sign in sign up
<<
Home , Gynecologic oncology

Diagnostic and Interventional in Gynecologic

BJÖRN THORVINGER

LUND 1990 OifMåatioa Document name LUND UNIVERSITY DOCTORAL DISSERTATION Department of Diagnostic Radiology Date of issue University May 1990 S-221 85 Lund, Sweden ÖBS^DW/fMEXL-1021) /199 0 Authors) Björn Thorvinger

Title and nblitle Diagnostic and in gynecologic neoplasms

Abstract The role and clinical value of the modern radiologic methods for evaluation of gyne- cologic tumors is not finally settled. The aims of our investigation were therefore tg, compare clinical examination with CT in patients with possible recurrence of cervical carcinoma; to evaluate th-> usefulness of CT in patients with fistulas following gynecologic tumors or their treatment; ^evaluate the ability of transabdominal US and MR imaging in intrauterine staging including myometrial invasion in patients with endometrial carcinoma; to evaluate CT in the capacity of monitoring response, probable recurrence or clinical remission in patients with ovarian carcinoma; and J|o evaluate the effect of intraarterial occlusion in facilitating and in evalu.;ing the role of the intraarterial infusion in gynecologic tumors otherwise refractory to all therapy given. CT was more accurate (91%) than clinical pelvic examination (78%) in revealing extensive disease after radiation and/or surgical treatment. CT was also a most valuable tool in demonstrating genital fistulas following gynecologic malignancy or its treatment. Transabd nninal US did not improve staging in early endometrial carcinoma while MR had potential for delineating intrauterine tumor growth (accuracy for myometrial invasion 95%). CT was most valuable in the evaluation of therapeutic response of ovarian malignancy. For possible recurrence or in clinical remission, only positive CT was of clinical significance. The potentials of transcatheter intraarterial management in order to facilitate operability are also discussed.

Keywords Gynecology, Neoplasms; -, fistulas; Uterus, neoplasms; , -, CT; MRI; Interarterial therapy. Classification system and/or index terms (if any)

Supplementary bibliographical information Language English

ISSN and key tide ISBN

Recipient's notes Number of pages Price

Security classification

Distribution by (name and address) Dr Björn Thorvinger, Department of Diagnostic Radiology, University Hospital, S-221 85 V.TnVunäenigne»?, being the copyright owner of the abstract of the above-mentioned dissertation, hereby grant to all reference sources permission to publish and disseminate the abstract of the above-mentioned dissertation.

March 1st, 1990 3 Date DIAGNOSTIC AND INTERVENTIONAL

R.AD X OLOGY

IN GYNECOLOGIC NEOPLASMS

BJÖRN THORVINGER leg. läk., Malmö

Akademisk ^ *.ndling som med vederbörligt tillstånd av Medicin? -. 1'akulteten vid Universitetet i Lund för avläggande ?/ jK'.*dicine doktorsgraden kommer att offentligen försvaras i t .reläsningssal 1, Centralblocket, Lasarettet i Lord tisdagen den 8 maj 1990 kl 09 15. FROM THE DEPARTMENT OF DIAGNOSTIC RADIOLOGY. LUND UNIVERSITY. SWEDEN.

Diagnostic and Interventional Radiology in Gynecologic Neoplasms

BJÖRN THORVINGER

Lund 1990 This thesis is based on the following papers:

I. Thorvinger B., Hauksson A.. Samuelsson L. and Tropé C: Computed tomography in advanced carcinoma of the iterine . Acta Radiol. Diagnosis 25 (1984), 423. II. Thorvinger B.. Horvath G. and Samuelsson L.: CT demonstration of fistulas in patients with gynecologic neoplasms. Acta Radiol. Diagnosis 31 (1990). III. Thorvinger B.. Gudmundsson T., Horvath G., Forsberg L. and Holtås S.: Staging in local endometrial carcinoma - Assessment of magnetic resonance and ultrasound examinations. Acta Radiol. Diagnosis 30 (1989), 525. IV. Thorvinger B., Samuelsson L. and Skjaerris J.: Computed tomography of malignant ovarian disease. Acta Radiol. Diagnosis 28 (1987), 739. V. Thorvinger B.. Jörgensen C. W., Samuelsson L. and Tropé C: Transcatheteral intraarterial management of gynecologic tumors. Acta Radioi. Diagnosis 26 (1985). 701.

•-908090 DIAGNOSTIC AND INTERVENTIONAL RADIOLOGY IN GYNECOLOGIC NEOPLASMS

BJÖRN THORVINC.FR

Modern gynecologic is concerned with the dent. A potential problem with CT is the evaluation of the exact delineation of disease extent (staging) before treat- sagittal structures of the pelvis, because they cannot be ment, better identification of the sites of treatment failure, obtained without a computed reconstruction, which gives and more aggressive curative, as well as palliative treat- poor imaging. ment for tumors of the female reproductive tract. Current MR is a relatively new technique, which involves the dia. istic radiology may significantly facilitate solutions use of radiofrequency waves in varying magnetic fields to in all three major areas of concern. produce cross-sectional images of the body with high soft tissue contrast discrimination. The examination is non-in- I. Radiologic techniques used in evaluation of vasive, without ionizing radiation and has no observable gynecologic tumors hazard (73). The images can easily be obtained in the Until the last decade diagnostic radiology relied on coronal, sagittal, and transverse planes. The pelvis is conventional techniques, including conventional abdomi- particularly well suited to MR imaging because of the nal , urography, various barium examinations abundant natural contrast provided by fat. urine, and gas and hysterosalpingography (HSG) in the evaluation of as well as the absence of respiratory motion. At present, patients with gynecologic tumors. With the introduction the main disadvantages of MR is its long scanning time of the modern imaging techniques, i.e. ultrasonography and high cost of equipment and performance of examina- (US), computed tomography (CT), and magnetic reso- tions. nance (MR) imaging, the diagnostic spectrum has been Interventional radiology. In addition to. among others, changed entirely. In addition, the variety and possibilities percutaneous biopsy, nephrostomy, and ureteral stents. of interventional radiologic techniques have rapidly in- an increased interest in transcatheter management of ma- creased. lignant gynecologic tumors has been stimulated by the Ultrasonography. Two types of sonographic imaging development techniques of percutaneous transcatheter in- devices are available for evaluation of a patient with a fusion and occlusion. The rationale for arterial infusion is pelvic mass: static B-mode scanners and real-time equip- to expose the to a high local concentration of a ment. The static device affords serial tomographic imag- chemotherapeutic agent (as compared with the intrave- ing of the pelvis and the surrounding structures. The nous administration) without any increase in toxicity. The real-time scanner allows far more flexibility in the scan- control of a hemorrhage in patients with gynecologic ma- ning technique with wide-angle high resolution in trans- lignancy by transcatheter occlusion of tumors may at verse, sagittal, or oblique planes. A full urinary bladder is times be life-saving and allow more aggressive debulking required to create a sonic fluid-filled window to visualize surgery (II). the pelvic structures. The main advantages of US are its low cost compared with CT and the absence of ionizing II. Evaluation of gynecologic tumors. State of the art radiation. There is no tissue damage or risk for the patient 1. Carcinoma of the cervix in the diagnostic range of 2.5 MHz to 10 MHz (22). A Cervical carcinoma accounts for about 3 per cent of all major disadvantage of 'IS is that it is extremely operator- new cases of in the Swedish female population dependent. (77). The incidence varies on an international basis, as do CT. A compensation for the exposure of ionizing radi- other gynecologic (I). A vast majority of cervical ation with CT is that it has a better over-all contrast cancers are squamous cell carcinomas (95%); the remain- discrimination than US and it is far less operator-depen- der consisting of adenocareinomas. mixed carcinomas BJÖRN THORVIMiER

