Lubin, E. N.: J. Urol. 57:542-551,1947. 12. Liebow, A. A., Lindskog, G. E. and Bloomer, W. E.: Cancer 1 :223-233,1948. 13. Herbut, P. A. and Clerf, L. H.: J.A.M.A. 130:1006-1012, 1946. 14. McKay, D. G., Ware, P. F., Atwood, D. A. and Harken, D. E.: Cancer 1 :208222,1948. 15. Gladstone, S. A.: Cancer 2 :604-894, 1949. 16. Gladstone, S. A.: Am. J. Clin. Path. 19:891~894, 1949.