Neoadjuvant chemotherapy for advanced ovarian cancer: the role of cytology in pretreatment diagnosis

Neoadjuvant chemotherapy for advanced ovarian cancer: the role of cytology in pretreatment diagnosis

Available online at www.sciencedirect.com R Gynecologic Oncology 90 (2003) 644 – 650 www.elsevier.com/locate/ygyno Neoadjuvant chemotherapy for adv...

236KB Sizes 0 Downloads 62 Views

Available online at www.sciencedirect.com R

Gynecologic Oncology 90 (2003) 644 – 650

www.elsevier.com/locate/ygyno

Neoadjuvant chemotherapy for advanced ovarian cancer: the role of cytology in pretreatment diagnosis Peter E. Schwartz, M.D.a,* and Wenxin Zheng, M.D.b a

Department of Obstetrics and Gynecology, Yale University School of Medicine 333 Cedar Street New Haven, CT 06520, USA b Department of Pathology, Yale University School of Medicine 333 Cedar Street New Haven, CT 06520, USA Received 4 February 2003

Abstract Objective. The objective was to determine the role of cytology in the pretreatment evaluation of women with clinical findings consistent with ovarian cancer who are being considered for neoadjuvant chemotherapy. Methods. Pretreatment cytology slides were available for review from 60 of 72 consecutive patients treated with platinum-based neoadjuvant chemotherapy who were believed to have ovarian cancer based on clinical findings. Fifty of the 72 patients had evidence of both intraabdominal and extraabdominal tumor spread prior to treatment. Fifty-three of 66 patients had CA125 values ⬎500 U/mL, 34 being ⬎1500 U/mL. Pretreatment cytology was compared to surgical specimens obtained following chemotherapy. Results. Cytologic findings were consistent with ovarian cancer in 55 patients, not consistent with ovarian cancer in 4 cases, and insufficient for diagnosis in one case. Forty-seven of the 60 patients underwent surgery. Forty-two of 43 patients with cytology consistent with ovarian cancer had epithelial ovarian cancers at surgery. One had no pathologic evidence of disease. Three of the 4 patients thought not to have cytology consistent with ovarian cancer underwent surgery following neoadjuvant chemotherapy. Two had ovarian epithelial cancers and one had a mesonephric adenocarcinoma. The one patient with cytology insufficient for diagnosis also had an epithelial ovarian cancer at diagnosis. Conclusions. Cytology proved to be extremely helpful in supporting the clinical impression of an apparent advanced ovarian cancer. When the cytologic diagnosis does not match the clinical impression, communication between the cytologist or pathologist and the clinician is essential. © 2003 Elsevier Inc. All rights reserved.

Introduction Approximately 70% of epithelial ovarian cancers (EOC), the most common form of ovarian cancers, are not diagnosed until the disease has involved the upper abdomen or spread beyond the abdominal cavity [1]. Conventional treatment for advanced stage EOC is surgery to confirm that the cancer is of ovarian origin, aggressive “optimal” surgical tumor debulking, i.e., cytoreductive surgery, to remove all macroscopic disease, and postoperative combination chemotherapy, usually consisting of a platinum-type drug and a taxane to achieve prolonged progression-free and overall survival [2– 4]. Despite this aggressive approach, the overall * Corresponding author. Fax: ⫹1-203-785-4135. E-mail address: [email protected] (P.E. Schwartz). 0090-8258/03/$ – see front matter © 2003 Elsevier Inc. All rights reserved. doi:10.1016/S0090-8258(03)00376-7

5-year survival for patients with International Federation of Gynecology and Obstetrics (FIGO) Stage IIIC (the most common stage of this disease) and Stage IV disease generally remains in the 10 –20% range, with women continuing to succumb to the disease after 5 years [4 – 6]. Alternative strategies for the initial management of advanced ovarian cancer started in the late 1970s. One approach, interval cytoreduction surgery, has been proposed for women whose initial surgical cytoreduction was “suboptimal.” The latter is defined as leaving residual disease at the initial surgery that has a maximum diameter greater than 1 cm at any site in the abdominal cavity [7,8]. Suboptimally cytoreduced patients subsequently undergo “interval surgical cytoreduction” after receiving three cycles of chemotherapy. Postoperatively additional chemotherapy is administered. Unfortunately, patients who were not optimally

