- Email: [email protected]
The use of ultrasound and computed tomography scanning in the management of gynecologic cancer patients
GYNECOLOGIC
ONCOLOGY
18,
54-61 (1984)
The Use of Ultrasound and Computed Tomography Scanning in the Management of Gynecologic Cancer Patients R. H. J. KERR-WILSON, FRCS, MRCOG, HUGH M. SHINGLETON, M.D., JAMES W. ORR Jr., M.D., AND KENNETH D. HATCH, M.D. Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, University of Alabama, University Station, Birmingham, Alabama 35294 Received November 4, 1982 The contribution of abdominal-pelvic computed tomography (CT) and ultrasound scanning to the management of gynecologic malignancy patients was assessed retrospectively in 62 scans that were performed immediately preceding laparotomy. A total of 55.5% CT and 71.4% ultrasound scans were found to be helpful. Diagnostic problem areas included the detection of lymph node metastases, peritoneal implants, omental metastases, and parametriat extension. An additional 75 scans performed in patients not undergoing laparotomy were received to assess any alteration in management resulting from the scans. It is concluded that clinicians should be more discriminating in their requests for scanning procedures and that appropriate communication between the clinician and the radiologist may avoid unproductive scans.
INTRODUCTION
Investigations in medicine are seldom warranted unless their results affect patient care. Unfortunately, many diagnostic tests are undertaken before their benefit has been adequately evaluated; new methods of diagnosis are welcomed uncritically, contributing to the rapidly increasing costs of medical care. Computed tomography (CT) has made a marked difference in the management of head injuries. However, its application in gynecology and, in particular, gynecologic oncology is not clear. Some authors [l] conclude that abdominal CT scanning “contributes greatly in the management of gynecologic cancer,” others [2] state that CT scans are “beneficial.” However, other authors [3] report that scanning makes little difference to patient management and that it is not useful as a staging tool [4] in patients with cervical cancer. Ultrasonography was originally developed for use in obstetrics, where its benefits are undisputed. Ultrasound provides correct diagnosis in 92% of patients with hydatidiform mole [5], but has only 67% agreement with surgical findings in other gynecologic malignancies [6]. Walsh et al. [7] considered that ultrasound was slightly more accurate than CT scanning in the evaluation of patients with pelvic masses. In assessing the efficacy of both CT scanning and ultrasound evaluation of pelvic masses, Sommer et al. [8] concluded that neither technique is accurate enough to substitute for exploratory surgery. 54 0090-8258184$1.50 Copyright All rights
0 1984 by Academic Press, Inc. of reproduction in any form reserved.
ULTRASOUND
AND
COMPUTED
TOMOGRAPHY
SCANNING
55
For any imaging technique to be of benefit in gynecologic oncology, it should have the ability to detect tumor spread. Involvement of neighboring structures and the presence of distant metastases is of crucial importance in treatment planning. A reliable assessment of tumor spread has considerable benefit in avoiding potentially unnecessary surgery. This investigation was undertaken to assess the use of ultrasound and CT scanning in the management of gynecologic cancer, with particular attention to the accuracy of detecting metastases. PATIENTS AND METHODS:
The charts of patients with abdominal and pelvic malignancies undergoing a laparotomy on the gynecologic oncology service at the University of Alabama, Birmingham, between January 1978 and July 1982 were reviewed. All those who had either a CT and/or ultrasound scan within 3 days prior to surgery were included in the study (Table 1). Forty-four patients had preoperative ultrasound scans and 20 had CT scans before surgery, from a total of 1109 ultrasound and in excess of 250 CT scans performed on gynecologic patients during this period. The study group undergoing laparotomy reflects situations in which it was thought CT or ultrasound might have provided additional information before surgery performed during this interval, when 1433 major operative procedures occurred. Total abdominal hysterectomy (TAH), bilateral salpingo-oophorectomy (BSO), and a staging procedure were carried out on three patients with uterine sarcoma (Table 2). A second-look laparotomy was performed on one patient with tubal carcinoma. One patient with metastatic breast carcinoma and one patient with both squamous cell carcinoma of the cervix and carcinosarcoma of the uterus had TAH and BSO. Two patients with recurrent uterine sarcoma had a laparotomy and biopsies carried out. The six benign conditions included two leiomyomatas, two benign ovarian neoplasms, one TAH for hydatidiform mole, and one atypical case of endometriosis. Four patients had more than one gynecologic malignancy at the time of surgery; in addition to carcinoma of the cervix, one patient had adenocarcinoma of the ovary, one had carcinoma of the fallopian tube, one had carcinoma of the endometrium, and another had a coexisting carcinosarcoma of the uterus. Ultrasound was performed almost exclusively by an experienced ultrasonographer using a Unirad digital machine employing a gray-scale B scanning technique. TABLE 1 NUMBER
CT scan Ultrasound scan CT and ultrasound scans Total
OF SCANS PERFORMED PRECEDING LAPAROTOMY
Abdomen Pelvis Abdomen and pelvis Abdomen Pelvis Abdomen and pelvis Abdomen
2 0 16 7 9 26 2 62
56
KERR-WILSON
ET AL.
