The impact of CA-125 on the sensitivity of abdominal/ pelvic CT scan before second-look laparotomy in advanced ovarian carcinoma

Transcription

1 Int J Gynecol Cancer 1996, 6, The impact of CA-125 on the sensitivity of abdominal/ pelvic CT scan before second-look laparotomy in advanced ovarian carcinoma P. G. ROSE*, K. L. REUTERt, B. E. NELSON*, J. SIROIS*, L. FOURNIER:~, F. R. REALE & R. E. HUNTER* *Department of Obstetrics and Gynecology, t Department of Radiology, ~Department of Nuclear Medicine, and Department of Pathology, University of Massachusetts Medical Center, 55 Lake Avenue North, Worcester, MA 01655, USA Abstract. Rose PG, Reuter KL, Nelson BE, Sirois J, Fournier L, Reale FR, Hunter RE. The impact of CA-125 on the sensitivity of abdominal/pelvic CT scan before second-look laparotomy in advanced ovarian carcinoma. Int J Gynecol Cancer 1996; 6: For ovarian carcinoma patients with an elevated CA-125 level at diagnosis, elevation of the antigen at the time of second-look laparotomy is consistently associated with persistent disease. This study was undertaken to determine the sensitivity and specificity of abdominal/pelvic CT scans for persistent ovarian carcinoma in patients with normal CA-125 levels before second-look laparotomy. Forty-five patients with stage III and IV ovarian carcinoma who had CA-125 levels obtained prior to initial surgery, CA-125 values <35 IU m1-1 after chemotherapy and underwent a second-look laparotomy, were studied. Forty patients with initially elevated CA-125 levels normalized their CA- 125 levels during chemotherapy. Five patients with normal initial CA-125 levels had values < 35 IU m1-1 at the completion of chemotherapy. CT scans were classified as definitively positive, suspicious or negative and were compared with second-look laparotomy results. Only two of the 45 patients (4.4%) had a positive scan which could be confirmed by CTdirected biopsy. In the 40 patients with initially elevated CA-125 levels, the sensitivity for abdominal/pelvic CT scans was only 10%. The negative predictive value was not altered by analyzing initial CA-125 values at critical values of 35, 100 and 500 IU m1-1. Among patients with CA-125 levels <35 IUm1-1 prior to initial treatment, four had no evidence of persistent disease on CT scan or second-look surgery and one patient with a suggestive CT scan had small volume disease (2 mm) at second-look laparotomy. For all 45 patients, when scans suggestive for persistent disease were included, CT scans had a sensitivity of 52% and a specificity of 75%. The addition of CA-125 testing decreased the sensitivity of abdominal/pelvic CT scanning for persistent disease. CT scanning is most likely to be of assistance in patients with liver or nodal disease or bulky residual disease after primary cytoreduction, since this disease is more likely accessible to CT-directed biopsy. Patients with negative CA-125 levels prior to initial surgery may also benefit. KEYWORDS: abdominal/pelvic CT scan, CA-125, ovarian carcinoma, secondlook laparotomy. Address for correspondence: Dr P. G. Rose, Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, University MacDonald Womens Hospital, 2074 Abington Road, Cleveland, Ohio 44106, USA IGCS Ovarian carcinoma is both a difficult malignancy to diagnose and to follow clinically, except in patients who demonstrate progressive disease. The absence of

2 214 P. G. Rose et. al. physical abnormalities is not a reliable indicator of the absence of disease. The second-look laparotomy, an invasive operative procedure, remains the most sensitive method for determining the disease status of patients with ovarian carcinoma following primary chemotherapy. With the advent of computerized tomography in the mid-1970s, an important noninvasive technique for evaluating the abdominal and pelvic contents became available. Studies have shown that abdominal pelvic CT scans prior to second-look laparotomy can detect persistent disease in 11-84% of cases (1-9). In 1983, Bast et al. developed CA-125, a monoclonal antibody against an ovarian cancer antigen (1 ). In patients