|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2003 American Society for Clinical Oncology Ultrasound Screening for the Early Detection of Ovarian CancerFrom the Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, University of Kentucky Medical Center, Lexington, KY. Address reprint requests to Paul D. DePriest, MD, Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, University of Kentucky Medical Center, 331B Whitney-Hendrickson Bldg, 800 Rose St, Lexington, KY 40536; email: pddepr0{at}email.uky.edu.
Ovarian cancer screening in the general population has been performed using ultrasound examination of the female pelvis and serum tumor marker determinations. Ultrasound examinations, particularly transvaginal sonography (TVS), have been advocated as potentially useful modalities. Investigators from the University of Kentucky (Lexington, KY) and Hirosaki University (Hirosaki, Japan) have recently published results from ultrasound-based ovarian cancer screening studies. The Kentucky trial screened 14,469 women using TVS on an annual basis. One hundred eighty women underwent surgery, and 17 ovarian cancers were detected, 11 of which were invasive epithelial lesions. The Hirosaki trial reported the results of an ultrasound-based screening study among 51,550 women who were first-time participants. Three hundred twenty-four women underwent surgery, and 22 ovarian cancers were detected. In each of these trials, the positive predictive value of gray-scale sonography was low. Morphologic tumor indexing and Doppler examinations have both been proposed as potential second-line studies, which could increase the positive predictive value of gray-scale ultrasound. A review of these techniques is presented. At present, ovarian cancer screening in the general population using ultrasound examinations is an experimental technique. Further studies are needed to determine whether second-line testing can improve the positive predictive value of gray-scale sonography such that asymptomatic women do not undergo unnecessary surgery for benign masses.
OVARIAN CARCINOMA is a leading cause of cancer-related death among women in the United States. It is estimated that in the year 2002 approximately 23,300 women were diagnosed as having ovarian cancer and that 13,900 women will die of the disease.1 More than 70% of all patients with ovarian cancer seek treatment after there has been regional or distant spread of disease. The 5-year survival rate among women with advanced-stage disease is less than 30%. These poor long-term survival rates have not improved significantly, even with the advent of cisplatin/taxane chemotherapy regimens and aggressive surgical debulking. Only 25% of patients present with stage I disease, when treatment is associated with survival rates of 85% to 90% at 5 years. If screening strategies could identify women with early-stage ovarian cancer, then long-term survival may be improved. Many investigators have advocated screening for ovarian cancer in an attempt to increase survival through early detection. Ultrasonography and serum marker evaluations have both been used in screening studies.2,3 Campbell et al4 published the results of a trial using abdominal ultrasound to screen for ovarian cancer. In this study, abdominal ultrasound examinations were shown to be capable of detecting early-stage ovarian cancers. Bourne et al5 and van Nagell et al6 published results of studies using transvaginal sonography (TVS) to screen for ovarian cancer. These studies revealed that TVS was safe, well tolerated, and effective in the identification of ovarian cancer. Studies are currently under way in the United States and the United Kingdom to evaluate the efficacy of sonography in the early detection of ovarian cancer. The purpose of this report is to review the current status of ultrasound-based ovarian cancer screening and the use of morphologic indexing and Doppler sonography as second-line procedures to improve the accuracy of ultrasound screening.
Since 1986, the University of Kentucky (Lexington, KY) has performed an ongoing study using TVS to screen for ovarian cancer. The most recent publication from this trial reported on results from more than 14,000 screened women.7 The purpose of the trial was to determine whether TVS could detect ovarian cancer at an earlier stage, thus making the cancer more amenable to therapy and possibly cure. From 1987 to 1999, 14,469 women were enrolled onto the University of Kentucky Ovarian Cancer Screening Program. Women participating in the screening study were recruited through media campaigns and via communications through civic organizations. Eligible women included all women 50 years of age or older and women 25 years of age or older with a documented family history of ovarian cancer in at least one primary or secondary relative. BRCA1 and BRCA2 testing was not performed as part of this trial. Postmenopausal was defined as the absence of menses for a minimum of 12 months. All study participants completed a questionnaire regarding medical history and family history of cancer. Any woman with a known ovarian tumor or a personal history of ovarian cancer was excluded from this investigation. After informed consent was obtained, patients underwent screening by a certified technologist. Screening was provided at no cost to the patient.