TABLE 2 SURGICAL
PROCEDURES ACCORDING TO DIAGNOSTIC
Cervical cancer
Surgery TAH, BSO TAH, BSO, and staging laparotomy Staging laparotomy Second-look Radical hysterectomy Exenteration attempted Exenteration performed Bowel bypass Urinary diversion Laparotomy Benign condition
CATEGORY
Endometrial cancer
Ovarian cancer
Other (see text)
11
12
2 3
6
1
7 1 5 1 1 1
I 2
2 6
CT scanning was carried out with and without contrast at sections from 1 or 2 cm on a General Electric 8800 CT scanner. Assessment of local and distant disease involvement was requested, depending on the nature of the original tumor. In addition, the reports of 54 patients having scans without laparotomy were reviewed to determine if there was any subsequent alteration in management (Table 3). Thirty patients had a total of 41 CT scans; 34 ultrasound scans were carried out on 24 patients. In most instances, the scans were performed to detect tumor recurrence, progression or regression, or metastases before or during chemotherapy. RESULTS
Scan results were divided into whether they were helpful or misleading. Helpful scans were further subdivided into those which were (a) reported as normal and found to be normal at surgery, (b) diagnostic, and (c) confirmatory. The scan was considered diagnostic if it altered the patient’s management or provided information not previously known, such as detecting abdominal massesor enlarged para-aortic nodes. A confirmatory scan was one which confirmed findings already known (such as pelvic masses) or gave new but incomplete information such as identifying enlarged para-aortic nodes, but missing an omental “mass.” Misleading scans were further subdivided into false positive and false negative. The results are summarized in Tables 4 and 5. TABLE 3 SCANS WITHOUT
LAPAROTOMY
Management Procedure
Changed
No change
Total
CT scan Ultrasound scan
7 (17.1%) 2 (5.8%)
34 (82.9%) 32 (94.1%)
41 (100%) 34 (100%)
57
ULTRASOUND AND COMPUTED TOMOGRAPHY SCANNING TABLE 4 CT SCANS
RESULTS OF
Helpful Normal Abdomen Pelvis Total (%)
6 0 6 (16.7)
Diagnostic” 1 2 3 (8.3)
Misleading Confirmatoryb 5 6 11 (30.5)
False +
False -
3 1 4 (11.1)
5 7 12 (33.3)
Total 20 16 36 (100)
” Diagnostic: either altered patient management or provided new accurate information. b Confirmatory: either confirmed findings already known or provided new but incomplete information.