with elevated CA-125 levels at the time of diagnosis, elevation of this antigen at the time of second-look laparotomy was consistently associated with persistent disease (1~-13). In 1987, the FDA approved CA-125 as a tumor marker which, if elevated above 35 IU ml-1, could be used to obviate second-look laparotomy. CA-125 has assumed a pivotal role in the management of patients whose levels are elevated at the time of diagnosis. However, despite clinical absence of disease and normalization of CA-125 levels, 56-73% of patients will be found to have persistent ovarian cancer at second looklaparotomy (11-14). The role of abdominal pelvic CT scanning in a population of patients who have completed chemotherapy with normalization of initially elevated CA-125 has not been studied. In the current report, we examine the impact of CA-125 on the sensitivity of CT-scanning before second-look laparotomy. Patients and methods Patients treated at the University of Massachusetts Medical Center for stage III and IV ovarian carcinoma who had: (a) an initial CA-125 drawn between 1984 and 1992; (b) a complete clinical response after a prescribed course of chemotherapy, and (c) were candidates for and consented to a second-look laparotomy; were the subjects of this study. As a prerequisite, all patients with elevated CA-125 levels normalized by the time of their proposed second-look surgery. Additionally, all patients had to have undergone a CT scan of the abdomen and pelvis after the completion of chemotherapy and prior to the proposed second-look laparotomy which was performed within 4-6 weeks of completion of chemotherapy. CT scans were obtained within 3 weeks of surgery and, in cases in which a definitive mass was identified and could be biopsied, a percutaneous needle biopsy was performed. In cases where a definitive diagnosis of persistent disease could not be obtained, second-look laparotomy was performed. Second-look laparotomy findings were classified as negative, macroscopically positive if a biopsy of a suspicious area was histologically positive or microscopically positive if only a non-suspicious area was histologically positive. CA-125 assays were performed using the Centocor CA-125 RIA kit (Centocor, Malvern, PA). Patients were classified as CA-125 positive if their serum levels were elevated above 35 IU ml-1 or CA-125 negative if their values were < 35 IU m1-1 prior to initial treatment. Early in our study CA-125 levels > 500 IU ml 1 were not diluted to determine the exact value. To determine if the degree to which CA-125 was positive was important, the significance of a pretreatment CA-125 was analyzed at the following critical values; 35, 100, and 500 IU ml--1 Of the 45 patients in our study, 33 CT scans were obtained for a second review. Of those 33, 30 were performed at the University of Massachusetts Medical Center. These studies were performed on a GE 9800 unit. For opacification of the gastrointestinal tract, 'redi-cat 2' (barium suspension) was given at least 4 h before imaging, followed by a repeat dose 30 min prior to the CT scan. Images were obtained of the liver before intravenous (IV) contrast was administered. However, intravenous contrast was given to all but one of these patients who had a prior contrast reaction. The IV contrast used was Renografin 60% or Hypaque 60% as a 150 cc bolus, injected in a biphasic technique at 1.2 cc/s for 100 cc and 0.7 cc/s for 50 cc. Dynamic imaging of the liver was obtained. A vaginal tampon was used for localization of the vagina by trapped air. Imaging was done at 10 mm intervals from the dome of the diaphragm to the symphysis pubis. The three cases done at outside institutions were performed with comparable equipment, with oral and IV contrast. The CT scans were reviewed by one author, a radiologist, without knowledge of second-look laparotomy findings. CT scans suggestive of persistent disease including scans demonstrating mesenteric thickening, bowel wall thickening and small amounts of ascites were classified as suggestive. In patients with suggestive scans with lesions not amenable to biopsy confirmation, secondqook laparotomy was performed. True positive results were defined as suspected disease on CT scanning which was histologically confirmed. The CT findings and results (but not CA-125 values) of 12 of the 45 patients at secondqook laparotomy were previously reported (8~ GCS, International Journal of Gynecolo~ical Cancer 6, t

