bs_bs_banner
doi:10.1111/jog.12436
J. Obstet. Gynaecol. Res. Vol. 40, No. 7: 1931–1939, July 2014
Prediction of histological types of endometrial cancer by endometrial cytology Masao Okadome1, Toshiaki Saito1, Naoko Nishiyama2, Kazuya Ariyoshi1, Kumi Shimamoto1, Takako Shimada1, Keisuke Kodama1, Shogo Imamura2, Ken-ichi Nishiyama3 and Kenichi Taguchi2 1
Gynecology Service, National Kyushu Cancer Center, 2Department of Pathology, National Kyushu Cancer Center, and Department of Pathology, Japanese Red Cross Fukuoka Hospital, Fukuoka, Japan
3
Abstract Aim: Few studies have examined the accuracy of preoperative endometrial cytology in diagnosing low- and high-risk histology in women with endometrial cancer (EC). This single-institutional retrospective study compared the accuracy of endometrial cytology and biopsy in preoperatively predicting low-risk and high-risk histology of EC. Methods: Between January 2006 and March 2013, 198 women with EC were examined by endometrial cytology, endometrial biopsy and hysterectomy specimen in National Kyushu Cancer Center. Among these women, 110 had endometrial cytology samples available to compare with endometrial biopsy, and were enrolled in our study (mean age ± standard deviation: 59.57 ± 10.32 years). Single-use plastic endometrial suction curettes were used in 12 of the 110 cases and thin metallic curettes for the rest. Results: For type 2 EC, which includes grade 3 endometrioid adenocarcinoma and non-endometrioid histology, biopsy was 67.6% sensitive (25/37) and 84.9% specific (62/73); whereas cytology was 70.3% sensitive (26/37) and 91.8% specific (67/73). Cytology precisely diagnosed only one of 14 cases of serous carcinoma, but it diagnosed 11 of the 14 cases as type 2 EC, and its accuracy in distinguishing EC types was not inferior to endometrial biopsy (10/14). For EC, 9.1% (10/110) were unevaluable using biopsy, significantly more than the 0% (0/110) by cytology (P = 0.002). Conclusion: Although preoperative prediction of serous carcinoma was difficult, endometrial cytology had a higher evaluable rate for EC types. Endometrial cytology may complement endometrial biopsy in preoperative women with EC. Key words: endometrial cytology, endometrial neoplasms/pathology/surgery, endometrial sampling, neoplasm grading, preoperative care.
Introduction Preoperative endometrial sampling with a Pipelle is sensitive and accurate for diagnosis of high-grade endometrial tumors, including tumors with nonendometrioid histology.1 In Japan, endometrial cytology is prevalently used to detect endometrial cancer
(EC) as it is less painful and more cost-effective.2,3 Lymphadenectomy is proposed to be advantageous in patients with EC when preoperative histology indicates high-risk disease, for example, International Federation of Gynecology and Obstetrics (FIGO) grade 3 endometrioid or non-endometrioid histology if accurate frozen sections are unavailable.4
Received: October 17 2013. Accepted: February 23 2014. Reprint request to: Dr Masao Okadome, Gynecology Service, National Kyushu Cancer Center, 3-1-1 Notame, Minami-ku, Fukuoka 811-1395, Japan. Email: [email protected]
© 2014 The Authors Journal of Obstetrics and Gynaecology Research © 2014 Japan Society of Obstetrics and Gynecology
1931
M. Okadome et al.
Few studies have examined the accuracy of preoperative endometrial cytology in diagnosing low- and high-risk histology in patients with EC. A cytological scoring system for endometrioid adenocarcinoma was useful for predicting histological grade.3 Here, in a single-institution retrospective study, we compared endometrial cytology and endometrial biopsy in predicting low- and high-risk histology in preoperative women with EC.