Just over half (55.5%) of the CT scans were considered to have been helpful; 33.3% of the reports were false negative, the main difficulty being detection of enlarged nodes and peritoneal implants by tumor. Metastatic cancer in nodes 2 cm in diameter or less was not detected at CT scan. Likewise, widespread peritoneal implants of less than 2 cm on parietal peritoneum, diaphragm, mesentery, and small and large bowel remained undetected by CT scan. In one patient a node measuring 5 x 2.5-cm was identified on CT scan but the directed biopsy was negative; however, biopsy at laparotomy proved to be positive for metastatic disease. Parametrial extension of tumors was only recognized on one CT scan out of three patients in whom it was found at subsequent laparotomy. In those patients who had scans without laparotomy (Table 3), the scan results did not change management following 34 of the 41 CT scans (82.9%) and in 32 out of the 34 ultrasound scans (94.1%). Seven (17.1%) CT scans either had reports which were considered helpful in deciding on future management, or they offered new information. Para-aortic node enlargement was reported in one patient, a perinephric abscess in another, and a pelvic mass in another which had been undetected clinically. Of the ultrasound scans, the two which helped in the patient’s management were the ones which confirmed suspected liver metastases in a jaundiced patient with abnormal biochemical liver function following radiotherapy for stage IIIB cervical carcinoma, and another which confirmed the presence of a hydatidiform mole. Because the results of the CT scans were not as good as expected, 15 of those which preceded laparotomy were reviewed by an independent radiologist and TABLE 5 RESULTS OF ULTRASOUND
SCANS
Helpful Normal Abdomen Pelvis Total (%)
12 2 14 (20)
Misleading
Diagnostic”
Confirmatoryb
False +
2 6 8 (11.4)
6 22 28 (40)
0 0 0 (0)
False -
Total
15 5 20 (28.6)
35 3.5 70 (100)
a Diagnostic: either altered patient management or provided new information. ’ Confirmatory: either confirmed findings already known or provided new but incomplete information.
58
KERR-WILSON
ET AL.
one of the clinicians (R. K. W.). The radiologist was given a clinical presentation of each patient but did not have the previous report available nor the laparotomy findings. Four out of fifteen patients had a more accurate diagnosis made at review. A false-positive report of para-aortic nodes was reviewed, the presence of a fistula, a uterine tumor, and pelvic nodes were diagnosed. However, a 6-cm omental mass which contained tumor was not visible on the CT scan. Tumor extension to pelvic sidewall in one patient and small (<2 cm) para-aortic nodes containing metastatic tumor in patients were also undiagnosed at review. The number of misleading reports with ultrasound was 28.6% (Table 5), difficulty again being found with detection of involved lymph nodes and peritoneal implants having less than a 2-cm diameter. Peritoneal implants were undetected by ultrasound in all of 10 patients who were later found to have them at surgery. Only one patient with parametrial extension of disease at surgery had an ultrasound scan beforehand, and this escaped detection by the scan. Seven patients had gross omental metastases which were not recognized on the ultrasound scan prior to surgery. Ultrasound was correct in diagnosing liver metastases on one occasion and enlarged para-aortic nodes in another patient. The ultrasonographer diagnosed ovarian carcinoma correctly on two occasions and uterine sarcoma on three. DISCUSSION
The findings reported here are the results of our initial experience with CT scanning, together with the findings of an experienced ultrasonographer having a particular interest in the field of gynecologic oncology. This may explain some of the unexpectedly poor results with CT scans. We acknowledge that technological development and increased expertise in CT imaging has made the technique more valuable in our institution since this report was started. This study is imperfect in that it is retrospective and relies on reports of scans rather than the scans themselves, with later subjective assessment by clinicians who may have been excessively critical. However, later review of the CT scans by both clinician and radiologist indicated the limitation of CT scans with the technology available; the resolution of the scan did not accurately predict disease spread to those areas which are particularly important to gynecologic oncologists. Although the results were disappointing, they are not dissimilar to the findings of other authors on the subject [4,8]. The main areas of difficulty are in detection of small peritoneal implants, minimally enlarged metastatic lymph nodes, omental metastases, and parametrial extensions of tumor. Lesions less than 2 cm are difficult to detect by both ultrasound [9,10] and CT [4,7,X]. Using both scanning techniques, a correct diagnosis was made in only one out of three patients with omental metastases by Sommer et al. [8] and Chen et al. [l] agree that omental involvement is not always easy to diagnose by CT. Grumbine et al. [4] found only 58% accuracy of CT scans in diagnosing parametrial extension, the most common error being false-positive prediction of extension when none was found. Some of these problems are already being resolved with increased resolution by more sophisticated equipment. However, there will always be a problem with lymph node metastases, since even if the nodes can be identified without histological
ULTRASOUND
AND
COMPUTED
TOMOGRAPHY
SCANNING
59
diagnosis it may be impossible short of laparotomy to differentiate between hyperplasia and malignancy. This can be overcome to some extent by directed needle aspiration, which may be undertaken under both ultrasound and CT guidance. A positive result of needle aspiration (see Figs. 1 and 2) can be of considerable advantage to the clinician and avoid the necessity for laparotomy. Much of the problem with both ultrasound and CT is the result of the clinician’s uncritical approach to requesting the scan and failure to communicate properly with the radiologist. Ginaldi et al. [3] although they found new information in 10 out of 17 CT scans in patients with carcinoma of the cervix, found that their management was only changed in one case. Our results were similar for both ultrasound and CT in those patients who did not have laparotomy but were assessed purely to see if the investigation changed their management (Table 3). Closer communication between clinician and radiologist is important if scans which do not alter patient management are to be avoided. The clinician should be informed of the limitation of the technique requested and as a result may decide that his management of a patient would not be altered by a scan. In exchange, he is able to inform the radiologist of the patient’s clinical condition and tell him in which areas he is most interested. Although we found the effectiveness of CT and ultrasound to be comparable, this may have been because of our in-depth experience with ultrasonography compared to less experience with CT during this period. However, other authors [6] have also found both ultrasound and CT to provide similar information.