3 Second-look laparotomy in advanced ovarian carcinoma 2I 5 Results Forty-five patients were studied. Forty had CA-125 levels elevated above 35 IU ml-1, and five had levels < 35 IU m1-1 at diagnosis. The characteristics of the population including CA-125, stage, histology, postoperative residual disease, and disease status at second-look evaluation are listed in Table 1. Among the 40 CA-125 positive patients, initial CA-125 levels varied from IU ml- 1 with a median value of 326 IU ml-1. Table 1. Characteristics of study population CA-125 positive CA-125 negative n patients n patients Stage 40 5 III 35 4 IV 5 1 Histology Serous 31 3 Endometrioid 5 1 Mucinous 3 0 Other 1 1 Grade Postoperative residual disease None 4 3 < 2 cm 32 2 >2 cm 4 0 Results of second-look 20 4 Microscopic positive 4 0 Macroscopic positive 16" 1 *Disease status determined in two patients with CT-directed needle biopsy. On abdominal/pelvic CT scan only two of the 45 patients (4.4%) had a positive scan which could be confirmed by CT-directed biopsy. In both cases persistent disease involving the liver was histologically confirmed. Both patients had stage IIIC ovarian carcinoma and had no evidence of hepatic metastases on the CT scan obtained prior to their primary surgery. One patient experienced significant bleeding after her liver biopsy, requiring hospitalization for 4 days and transfusion of two units of red blood cells. The sensitivity of CT scans for persistent disease for all patients was 9.5%. Twenty of the 40 patients (50%) who normalized their CA-125 levels after chemotherapy had persistent disease at second-look laparotomy. Sixteen of these 20 had gross evidence of persistent disease, nine of whom had tumor implants measuring greater than 2 cm. Only two of nine patients could be diagnosed by CT IGCS, International Journal of Gynecological Cancer 6, directed needle biopsy. In this group of patients, the sensitivity of abdominal/pelvic CT scan for persistent disease that could be evaluated by biopsy was 10%. For those patients who had normalized their CA-125 level, the positive predictive value and negative predictive value were 100% and 53% respectively. When analyzed at an initial critical threshold CA-125 level of 100 IU m1-1, 6.3% of patients had positive CT scans, and the sensitivity was 11.8%. The negative and positive predictive values were 50% and 100%, respectively. At an initial critical threshold value of 500 IU ml-1, none of patients had positive CT scans and the negative predictive value was 50%. Five patients with normal CA-125 levels prior to initial treatment underwent abdominal/pelvic scans before second-look laparotomy. In four of the five patients, CT scans and second-look laparotomy were negative. One patient with a suggestive CT scan was found to have persistent 2 mm disease at second-look surgery. For all 45 patients, when CT findings suggestive of persistent disease were included, CT scans had a sensitivity of 52% and specificity of 75% (Table 2). Eleven cases had a positive CT scan and positive findings at second-look surgery. Six patients had CT findings suggestive of persistent disease but were found to be negative at second-look surgery. Using findings suggestive of disease, the positive and negative predictive values of abdominal/pelvic CT scan were 65% and 64%, respectively. Comment Prior studies concerning the use of abdominal/pelvic CT scans before second-look laparotomy have reported sensitivities and predictive values that have varied widely (Table 3). Many previous publications addressing this issue have classified findings suggestive of persistent disease as positive. Additionally, a suggestive CT scan and a positive second-look laparotomy have been considered a true positive, irrespective of the correlation of sites of the abnormal CT and histologic findings. In the current study the site of suspected disease on CT scanning and sites and sizes of persistent disease at second-look laparotomy are compared. The limitation of CT scanning in the detection of small peritoneal implants has previously been recognized by many authors O'2"4"5"7'9). The introduction of CA-125 into the management of ovarian cancer patients has allowed greater precision in determining response to therapy and the presence of persistent disease (1 ). Although this marker has allowed better