P.E. Schwartz, W. Zheng / Gynecologic Oncology 90 (2003) 644 – 650

cytoreduced at the initial operation failed to achieve the same survival as those who were optimally surgically cytoreduced at the initial surgery [7]. A third strategy has evolved to treat women with advanced stage ovarian cancer, i.e., neoadjuvant chemotherapy, which is the administration of chemotherapy prior to any attempt to perform cytoreductive surgery [9 –14]. This strategy was employed by us for a decade (1979 –1989) exclusively for those women who were medically too compromised at the time of diagnosis of an apparent advanced stage ovarian cancer to tolerate primary surgical cytoreduction [9 –11]. Beginning in 1989, and based on dramatic clinical responses frequently observed in medically compromised patients, diagnostic imaging criteria were developed to identify patients with advanced stage ovarian cancer who were unlikely to be optimally surgically cytoreduced at the initial surgery [11,15]. While strong concern had been expressed that the patient’s lives might be compromised with the neoadjuvant chemotherapy strategy, the long-term follow-up data now available demonstrate that the progression-free and overall survival is not impaired with this approach [11–15]. The selection criteria employed at Yale University to identify patients to be considered candidates for neoadjuvant chemotherapy for apparent advanced ovarian cancer are a physical examination consistent with an advanced stage ovarian cancer, diagnostic imaging studies consistent with an advanced stage ovarian cancer that is unlikely to be optimally cytoreduced at the initial operation, and cytologic or histologic specimens consistent with an EOC [11,15]. This article will review the pretreatment cytologic specimen findings from a series of patients referred to us, who then received neoadjuvant platinum-based combination chemotherapy for apparent advanced-stage ovarian cancers. The findings of the surgical specimen histologic evaluations obtained after neoadjuvant chemotherapy was completed are presented and compared to the pretreatment cytology diagnoses.

Material and methods Pretreatment cytology slides were available for review from 60 of 72 consecutive patients who were treated with neoadjuvant chemotherapy for clinically apparent advanced stage ovarian cancers (consistent with stage IIIC or IV disease) at the Yale University School of Medicine between 1979 and 1995. These slides were reviewed by a gynecologic pathologist (W.Z.). The patients’ characteristics are presented in Table 1. Fifty of the 72 (69.4%) patients had evidence of extraabdominal tumor spread prior to treatment, 55 had ascites, 24 had unilateral pleural effusions, and 13 had bilateral pleural effusions. Fifty-three of the 66 patients with known preoperative serum CA125 levels had values ⬎500 U/ml, 34 being ⬎1500 U/ml. Ascitic fluid was tapped from every patient. An average

645

Table 1 Patient characteristics Age (median range in years) Site of cancer Intraabdominal Extraabdominal Ascites Yes No Pleural effusions None Unilateral Bilateral Preoperative serum CA125 (U/ml) ⬍35 35–200 201–500 501–1500 1501–5000 ⬎5000 Unknown

68 (19–85) 22 50 55 17 35 24 13 5 5 3 19 27 7 6

of 30 to 60 ml of ascites was sent to the cytology laboratory at the Yale–New Haven Hospital. The samples were treated with Cell-Lite to lyse red blood cells and then spun down in 50-ml sterile plastic tubes. An average of one slide with Pap staining was made from each sample for cytologic examination. A cell block was also made when a cell pellet was visible. All pretreatment cytology slides were reviewed in a blinded fashion without knowing subsequent pathology diagnosis. There were characterized as (1) being consistent with EOC, (2) not being consistent with EOC, or (3) having an insufficient amount of material for evaluation. The following criteria were used to identify the presence of malignant cells consistent with an ovarian primary cancer based on well-established cytologic features published previously and our accumulated experience [16,17]: (a) Malignant cells with abundant cytoplasm not typical of mucinous carcinoma suggested serous carcinoma (Fig. 1). (b) Micropapillary structures or minimal architectural features suggesting papillary formation were viewed to be consistent with a serous carcinoma (Fig. 2), particularly when psammoma bodies were seen. (c) Malignant cells with vacuoles or a hint of clear cell differentiation suggested clear cell carcinoma or a serous or an endometrioid carcinoma with clear cell features (Fig. 3). (d) The presence of prominent nucleoli were looked for as they frequently were present in serous and clear cell carcinomas (Fig. 4). (e) Malignant cells with ciliation were also compatible with a primary ovarian cancer. (f) When endocervical type malignant cells were present, they were most compatible with a mullerian or ovarian primary cancer. When malignant cells with intestinal differentiation were present the differential diagnosis was broader.