(adenosquamous) and rare sarcomas (mixed mesodermal Table 1 tumors, lymphosarcomas). The earliest squamous cell C lirtuul .stttiiini; wstem tor ctrvictil cart iru>rnu carcinoma is confined to the epithelial layers (carcinoma in situ) and it is considered that the disease remains Stage Extent of disease Subclussiticution confined to the mucosa membrane lor several years be- fore invading the subjacent stroma. Carcinoma in situ 1) In Mlu occurs most frequently in women aged between 30 and 40 1 Confined to cervix a. Microinvasion b. All other years of age. whereas invasive carcinoma is encountered II Extending beyond a. Vaginal fornix involved most often in women aged between 40 and 50 years (30. cervix (central b. Paranietrium. but not 46.67.91). pelvic disease) extending to pelvic Following penetration of the basement membrane and Mi.leu.ill III Extending beyond a. Lower involved involvement of the cervical stroma. the disease spreads cervix (advanced b. Nodular pelvic sidewall by direct contiguity to the vagina and the adjacent para- pelvic disease) involvement metrium. and via the lymphatic drainage (which is abun- IV Invading adjacent a. Clinical bladder or dant in this area) to the regional lympii nodes of the pelvis viscera and or ex- rectum involvement (parametrial, obturator, iliac and presacral) and to the trapelvic extension b. Outside juxta-regional (para-aortic) lymph nodes (33). Since the introduction of the Papanicolaou smear tech- sis. and superimposition of the adjacent normal nodes (47. nique the mortality rate has decreased. However, ap- 58. 80. 87). proximately one fifth of the patients seeking medical at- Apart from detecting hydronephrosis (indicating stage tention have been experiencing the symptoms (bleeding III) US offers little in patients with early stages of cervical and leukorrhea) for over six months (46). carcinoma (69). Tumor extension into the parametrium Mapping the extent of the malignant process is crucial with enlarged pelvic lymph nodes may be demonstrated in determining the most appropriate mode of therapy and but infiltration into the bladder and rectum may be diffi- estimating the prognosis. Staging is made clinically and is cult to define by US (47). defined according to the classification of the International The role of CT in patients with cervical carcinoma is to Federation of Gynecology and (FIGO) (74) determine tumor size, parametrial and pelvic wall exten- (Table I). The staging is based on a general physical sion, and pelvic adenopathy. The axial images obtained at examination, a pelvic examination during general anesthe- CT give an incomplete information about the early stages sia with histologic verification, cysto- and rectoscopy. (especially IB and IIB) of the disease (43). CT has (com- and a thorax radiogram. Additional bone radiograms may pared with surgical evaluation), an overall accuracy of also be used. All other radiologic diagnostic techniques more than 65 per cent and over 90 per cent in the staging employed may be used to select the choice of therapy but of advanced lesions (43. 84). A clinical diagnostic problem not the staging. is the differentiation between tumor recurrence and radi- l)iaf{r'istic imaging. Conventional radiography, includ- ation fibrosis. The development of ureteral obstruction ing abdominal survey and urography may occasionally after treatment has been reported almost invariably as reveal metastases from a cervical carcinoma. Hydrone- being due to the recurrent tumor (95%), rather than radi- phrosis due to parametrial invasion is present in 20 per ation fibrosis. Thus if hydronephrosis is the only finding at cent of the cases at the time for diagnosis (47). Sequelae CT. this gives additional information (47). after may include a thickened and rigid MR imaging may well play a role in the distinction bladder and rectum with a widened presacral space. Rec- between early stages of cervical carcinoma. This is of tovaginal and vesicovaginal fistulas arise in approximately clinical importance in choice of treatment. The ability of one per cent of the patients with somewhat higher occur- sagittal planes may outline uterine or vaginal tumor exten- rence after hysterectomy and radiation therapy (76). Uro- sion (8, 9). With MR imaging differentiation between a graphy, barium studies and cystography may reveal fistu- uterine invasion versus an obstructed uterus with hemato- las. Lower extremity edema secondary to compression of metria can be obtained (36). The differentiation between the iliac veins and radionecrosis of the bony pelvis, and benign versus malignant lymph nodes, however, cannot aseptic necrosis of the femoral heads may occasionally up to date be determined by MR examinations (37. 41, need diagnostic radiologic evaluation (13). 90). Lymphography has been the method of choice for de- Clinical problems of radiologic interest. To compare tecting extensive disease. The sensitivity of lymphogra- the results of therapy among different institutions. FIGO phy is, however, reported variable (28-83%). The speci- has limited the tools of clinical staging. ficity is found to be somewhat higher (47-100%) so only a a) In studies conducted examining surgically staged positive examination is of clinical significance. However, there appears to be a considerable inaccu- some non-malignant cases of nodal filling occur including racy in the FIGO system, which may influence the choice fatty infiltration, immunologic disorders, , fibro- of treatment (2). RADIOLOGY IN GYNECOLOGIC NEOPLASMS