The screening algorithm used in the Kentucky trial is depicted in Fig 1
Seventeen women were found to have ovarian cancer. Histopathologic findings in these patients are listed in Table 2  . Eleven women had epithelial ovarian cancer. Five of these epithelial cancers were stage I, three were stage II, and three were stage III. Four patients had false-negative findings on TVS. These patients developed interval ovarian cancer within 12 months of a normal scan result. Two of these patients had stage II disease, and two patients had stage III disease.
Ovarian cancer screening with TVS had a sensitivity of 81%. Statistical performance of TVS for the early detection of ovarian cancer is shown in Table 3 . Survival among ovarian cancer patients who presented for annual screening was 93% at 2 years and 84% at 5 years. Survival statistics for patients presenting for screening are listed in Table 4. The authors concluded that TVS screening for ovarian cancer was well tolerated by patients and that it produced a moderate protective effect against death from ovarian cancer.
Investigators from Hirosaki University (Hirosaki, Japan) have reported on a large, population-based ovarian cancer screening program using TVS followed by second-line diagnostic testing for patients with abnormal screen results.8 In this study, TVS was offered to asymptomatic women who presented for cervical cytologic screening. Women were eligible for the trial if they were older than 30 years of age and resided in the Aomori Prefecture. The study algorithm is depicted in Fig 2 .
The number of women participating in the screening study was 183,043. Of these women, 51,550 were first-time participants. The results of primary screens were abnormal in 5,309 first-time participants (10%). Only 4,452 of these women underwent secondary screening with additional TVS and serum tumor marker evaluation. Closer examinations with computed tomographic scanning, magnetic resonance imaging, and serial tumor marker evaluations were necessary in 2,554 women. On the basis of these findings, 324 women underwent surgery. The histopathologic findings are listed in Table 5 . Twenty-two patients were found to have ovarian cancer, and of these patients, 17 had primary invasive epithelial cancer. The stage distribution of the 17 epithelial cancers was as follows: stage I, 12 cancers; stage II, two cancers; stage III, two cancers; and stage IV, one cancer. The authors concluded that TVS screening was a viable method for the early detection of ovarian cancer.
Ultrasonography using simplistic abnormality criteria, as defined in the Kentucky and Hirosaki trials, can lead to the detection of ovarian cancer in asymptomatic women. Unfortunately, use of these simplistic abnormality criteria results in the identification of many benign and malignant ovarian tumors. The detection of these benign tumors could lead to unnecessary surgery in healthy, asymptomatic women. In the Kentucky trial, the positive predictive value for TVS was 9.4%. The Hirosaki trial showed TVS to produce a positive predictive value of 6.8%, even with the application of multiple secondary and tertiary studies. The differences in the positive predictive value between the two studies can be attributed to the large number of premenopausal women enrolled onto the Japanese trial. Premenopausal women have active ovarian follicle formation and are more likely to have benign cysts, such as endometriomas and dermoids. Benign ovarian cysts can be difficult to differentiate from cancer when relying on gray-scale sonography images. The application of second-line studies, such as morphology tumor indexing and Doppler examinations, to sonographic screening may increase the accuracy of TVS.9