4 216 P. G. Rose et. al. Table 2. Suggestive CT scan and surgico-pathologic findings Patient number/ct path correlation CT scan Surgico-pathology 1 True negative 2 True positive Positive, thick mesentary 3 True negative 4 True positive 5 False positive 6 True negative 7 False positive 8 False negative 9 False positive 10 True negative 11 True negative 12 True positive 13 False positive 14 True positive 15 True positive 16 False negative 17 True negative 18 False negative 19 False positive 20 True negative 21 True negative 22 True positive 23 True negative 24 True negative 25 True negative 26 True negative 27 True negative 28 False negative 29 True positive 30 False negative 31 True positive 32 True negative 33 False negative 34 True positive 35 True negative 36 True negative 37 False negative 38 True positive 39 True negative 40 False negative 41 False positive 42 True negative continued Positive, loculated ascites Positive, pleural effusions, thickened mesentery, ascites Positive, liver calcifications Positive, ascites Positive, liver Positive, cecum and descending colon Positive, liver and ascites Positive, right hemidiaphram, iliac regions Positive, left presacral mass Positive, liver Positive, thick mesentery, ascites Positive, pelvic mass Positive, right hemidiaphragm, liver lesions, adrenal glands Positive, loculated ascites, liver lesions, omentum Positive, pelvic mass Positive, pleural effusions, ascites in pelvis Macroscopic right diaphragm 1.2 cm, omentum 5 cm, mesenteric implant 3 cm Microscopic positive washings Macroscopic left and right infundibular pelvic ligament (3 cm), left and right parocolic gutter transverse colon, biopsy of liver calcifations negative Macroscopic multiple nodules 3 cm Macroscopic, evaluated by CT directed biopsy Macroscopic 1 mm cecum, sigmoid, washings Macroscopic, largest implant 2 mm Macroscopic, 2 cm rectal tumor, small nodules in ileum, omentum, and peritoneum Microscopic, positive washings only Macroscopic, 1-3 mm nodule of small bowel adhesion, right pelvic side wall nodule, left paracolic gutter Macroscopic, evaluated by CT-directed biopsy Macroscopic, right diaphragm (2 mm), sigmoid colon (2 mm), gastrocolic omentum (3 mm), appendix, cul de sac, left pelvic node Macroscopic, 3 cm pelvic mass, 4 mm small and large bowel Macroscopic, cul de sac (1 cm) Macroscopic, cul de sac (2 cm) Macroscopic, < 1 cm in greatest diameter in bladder serosa, ileum and sigmoid Microscopic, 2 mm right diaphragm Microscopic, serosal surface ascending colon Macroscopic, 10/89 positive + right peritoneum + liver capsule + peritoneum right gutt + small bowel serosa 1 cm cystic lesion of omentum + on path + omentum + rectal serosa (0.6 cm rectal nodule) + parasigmoid region K" 1996 IGCS, International Journal of Gynecological Cancer 6,

5 Second-look laparotomy in advanced ovarian carcinoma 2I' 7 Table 2. Suggestive CT scan and surgico-pathologic findings, continued Patient number/ct path correlation CT scan Surgico-pathology 43 False negative 44 False negative 45 True positive Positive, thick small bowel mesentary on left liver calcification, thick fascia on left Microscopic, omentum Macroscopic, 3 cm nodule small bowel mesentery Macroscopic, 1 cm nodule ant abdominal wall, 4 x 3 cm tumor mass in gastro-colic ligament Table 3. Predictive value of CT scan before second-look laparotomy Name (Ref) Year n Stage % Positive Sensitivity NPV PPV Stern et al. (1) I-III Goldhirsh et al. (2) II-IV NC NC NC NC Brenner et al. (3) I-IV