Table 2 ac, presacral and para-aortic) and distant metustases oc- Clinical surgical staging for endometriul carcinoma cur relatively late in the disease, so that most lesions are detected before such spread has occurred. This is espe- Stage Extent of disease Subclassification cially true if the lesion historically is well differentiated. In eaily lesions the five year survival rate is high: about 80 I Confined to corpus a. Uterus sounds up to 8 cm per cent. Anaplastic tumors, on the other hand, are much of uterus b. Uterus sounds above 8 cm more aggressive, even in the early stages (3. 5. 45). II Extends to endocer- vix or cervix The classification (only endometrial adenocarcinomas: III Extends outside a. Adnexal mass not sarcomas) according to FIGO (74) is shown in Table 2. uterus, but confined b. Parametrial spread The staging is based on a clinical examination including to true pelvis palpation, inspection, fractional currettage. hysteroscopy. IV Extends outside pelvis cystoscopy. proctoscopy. and a thoracic radiogram. HSG and radiograms of the skeleton may also be included. Table 3 Findings from additional examinations are allowed to in- Surgic pathologic staging for oiurian carcinoma fluence the therapy but not the clinical staging. Diagnostic imaging. HSG was previously the method Stage Extent of disease Subclassification of choice in the radiologic diagnostic approach. If used today it is only for outlining the uterine cavity to facilitate I Growth limited to I. Tumor has smooth surface one ovary (a), or to 2. Malignant papillations intracavitary radiation treatment. The yield of lymphogra- both Ib)' on surface phy is small unless reserved for the more advanced bulky II Extra-ovarian a. Confined to gynecologic cancer or one extending into the cervix (47). growth, but confined organs The reports on transabdominal US have been contra- within pelvis b. Extragynecologic extension* dictory in distinguishing between early stages of the dis- HI Intra-abdominal spread ease, including possible deep myometrial invasion (20. IV Distant metastases 64). CT is found to be unreliable in determining the extent of the disease in the early stages, but is helpful in the * The additonal presence of malignant ascites is subclassification "c\ advanced stages of the disease and can accurately differ- entiate stages I and II from stages HI and IV by demon- b) After radiation therapy the parametrium often has a strating pelvic wall extension and extrapelvic spread (31. rubber-like consistency making palpation most unreliable. 85). c) The infrequent but significant complication with gy- The normal uterine cavity is well defined by MR scan- necologic fistulas. Their presence often gives diffuse ning. There are three definite zones within the uterus - the symptoms and conventional verification (e.g. barium en- inner high intensity zone, representing the endometrium. ema) may be most painful after irradiation and may not the junctional zone of low intensity and the reveal an accompanying mass. of medium intensity. Therefore, MR imaging is supposed d) Severe bleeding at surgery and to amplify chemo- to determine early myometrial invasion in patients with therapy. endometrial carcinoma, as well as an early parametrial extension (32. 39. 90). 2. Endometrial carcinoma Clinical diagnostic problems of radiologic interest, a) Carcinoma of the endometrium is the second most com- The diagnosis of deep myometrial invasion. Pre-operative mon malignancy of the female pelvis in Sweden (approxi- radiation therapy may be deleted or extensive adenec- mately 5r/r of all female cancers) (77). It is primarily a tomy at surgery may not be needed when there is no. or disease of postmenopausal women, with a peak incidence only superficial myometrial penetration (20). b) Severe in the decade from 55 to 65 years of age. The coincidence bleeding al surgery and to amplify chemotherapy. of obesity, hypertension, and diabetes in many patients with this disease is indicative of an underlying endocrine 3. Ovarian carcinoma disorder (49, 71). Because of the complex embryologic and histogenetic Benign cystic hyperplasia progressing to adenomatous development, the ovaries are a source of a greater variety hyperplasia. then anaplasia. and finally, nennlasia has of tumors, both benign and malignant, than any other been demonstrated as the preliminary sequenc; in a num- organ in the body. Ovarian tumors may be benign or ber of patients with endometrial carcinoma (45). The dis- malignant, solid or cystic, they may be of mixed types, ease probably has an in situ first stage, followed by an and they may produce hormones. Of great clinical signifi- invasion of the surrounding endometrial stroma before it cance is the fact that, whether benign or malignant, they involves the underlying myomctrium. Fortunately, deep are often clinically silen: until late in the course of devel- myometrial infiltration, an extension beyond the corpus of opment, unless there is ovarian torsion or rupture. The the uterus, local lymph node involvement (obturator, ili- most common malignant ovarian neoplasms in women are BJÖRN THORVINGER various forms of epithelial tumors IWf). The regional The role of MR in the diagnosis of patients with ovarian lymph nodes are the iliac, lateral sacral, para-aortic and carcinoma is yet to be determined. Currently, a differenti- inguinal nodes. is the leading cause of ation between benign and malignant disease cannot be gynecologic malignancy death (4. 15. 68. 75). Approxi- made by MR imaging, and the resolution and definition of mately 5.5 per cent of all new cases of cancer in Swedish most structures by MR are equivalent to those of CT and women are of ovarian origin (77). The staging system of US. An advantage for MR is. however, its potential to ovarian malignancy recommended by FIGO (70) is shown distinguish bowel loops from abnormal masses (16. 36). in Table 3. The staging is carried out by a physical exami- Another advantage is its ability to scan in any desirable nation, including pelvic inspection and palpation. Lapar- plane, especially in the pelvis. oscopy and/or surgical exploration with histologic typing Angiography may be used for the delineation of hepatic is mandatory. metastases. Most ovarian malignancies and their metasta- Diagnostic imaging. The conventional radiography of ses are relatively hypovascular. A combination of angiog- the abdomen may reveal soft tissue masses. The presence raphy and CT may. however, exquisitely deal with specif- of calcifications may indicate dermoid or cystadenocar- ic difficult problems. Hepatic arterial infusion and emboli- cinoma. The latter develops calcifications in the primary zation are well established and successful (47. 59. 86). tumor or in the metastases in approximately 10 to 12 per Clinical diagnostic problems of radiologic interest, a) cent. Pulmonary metastases and pleural effusion (often To evaluate therapy response: b) to reveal tumor recur- rightsided) can be detected on chest radiographs and may rence: c) to verify complete clinical remission without be used to evaluate response to chemotherapy. Metasta- repeated laparotomy: d) severe bleeding at debulking sur- ses to the small and large bowel may occur by direct gery and to amplify chemotherapy. extension, or serosal seeding, distributed by ascites. The transverse colon may be involved by contiguous spread 4. Other gynecologic neoplasms from the greater omentum (47. 75). Leiomyomas (fibroids) are the most common benign The vast majority of the radionuclide scans performed tumors of the female genital tract and are found in almost in the staging of ovarian malignancies (like in other gyne- one third of women during their reproductive years. The cologic tumors) have now become redundant (47). This leiomyomas are classified by their location and may be includes screening for liver metastases. The major pitfall serosal. intramural, or pedunculated. Pedunculated fi- of hepatic scintigraphy is its low specificity, as well as a broids may simulate an adnexal mass. Complications in- high false positive rate. Osseous metastases are rare and clude degeneration (especially in pregnancy), infection, radionuclide bone scanning should be reserved specifical- bleeding and rarely sarcomatous change, which may be ly for patients with symptoms from the skeleton (51). suggested by rapid growth (55. 75). Lymphography has no crucial role in the routine diag- On conventional radiographs, fibroids may appear as a nostic work-up in patients with ovarian cancer, but may pelvic soft tissue mass displacing the bladder and the be a complement if CT and US are negative (47). bowel. Irregular calcification may be seen, especially in Real-time US is a valuable tool in the hand of an experi- elderly women. In the diagnostic work-up of suspected enced operator for detecting a pelvic mass, presence of fibroids, US is often used, the accuracy of which is about ascites, obstructive uropathy. hepatic metastases. and tu- 90 per cent (25). The examination is of less value when mor implants (>2-3 cm) at the peritoneal surface (52. 54). tumors are small, when the uterus is retroverted or retro- Shortcomings of US is its inability to detect involvement displaced. Fibroids become more sonolucent in pregnancy of the bowel and bladder or bony metastases (6. 54. 56. and care should be taken not to misinterpreting the echo- 63). free fundus of a retrovertcd uterus (21). The CT appear- With the increased resolution of CT it has been reported ance usually shows a large lobulated uterus which may to be a most useful tool as a non-invasive imaging method contain areas of calcification (78). MR can provide a more to delineate the entire extent of tumor (70. 88). Since the accurate assessment of the number, size, and precise site vast majority of the patients, at the time of presentation, of leiomyomas (40). The optimal imaging sequence for the are either of stages III or IV, the greatest advantages of diagnosis of a leiomyoma is the T2 weighted image which CT is its ability to detect disease within the pelvis, as well yields the best contrast between the tumor and the sur- as extrapelvic disease, i.e. in the liver, retroperitoneum rounding tissue (27. 40). and omentum (70). There are. however, reports of low Uterine sarcomas are rare (3-5 r/r of all uterine malig- sensitivity in patients undergoing a second look laparoto- nancies). They may arise from leiomyomas, from the my. because residual microscopic tumor cannot be shown myometrium itself, or from the endometrial stroma. Car- by CT (7). In assessing the response to treatment. CT is cinosarcoma and mixed mcsodermal tumors containing preferable to US. since CT allows recognition of more both epithelial and connective tissue malignant cells arc disease sites without interference from various artifacts also encountered. All the sarcomas have a poor prognosis (88). In addition, the CT examination is more easily com- often with early spread via the blood stream, lymphatics pared with previous scans. and by contiguity (',2. 49). CT can demonstrate uterine RADIOKXiY IN uVNK'OLOGIC NEOPLASMS enlargement with inhomogeneity and ureas of low attenu- ID To ev:>!nate the usefulness of CT in patients with ation. Occasionally, calcifications may he present (47). fistulas following gynecological tumors or their treatment. Gestational trophohlastic disease includes hydatidi- HI) To evaluate the ability of transabdominai US and form mole, where the mole tissue is only in the endometri- MR in intra-uterine staging including myometrwl invasion al cavity: chorio-adenoma destruence. where the mole in patients with endometrial carcinoma. tissue invades the myometrium locally, and choriocarci- IV) To evaluate CT in the capacity of monitoring thera- noma. which may metastasize. Most choriocarcinomus py response, probable recurrence or clinical remission in follow a molar pregnancy, hut they may also he seen after patients with ovarian carcinoma. normal term pregnancy, an abortion or a tuhal pregnancy V) To evaluate the effect of intra-arterial occlusion in (47. 66). US is the best tool for evaluating a suspected facilitating surgery and in evaluating the role of intraarte- molar pregnancy, with its "grape-like" pattern (62). The rial infusion in gynecologic tumors other» ise refractory to appearance of invasive mole or is that of all given therapy. abnormal echoes within the myometrium. CT is the best method of detecting liver and brain metastases of chorio- Material and Methods carcinoma which occurs in about 10 per cent of the pa- Patients. Thirty-two patients with cervical carcinoma, tients (24). aged 30 to 74 (mean 51) years were evaluated by pelvic Vaginal carcinoma is a disease found in older women. It examination and CT (I). According to FIGO they were comprises slightly more than one per cent of all gyneco- initially staged IB 114 patients). HA (9 patients). IIB 14 logic cancers. A vast majority are squamous cell carcino- patients). HI A (I patient), and IIIB (4 patients). Thirty mas, while the remainder include clear cell carcinoma, patients had clinically suspected recurrence following pre- , and . Metastatic involve- vious treatment which included radiation, surgery, and ment of the vagina (from cervix or ) is more common chemotherapy in various combinations. The remaining 2 than primary . The disease spreads primar- women had symptoms and signs of possible advanced ily by local invasion into adjacent organs (bladder, rec- primary disease, but the clinical examination was consid- tum, paracolpial tissue), and to the pelvic lymph nodes. ered unreliable due to extreme obesity. The time interval Upper vaginal lesions spread similarity to cervical lesions between the clinical and CT examinations was at the most via the lymphatic drainage to the obturator and iliac nodes 18 days and in most cases less than a week. (60). A retrospective analysis of all CT examinations of pa- Cancer of the vulva is also a neoplastic form which rents with gynecologic malignancy over a four year peri- affects older women and accounts for approximately 4 per od was performed in order to reveal genital fistulas (II). cent of all cases. Over 90 per cent of invasive vulva Ten patients (aged between 40-74 years, mean 59) were carcinomas are of squamous cell type. The principal mo- found positive. Eight women had been treated for cervical dalities of spreading are lymphatic, usually to the inguinal carcinoma, one for vaginal, and one for endometrial carci- and femoral nodes (14). Since the tumors of the vagina noma, respectively. The FIGO staging was IIIA or more and vulva spread locally and by the lymphatic system. (cervical carcinoma), and II (vaginal and endometrial car- US, CT and lymphography are useful in determining the cinoma). nodal involvement and to assist in radiotherapy planning Twenty women with historically proven endometrial (47). carcinoma (aged 44-77 years: mean 64) were examined Carcinoma of the fallopian tube is the rarest of gyneco- with MR and 10 (aged 36-73 years: mean 58) with US (HI). logic malignancies. It imitates the histology and the nodal Quite accidentally, all 30 patients were clinically found to spread of ovarian cancers. Most forms are pure adenocar- be in either stage I (21 women) or stage II (9 women). cinomas. It extends to the omentum and the retroperito- Over a four year period 105 patients with historically neal nodes (72). The radiologic work-up in patients with proven ovarian malignancy were evaluated with one or is similar to that of ovarian malig- more abdomino-pelvic CT examinations (IV). The pa- nancy (47), and the diagnostic problems of radiologic tients. 10 to 78 years old (mean 55 years) were divided interest are also the same which also can be stated about into three groups depending on the purpose of the evaluat- vaginal or vulva cancer in relation to cervical cancer. ed CT Thus, 44 patients were evaluated for therapy Several studies in the past have dealt with some of the response (group A). 30 patients had suspected recurrence clinical problems of radiologic interest discussed above, (group B) and 31 patients for-verification of complete however, often with varying or even contradictory re- clinical remission (group C). The FIGO classification was: sults. stage I 24 patients: stage II 15 patients: stage III 55 The aims of the present investigation (I-V) were: patients, and stage IV II patients. Ninety-seven of the I) To evaluate CT in comparison with the clinical ex- patients had epithelial tumors. 3 patients had dysgermin- amination in patients with cervical carcinoma after radi- oma. 3 patients had malignant . one patient had a ation with possible recurrence in the parametrium and granulosa cell, and one a mixed mesodermal tumor, re- along the pelvic wall. spectively. 10 BJÖRN THORVINCER