FIG. 1. CT scan showing enlarged lymph nodes anterior to the vena cava (closed arrow) and lateral to the aorta (open arrow). An aspiration needle is positioned anterior to the paracaval node. This was subsequently aspirated and the cytology was positive for metastatic squamous cell carcinoma.
60
KERR-WILSON
ET AL.
FIG. 2. Pelvic CT scan indicating recurrent tumor (arrow) extending to the right pelvic sidewall. Following documentation by fine needle aspiration, she was not considered a candidate for exenteration.
Ultrasound has the additional advantage over CT of being less costly and avoiding any risk from radiation [9]. In conclusion, although this is a report of our early experience with CT scans in gynecologic oncology, we wish to emphasize that both CT and ultrasound scans should not be used indiscriminately by clinicians to preempt their clinical judgment, nor should scans be considered as authoritative in themselves. As with other investigations, they should be assessed together with the information provided by a history and a physical examination. They should only be performed if they have potential therapeutic implications. Close communication between radiologist and clinician is mandatory. REFERENCES 1. Chen, S. S., Kumari, S. and Lee, L. Contribution of abdominal CT in the management of gynecologic cancer: Correlated study of CT image and Gross Surgical Pathology, Gynecol. Oncol. 10, 162-172 (1980). 2. Photopulos, G. J., McCartney, W. H., Walton, L. A., and Staab, E. V. Computerized tomography applied to gynecologic oncology, Amer. J. Obstet. Gynecol. 135, 381-383 (1979). 3. Ginaldi, S., Wallace, S., Jing, B-S, Berardino, M. E. Carcinoma of the cervix: Lymphangiography and computed tomography, Amer. J. Roenrgenol. 136, 1087-1091 (1981). 4. Grumbine, F. C., Rosenshein, N. B., Zerhouni, E. A., and Siegelman, S. S. Abdominal computed tomography in the preoperative evaluation of early cervical cancer, Gynecol. &co/. 12, 286290 (1981).
ULTRASOUND
AND COMPUTED
TOMOGRAPHY
SCANNING
61
5. Kobayashi, M. Use of diagnostic ultrasound in trophoblastic neoplasms and ovarian tumors, Cancer 38, 441-452 (1976). 6. Nash, C. G., Alberts, D. S., Suciu, T. N., Giles, H. R., Tobias, D. A., and Wardman, R. S. comparison of B-mode ultrasonography and computed tomography in gynecologic cancer, Gynecol. Oncol. 7, 172-179 (1979). 7. Walsh, J. W., Rosenfield, A. T., Jaffe, C. C., Schwartz, P. E., Simeone, J., Dembner, A. G., and Taylor, K. J. W. Prospective comparison of ultrasound and computed tomography in evaluation of gynecologic pelvic masses, Amer. J. Roenrgenol. 131, 955-960 (1978). 8. Sommer, F. G., Walsh, J. W., Schwartz, P. E., Viscomi, G. N., Jaffe, C. C., Taylor, K. J. W., and Rosenfield, A. T. Evaluation of gynecologic pelvic masses by ultrasound and computed tomography, J. Reprod. Med. 27, 45-50 (1982). 9. Brascho, D. J. Gynecologic malignancy, in Abdominal ultrasound in the cancer paiient, (D. J. Brascho and T. H. Shawker, Eds.), Wiley, New York, pp. 209-245 (1980). 10. Levi, S., and Delva, R. Value of ultrasonic diagnosis of gynecological tumors in 370 surgical cases, Acta Obstet. Gynecol. Stand. 55, 261-266 (1976).