646

P.E. Schwartz, W. Zheng / Gynecologic Oncology 90 (2003) 644 – 650

Fig. 1. Peritoneal fluid sample derived from a 65-year-old woman with an advanced ovarian cancer prior to receiving neoadjuvant chemotherapy. Clusters and individual malignant cells show abundant eosinophilic cytoplasm. These features are consistent with an ovarian serous carcinoma. Post-neoadjuvant chemotherapy surgical specimens confirmed that the patient had an ovarian serous adenocarcinoma.

Fig. 2. Peritoneal fluid sample derived from a 47-year-old woman with the clinical impression of an advanced ovarian cancer prior to receiving neoadjuvant chemotherapy. Clusters of serous type tumor cells form a pattern reminiscent of papillary structures. They are characteristic of papillary serous carcinoma, although the nuclear grade for this case was low. Psammoma bodies were seen in a different field (see insert in upper right corner). With the clinical setting mentioned in the text, the finding is basically diagnostic for ovarian papillary serous carcinoma.

P.E. Schwartz, W. Zheng / Gynecologic Oncology 90 (2003) 644 – 650

647

Fig. 3. Peritoneal fluid sample derived from a 65-year-old woman prior to her receiving neoadjuvant chemotherapy. Clusters of tumor cells show abundant vacuoles in the cytoplasm and form a pattern reminiscent of glandular and papillae-like structures. The findings are suggestive of clear cell carcinoma or an adenocarcinoma with clear cell features. The findings are strongly compatible with primary ovarian cancer. A clear cell carcinoma of the ovary was confirmed in post-neoadjuvant chemotherapy specimens.

Fig. 4. Peritoneal fluid sample derived from a 73-year-old woman with a clinical impression consistent with ovarian cancer prior to her receiving neoadjuvant chemotherapy. Clusters of serous-type tumor cells show an open nuclear chromatin pattern with prominent nucleoli. These features are consistent with an ovarian serous carcinoma. Post-neoadjuvant chemotherapy samples confirmed the presence of ovarian serous carcinoma.

648

P.E. Schwartz, W. Zheng / Gynecologic Oncology 90 (2003) 644 – 650

Results

Discussion

Cytologic findings in specimens obtained from women with physical examinations and diagnostic imaging findings compatible with an advanced stage ovarian cancer were consistent with an EOC in 55 cases, not consistent with ovarian cancer in 4 cases, and insufficient for diagnosis in one case. Forty-seven of the 60 patients (78.3%) with cytologic specimens available for review underwent surgery following neoadjuvant chemotherapy. Thirteen patients (21.7%) were considered medically unfit to tolerate surgery. The histologic findings from surgical specimens obtained following neoadjuvant chemotherapy were compared to the preoperative findings in the cytology specimens (Table 2). Serous cancers were the dominant type followed by undifferentiated epithelial cancers. The pathology specimens from one operation were unavailable for review. Three of the four patients interpreted cytologically to have a carcinoma not compatible with ovarian cancer did undergo surgery following neoadjuvant chemotherapy for presumed ovarian cancer. Their treatment was based on their clinical and diagnostic imaging findings. All had pelvic masses consistent with an ovarian cancer. Pretreatment, two of these patients had evidence of intra- and extraabdominal spread of disease. One had ascites, bilateral pleural effusions, and a CA125 of 1025 U/ml. The second patient had intra- and extraabdominal disease, bilateral pleural effusions, and a CA125 of 23 U/ml. Two patients had intraabdominal disease only, one of whom had clinical ascites associated with a serum CA125 of 870 U/ml. The remaining patient had no ascites but her CA125 was 4860 U/ml. At surgery one patient was found to have evidence of a poorly differentiated serous cancer, one had an undifferentiated adenocarcinoma, and the final patient had a mesonephric (wolffian) adenocarcinoma arising in the left paracervical area and involving both ovaries, fallopian tubes, uterine and cervical serosa, the left ureter, and the appendix. Forty-two of the remaining 43 patients were found to have EOC at the time of surgery. One patient had a pathologic complete response but subsequently recurred and died of disease 46 months following the original diagnosis.