Methods By reviewing the database, we found 270 women who underwent surgery for EC by the Gynecology Service of the National Kyushu Cancer Center from January 2006 to March 2013. A total of 198 women were examined by preoperative endometrial cytology, endometrial biopsy and hysterectomy specimen. Among these, we enrolled 110 women for whom endometrial cytology samples were available for review. We used thin metallic curettes when preoperative endometrial biopsies were performed from 2006 through March 2012 and began to use Pipet Curets (Cooper Surgical, Trumbull, CT, USA) which is similar to Pipelle, a single-use flexible plastic endometrial suction curette, from April 2012. Attending physicians collected biopsy specimens using thin metallic curettes in uterine cavities, with 2–4 passes according to the patient’s discomfort. Biopsy specimens were preoperatively diagnosed by two pathologists (K. N. and K. T.) using hematoxylin–eosin (HE)-stained sections. One gynecologic pathologist/ oncologist (K. A.) reviewed the biopsy specimens that were originally inspected by the two general pathologists (K. N. and K. T.). Endometrial cytology was performed by using Endocyte (Laboratoire CCD, Paris, France). The Endocyte was inserted into the uterine cavity and cells on the tips were spread onto glass slides. Endometrial cell slides were placed in 95% ethanol and then were processed by Papanicolaou stain. Cytological review was performed by two cytotechnologists (N. N. and S. I.) at National Kyushu Cancer Center, Department of Pathology, who were blinded to information about final histological diagnoses of hysterectomy specimens. When the results differed between the cytotechnologists, a final diagnosis was determined after discussion between them. Histological grade was assigned using published criteria. Grade 1–3 endometrioid adenocarcinoma was diagnosed by a cytological scoring system that used rates of cluster formation, nuclear pleomorphism, nuclear size,
1932
chromatin structure, chromatin distribution, presence of euchromatin, large nucleoli, histiocytes and necrosis, and absence of normal endometrial cells.3 Cytological diagnosis of uterine serous carcinoma was based on findings of tumor diathesis, psammoma bodies and papillary clusters composed of tumor cells with enlarged nuclei and numerous nucleoli.5 Cytological diagnosis of clear cell carcinoma is based on findings of adenocarcinoma cells with a translucent clear cytoplasm, centrally located nuclei, large and prominent nucleoli, and spherical aggregates of cells arranged in a single layer to a so-called mirror ball pattern.6 Histological types were divided into two groups: type 1 EC, including grade 1 and 2 endometrioid adenocarcinoma; and type 2, EC including grade 3 endometrioid adenocarcinoma, serous carcinoma and clear cell adenocarcinoma.7 Carcinosarcoma, squamous cell carcinoma and poorly differentiated adenocarcinoma of non-endometrioid type were classified as type 2 EC in the present study. Each endometrial cell slide was evaluated and categorized into type 1 EC or type 2 EC; specific histological types were also predicted for each slide.
Immunostaining for p53 Immunohistochemistry (IHC) for p53 was analyzed in 12 of 14 specimens of serous carcinoma and one specimen of serous carcinoma + clear cell carcinoma. The IHC used mouse monoclonal anti-p53 antibody (DO-7; Novocastra, Newcastle, UK), and was processed by Leica BOND-III automated system (Leica Microsystems, Tokyo, Japan) according to manufacturer’s instructions. Positive and negative controls used for IHC were histological sections that already had been established as either positive or negative. Statistical analyses Fisher’s exact test was used to compare endometrial biopsies with endometrial cytology on agreement of EC types and specific histologies, and on rates of overdiagnosis, underdiagnosis and unevaluable cases of EC types. P < 0.05 was considered significant. Analyses used Dr SPSS II software, version 11.0.1 (SPSS Japan, Tokyo, Japan).
Results The age distribution of 198 women of the study population was 27–85 years (mean age ± standard deviation [SD], 57.33 ± 10.77 years). Of the 198 women who provided preoperative endometrial cytology, endometrial
© 2014 The Authors Journal of Obstetrics and Gynaecology Research © 2014 Japan Society of Obstetrics and Gynecology
Prediction of endometrial cancer types