Intra-urteriul therapy was performed in 17 women aged between 28 and 75 years (mean 55 years) (V). Four pa- tients had uterine cervical carcinoma. 10 ovarian carcino- ma, one. sarcoma of ihe uterine body, one synchronous ovarian and cervical carcinoma, and one ovarian mctasta- ses from a sigmoid carcinoma resected three years earlier. The indications for intra-arterial management were to fa- cilitate surgery in 13 of the patients and for palliation in the remaining 4 women refractory to conventional treat- ment.

Radiologic techniques employed CT it. II. V). All 147 patients were examined with a third generation scanner (Philips 300. 310. or 350). A slice thickness of 12 mm was used, with a slice interval of 18 or Fig. I. A central recurrent muss and rightsided iliac udenopathv 24 mm in the abdomen, and consecutively in the pelvis. (—») in a patient FIGO-staged IIA for cervical carcinoma. The patients were given 1000 ml of diluted meglumine diatrizoate solution orally. Twenty ml or more of 60 per cent meglumine metrizoate. or iohexol (240 mg I/ml) was given intravenously for urinary tract enhancement. Two patients with known vesicovaginal fistulas did not receive enteric contrast medium (II). In selected cases a tampon was introduced to outline the vaginal space (I. II). MR till) examinations were performed on a 0.3T Fonar /J-3000 M-system with a vertical magnetic field. Solenoid surface coil wrapped around the pelvis was used in 14 patients while 6 patients had a body coil. Multislice im- ages were obtained in the sagittal planes using long TR/TE (2000/60 msec) and short TR/TE (400/16 msec) spin echo sequences. In the transverse plane images were obtained using a long TR/TE (2000/60 msec) spin echo sequence. The slice thickness/interval was 5.0/7.1 mm and the pixel Fig. 2. Extensive presacral adenopathy in a patient uith a IB size was I mm in all patients (a matrix of 256x256 pixels). cervical carcinoma. The images of one patient were also obtained in the cor- onal plane. US till) was performed with a real-time scanner occlusion therapy added during the lust setting. Gelfoam equipped with a 3.5 or 5.0 MHz sector transducer. The and/or the occluding "spring embolus' coil system (Cook) examinations were carried out traiisabdominally with dis- was used for occlusion. In addition, one patient had a left tended urinary bladder. In 3 cases a urinary catheter was ovarian artery occluded with 2 ml of ethanol after selec- inserted to facilitate filling of the urinary bladder with tive intra-arterial and in another patient five sterile solution of sodium chloride. The examinations lumbar arteries were embolized. The catheters used had were performed and read by experienced staff members. an OD/ID of 2.2/1.4 mm. HSG Illli. The instruments used were a two catheter Altogether 7 patients had both chemotherapy ;ind oc- suction cup device according to Furkenscher & Sem. or clusion therapy. 8 patients received only occlusion thera- Shulze. with injection of 10 to 20 ml of meglumine metri- py and 2 only chemotherapy. zoate (200 mg I/ml) under fluoroscopic control. Antero- posterior. oblique, and lateral projections were recorded on a 100 mm film system. Intra-arterial therapy IV). After an initial lumbar and Results pelvic angiogram, the infusion and occlusion therapy were CT in possible recurrent cervical carcinoma (I). All performed from the femoral artery bilaterally in the inter- patients were followed up with repeat clinical pelvic ex- nal iliac artery or in its anterior division, as distal as aminations. Furthermore, a follow-up by surgery was car- possible. The chemotherapeutic agent used was 10 mg of ried out in 9 patients, by biopsy in 11. autopsy in 13. and mitomycin-C in 20 ml sterile water, and 100 ml isotonic conventional radiography in 26 patients. Ten patients had saline with an infusion rate of 6 to 8 ml/min up to a total of further CT examinations in order to confirm normal find- 60 ml. The procedure was repeated twice fortnightly with ings or to check the treatment. Ttie average follow-up RADIOLOGY IN liYNKCOI.OGIC NEOPLASMS 11

All patients, except ? had been treated with irradiation for a primarily extensive cervical carcinoma (sluge 1IA or more). Five of these patients had experienced additional surgery. The remaining patients had been treated with radiation and surgery for a vaginal and an endometrial carcinoma, respectively. All patients, except one. had pain as the main symptom ;snd only 5 of the 10 patients had clinical signs of a fistula (Fig. 3). CT depicted admin- istrated contrast material in the vaginal space in 8 patients and 2 of these also had contrast medium in the uterine cavity. Two patients had air in the bladder and the re- maining patient had air in the pubic area. Three patients had recurrent disease. Staging in early endometrial carcinoma (III). The MR and US staging of endometrial carcinoma were compared with sounding (length of the uterine cavity), hysteroscopy Itumor extension into the cervix), histopathological ex- amination (myometrial invasion). The HSG examinations Fig. 3. A vesico-vaginal fistula (—»1 in u 65-year old patient after inlracavitary radium treatment and total hysterectomy tor an of 28 of the patients were evaluated in the same way endomeiriai carcinoma stage II. except for myometrial invasion. The concoidance of a possible differentiation between stage IA and IB (measuring the length of the uterine Table 4 cavity from the dome to the cervical os) was between MR The results of CT and clinical examination in possible recurrent and clinical staging (sounding) 90 per cent, between HSG cervical carcinoma tn=32) and sounding 85 per cent, and between US and sounding 50 per cent. The concordance of differentiating stage I CT Clinical from stage II (cervical tunu r extension) was between MR examination and hysteroscopy 85 per cent, between HSG and hyster- oscopy 82 per cent, and between US and hysteroscopy 50 91 r/ 78r; Accuracy r per cent. Sensitivity and specificity of outlining cervical Sensitivity n.v; Specificity UW'r 67^; tumor extension are presented in Table 5. MR imaging and US suggested 6 and 4 patients, respec- tively, to have deep myornetrial tumor invasion (more time was approximately 15 months. All the information than one third of the myometrial wall), and all except for obtained at the follow-up was then compared with the CT one in each group were confirmed at pathologic anatomic and clinical examination at the time for (he suggested examination (Fig. 4). There were no false negative US or recurrence. MR examinations. Hematopyometra was suggested at MR CT depicted 17 out of 20 patients with advanced disease examination in 5 of the patients, but it was not possible to (growth beyond the cervix, upper vagina and the medial verify clinically. parametria) (Figs I. 2). CT ofmalifinuiit ovarian disease IIV). The CT findings There were 3 false negative examinations in which later were compared with the results of the follow-up proce- lymph node metastases were confirmed. No false positive dures. Laparotomy was performed in 85 patients within CT were obtained. At CT all 12 patients were correctly three months after CT. Histologic verification, using fine diagnosed with either none or only local (in connection needle biopsy was made in the remaining 20 patients. All with the vaginal cuff) disease. The clinical examination at patients except for the 3 youngest had repeated rectova- the time of the CT examination had. according to the ginal bimanual pelvic examinations. A radiologic follow- follow-up. 7 false results. The concordance between CT up included 37 further CT and 9 US examinations. The and clinical pelvic examinations was 68 per cent and all in follow-up time was from nine to 49 months, the average conformity with the follow-up results. Accuracy, sensitiv- being 29 months. ity, and specificity compared with the follow-up are In the 44 patients (group A) who were evaluated for shown in Table 4. therapy response. CT provided further information in ad- Genital fistulas III). Verification of the fistulas suggest- dition to the clinical examination of 24 women. Such ed at CT was made by cystoscopy in 5 patients, clinical additional findings were progression, regression, or new pelvic examination (3 patients), and barium enema in one tumor sites (Fig. 5). In the remaining 20 patients CT gave patient. In one patient no verification was possible (a no additional information but was never contradictory to vulvo-pubical fistula). follow-up. 12 BJÖRN IIIOKV1N(,IK