Ovarian cancer remains the major pelvic reproductive organ cancer health hazard for American women [1]. A lack of early warning signals and a lack of sensitive, clinically available diagnostic tests to detect early stage disease results in most women not being diagnosed until the cancer is in an advanced stage, and cure is frequently elusive [1,17]. Neoadjuvant chemotherapy is now gaining popularity as a means of improving the physical and emotional trauma associated with initial ovarian cancer treatment [9 –14]. The subjective improvement in the sense of well-being when patients are given chemotherapy first is well recognized compared to when aggressive cytoreductive surgery is the first step in the treatment regimen [11–14]. The reduction in operating room time, estimated blood loss, intensive care unit stays, and overall hospital stays is substantial in women who undergo neoadjuvant chemotherapy prior to cytoreductive surgery [11,12]. Most important is that the overall and progression-free survivals are not impaired with this approach [11–14]. Concern has been raised about the reliability of using cytology to diagnose ovarian cancer. The present study represents the largest one to record cytology as a means to identify malignant cells consistent with ovarian cancer and then treating the patient with neoadjuvant chemotherapy. It has been our view that the pathologist is not required to confirm the diagnosis of ovarian cancer based on cytologic material. Rather it has been our approach to ask the pathologist to confirm that the cytologic findings are consistent with ovarian cancer. It would be ideal if our pathologist or cytologist could exclude the possibility of ovarian primary or definitely confirm the ovarian primary. However, the reality is the opposite simply because of the limitations of cytology. Malignant cells were readily apparent in the cytology samples from all but one patient in this study. The malignant features included clusters and single pleomorphic tumor cells with high nuclear-to-cytoplasm ratios. There were no particular features distinguishing the malignant cells originating from ovarian cancers from those arising from nongynecologic organs. It is well known that the cytologic

Table 2 Cytologic assessment of pretreatment specimens vs histologic findings from surgery following neoadjuvant chemotherapy for apparent advanced-stage ovarian cancer Cytology

Total patients

Pathology from surgical staging Serous

Undifferentiated

Endometrioid

Othera

NEDb

2

3

1

Consistent with ovarian cancer Not ovarian cancer Insufficient

43

29

8

3 1

1 1

1

Total

47

31

9

a b

— —



Other, mixed epithelial (2); mesonephric adenocarcinoma (1); mixed mesodermal (1). No evidence of disease.

2

1

— —

— 4

1

P.E. Schwartz, W. Zheng / Gynecologic Oncology 90 (2003) 644 – 650

diagnosis of ovarian cancer is difficult [16]. However, the clinical setting presented in this study does not seek a definitive pathologic diagnosis but rather to gain additional support for the clinical impression derived from the physical examination, serology studies, and diagnostic imaging findings. Based on this particular pinpoint clinical managementrelated question, i.e., compatible or noncompatible with ovarian primary, we summarized the following cytologic features to lead us to the conclusion that a cytologic specimen is “compatible with an ovarian primary.” Ovarian adenocarcinomas are characterized by a predominance of smaller, less cohesive cancer cell clusters and papillae with irregular outlines and many single cells. The tumor cells are typically pleomorphic and appear syncytial with indistinct cell borders. The tumor cell cytoplasm is usually abundant and vacuolated [16]. Malignant cells with the above features are compatible with serous carcinoma, particularly when papillary architecture and/or psammoma bodies can be identified. We found the latter findings to be very useful even when a minimal amount are present in either smear slides or sections from cell blocks. Malignant cells with ciliation are rare but have been reported [18]. Histologically, ovarian serous carcinomas may be focally ciliated. Although we did not see any case with ciliated malignant cells in this series, our pathologists or cytologists are aware that the presence of ciliation does not necessarily rule out a malignancy being compatible with an ovarian serous adenocarcinoma. It is almost impossible to use cytology alone to differentiate an ovarian serous carcinoma from a peritoneal serous carcinoma or a peritoneal mesothelioma. Clinically, this may not be an issue since the management plans for ovarian or peritoneal serous carcinomas are similar. Peritoneal mesotheliomas are unlikely to be a concern because they have different clinical findings which are less likely to lead to a concern for an ovarian cancer being present. Similarly, cytology findings are useless to differentiate malignancies arising in the Mullerian system or between welldifferentiated carcinomas and borderline tumors. Fortunately, pelvic examinations, direct biopsies, and diagnostic imaging studies are able to exclude endometrial or cervical primaries in most scenarios. We pay less attention to mitotic figures in these cytology samples since all advanced-stage cancers have high mitotic counts. In the current study, we did not use any immunohistochemical markers to help confirm the clinical impression in difficult cases or those cases that were not consistent with an ovarian primary. This was mainly because these cases had no cell block(s) available for additional slide cutting. This may be related to the way we collect specimens in our clinical practice. Classically, for a cytology laboratory to make a cell block from an ascites sample, one needs more than a half million cells in the specimen. The majority of our samples did not contain enough cells to make a cell block. Cytokeratin markers (CK7 and CK20) are useful to differentiate ovarian and gastrointestinal cancers, mucin markers (MUC2 and MUC5A) are useful for ovarian versus appen-