ONCOLOGY
18,
54-61 (1984)
The Use of Ultrasound and Computed Tomography Scanning in the Management of Gynecologic Cancer Patients R. H. J. KERR-WILSON, FRCS, MRCOG, HUGH M. SHINGLETON, M.D., JAMES W. ORR Jr., M.D., AND KENNETH D. HATCH, M.D. Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, University of Alabama, University Station, Birmingham, Alabama 35294 Received November 4, 1982 The contribution of abdominal-pelvic computed tomography (CT) and ultrasound scanning to the management of gynecologic malignancy patients was assessed retrospectively in 62 scans that were performed immediately preceding laparotomy. A total of 55.5% CT and 71.4% ultrasound scans were found to be helpful. Diagnostic problem areas included the detection of lymph node metastases, peritoneal implants, omental metastases, and parametriat extension. An additional 75 scans performed in patients not undergoing laparotomy were received to assess any alteration in management resulting from the scans. It is concluded that clinicians should be more discriminating in their requests for scanning procedures and that appropriate communication between the clinician and the radiologist may avoid unproductive scans.
INTRODUCTION
Investigations in medicine are seldom warranted unless their results affect patient care. Unfortunately, many diagnostic tests are undertaken before their benefit has been adequately evaluated; new methods of diagnosis are welcomed uncritically, contributing to the rapidly increasing costs of medical care. Computed tomography (CT) has made a marked difference in the management of head injuries. However, its application in gynecology and, in particular, gynecologic oncology is not clear. Some authors [l] conclude that abdominal CT scanning “contributes greatly in the management of gynecologic cancer,” others [2] state that CT scans are “beneficial.” However, other authors [3] report that scanning makes little difference to patient management and that it is not useful as a staging tool [4] in patients with cervical cancer. Ultrasonography was originally developed for use in obstetrics, where its benefits are undisputed. Ultrasound provides correct diagnosis in 92% of patients with hydatidiform mole [5], but has only 67% agreement with surgical findings in other gynecologic malignancies [6]. Walsh et al. [7] considered that ultrasound was slightly more accurate than CT scanning in the evaluation of patients with pelvic masses. In assessing the efficacy of both CT scanning and ultrasound evaluation of pelvic masses, Sommer et al. [8] concluded that neither technique is accurate enough to substitute for exploratory surgery. 54 0090-8258184$1.50 Copyright All rights
0 1984 by Academic Press, Inc. of reproduction in any form reserved.
ULTRASOUND
AND
COMPUTED
TOMOGRAPHY
SCANNING
55
For any imaging technique to be of benefit in gynecologic oncology, it should have the ability to detect tumor spread. Involvement of neighboring structures and the presence of distant metastases is of crucial importance in treatment planning. A reliable assessment of tumor spread has considerable benefit in avoiding potentially unnecessary surgery. This investigation was undertaken to assess the use of ultrasound and CT scanning in the management of gynecologic cancer, with particular attention to the accuracy of detecting metastases. PATIENTS AND METHODS:
The charts of patients with abdominal and pelvic malignancies undergoing a laparotomy on the gynecologic oncology service at the University of Alabama, Birmingham, between January 1978 and July 1982 were reviewed. All those who had either a CT and/or ultrasound scan within 3 days prior to surgery were included in the study (Table 1). Forty-four patients had preoperative ultrasound scans and 20 had CT scans before surgery, from a total of 1109 ultrasound and in excess of 250 CT scans performed on gynecologic patients during this period. The study group undergoing laparotomy reflects situations in which it was thought CT or ultrasound might have provided additional information before surgery performed during this interval, when 1433 major operative procedures occurred. Total abdominal hysterectomy (TAH), bilateral salpingo-oophorectomy (BSO), and a staging procedure were carried out on three patients with uterine sarcoma (Table 2). A second-look laparotomy was performed on one patient with tubal carcinoma. One patient with metastatic breast carcinoma and one patient with both squamous cell carcinoma of the cervix and carcinosarcoma of the uterus had TAH and BSO. Two patients with recurrent uterine sarcoma had a laparotomy and biopsies carried out. The six benign conditions included two leiomyomatas, two benign ovarian neoplasms, one TAH for hydatidiform mole, and one atypical case of endometriosis. Four patients had more than one gynecologic malignancy at the time of surgery; in addition to carcinoma of the cervix, one patient had adenocarcinoma of the ovary, one had carcinoma of the fallopian tube, one had carcinoma of the endometrium, and another had a coexisting carcinosarcoma of the uterus. Ultrasound was performed almost exclusively by an experienced ultrasonographer using a Unirad digital machine employing a gray-scale B scanning technique. TABLE 1 NUMBER
CT scan Ultrasound scan CT and ultrasound scans Total
OF SCANS PERFORMED PRECEDING LAPAROTOMY
Abdomen Pelvis Abdomen and pelvis Abdomen Pelvis Abdomen and pelvis Abdomen
2 0 16 7 9 26 2 62
56
KERR-WILSON
ET AL.