Most ovarian masses detected by ultrasound screening are benign.7 It is essential that ultrasound images are interpreted in a manner that decreases observer variation and false-positive results. Several investigators have advocated the use of morphologic scoring systems, which aid in standardizing and quantifying the interpretation of ultrasound images.9–12 Granberg et al10 reported that ovarian cyst characteristics were reliably predicted by sonographic examination. Furthermore, the authors found that papillary projections on the internal cyst wall were most predictive of malignancy. Sassone et al11 reported an index that scored four different morphologic characteristics of ovarian cyst architecture, including wall structure, cyst wall thickness, septation, and echogenicity. Each category was assigned discrete scoring possibilities, thereby limiting variability among observers. The scoring index was evaluated by analyzing 143 ovarian tumors. The index was highly sensitive (100%) and moderately specific (83%) in the differentiation of benign masses from malignant masses. DePriest et al9 reported a morphologic index system, which scored only three structural characteristics (ovarian volume, cyst wall, and septae). This simplified system was tested on 213 ovarian masses. The sensitivity for ovarian cancer detection was 89%, and the specificity was 70%. Lerner et al12 used the Sassone classification system data to perform a multiple regression analysis to define more accurately the relative importance of each structural component. The authors were able to simplify the index and apply weighted scoring to the structural components. The revised scoring system was tested in the evaluation of 350 patients with ovarian tumors. The sensitivity of the index was 97%, and the specificity was 77%. Using a similar strategy, Ferrazzi et al13 produced a weighted scoring system that was tested on 330 ovarian tumors. The sensitivity of this index was 87%, and the specificity was 67%.
The above-referenced morphologic scoring systems were developed and validated within the parent institution. Internal validation is prone to observer bias. Mol et al14 performed an external validation of several sonographic scoring systems, including the Finkler, Granberg, Sassone, Lerner, DePriest, and Ferrazzi indices. The authors evaluated 170 ovarian tumors and compared the diagnostic accuracy of the various scoring systems (Table 6
Persistent ovarian cysts or masses identified in the Kentucky and Hirosaki trials had a low likelihood of malignancy (9.4% and 7.0%, respectively). Application of morphologic scoring systems to sonographic screening protocols could allow the maintenance of a high sensitivity, approaching 80% to 90%, and improve the positive predictive value to greater than 20%. Morphologic indexing would also serve to standardize ultrasound reading without adding costs. Finally, the addition of morphologic tumor indexing can be expected to decrease unnecessary surgery for simple, benign ovarian cysts.
Doppler sonography has been proposed by many investigators as a possible second-line strategy in the differentiation of benign versus malignant ovarian masses.15–18 Kurjak et al19 have published extensively on the accuracy of Doppler examination in differentiating benign masses from malignant ovarian masses. However, in several reports,20,21 Doppler studies have not been shown to add useful information beyond gray-scale imaging. Stein et al20 evaluated 170 adnexal masses using gray-scale and Doppler analysis. The authors concluded that gray-scale sonography was equally as effective as Doppler sonography in the prediction of malignancy, with positive predictive values of 50% versus 49%, respectively. The sensitivity of gray-scale sonography in the detection of ovarian cancer was superior to that of Doppler sonography (98% v 67%, respectively). Even when Doppler examinations were simplified and limited to the expression of internal color flow, gray-scale sonography was a more sensitive indicator of malignancy than Doppler sonography (98% v 77%, respectively). Valentin21 studied 173 ovarian masses with gray-scale and Doppler ultrasound. Valentin concluded that gray-scale sonography yielded an accurate diagnosis in 51% of patients and that Doppler sonography contributed little to the specific diagnosis of pelvic tumors. Recently, some studies have shown that three-dimensional–power Doppler examinations may be much more accurate than two-dimensional Doppler examinations.19,22 Kurjak et al19 described a series of 120 patients with ovarian tumors studied with two-dimensional– versus three-dimensional–power Doppler sonography. The authors found that three-dimensional