649

diceal cancers, and breast marker (GCDFP 15) may identify breast metastasis to the ovaries. In certain situations, particularly when clinical findings also raise concern for metastasis, a panel of immunomarkers should be performed. Therefore, we recommend that clinicians try their best to provide more cellular samples in order to yield an adequate amount of cells if a cell block is needed for further evaluation. The clinician must integrate clinical and laboratory data to formulate the most likely diagnosis when confronting a patient with intraabdominal carcinomatosis. A careful medical history, physical examination, and diagnostic imaging studies are invaluable in diagnosing an apparently advanced-stage ovarian cancer. Obtaining cytologic or histologic evidence of a disease process compatible with ovarian cancer further supports the clinical impression. Four patients in this series did not have cytologic evidence compatible with ovarian cancer but did not, by thorough diagnostic evaluation, have an obvious alternative origination site for the cancer. Two of the 4, subsequent to receiving neoadjuvant chemotherapy, underwent surgery that confirmed the clinical diagnosis of an ovarian cancer and one was found to have a mesonephric adenocarcinoma. The remaining patient was not physically able to tolerate a surgical procedure. Only one of the 60 patients was believed to have insufficient cytologic material to make a diagnosis of a cancer consistent with an ovarian primary. She received neoadjuvant chemotherapy for an apparent advanced-stage ovarian cancer and subsequently was found to have a serous cancer of the ovary when she did undergo surgery. In our experience successful use of cytologic material to confirm the presence of an advanced ovarian cancer requires excellent communication between the clinician and the cytologist or pathologist. Additionally, a cytologist or pathologist with special training in gynecologic cytology and pathology may prove to be critical for making this approach successful. When the cytologic diagnosis does not match the clinical impression, communication between the cytologist or pathologist and the clinician is essential. It has been advocated by some investigators that all women with advanced-stage disease that appears compatible with ovarian cancer undergo laparoscopic assessment to confirm the diagnosis of ovarian cancer through laparoscopic-directed biopsies and to use laparoscopy to assess surgical cytoreducibility [12,13]. Since most patients with CT evidence of advanced intraabdominal carcinomatosis compatible with nonresectable ovarian cancer tend to have poorly differentiated carcinomas, laparoscopic-directed biopsies may be of limited use. If one considers the costs of laparoscopic procedures, there may be little gain at a high price. However, laparoscopy may have a role when cytology is clearly at odds with the clinical data. The latter circumstance may have been present in five cases in the current series. More important was the fact that advancedstage nonovarian cancers were not being mistaken for ovar-

650

P.E. Schwartz, W. Zheng / Gynecologic Oncology 90 (2003) 644 – 650

ian cancer and patients were not receiving inappropriate chemotherapy. At present a prospective randomized trial comparing six cycles of platinum-based regimens following cytoreductive surgery or three cycles of platinum-based neoadjuvant chemotherapy followed by surgery and then three additional chemotherapy cycles is being conducted in Europe under the direction of the European Organization for Research and Treatment of Cancer (EORTC) [13]. A total of 742 women will be randomized to two arms of this study. The EORTC research trial may provide significant information on the value of neoadjuvant chemotherapy and may help us further in understanding the role of cytology in the pretreatment diagnosis of advanced ovarian cancer. The authors remain comfortable in continuing to use cytology as a means of supporting the pretreatment clinical diagnosis of advancedstage cancer consistent with EOC in women being considered for neoadjuvant chemotherapy.