TABLE 2 SURGICAL
PROCEDURES ACCORDING TO DIAGNOSTIC
Cervical cancer
Surgery TAH, BSO TAH, BSO, and staging laparotomy Staging laparotomy Second-look Radical hysterectomy Exenteration attempted Exenteration performed Bowel bypass Urinary diversion Laparotomy Benign condition
CATEGORY
Endometrial cancer
Ovarian cancer
Other (see text)
11
12
2 3
6
1
7 1 5 1 1 1
I 2
2 6
CT scanning was carried out with and without contrast at sections from 1 or 2 cm on a General Electric 8800 CT scanner. Assessment of local and distant disease involvement was requested, depending on the nature of the original tumor. In addition, the reports of 54 patients having scans without laparotomy were reviewed to determine if there was any subsequent alteration in management (Table 3). Thirty patients had a total of 41 CT scans; 34 ultrasound scans were carried out on 24 patients. In most instances, the scans were performed to detect tumor recurrence, progression or regression, or metastases before or during chemotherapy. RESULTS
Scan results were divided into whether they were helpful or misleading. Helpful scans were further subdivided into those which were (a) reported as normal and found to be normal at surgery, (b) diagnostic, and (c) confirmatory. The scan was considered diagnostic if it altered the patient’s management or provided information not previously known, such as detecting abdominal massesor enlarged para-aortic nodes. A confirmatory scan was one which confirmed findings already known (such as pelvic masses) or gave new but incomplete information such as identifying enlarged para-aortic nodes, but missing an omental “mass.” Misleading scans were further subdivided into false positive and false negative. The results are summarized in Tables 4 and 5. TABLE 3 SCANS WITHOUT
LAPAROTOMY
Management Procedure
Changed
No change
Total
CT scan Ultrasound scan
7 (17.1%) 2 (5.8%)
34 (82.9%) 32 (94.1%)
41 (100%) 34 (100%)
57
ULTRASOUND AND COMPUTED TOMOGRAPHY SCANNING TABLE 4 CT SCANS
RESULTS OF
Helpful Normal Abdomen Pelvis Total (%)
6 0 6 (16.7)
Diagnostic” 1 2 3 (8.3)
Misleading Confirmatoryb 5 6 11 (30.5)
False +
False -
3 1 4 (11.1)
5 7 12 (33.3)
Total 20 16 36 (100)
” Diagnostic: either altered patient management or provided new accurate information. b Confirmatory: either confirmed findings already known or provided new but incomplete information.
Just over half (55.5%) of the CT scans were considered to have been helpful; 33.3% of the reports were false negative, the main difficulty being detection of enlarged nodes and peritoneal implants by tumor. Metastatic cancer in nodes 2 cm in diameter or less was not detected at CT scan. Likewise, widespread peritoneal implants of less than 2 cm on parietal peritoneum, diaphragm, mesentery, and small and large bowel remained undetected by CT scan. In one patient a node measuring 5 x 2.5-cm was identified on CT scan but the directed biopsy was negative; however, biopsy at laparotomy proved to be positive for metastatic disease. Parametrial extension of tumors was only recognized on one CT scan out of three patients in whom it was found at subsequent laparotomy. In those patients who had scans without laparotomy (Table 3), the scan results did not change management following 34 of the 41 CT scans (82.9%) and in 32 out of the 34 ultrasound scans (94.1%). Seven (17.1%) CT scans either had reports which were considered helpful in deciding on future management, or they offered new information. Para-aortic node enlargement was reported in one patient, a perinephric abscess in another, and a pelvic mass in another which had been undetected clinically. Of the ultrasound scans, the two which helped in the patient’s management were the ones which confirmed suspected liver metastases in a jaundiced patient with abnormal biochemical liver function following radiotherapy for stage IIIB cervical carcinoma, and another which confirmed the presence of a hydatidiform mole. Because the results of the CT scans were not as good as expected, 15 of those which preceded laparotomy were reviewed by an independent radiologist and TABLE 5 RESULTS OF ULTRASOUND
SCANS
Helpful Normal Abdomen Pelvis Total (%)
12 2 14 (20)
Misleading
Diagnostic”
Confirmatoryb
False +
2 6 8 (11.4)
6 22 28 (40)
0 0 0 (0)
False -
Total
15 5 20 (28.6)
35 3.5 70 (100)
a Diagnostic: either altered patient management or provided new information. ’ Confirmatory: either confirmed findings already known or provided new but incomplete information.