Doppler sonography was superior to two-dimensional Doppler sonography in the detection of ovarian cancer, with a superior sensitivity, specificity, positive predictive value, and negative predictive value. In the evaluation of 71 patients with ovarian masses, Cohen et al22 similarly found that three-dimensional–power Doppler sonography was more accurate than two-dimensional Doppler sonography or three-dimensional gray-scale sonography. The authors reported that three-dimensional–power Doppler sonography had a sensitivity of 100%, a specificity of 75%, and a positive predictive value of 50%. Unfortunately, the superiority of three-dimensional–power Doppler sonography has not been uniformly accepted by other experts. Guerriero et al,23 who have published extensively in the field of Doppler analysis, responded to the Cohen study in a letter to the editor stating that current data do not support the assumption that three-dimensional–power Doppler evaluations are superior to two-dimensional Doppler evaluations. Many questions remain concerning Doppler examination as a possible second-line study for ultrasound-based ovarian cancer screening protocols. Should the examination focus on the quantitative differences in blood flow between benign and malignant masses by measuring the pulsatility index or resistive index? Should the Doppler study aim to identify qualitative differences in blood flow within complex masses? Can Doppler analysis maintain a high sensitivity in the detection of ovarian cancers while improving specificity and the positive predictive value of gray-scale sonography? Is this highly technical examination prone to interobserver variation and difficulties of standardization? Is the equipment needed for Doppler examination too expensive to be incorporated into mass screening algorithms? At present, it does not seem that there is a consensus of opinion that Doppler evaluations of ovarian masses will significantly improve the sensitivity and positive predictive value of gray-scale sonography as applied to screening strategies. However, further analysis of this technology is warranted.
TVS can detect ovarian cancer among asymptomatic women. Most cancers detected by sonography are early-stage lesions. Unfortunately, ultrasound examinations detect both benign and malignant ovarian masses. The identification of benign ovarian masses could lead to excessive surgical interventions among healthy, active, asymptomatic women. Second-line studies, such as morphologic tumor indexing or Doppler sonography, may prove beneficial in differentiating benign from malignant masses and increasing the positive predictive value of a screening algorithm while maintaining a high sensitivity. Ovarian cancer has a low prevalence in the population. Therefore, any proposed screening strategy must have high sensitivity and specificity. At present, ultrasound ovarian cancer screening in the general population should only be performed as part of an institutional review board–approved investigational protocol. Future studies are needed to determine the patient populations who would most benefit from screening and the exact algorithm that should be applied. In the United Kingdom, a randomized controlled trial for the early detection of ovarian cancer is underway, comparing serum marker screening, ultrasound screening, and no screening. This study may answer many lingering questions regarding ovarian cancer screening in the general population.
DR. CANNISTRA: In order to assess the value of a screening test, we generally would like to see randomized data that show a reduction in cancer-specific mortality. What current or planned studies will help to address this point? DR. DePRIEST: There are currently two different ways to look at screening. One is ultrasound or imaging screening and the other is tumor marker or serum screening. Of the two, I think serum markers can be applied to a broader population with much lower interobserver variation. The way the serum marker studies have been developed is that the participants have a primary screening with serum, and then if their abnormality is noted, they have an imaging study. The biggest application of what we’ve learned has