biopsy and hysterectomy specimens, 10 women showed no atypical cells and no atypical tissues in their cytology or biopsy specimens. For nine women, no atypical cells were found in their cytology specimens but atypical tissues were found in their biopsy specimens. For four women, no atypical tissues were found in their biopsy specimens but atypical cells were found in their cytology specimens. In all, no atypical cells were found in 19 of 198 (9.60%) cytology specimens and no atypical tissues were found in 14 of 198 (7.07%) biopsy specimens. The difference was not significant (P = 0.468). Of the remaining 175 women, we were able to review both cytology specimens and biopsy specimens from 110 women (mean age ± SD, 59.57 ± 10.32 years). The other specimens were not available for review because they had been returned to the hospitals or clinics where patients were initially examined. Figure 1 shows one of 13 cases of grade 3 endometrioid adenocarcinoma. In this case, preoperative diagnosis by endometrial biopsy was grade 1 endometrioid adenocarcinoma (Fig. 1a) but the diagnosis for the resected uterus was grade 3 endometrioid adenocarcinoma (Fig. 1b). The resected specimen had some well-differentiated adenocarcinoma but more
(a)
(c)
than 50% of its cells indicated solid proliferation of carcinoma cells (Fig. 1c). A review of the preoperative endometrial cytology of this case using a cytological scoring system3 was grade 3 endometrioid adenocarcinoma (Fig. 1d,e). Table 1 shows detailed histological types predicted by biopsy and cytology. Final diagnoses were histological diagnoses of resected uterus specimens. Endometrial biopsies provided 10 cases that could not be classified for detailed histological type, including five for adenocarcinoma not otherwise specified (AC NOS), one for carcinoma not otherwise specified (Ca NOS), and three for atypical cells or glands. ‘AC NOS’ means that adenocarcinoma tissue was seen but the collected tissue was scant and pathologists could not diagnose it as endometrioid adenocarcinoma or non-endometrioid adenocarcinoma. ‘Ca NOS’ means that carcinoma tissue was seen but the collected tissue was too small to be specified. ‘Atypical cells or glands’ mean that atypical cells or glands were seen but most of the specimens were mucous materials or necrosis. Endometrial cytology showed no case to be unevaluable (Table 1). Of 14 patients with serous carcinoma ± endometrioid adenocarcinoma, one case was precisely
(b)
(d)
(e)
Figure 1 Histological and cytological appearance of one case with grade 3 endometrioid adenocarcinoma. (a) Histological specimen preoperatively obtained by a thin metallic curette. (b) Histological specimen from hysterectomized uterus. (c) Specimen from hysterectomized uterus: low-powered appearance. More than 50% of the tumor showed solid proliferation, as seen in the right side of the figure. (d,e) Cytological appearances of the endometrial smear for the same case, which was preoperatively obtained by Endocyte. Nuclear pleomorphism is mild to moderate, but tumor diathesis is remarkable and euchromatin is noticeable. The cytological score of this case was >20; cytological diagnosis was grade 3 endometrioid adenocarcinoma (Papanicolaou stain, original magnification ×400).
© 2014 The Authors Journal of Obstetrics and Gynaecology Research © 2014 Japan Society of Obstetrics and Gynecology
1933
1934
1 1
1
1
1
1 2 2
1 2 1
Numbers in boldface agree with final histological diagnoses of hysterectomized specimens. Underscored numbers were within same endometrial types (e.g. type 1 endometrial cancer and type 2 endometrial cancer). AC, adenocarcinoma; AEH, atypical endometrial hyperplasia; C, clear cell carcinoma; Ca, carcinoma; CS, carcinosarcoma; EM, endometrioid adenocarcinoma; G1, grade 1 EM; G2, grade 2 EM; G3, grade 3 EM; NOS, not otherwise specified; S, serous carcinoma; Sm, small cell carcinoma.
1
7 7
1
1
4 1
1 1
2
3 2 1
5 15 5 3 1 36 10 1 1 1 1 1 2 1
2 2 2 7 1 1 1 8 1 1 7 1 38 13 2 1 3 G1 G2 G3 S ± EM S+C C ± EM CS
46 27 13 14 1 4 5
AEH
Diagnostic measures Endometrial biopsy diagnosis (reviewed diagnosis) Endometrial cytology diagnosis (reviewed diagnosis) G1 G2 G3 S ± EM S + C C ± AC Ca Atypical AEH G1 G2 G3 G3 + Sm S ± EM C ± EM CS EM NOS NOS cells or glands Final diagnosis
Table 1 Detailed histological types predicted by endometrial biopsy and endometrial cytology