Clarke-Pearson et al. (4) I-IV Silverman et al. (s) NS NC 0.40 NC NC Megibow et al. (6) III-IV Stehman et al. (7~ I-IV NC 0.26 Reuter et al. (s~ II-IV Lund et al. (9) IIB-IV Current CT positive 45 III-IV CT suggestive 45 III-IV NC = not able to be calculated; NS = not stated; NPV=negative predictive value; PPV = positive predictive value. selection of patients for second-look laparotomy, approximately one-half of patients with normal CA- 125 levels have persistent disease at second-look laparotomy (11-14). Even with normalization of CA-125 levels within 3 months of initial surgery, 36% of patients have persistent disease 5). In the current study, which was limited to patients who had normal CA-125 levels before second-look laparotomy, only two patients (4.4%) had CT evidence of definitive persistent disease which would obviate second-look surgery. Both of these patients had liver metastases. Since 47% of patients had persistent disease at surgery, the sensitivity of CT for definitive persistent disease was only 9.5%. Abdominal/pelvic CT scanning was highly specific (100%) for definitive persistent disease which could be histologically confirmed in both patients. Unfortunately, one patient developed significant bleeding following fine needle liver biopsy. This complication is rare and seen in only four of 1060 cases reported by Bret et al. (16). The use of different critical threshold values for CA-125 did not affect the negative predictive value significantly. When CT-findings suggestive of persistent disease were included, the sensitivity of abdominal/pelvic CT scans for persistent disease increased from 9.5% to 52%. However, the specificity decreased from 100% to 75%. Fifteen of the 20 CA-125 positive patients who had a positive second-look laparotomy, had gross disease which in nine cases was greater than 2 cm. In 1996 IGCS, International Journal of Gynecological Cancer 6, previous studies, patients who normalized their CA- 125 uniformly had less than 2 cm of persistent disease found at second-look laparotomy (11"12). Persistent disease of this size would be difficult to detect with computerized tomography. Patsner et al. reported the collaborative results of 125 patients who prior to diagnosis had elevated CA-125 levels and after completing chemotherapy underwent second-look laparotomy (17). Thirteen of the 102 patients (12.7%) who normalized their CA-125 had persistent disease at second-look measuring greater than 2 cm. Disease of this size is more likely to be detectable by CT scan, particularly if it is located in the retroperitoneum or organ parenchyma. Five patients in our study had normal CA-125 levels before diagnosis. Because of the small number of patients in this subgroup, no definite conclusions can be drawn. However, the accuracy of CT scans in this population should be similar to that reported before the introduction of the CA-125 assay. Our current results differ from our previously published results regarding the sensitivity of abdominal/pelvic CT scans in ovarian cancer (s). In the previous study, CA-125 levels were not utilized to determine eligibility for second-look laparotomy. The sensitivity of CT scans for persistent disease in this population was 84%. With recognition of persistent disease in patients with elevated levels of CA-125, our current study has a more select group of patients for

6 2 1 8 P.G. Rose et. al. second-look laparotomy. A second important factor is the amount of residual disease following initial cytoreductive surgery. In three consecutive studies from our institution, the percentage of patients optimally cytoreduced has increased significantly; 27%, 69%, 86% during , , respectively ~1s-2 ). The current study suggests that the sensitivity of abdominal/pelvic CT scans in a population of patients with normalization of their CA-125 is markedly less than reported in the literature when considering abdominal/pelvic CT scans that are definitively positive. With the more general categorization of positive for persistent disease including suggestive findings, as used in previous publications, the