gone into developing ways of differentiating benign from malignant, even if ultrasound is going to be used as a second-line screen to, for example, serum CA-125 or other markers that are being developed. Will there be a randomized prospective trial with ultrasound? I doubt it. DR. BAST: I believe that the PLCO trial is still ongoing, although it is probably inadequately powered to detect ovarian cancer [Control Clin Trials 21:273S–309S, 2000]. The UK trial involves 200,000 women: 100,000 women followed in the community by conventional pelvic exams, 50,000 women followed by CA-125 serially, and 50,000 women followed by ultrasound. The women who were previously being followed in a randomized study with ultrasound versus conventional observation have been folded into that larger UK study [Int J Biol Makers 13:216–220, 1998]. DR. SKATES: The PLCO approach is to do simultaneously ultrasound and CA-125 in the screen group and nothing in the nonscreen group. After that, there is no definition or guidance as to what should happen other than referral to a gynecologist. The UK trial that Ian Jacobs is running does ultrasound on an annual basis only and then follows up with a second-line ultrasound for those who appear positive in the first set [Lancet 353:1207–1209, 1999]. My understanding is that they use the Kentucky criteria to signify what is positive and negative in an ultrasound screen, and then it’s a longitudinal CA-125 that they use in the CA-125 arm. DR. CANNISTRA: Are these standard risk patients? DR. SKATES: Yes, and that means at most one relative with ovarian or breast cancer. DR. BEREK: One of the criticisms that often comes up with regards to the ultrasound screening is that what you’re really detecting is indolent tumors and low-grade lesions that might otherwise easily be detected clinically. DR. CANNISTRA: Length time bias. DR. BEREK: What was the grade of the tumors that was detected in yours and the Sato study, and also, did this include or exclude borderline lesions? [Cancer 89:582–588, 2000] DR. DePRIEST: I tried to differentiate the invasive epithelial cancers from the borderline tumors and granulosa cell tumors. In the published trial from 2000 [Gynecol Oncol 77:350–356, 2000], there were 11 epithelial cancers, and there was a distribution of grades. I don’t doubt that there’s going to be some lead bias there. How long does ovarian cancer sit there in the patient before it becomes clinically evident? I don’t think anybody has the answer to that. DR. VASEY: Are there studies involving a formula of ultrasound and CA-125 like the RMI that surgeons use prior to surgery? Is that something that’s been locked into a screening program? DR. DePRIEST: We are not doing that. DR. BAST: I believe that the PLCO triggers referral to a gynecologist if either the CA-125 is elevated or if the ultrasound is abnormal. Both CA-125 and ultrasound are done simultaneously, so that they are parallel single-stage procedures. In the UK trial, the CA-125 trend or velocity determines whether you get an ultrasound, and then the morphology of the ultrasound determines referral. So this is a two-stage design. DR. VASEY: So this is different from the RMI score. DR. BAST: Yes, and regarding the pelvic mass, a number of papers have looked at a complex algorithm or formula that integrates the impact of different morphologic criteria with different tumor markers, including CA-125, 19-9, and 72-3. DR. BOOKMAN: I was wondering if you’ve run into any primary peritoneal tumors, and if so, how are they scored, and why is Doppler flow imaging less effective than gray scale? DR. DePRIEST: As far as peritoneal cancer, no, we haven’t detected any, and we don’t think that ultrasound is capable of detecting those, except finding ascites. I think you can blame ineffectiveness of Doppler flow imaging on technical difficulty in performing the exam. It takes a high level of skill to find the blood vessel, identify the flow, and then create a waveform and calculate the flow.