References [1] Cannistra SA. Cancer of the ovary. N Engl J Med 1993;329:1550 –9. [2] Schwartz PE. Surgical management of ovarian cancer. Arch Surg 1981;116:99 –106. [3] Hoskins WJ. Surgical staging and cytoreductive surgery of epithelial ovarian cancer. Cancer 1993;71:1534 – 40. [4] McGuire WP, Hoskins WJ, Brady MF, et al. Cyclophosphamide and cisplatin compared with paclitaxel and cisplatin in patients with stage III and stage IV ovarian cancer. N Engl J Med 1996;334:1– 6. [5] Hoskins WJ, Bundy BN, Thigpen JT, Omura GA. The influence of cytoreductive surgery on recurrence-free interval and survival in small volume stage III epithelial ovarian cancer: a Gynecologic Oncology Group study. Gynecol Oncol 1992;47:159 – 66.

[6] Harries M, Gore M. Part I: chemotherapy for epithelial ovarian cancer—treatment at first diagnosis. Lancet Oncol 2002;3:529 –36. [7] Neijt JP, ten Bokkel Huinik WW, van der Burg MEL, et al. Randomized trail comparing two combination chemotherapy regimens (CHAP-5 v CP) in advanced ovarian cancer. J Clin Oncol 1987;5: 1157– 68. [8] van der Burg MEL, van Lent M, Buyse M, et al. The effects of debulking surgery after induction chemotherapy on the prognosis in advanced epithelial ovarian cancer. N Engl J Med 1995;332: 629 –34. [9] Chambers JT, Chambers SK, Voynick IM, Schwartz PE. Neoadjuvant chemotherapy in Stage X ovarian carcinoma. Gynecol Oncol 1990; 37:327–331. [10] Schwartz PE, Chambers JT, Makuch R. Neoadjuvant chemotherapy for advanced ovarian cancer. Gynecol Oncol 1994;53:33–7. [11] Schwartz PE, Rutherford TJ, Chambers JT, Kohorn EI, Thiel RP. Neoadjuvant chemotherapy for advanced ovarian cancer: Long-term survival. Gynecol Oncol 1999;72:93–9. [12] Surwit E, Childers J, Atlas M, et al. Neoadjuvant chemotherapy for advanced ovarian cancer. Int J Gynecol Cancer 1996;6:356 – 61. [13] Vergote IB, DeWever I, Decloedt J, et al. Neoadjuvant chemotherapy versus primary debulking surgery in advanced ovarian cancer. Semin Oncol 2000;27(Suppl 7):31– 6. [14] Kayikciog Lu F, Kose MF, Boran N, Caliskan E, Tulunay G. Neoadjuvant chemotherapy or primary surgery in advanced epithelial ovarian cancer. Int J Gynecol Cancer 2001;11:466 –70. [15] Nelson BE, Rosenfield AT, Schwartz PE. Preoperative abdominopelvic computed tomographic prediction of optimal cytoreduction in epithelial ovarian carcinoma. J Clin Oncol 1993;11:166 –72. [16] Sneige N, Fanning CV. Peritoneal washing cytology in women: diagnostic pitfalls and clues for correct diagnosis. Diagn Cytopathol 1992;8:632– 40 Discussion 640 –2. [17] Taylor KJW, Schwartz PE. Cancer screening in a high risk population: a clinical trial. Ultrasound Med Biol 2001;27:461– 6. [18] Kobayashi TK, Teraoka S, Tsujioka T, Yoshida Y. Ciliated ovarian adenocarcinoma cells in ascitic fluid cytology: report of a case with immunocytochemical features. Diagn Cytopathol 1988;4:234 – 8.