sensitivity of CT in this study is also lower when CA-125 values are considered. Abdominal/pelvic CT scans are extremely specific for the presence of definitive-appearing persistent disease before secondlook laparotomy. Patients with nodal disease or organ parenchymal disease are most likely to benefit from CT studies, since a positive biopsy could preclude second-look surgery. Patients with ovarian malignancies who do not have elevated CA-125 levels prior to initial treatment may also benefit from abdominal/ pelvic CT scanning as part of their disease status evaluation. References 1 Stern l, Buschema J, Rosenshein M, Siegelman S. Can computed tomography substitute for second-look operation in ovarian carcinoma? Gynecol Oncol 1981; 11: Goldhirsch A, Triller JK, Greiner R, Dreher E, Davis BW. Computed tomography prior to secondqook operation in advanced ovarian cancer. Obstet Gynecol 1983; 62: Brenner DE, Shaft MI, Jones HW, Grosh WW, Greco FA, Burnett LS. Abdominopelvic computed tomography: evaluation of patients undergoing second-look laparotomy. Obstet Gynecol 1985; 65: Clarke-Pearson DL, Bandy LC, Dudzinski M, Heaston D, Creasman WT. Computed tomography in evaluation of patients with ovarian carcinoma in complete clinical remission. J Am Med Assoc 1986; 255: Silverman PM, Osborne M, Dunnick NR, Bandy LC. CT prior to secondqook operation in ovarian cancer. Am J Radiol 1988; 150: Megibow AJ, Bosniak MA, Ho AG, Beller U, Hulnick DH, Bechman EM. Accuracy of CT in detection of persistent or recurrent ovarian carcinoma: correlation with second-look laparotomy. Radh~l 1988; 166: Stehman FB, Calkins AR, Wass JL, Smirz LR, Sutton GP, Ehrlich CE. A comparison of findings at secondqook laparotomy with preoperative computed tomography in patients with ovarian cancer. Gynecol Oncol 1988; 29" Reuter KL, Griffin T, Hunter RE. Comparison of abdominopelvic computed tomography results and findings at secondlook laparotomy in ovarian carcinoma patients. Cancer 1989; 63: Lund B, Jacobson K, Rasch L, Jenson F, Olesen K, Feldt- Rasmussen K. Correlation of abdominal ultrasound and computed topography scan with second- or third-look laparotomy in patients with ovarian carcinoma. Gynecol Oncol 1990; 37: Jacobs I, Bast RC. The CA-125 tumor associated antigen: A review of the literature. Hum Reprod 1989; 4: Niloff JM, Bast RC, Schaetzl EM, Knapp RC. Predictive value of CA-125 antigen levels in second-look procedures for ovarian cancer. Am J Obstet Gynecol 1985; 151: Berek JS, Knapp RC, Malkasain GD, Lavin PT, Whitney C, Niloff JM, Bast RC. CA 125 serum levels correlated with second-look operations among ovarian cancer patients. Obstet Gynecol 1986; 67: Rubin SC, Hoskins WJ, Hakes TB, Markman M, Reichman BS, Chapman D, Lewis JL. Serum CA 125 levels and the surgical findings in patients undergoing secondary operations for epithelial ovarian cancer. Am ] Obstet Gynecol 1989; 160: Patsner B, Day TG. Predictive value of CA-125 levels in advanced ovarian cancer. Am J Obstet Gynecol 1987; 156: Lavin PT, Knapp RC, Malkasian G, Whitney CW, Berek JC, Bast RC. CA 125 for the monitoring of ovarian carcinoma during primary therapy. Obstet Gynecol 1987; 69: Bret PM, Labadie M, Bretagnolle M, Paliard P, Fond A, Valette PJ. Hepatocellular carcinoma: diagnosis by percutaneous fine needle biopsy. Gastrointest Radiot 1988; 13: Pastner B, Orr JW, Mann WJ, Taylor PT, Partridge E, Allmen T. Does serum CA-125 level prior to second-look laparotomy for invasive ovarian adenocarcinoma predict size of residual disease? Gynecol Oncol 1990; 37: Griffin TW, Hunter RE, Cederbaum AI, et al. Treatment of advanced ovarian cancer with sequential combination chemotherapy. Cancer 1987; 60: Hunter RE, Griffin TW, Stevens S, et al. High dose/short duration cisplatinum/doxorubicin combination chemotherapy for advanced ovarian epithelial cancer. Cancer 1991; 68: Rose PG. The cavitational ultrasonic surgical aspirator for cytoreduction in advanced ovarian cancer. Am J Obstet Gynecol 1992; 166: Accepted for publication October 12, 1995 :! 1996 IGCS, International Journal qf Gynecological Cancer 6, ~