M. Okadome et al.
diagnosed with serous carcinoma using endometrial cytology (Table 1); in that case, psammoma bodies were seen on the slide. Other cytological diagnoses were seven cases of grade 3 endometrioid adenocarcinoma, two of clear cell carcinoma ± endometrioid adenocarcinoma, one of carcinosarcoma and three of grade 2 endometrioid adenocarcinoma. A case of clear cell carcinoma + serous carcinoma was also misdiagnosed as grade 2 endometrioid adenocarcinoma on cytology. All misdiagnosed serous carcinoma cases lacked psammoma bodies on the slides. Seven of 14 cases of serous carcinoma were correctly diagnosed by the gynecological pathologist/oncologist (K. A.). Only one of the 14 cases was correctly diagnosed as serous carcinoma by the cytotechnologists but 11 of 14 cases were diagnosed as type 2 EC and their accuracy on distinguishing type 2 EC was not inferior to endometrial biopsy (10/14). Three of the 14 cases of serous carcinoma ± endometrioid adenocarcinoma were not evaluated for detailed histological types on biopsy; they were two cases of AC NOS and one case of atypical cells. Of four cases of clear cell carcinoma ± endometrioid adenocarcinoma, two cases were diagnosed as having clear cell carcinoma components by endometrial biopsy, whereas all cases were correctly diagnosed by endometrial cytology. None of five cases of carcinosarcoma were correctly diagnosed by either endometrial biopsy or endometrial cytology (Table 1). Five cases of type 1 EC were overdiagnosed as clear cell carcinoma ± endometrioid adenocarcinoma by endometrial cytology, and two of the five cases were unevaluable by biopsy. One case of type 1 EC was overdiagnosed as clear cell carcinoma ± endometrioid adenocarcinoma by endometrial biopsy. Four cases of type 1 EC were overdiagnosed as serous carcinoma by endometrial biopsy (Table 1). For type 1 EC, endometrial biopsy was 80.8% (59/73) sensitive and 67.6% (25/37) specific; whereas cytology was 90.4% (66/73) sensitive and 70.3% (26/37) specific. For type 2 EC, endometrial biopsy was 67.6% sensitive (25/37) and 84.9% (62/73) specific; whereas cytology was 70.3% (26/37) sensitive and 91.8% (67/73) specific (Table 2). Agreement between biopsy results and final diagnosis for type 2 EC was 67.6% (25/37); agreement between cytology results and final diagnosis was 70.3% (26/37; P = 1.0; Tables 2,3). For EC types, the rate of unevaluable cases by biopsy was 9.1% (10/110)
© 2014 The Authors Journal of Obstetrics and Gynaecology Research © 2014 Japan Society of Obstetrics and Gynecology
Prediction of endometrial cancer types
Table 2 Type 1 and type 2 endometrial cancer predicted by endometrial biopsy and endometrial cytology Final histological type
n
Type 1 Type 2
73 37
Diagnostic measures Endometrial biopsy AEH Type 1 3
59 7
Type 2
NE
AEH
6 25
5 5
1
Endometrial cytology Type 1 Type 2 66 11
NE
6 26
Numbers in boldface indicate agreement between diagnosis and final histological type. AEH, atypical endometrial hyperplasia; NE, not evaluable.
Table 3 Agreement rates for endometrial cancer types, detailed histological types, overdiagnosis rate, underdiagnosis rate and unevaluable rate
Agreement of endometrial cancer types† Agreement of type 2 endometrial cancer Agreement of detailed histological types Overdiagnosis of endometrial cancer types Underdiagnosis of endometrial cancer types Unevaluable cases on endometrial cancer types
Endometrial biopsy
Endometrial cytology
P
84/110 (76.4%) 25/37 (67.6%) 61/110 (55.5%) 6/110 (5.5%) 10/110 (9.1%) 10/110 (9.1%)
92/110 (83.6%) 26/37 (70.3%) 63/110 (57.3%) 6/110 (5.5%) 12/110 (10.9%) 0/110 (0%)
0.238 1.0 0.892 1.0 0.823 0.002*
*Significant difference. †Endometrial cancer types are type 1 endometrial cancer and type 2 endometrial cancer. Type 1 endometrial cancer includes G1 endometrioid adenocarcinoma and G2 endometrioid adenocarcinoma. Type 2 endometrial cancer includes G3 endometrioid adenocarcinoma and non-endometrioid type cancer.
Table 4 Detailed histological types predicted by endometrial biopsy using Pipet Curets and endometrial cytology Final histological type G1 G2 G3 CS
G1 1 6 3 2
1 4 2
Pipet Curets G2 G3
Diagnostic measures Endometrial cytology C G1 G2 G3 G3 + Sm 1
1 1 2
1 1
5 1
2 1
1
Numbers in boldface indicate agreement between diagnosis and final histological type. C, clear cell carcinoma; G1, grade 1 endometrioid adenocarcinoma; G2, grade 2 endometrioid adenocarcinoma; G3, grade 3 endometrioid adenocarcinoma; Sm, small cell carcinoma.
and 0% (0/110) by cytology (Tables 2,3), which was significant (P = 0.002, Table 3). Agreements for EC types and detailed histological types, overdiagnosis and underdiagnosis of EC types were not significant (P = 0.238, 0.892, 1.0 and 0.823, respectively, Tables 1–3). The 10 cases unevaluable by biopsy included five cases of type 1 EC and five cases of type 2 EC. Three of the five type 1 EC cases and four of the five type 2 EC cases were correctly diagnosed by cytological reviews. The two misdiagnosed cases of type 1 EC were both grade 2 endometrioid adenocarcinoma; they were overdiagnosed as clear cell carcinoma. The misdiagnosed case of type 2 EC was serous carcinoma, which was underdiagnosed as grade 2 endometrioid adenocarcinoma.