58
KERR-WILSON
ET AL.
one of the clinicians (R. K. W.). The radiologist was given a clinical presentation of each patient but did not have the previous report available nor the laparotomy findings. Four out of fifteen patients had a more accurate diagnosis made at review. A false-positive report of para-aortic nodes was reviewed, the presence of a fistula, a uterine tumor, and pelvic nodes were diagnosed. However, a 6-cm omental mass which contained tumor was not visible on the CT scan. Tumor extension to pelvic sidewall in one patient and small (<2 cm) para-aortic nodes containing metastatic tumor in patients were also undiagnosed at review. The number of misleading reports with ultrasound was 28.6% (Table 5), difficulty again being found with detection of involved lymph nodes and peritoneal implants having less than a 2-cm diameter. Peritoneal implants were undetected by ultrasound in all of 10 patients who were later found to have them at surgery. Only one patient with parametrial extension of disease at surgery had an ultrasound scan beforehand, and this escaped detection by the scan. Seven patients had gross omental metastases which were not recognized on the ultrasound scan prior to surgery. Ultrasound was correct in diagnosing liver metastases on one occasion and enlarged para-aortic nodes in another patient. The ultrasonographer diagnosed ovarian carcinoma correctly on two occasions and uterine sarcoma on three. DISCUSSION
The findings reported here are the results of our initial experience with CT scanning, together with the findings of an experienced ultrasonographer having a particular interest in the field of gynecologic oncology. This may explain some of the unexpectedly poor results with CT scans. We acknowledge that technological development and increased expertise in CT imaging has made the technique more valuable in our institution since this report was started. This study is imperfect in that it is retrospective and relies on reports of scans rather than the scans themselves, with later subjective assessment by clinicians who may have been excessively critical. However, later review of the CT scans by both clinician and radiologist indicated the limitation of CT scans with the technology available; the resolution of the scan did not accurately predict disease spread to those areas which are particularly important to gynecologic oncologists. Although the results were disappointing, they are not dissimilar to the findings of other authors on the subject [4,8]. The main areas of difficulty are in detection of small peritoneal implants, minimally enlarged metastatic lymph nodes, omental metastases, and parametrial extensions of tumor. Lesions less than 2 cm are difficult to detect by both ultrasound [9,10] and CT [4,7,X]. Using both scanning techniques, a correct diagnosis was made in only one out of three patients with omental metastases by Sommer et al. [8] and Chen et al. [l] agree that omental involvement is not always easy to diagnose by CT. Grumbine et al. [4] found only 58% accuracy of CT scans in diagnosing parametrial extension, the most common error being false-positive prediction of extension when none was found. Some of these problems are already being resolved with increased resolution by more sophisticated equipment. However, there will always be a problem with lymph node metastases, since even if the nodes can be identified without histological
ULTRASOUND
AND
COMPUTED
TOMOGRAPHY
SCANNING
59
diagnosis it may be impossible short of laparotomy to differentiate between hyperplasia and malignancy. This can be overcome to some extent by directed needle aspiration, which may be undertaken under both ultrasound and CT guidance. A positive result of needle aspiration (see Figs. 1 and 2) can be of considerable advantage to the clinician and avoid the necessity for laparotomy. Much of the problem with both ultrasound and CT is the result of the clinician’s uncritical approach to requesting the scan and failure to communicate properly with the radiologist. Ginaldi et al. [3] although they found new information in 10 out of 17 CT scans in patients with carcinoma of the cervix, found that their management was only changed in one case. Our results were similar for both ultrasound and CT in those patients who did not have laparotomy but were assessed purely to see if the investigation changed their management (Table 3). Closer communication between clinician and radiologist is important if scans which do not alter patient management are to be avoided. The clinician should be informed of the limitation of the technique requested and as a result may decide that his management of a patient would not be altered by a scan. In exchange, he is able to inform the radiologist of the patient’s clinical condition and tell him in which areas he is most interested. Although we found the effectiveness of CT and ultrasound to be comparable, this may have been because of our in-depth experience with ultrasonography compared to less experience with CT during this period. However, other authors [6] have also found both ultrasound and CT to provide similar information.