1. Jemal A, Thomas A, Murray T, et al: Cancer statistics 2002. CA Cancer J Clin 52:23–47, 2002 2. DePriest P, Gallion HH, Pavlik EJ, et al: Transvaginal sonography as a screening method for the detection of early ovarian cancer. Gynecol Oncol 65:408–414, 1997[CrossRef][Medline] 3. Jacobs I, Skates S, MacDonald N, et al: Screening for ovarian cancer: A pilot randomised controlled trial. Lancet 353:1207–1209, 1999[CrossRef][Medline] 4. Campbell S, Bhan V, Royston P, et al: Transabdominal ultrasound screening for early ovarian cancer. Br Med J 299:1363–1367, 1989 5. Bourne T, Whitehead M, Campbell S, et al: Ultrasound screening for familial ovarian cancer. Gynecol Oncol 43:92–97, 1991[CrossRef][Medline] 6. van Nagell J, Higgins RV, Donaldson ES, et al: Transvaginal sonography as a screening method for ovarian cancer: A report of the first 1000 cases screened. Cancer 65:573–577, 1990[CrossRef][Medline] 7. van Nagell J, DePriest PD, Reedy MB, et al: The efficacy of transvaginal sonographic screening in asymptomatic women at risk for ovarian cancer. Gynecol Oncol 77:350–356, 2000[CrossRef][Medline] 8. Sato S, Yokoyama Y, Sakamoto T, et al: Usefulness of mass screening for ovarian carcinoma using transvaginal ultrasonography. Cancer 89:582–588, 2000[CrossRef][Medline] 9. DePriest PD, Varner E, Powell J, et al: The efficacy of a sonographic morphology index in identifying ovarian cancer: A multi-institutional investigation. Gynecol Oncol 55:174–178, 1994[CrossRef][Medline] 10. Granberg S, Wikland M, Jansson I: Macroscopic characterization of ovarian tumors and the relation to the histological diagnosis: Criteria to be used for ultrasound evaluation. Gynecol Oncol 35:139–144, 1989[CrossRef][Medline]
11. Sassone AM, Timor-Tritsch IE, Artner A, et al: Transvaginal sonographic characterization of ovarian disease: Evaluation of a new scoring system to predict ovarian malignancy. Obstet Gynecol 78:70–76, 1991 12. Lerner JP, Timor-Tritsch IE, Federman A, et al: Transvaginal ultrasonographic characterization of ovarian masses with an improved, weighted scoring system. Am J Obstet Gynecol 170:81–85, 1994[Medline] 13. Ferrazzi E, Zanetta G, Dordoni D, et al: Transvaginal ultrasonographic characterization of ovarian masses: Comparison of five scoring systems in a multicenter study. Ultrasound Obstet Gynecol 10:192–197, 1997[CrossRef][Medline] 14. Mol B, Boll D, DeKanter M, et al: Distinguishing the benign and malignant adnexal mass: An external validation of prognostic models. Gynecol Oncol 80:162–167, 2001[CrossRef][Medline] 15. Carter J, Lau M, Fowler JM, et al: Blood flow characteristics of ovarian tumors: Implications for ovarian cancer screening. Am J Obstet Gynecol 172:901–907, 1995[CrossRef][Medline] 16. Fleisher AC, Rodgers WH, Kepple DM, et al: Color Doppler sonography of ovarian masses: A multiparameter analysis. J Ultrasound Med 12:41–48, 1993[Abstract] 17. Kurjak A, Predanic M, Jupesic-Urek S: Transvaginal color and pulsed Doppler assessment of adnexal tumor vascularity. Gynecol Oncol 50:3–9, 1993[CrossRef][Medline] 18. Timor-Tristch IE, Lerner JP, Monteagudo A, et al: Transvaginal ultrasonographic characterization of ovarian masses by means of color flow directed Doppler measurements and a morphologic scorings system. Am J Obstet Gynecol 168:909–913, 1993[Medline] 19. Kurjak A, Kupesic S, Anic T, et al: Three-dimensional ultrasound and power Doppler improve the diagnosis of ovarian lesions. Gynecol Oncol 76:28–32, 2000[CrossRef][Medline]
20. Stein SM, Laifer-Narin S, Johnson MB, et al: Differentiation of benign and malignant adnexal masses: Relative value of gray-scale, color Doppler, and spectral Doppler sonography. Am J Roentgenol 164:381–386, 1995 21. Valentin L: Pattern recognition of pelvic masses by gray-scale ultrasound imaging: The contribution of Doppler ultrasound. Ultrasound Obstet Gynecol 14:338–347, 1999[CrossRef][Medline] 22. Cohen LS, Escobar PF, Scharm C, et al: Three-dimensional power Doppler ultrasound improves the diagnostic accuracy for ovarian cancer prediction. Gynecol Oncol 82:40–48, 2001[CrossRef][Medline] 23. Guerriero S, Ajossa S, GianBenedetto M: Is three-dimensional power Doppler ultrasound better than two-dimensional power Doppler? Gynecol Oncol 84:352–353, 2001 (letter) Submitted February 11, 2003; accepted February 28, 2003. This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2003 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