a b c Kig. 4. A 51-year old woman with a FKiO-staged IA endometrial weighted sagittal MR image ihl the uterine hody has sharp outer carcinoma. HSCi in lateral view la) reveals tumor growth of the margins. The T2 weighted image (el re\eals deep myometrial anterior wall of the uterine hods not including the cervix. On a II growth.

Table 5

S''n!>:liiit\. \pc( if'u it v it nil in < urin \ in tlt'tt'iniiimtion of < crvit til In nun t'xtttnicn < I'tnptUttt with li\ N

IN f!» IV.tal Sensi- Speci- Accuracv tivity ficity i' 11 I''' I

use; 17 2M Mi N2 MR 14 2(1 Ml IS 4 Mi

In the group of 30 patients with clinically suspected clinially. to be in complete remission after 15 to 31 months recurrence (group B). CT depicted 14 patients with recur- of follow-up, while 3 patients were still found to be non- rence which was eorr-.-ct according to the follow-up (I'ig. resectable. The remaining woman developed a rectova- 6). At C'Y there were, however. 5 false negative examina- ginal fistula which disqualified her for immediate surgery. tions giving a sensitivity of 74 per cent. In the clinical remission group of 31 patients (group C). CT suggested 27 patients to be in true remission of which 7 were found to Summary of Results be false at the follow-up, giving a sensitivity of 36 per It CT was more accurate (91'/) than clinical pelvic cent. Of those 12 patients with active disease not ob- examination (78'r) in revealing extensive disease after served on CT. surgical exploration revealed nodular dis- radiation and/or surgical treatment. ease in the abdominal cavity or pelvis. Sixty-one of 105 II) CT was a most valuable tool in demonstrating genital baseline CT examinations were positive lor malignant fistulas following gynecologic malignancy or its treatmciit. disease. The results of groups B and C are listed in Tables the clinical symptoms were often unspeeifie and possible ft and 7. respectively. Specific CT findings are summa- recurrence might be revealed. rized in fable 8. III) Transabdominal US did not refine staging in early Trwiscatlwter arterial mantifienwnt (Vi. The evaluati.): endometrial carcinoma (stages I and M). VIK. however, of transcatheter intra-arteiial management was made us- had a potential for delineating inira-uterine tumor growth ing surgery in 9 patients and clinical follow-up in S pa- (accuracy of demonstrating myomelrial invasion: 95'') in tien's. addition to other information such as hematopyometra. In the palliation group of 4 patients Ihe result was poor IV) CT was most valuable in the evaluation of therapeu- but 2 patients had some transitory tumor reduction. In the tic response of patients with ovarian malignancy. In pa- presurgical group of 13 patients who previously had been tients with possible recurrence (sensitivity: 74'') or in found non-resectable clinically (2 patients) or at laparoto- clinical remission (sensitivity: }h'i). However, only a my (II patients). 9 patients had surgery without major positive C'T was of clinical signficance. bleeding complications and 6 of these women were found. V) The potentials of transcatheter intra-arteiial manage- KADIOI.OUY IN (iYNKCOKKIU NEOPLASMS 13

Fig. V Omental infiltration I—•) in a 58-year old woman with an Fig. ft. A tumor implant (—») at the diaphragmatic dome in a endometrioid ovarian carcinoma verified at laparotomy. 53-year old woman with an undifferentiated ovarian carcinoma.

Table 6 spin-off of routine work-up with CT in patients treated Clinically suspected recurrence of aid rinn carcinoma in JO !>

Table 8 sity on T2 weighted images between late fibrosis and Specific CT findings in 105 patients with ovarian malignancies recurrent tumors. MR may therefore be useful in differen- tiating a neoplasm from posttreatment fibrosis (17). CT findings No. of patients Important prognostic factors of endometriul caninoma are the uterine size, cervical tumor extension, histologic Abdomino-pelvic mass 40 grading, and myometrial tumor invasion. The staging pro- Ascites 28 cedures are most reliable in determining all but myome- Adenopathy II Liver metastases 10 trial invasion. There is a very '.lose relationship between Hydronephrosis 9 lymph node metastases and thickness of the intact uterine Peritoneal implants 8 wall. If the intact uterine wall is more than 10 mm thick, Omental implants 8 lymph node metastases seldom are present (35). The mini- Abominal wall mass 6 mal thickness of the intact uterine wall is closely related Spleen metastases 2 Kidney metastases 2 to the depth of the myometrial invasion. Our study (HI) Bone metastases or invasion 2 indicates that MR is most reliable (accuracy 95%) in revealing deep myometrial infiltration with no false nega- tive MR examinations. A shortcoming of our investigation major problem, while the single indicator for lymph node was, however, that all patients were treated with radiation spread is the size of the nodes. Most authors regard lymph before surgery, so the only verification of myometrial nodes being more than 1.5 cm as enlarged, but WALSH et infiltration was scarring at the anatomic-pathologic exami- al. (83) used 2 cm as the minimum size of adenopathy. In nation. The results are, however, in agreement with other the lumbar area, the lymph nodes are cut at right angles, investigations (39). Leiomyoma was suggested at MR ex- but are often measured at oblique angles in the pelvis amination in 3 out of 5 patients who were positive at the making the CT demonstration of size often doubtful (48). pathologic examination. The remaining 2 patients with Another disturbing factor is that gynecologic tumors, es- verified leiomyomas had tumors with a diameter of about pecially cervical carcinomas are often accompanied by 5 mm and these were not observed even at a review of the , leading to enlarged reactive nodes. In addition, MR images. small nodes may contain cancer cells, and large nodes A drawback of MR is the rather high rate of anxiety or may be microscopically normal (44). Ii> spite of the dis- even claustrophobia because of the dimensions of the cussed difficulties of CT in demonstrating metnsiatic gantry, examination duration, coil noise and temperature lymph nodes the accuracy of CT is reported to be satisfac- within the bore (61). Seven patients were excluded from tory (74-83 per cent) (84, 89). the study when confronting the MR equipment, mainly As indicated (I), CT may be most helpful in the clinical because of obesity and/or claustrophobia. work-up of patients after radiation therapy. In such cases US was also found reliable in revealing deep myome- palpation may be unreliable because of parametrial fibro- trial infiltration (accuracy 90%), but it was recorded in a sis. Radiation fibrosis has been reported to be a main too small material. The recent study of FLEISCHER et al. cause of misinterpretation at CT (82), but this was not the (20) supports our data. The outlining of the cervical exten- case in our investigation. In concordance with other se- sion was poor using US and 2 out of 3 patients with ries, our data indicate that hydronephrosis is seldom due leiomyoma were missed. Other recent studies using trans- to radiation fibrosis. but often to tumor recurrence (13, vaginal US scanning, however, report even better tissue 47). In addition, complications such as fistulas with or characterization of the uterus, as well as of the extra-uter- without a recurrent mass may be revealed (II). ine areas (50). Using this approach there is no need for the Recent studies clearly indicate that MR imaging may be mandatory full bladder which is often discomforting the the most useful tool to assess the clinical staging of cervi- patient, and the technical difficulties due to adiposity are cal carcinoma (38, 79). MR is capable of demonstrating not unimportant. A study by HIRAI et al. (35) indicates the actual tumor mass in multiple planes. Cervical inva- that angio-CT after contrast medium injected into the sion, and tumor extension to the uterine corpus are accu- internal iliac arteries is also a most reliable tool in deter- rately documented. MR is reported to have a high accura- mining the minimal thickness of the intact uterine wall. cy (88-89%) in determining parametrial disease, which is The surgical staging of ovarian malignancy often out- a valuable contribution to clinical staging because lesions rules the need for additional assessment by CT or US in can be safely resected or precisely identified. Tumor ex- the staging procedure. The results of the present study tension to the bladder and rectum is also reliably ex- (IV) indicate, however, that CT has a major role in the cluded, which may eliminate the need for cystoscopy or follow-up work of ovarian malignancy in order to limit the sigmoidoscopy in a majority of patients. There are, how- numbers of repeated laparotomies. It is evident that the ever, no crucial evidence that MR is more accurate than major drawback of CT is that intraperitoneal and hepatic CT in imaging abnormal lymph nodes (37, 38). A recent metastases are not observed, especially if the nodules are study has found a significant difference in the signal inten- less than I cm (7, 12). The nodules are seeded in the RADIOLOGY IN GYNECOLOGIC NEOPLASMS 15