FIG. 1. CT scan showing enlarged lymph nodes anterior to the vena cava (closed arrow) and lateral to the aorta (open arrow). An aspiration needle is positioned anterior to the paracaval node. This was subsequently aspirated and the cytology was positive for metastatic squamous cell carcinoma.
60
KERR-WILSON
ET AL.
FIG. 2. Pelvic CT scan indicating recurrent tumor (arrow) extending to the right pelvic sidewall. Following documentation by fine needle aspiration, she was not considered a candidate for exenteration.
Ultrasound has the additional advantage over CT of being less costly and avoiding any risk from radiation [9]. In conclusion, although this is a report of our early experience with CT scans in gynecologic oncology, we wish to emphasize that both CT and ultrasound scans should not be used indiscriminately by clinicians to preempt their clinical judgment, nor should scans be considered as authoritative in themselves. As with other investigations, they should be assessed together with the information provided by a history and a physical examination. They should only be performed if they have potential therapeutic implications. Close communication between radiologist and clinician is mandatory. REFERENCES 1. Chen, S. S., Kumari, S. and Lee, L. Contribution of abdominal CT in the management of gynecologic cancer: Correlated study of CT image and Gross Surgical Pathology, Gynecol. Oncol. 10, 162-172 (1980). 2. Photopulos, G. J., McCartney, W. H., Walton, L. A., and Staab, E. V. Computerized tomography applied to gynecologic oncology, Amer. J. Obstet. Gynecol. 135, 381-383 (1979). 3. Ginaldi, S., Wallace, S., Jing, B-S, Berardino, M. E. Carcinoma of the cervix: Lymphangiography and computed tomography, Amer. J. Roenrgenol. 136, 1087-1091 (1981). 4. Grumbine, F. C., Rosenshein, N. B., Zerhouni, E. A., and Siegelman, S. S. Abdominal computed tomography in the preoperative evaluation of early cervical cancer, Gynecol. &co/. 12, 286290 (1981).
ULTRASOUND
AND COMPUTED
TOMOGRAPHY
SCANNING
61
5. Kobayashi, M. Use of diagnostic ultrasound in trophoblastic neoplasms and ovarian tumors, Cancer 38, 441-452 (1976). 6. Nash, C. G., Alberts, D. S., Suciu, T. N., Giles, H. R., Tobias, D. A., and Wardman, R. S. comparison of B-mode ultrasonography and computed tomography in gynecologic cancer, Gynecol. Oncol. 7, 172-179 (1979). 7. Walsh, J. W., Rosenfield, A. T., Jaffe, C. C., Schwartz, P. E., Simeone, J., Dembner, A. G., and Taylor, K. J. W. Prospective comparison of ultrasound and computed tomography in evaluation of gynecologic pelvic masses, Amer. J. Roenrgenol. 131, 955-960 (1978). 8. Sommer, F. G., Walsh, J. W., Schwartz, P. E., Viscomi, G. N., Jaffe, C. C., Taylor, K. J. W., and Rosenfield, A. T. Evaluation of gynecologic pelvic masses by ultrasound and computed tomography, J. Reprod. Med. 27, 45-50 (1982). 9. Brascho, D. J. Gynecologic malignancy, in Abdominal ultrasound in the cancer paiient, (D. J. Brascho and T. H. Shawker, Eds.), Wiley, New York, pp. 209-245 (1980). 10. Levi, S., and Delva, R. Value of ultrasonic diagnosis of gynecological tumors in 370 surgical cases, Acta Obstet. Gynecol. Stand. 55, 261-266 (1976).