dnical staging dnical staging 1 1 MR, chest radograph MR (or US. angio-CT), i chest radograph Therapy (surgical and/or radatton, I chemotherapy, hterventional management) Therapy (radatton, chemotherapy, I interventional management surgery) BaseineCT / / \ advanced stages (suspected) dntaal no clrtcal recurrence OHV) recurrence or symptoms baseineCT / \ I * negative CT positive CT dnicaly suspected recurrence i I or symptoms LAG biopsy I Fig. 7. Suggested protocol of cervical carcinoma. XT Fig. 8. Suggested protocol for endometrial carcinoma. intraperitoneal cavity by fluid from the pouch of Douglas up to the liver surface and the diaphragmatic domes (52). Another weakness of CT is detecting lymph node involve- iliac artery embolization is permanent neurologic damage, ment as in other types of gynecologic malignancy. The for instance drop-foot, anesthesia, and the Brown-Sé- internal structure of the lymph nodes is not available for quard syndrome (28). It should be clearly emphasized that CT analysis, making diagnosis entirely a matter of size. Dionosil (Glaxo, United Kingdom) and gelfoam powder, Several studies have found the sensitivity of CT being low although very effective in the pre-operative occlusion of in monitoring lymph node metastases compared with sur- renal tumors (18), should not be used in the internal iliac gical-pathologic evaluation (7, 12). Thus, there are indica- artery as they may produce an occlusion very close to the tions that 80 per cent of metastastic lymph nodes are less nerve tissue and block the collateral supply. If radiation or equal to I cm in diameter (12). As shown, CT is of high therapy has been given previously, there is probably a value in detecting small amounts of ascites, and this could decreased potential for developing an anastomosis. Be- be the only finding even if there is a co-existing extensive fore embolization of a lumbar artery feeding a tumor, the nodular disease. This finding has recenty been supported large radicular artery (artery of Adamkiewics) must be by the study of BUY et al. (10). It is, however, obvious that identified. Another source of complication is the inferior a negative CT (as lymphography) is not reliable enough to mesenteric artery. This vessel may be obliterated, espe- replace a laparotomy. cially in patients with arteriosclerosis and in cases where Contrary to cervical and endometrial carcinoma, no the distal colon may depend on the collateral supply from study has so far shown a crucial role for MR in patients the internal iliac arteries. Severe bowel ischemia has been with ovarian tumors. It may, however, provide additional reported after embolization in such cases (19). information about tissue characteristics and can therefore Recent studies confirm that intra-arterial management be used as a problem solving modality after US in the is certainly here to s.ay and should always be considered study of adnexal abnormalities (53). at the very onset of all types of gynecologic malignancy In addition to tumor detection the diagnostic radiologist (29, 57, 65, 92). has become increasingly involved in the therapy of pa- tients with gynecologic neoplasms. As shown (V) the results are more than promising in providing resectability Conclusions with intra-arterial transcatheter procedures. Although we Diagnostic imaging approach to a cervical carcinoma. did not see any major complications, such are frequently Clinical staging performed by an experienced examiner is reported including urinary bladder necrosis, skin necrosis, overall superior to CT, because of the high incidence of muscle necrosis, and reflux of embolic material (23, 34). early stages of the disease, although CT is more accurate Perhaps the most significant complication from internal in demonstrating the advanced stages and lymph node 16 BJÖRN THORVINGER

US. chest radiograph IIA there was 44 per cent error. In stage IIB the figure \ was 55 per cent and in stage III disease the clinical staging CT ( — mterventional management) was incorrect in 100 per cent of the cases. Furthermore, it is almost impossible to make a clinical diagnosis of deep surgical staging (with possible debuWng) myometrial infiltration in patients with endometrial carci- noma. This calls for the use of more accurate methods of therapy (radation. chemotherapy. evaluating tumor extension than the clinical FIGO stag- ing. These methods to refine clinical diagnosis must, hterventionaJ management) I therefore, basically rely on radiologic techniques. CT (therapy response) I ACKNOWLEDGEMENTS CT (dnical remission)' I wish to express my sincere gratitude for excellent guidance and support to: dnicaly suspected Professor Erik Boijsen, Professor Holger Pettersson. Doctor recurrence Lars Samuelsson; Miss Eva Jönsson for secretarial assistance: Mr Göran Eliasson for photographic illustrations; technicians and \ other friends at the Department of Diagnostic Radiology. CT

negative CT positive CT REFERENCES 1. AMERICAN CANCER SOCIETY. Cancer. Facts and figures.1980 . i i The American Cancer Society. New York 1980. laparotomy biopsy 2. AVERETTE H. E., DUDAN R. C. and FORD J. H JR: Explora- tory celiotomy for surgical staging of cervical cancer. Amcr. Fig. 9. Suggested protocol for ovarian malignancy. J. Obstet. Gynecol. 113 (1972), 1090. 3. BERMAN M. L.. BALLON S. C. LAGASSE L. K. and WATRINC. W. G.: Prognosis and treatment of . metastasis (26, 81, 84). There is, however, a discrepancy Amer. J. Obstet. Gynecol. 136 (1980). 679. between FIGO staging and surgical exploration in more 4. BLAUSTEIN A.: Metastatic carcinoma in (he ovary. In: Patho- than a third of the cases (2). MR is reported to have an logy of the female genital tract, p. 705. Edited by A. Blaus- accuracy of almost 90 per cent in demonstrating invasive tein. Springer-Verlag. New York 1982. disease compared with surgical staging (79). The sensitiv- 5. BORONOW R. C, MORROW C. P., CREASMAN W. T. e! al.: Surgical staging in endometrial cancer. Clinical pathologic ity of CT in detecting a recurrent tumor is 85 per cent (I) findings of a prospective study. Obstel. Gynecol. 63 (1984). or more (82), and perhaps even higher when there is a 825. baseline posttherapy CT to decrease the rate of false 6. BOWIE J. D.: Ultrasound of gynecologic pelvic masses. The positive or equivocal CT studies (47). A suggested proto- indefinite uterus and other patterns associated with diagnos- tic error. J. Clin. Ultrasound 5 (1977), 323. col for cervical carcinoma is shown in Fig. 7. 7. BRENNER D. E.. SHAFF M. I.. JONES H. W. et al.: Abdotnino- Diagnostic radiologic approach to an endometrial car- pelvic computed tomography. Evaluation in patients under- cinoma. The only variable of prognostic significance, going second-look laparotomy for ovarian carcinoma. Ob- which cannot be determined at FIGO staging of endome- stet. Gynecol. 65(1985), 715. trial carcinoma is the depth of myometrial penetration. 8. BRYAN J. P., BUTLER H. E., LIPUMA J. P. et al.: NMR scanning of the pelvis. Initial experience with a 0.3T system. MR, US (especially transvaginal) and angio-CT, are all Amer. J. Roentgenol. 141 (1983), 1111. found capable of depicting myometrial invasion. A sug- 9. BUTLER H. E.. BRYAN P. J., LIPUMA J. P et al.; Magnetic gested protocol is given in Fig. 8. resonance imaging of the abdominal female pelvis. Amer. J. Diagnostic radiologic approach to an ovarian tumor. Roentgenol. 143(1984), 1259. 10. BUY J. N., MOSS A. A.. GHOSSAIN M. A. et al.: Peritoneal The patient should be evaluated initially with US of the implants from ovarian tumors. CT findings. Radiology 169 pelvis, liver, peritoneal cavity, retroperitoneum, and the (1988), 691. kidneys. In the case of widespread disease CT examina- 11. CHUANC V. P. and WALLACE S.: Arterial infusion and occlu- tion of the entire abdomen and pelvis should be obtained. sion in cancer patients. Semin. Roentgenol. XVI, I (1981). Figure 9 shows a suggested protocol for ovarian malignan- 13. 12. CLARKE-PEARSON D. L., BANDY C. L., DUDZINSKI M. et al.: cy. Computed tomography in evaluation of patients with ovarian It is obvious that clinical staging is difficult and most carcinoma in complete clinical remission. JAMA 255 (1986). inaccurate, compared with surgery in the advanced stages 627. of the disease, even when performed by the most experi- 13. CUNNINGHAM J. J.. FUKS Z. Y. and CASTEU.INO R. A.: Radio- graphic manifestations of carcinoma of the cervix and com- enced examiner. In the study of cervical carcinoma by plications of its treatment. Radiol. Clin. N. Amer. 12(1974), AVERETTE et al. (2). 40 patients defined clinically as stage 93. IB cases, the error frequency was 20 per cent. In stage 14. CURRY S. L., WMARTON J. T. and RUTLF.DCE F.: Positive RADIOLOGY IN GYNECOLOGIC NEOPLASMS 17

lymph nodes in v ulvar squamous carcinoma. Gynecol. On- resonance imaging of the female pelvis. Initial experience. col. 9(1980), 63. Amer. J. Roentgenol. 141 (1983). 1119. 15. CZEKNOBILSKV B.: Primary epithelial tumors of the ovary. In 38. — LACEY C. G., SANDLES L. G. et a!.: Invasive cervical of the female genital tract, p. 511. Edited by A. carcinoma. Comparison of MR imaging and surgical findings. Blaustein. Springer-Verlag. New York 1982. Radiology 166 (1988). 623. 16. DOOMSG. C, HRICAK J. and TSCHOLAKOFF D.: Adnexal struc- 39. — STERN J. L.. FISHER M. R. et al.: Endometrial carcinoma tures. MR imaging. Radiology 158 (1986). 639. staging by MR imaging. Radiology 162 (1987). 297. 17. EBNER F., KRESSEL H. Y.. MINTZ M. C. et al.: Tumor recur- 40. — TSCHOLAKOFF D.. HEINRICHS L. et al.: Uterine leiomyo- rence versus fibrosis in the female pelvis. Differentiation mas. Correlation of MR. histopathologic findings, and symp- with MR imaging at 1.5T. Radiology 166 (1988). 333. toms. Radiology 158 11986). 385. 18. EKELUND L... EK A., FORSBERG L. et al.: Occlusion of renal 41. JOHNSON I. R.. SYMONDS E. M., WORTHINGTON D. M. et al.: arterial tumor supply with absolute ethanol. Acta Radiol. Imaging ovarian tumors by nuclear magnetic resonance. Brit. Diagnosis 25 (1984) 195. J. Obstet. Gynaecol. 91 (1984). 260. 19. FAYSAL M.: Angiographic management of post-prostatec- 42. KESTNER R. W.: Gynecology. Principles and practice. 3rd tomy bleeding. J. Urol. 122 (1979). 129. edition, p. 255. Year Book Med. Publ., Chicago 1979. 20. FLEISCHER A. C. DUDLEY B. S., ENTMAN S. S. et al.: Myome- 43. KILCHESKI T.. ARGER P., MULHERN C. et al.: Role of comput- trial invasion by endometrial carcinoma. Sonographic assess- ed tomography in the presurgical evaluation of carcinoma of ment. Radiology 162 (1987). 307. the cervix. J. Comput. Assist. Tomogr. 5 (1981), 378. 21. — JAMES A. E. JR. Minis J. B. and JULIAN C: Differential 44. KOLBENSTVEDT A.: Lymphography in the diagnosis of metas- diagnosis of pelvic masses by gray scale sonography. Amer. tases from carcinoma of the uterine cervix stages I and II. J. Roentgenol. 131 (1978), 469. Acta Radiol. Diagnosis 16 (1975), 81. 22. FRIGOLETTO F. (chairman): Diagnostic ultrasound imaging in 45. Koss L. G., SCHREIBF.R K., OBERLANDF.R S. G., MOUSSOURIS pregnancy. Consensus development conference. NIH Publi- H. F. and LESSER M.: Detection of endometrial carcinoma cation No. 84-667 (1984), 148. and hyperplasia in asymptomatic women. Obstet. Gynecol. 23. GUILIANO L., CARMAONANI G., BELGRANO E. and PUPPO P.: 64(1984), 1. Gelatin foam and isobutyl-2-cyanoacrylate in the treatment 46. LECART C: Cervical cancer. Epidemiology, histopathology. of life-threatening bladder hemorrhage by selective trans- diagnosis, j. Belg. Radiol. 71 (1988). 383. catheter embolization of the internal iliac arteries. Brit. J. 47. LEWIS E.: The use and abuse of imaging in gynecologic Urol. 51(1979). 125. cancer. Cancer 60 (1987), 1993. 24. GOLDSTEIN D. P. and DERKOWITZ R. S.: The management of 48. MAGNUSSON A.: Size of normal retroperitoneal lymph nodes. gestational trophoblastic neoplasms. Cur. Probl. Obstet. Acta Radiol. Diagnosis 24 (1983). 315. Gynecol. 4(1980). 1. 49. MARROW C. and TOWNSEND D. E.: Synopsis of gynecologic 25. GROSS B. H.. SILVER T. M. and JAFFE M. H.: Sonographic oncology. John Wiley and Sons, New York 1981. features of uterine leiomyomas. J. Ultrasound Med. 2 (1983). 50. MENDELSON E. B., BOHM-VELEZ M., JOSEPH N. and NEIMAN 401. H. L.: Gynecologic imaging. Comparison of transabdominal 26. GRUMBINE F. C, ROSENSHEIN N. B., ZERHOUNI E. A. and and transvaginal sonography. Radiology 166 (1988), 321. SIEGELMAN S. S.: Abdominopelvic computed tomography in 51. METTI.ER F. A., CHRISTIE J. H., CROW N. F. et al.: Radionu- the preoperative evaluation of early cervical cancer. Gyne- clide bone scan, radiographic bone survey, and alkaline col. Oncol. 12(1981). 286. phosphatase. Cancer 50 (1982), 1483. 27. HAMLIN D. J., PETTERSSON H.. FITZSIMMONS J. and MORGAN 52. MEYERS M. A.: Distribution of intraabdominal malignant L. S.: MR imaging of uterine leiomyomas and their complica- seeding. Dependency and dynamics of flow of ascitic fluid. tions. J. Comput. Assist. Tomogr. 9 (1985), 902. Amer. J. Roentgenol. 119(1973). 198. 28. HARE W. S. C, LOND F. R. C. R. and HOLLAND C. J.: 53. MITCHELL D. G., MINTZ M. C, SPRITZER C. E. etal.: Adnexal Pareses following internal iliac artery embolization. Radiolo- masses. MR imaging observations at I.5T with US and CT gy 146(1983), 47. correlation. Radiology 162 (1987), 319. 29. HARIMA Y.. SHIRAISHI T., HARIMA K. et al.: Transcatheter 54. MOYLE J. W., ROCHESTER D., SIDER L., SHROCK K. and arterial embolization therapy in cases of reurrent and ad- KRAUSE P.: Sonography of ovarian tumors. Predictability of vanced gynecologic cancer. Cancer 63 (1989), 2077. tumor type. Amer. J. Roentgenol. It. (1983). 985. 30. HARRIS R. W., BRINTON L. A., COWDELL R. H. et al.: Charac- 55. NOVAK E. R. and WOODRUFF J. D.: Novak's gynecologic and teristics of women with dysplasia or carcinoma in situ of the obstetrical pathology. W. B. Saunders, Philadelphia 1979. cervix uteri. Brit. J. Cancer 42 (1980). 359. 56. PALING M. R. and SHAWKER T. H.: Abdominal ultrasound in 31. HASUMI K., MATSUZAWA M.. CHEN H. F. et al.: Computed advanced ovarian carcinoma. J. ("lin. Ultrasound 9 (1981). tomography in the evaluation and treatment of endometrial 435. carcinoma. Cancer 50 (1982), 904. 57. Pisco J. M., MARTINS J. M. and CORREIA M. G.: Internal iliac 32. HEIKEN J. P. and LEE J. K. T.: MR imaging of the pelvis. artery. Embolization to control hemorrhage from pelvic neo- Radiology 166(1988), II. plasms. Radiology 172 (1989). 337. 33. HENRIKSEN E.: The lymphatic spread of carcinoma of the 58. PIVER S.. WALLACF. S. and CASTRO J.: The accuracy of lymph- cervix and the body of the uterus. A study of 420 necropsies. angiography in carcinoma of the uterine cervix. Amer. J. Amer. J. Obstet. Gynecol. 58 (1949), 924. Roentgenol. Ill (1971), 278. 34. HIETALA S.-O.: Urinary bladder necrosis following selective 59. PRANDO A., WALLACE S., BERNARDINO M. E. et al,: Comput- embolization of the internal iliac artery. Acta Radiol. Diagno- ed tomographic arteriography of the liver. Radiology 130 sis 19(1978), 316. (1979). 697. 35. HIRAI Y., KAKU S., TESHIMA H. et al.: Use of angiocomputed 60. PRIDE G. L., SCHULTZ A. E., CHUPRRVICH T. W. et al.: tomography to evaluate extent of endometrial carcinoma, Primary invasive squamous carcinoma of the vagina. Obstet. Gynecol. Oncol. 33(1989), 372. Gynecol. 53(1979), 218. 36. HRICAK H.: MRI of the female pelvis. A review. Amer. J. 61. QUIRK M. E., LETENDRF. A. J., CIOTTONE R. A. and LINOI.EY J. Roentgenol. 146(1986). 1115. F.: Anxiety in patients undergoing MR imaging. Radiology 37. — ALPERS C, CROOKS L, E. and SHELDON P. E.: Magnetic 170(1989), 463. 18 BJÖRN THORV1NGER

62. REID M. H.. MCGAHAN J. P. and Oi. R.: Sonographic evalua- features of uterine myoma. J. Comput. Assist. Tomogr. 5 tion of hydatidiform mole and its look alikes. Amer. J. Roent- (1981). 866. genol. 140(1983). 307. 79. TOGASHI K., NISHIMLRA K.. SAGOH T. et al.: Carcinoma of 63. REQUARD C. K.. METTLER F. A. JR and WICKS J. D.: Preoper- the cervix. Staging with MR imaging. Radiology 171 (1989). atiVe sonography of malignant ovanan neoplasms. Amer. J. 245. Roentgenol. 137 (1981). 79. 80. VIAMONTE M , AI.TMAN O.. PARKS R. et al.: Radiographic- 64. — WICKS J. D. and METTLER F. A.: Ultrasonography in the pathologic correlation in the interpretation of lymphangioa- staging of endometrial adenocarcinoma. Radiology 140 denograms. Radiology 80 (1963). 903. (1981). 781. 81. VICK C. W.. WALSH J. W.. WHEELOCK J. B. and BREWER W. 65. RETTENMAIER M. A.. MORÄN M. F., RAMSINOHANI N. F. et al.: H.: CT of the normal and abnormal parametria in cervical Treatment of advanced and recurrent squamous carcinoma cancer. Amer. J. Roentgenol. 143 (1984). 597. of the uterine cervix with constant intraarterial infusion of 82. WALSH J. W.. AMENDOLA M. A.. HALL D. J. et al.: Recurrent . Cancer 61 (1988). 1301. carcinoma of the cervix. CT diagnosis. Amer. J. Roentgenol. 66. REYNOLDS S. R. M.: Hydatidiform mole. A vascular congeni- 136(1981). 117. tal anomaly. Obstet. Oynecol. 47 (1976). 224. 83. KONERDING K. F. et al.: Computed tomographic detec- 67. ROTKIN I. and KINO R. W.: Environmental variables related tion of pelvis and inguinal lymph node metastases from pri- to cervical cancer. Amer. J. Obstet. Gynecol. 83 (1962). 720. mary and recurrent pelvic malignant disease. Radiology 137 68. RUTLEDGE F. N., FLETCHER G. H., SMITH J. P. et al.: Gyneco- (1980), 157. logic cancer. In: Cancer patient care. p. 263. Edited by R. L. 84. — and GOPLERUD D.: Prospective comparison between clip Clark and C. D. Hane. Year Book Med. Publ.. Chicago 1976. cal and CT staging in primary cervical carcinoma. Amer. > 69. SANDERS R. C. and JAMES A. E.: The principles and practice Roentgenol. 137 (1981), 997. of ultrasonography in obstetrics and gynecology. Appleton- 85. Computed tomography of primary persistent and recur- Century-Crofts. New York 1980. rent endometrial malignancy. Amer. J. Roentgenol. 139 70. — MCNEIL B. J.. FINBERO H. J. et al.: A prospective study of (1982), 1149. computed tomography and ultrasound in the detection and 86. WALLACE S. and CHUANG V. P.: The radiologic diagnosis and staging of pelvic masses. Radiology 146 (1983). 439. management of hepatic metastases. Radiologe 22 (1982), 56. 71. SCULLY R. E.: Cancer of the uterine corpus. Pathologic 87. — JING B. S.. ZORNOZA J. et al.: Is lymphangiography types. Int. J. Radiat. Oncol. Biol. Phys. 6 (1980). 361. whorthwhile? Current concepts in cancer. Updated cervix 72. SEDLIS A.: Carcinoma of the fallopian tube. Surg. Clin. N. cancer II. Stages IB and MB. Int. J. Radiat. Oncol. Biol. Amer. 58(1978), 121. Phys 5 (1979), 1873. 73. SOUNDERS R. D.: Biological hazards of NMR. Proceedings of 88. WHITI.EY N.. BRENNER D.. FRANCIS A. et al: Use of the an international symposium on nuclear magnetic resonance computed tomographic whole body scanner to stage and imaging, p. 65. Bowman Gray School of , Wake follow patients with advanced ovarian carcinoma. Invest. Forest University. Winston Salem. N. C. I98I. Radiol. 16(1981), 479. 74. STECKEI. R. J. and KACAN A.: Diagnosis and staging of can- 89. et al.: Computed tomographic evaluation of carcino- cer, p. 244. W. B. Saunders, Philadelphia 1976. ma of the cervix. Radiology 142 (1982). 439. 75. STEVENS G. M.: The female reproduction system. An atlas of 90. WORTHINGTON J. L.. BALFE D. M.. LEE J. K. T. et al: Uterine tumor radiology. Year Book Med. Publ.. Chicago 1971. neoplasms. MR imaging. Radiology 159(1986), 725. 76. STROCKBINE M. F., HANCOCK J. E. and FLETCHER G. H.: 91. WYNDER E. L. and CORNFIELD J.: A study of environmental Complications in 831 patients with factors in cancer of the cervix. Amer. J. Obstet. Gynecol. 68 of the intact uterine cervix treated with 3000 rads or more (1954). 1016. whole pelvis radiation. Amer. J. Roentgenol. 108 (1970). 293. 92. YAMASHITA H. Y.. TAKAHASHI M., BUSSAKA H. et al.: Bal- 77. SWEDISH NATIONAL CANCER REGISTRY. The Swedish National loon-occluded arterial infusion therapy in the treatment of Board of Health and Welfare. Stockholm 1984. primary and recurrent gynecologic malignancies. Cardio- 78. TADA S. TSUKIOKA M.. ISHII C. et al.: Computed tomographic vasc. Intervent. Radiol. 12 (1989). 188. Printed in Sweden by Almqvist & Wiksell Tryckeri, Uppsala 1990