There were 12 cases of EC for which Pipet Curets were used in endometrial biopsies among the cohort. Agreement between biopsy results using Pipet Curets and final diagnosis for specific histologies was 25.0% (3/12) and 66.7% (8/12) between endometrial cytology and final diagnosis (Table 4), which was not significant (P = 0.099).
p53 immunostaining Eleven of 12 tested specimens of serous carcinoma, and one tested specimen of serous carcinoma + clear cell carcinoma were p53+ (Table 5). The negative case (S6) was diagnosed as serous carcinoma by the gynecologic pathologist/oncologist (K. A.) who reviewed the HE-stained section and was diagnosed as grade 2
© 2014 The Authors Journal of Obstetrics and Gynaecology Research © 2014 Japan Society of Obstetrics and Gynecology
1935
M. Okadome et al.
Table 5 Immunocytochemical expression of p53 in preoperative endometrial biopsy specimens of serous carcinoma and serous carcinoma with clear cell carcinoma (reviewed diagnosis for both types)
S1 S2 S3 S4 S5 S6 S7 S8 S9 S 10 S 11 S 12 S 13 S 14 S+C
p53
Endometrial biopsy
Endometrial cytology
+ + + + + − + + + + + + ND ND +
AC NOS G1 G3 Serous Clear Serous Serous Serous Serous Clear AC NOS Serous Atypical cells Serous Serous
Clear + EM G3 Carcinosarcoma G2 G3 G2 Serous G3 Clear + EM G3 Clear + EM G3 G3 G3 G2
AC, adenocarcinoma; C, clear cell carcinoma; EM, endometrioid adenocarcinoma; G1, grade 1 EM; G2, grade 2 EM; G3, grade 3 EM; ND, not done; NOS, not otherwise specified; S, serous carcinoma.
endometrioid adenocarcinoma by the cytotechnologists (N. N. and S. I., Table 5). The three p53+ cases (S1, S2 and S11) were not diagnosed as serous carcinoma by review of HE-stained sections, but were diagnosed as type 2 EC by cytological review (Table 5). A p53+ serous carcinoma case and the p53– case are shown in Figure 2.
Discussion The Pipelle is generally considered to be the best device in detecting endometrial carcinoma, with detection rates of 99.6% and 91%, respectively, in postmenopausal and premenopausal women;8 it is 81% sensitive and more than 98% specific in detecting EC.8 Preoperative endometrial sampling with Pipelle is sensitive and accurate for the diagnosis of high-grade endometrial tumors, including tumors with non-endometrioid histology.1 In Japan, an attempt has been made to detect asymptomatic EC using endometrial cytology, and to clarify whether detection of asymptomatic EC correlates with reduced mortality.9 Another study evaluated the usefulness of cytological scoring in assessing morphological differentiation of endometrioid adenocarcinomas
1936
using endometrial smears, and concluded that the cytological scoring system applied for endometrioid adenocarcinoma was useful for predicting histological grade and tumor malignant potential.3 When no macroscopic extrauterine disease is present, tumor diameter is 2 cm or less, and preoperative findings show FIGO grade 1 or 2 endometrioid histology or complex and/or atypical hyperplasia (low-risk patients), the risk of positive lymph nodes or lymph node recurrence is less than 1% and the 3-year recurrence-free survival rate is 98.7%. In contrast, highrisk patients with grade 3 or non-endometrioid histology or mixed Müllerian tumors or intraoperative macroscopic extrauterine disease have significantly higher risk of positive lymph nodes or lymph node recurrence (27%) than low-risk patients and a lower 3-year recurrence-free survival rate (71.2%). In these high-risk patients, lymphadenectomy may be advantageous.4 In addition, multivariate analysis confirmed that tumor volume, high serum CA125 level and grade 3/serous adenocarcinoma were independent risk factors for lymph node metastasis among 211 patients with endometrial carcinoma. This validation study was carried out using only preoperative data.10 Consequently, preoperative diagnosis of type 2 EC disease is important. In the present study, we investigated whether endometrial cytology could predict grades of endometrioid adenocarcinoma and tumors of non-endometrioid histology or complement preoperative endometrial biopsy in diagnosing type 1 and type 2 EC. For type 1 and type 2 EC, endometrial cytology had a significantly lower rate of unevaluable cases than did endometrial biopsy when biopsy was done by a metallic curette. An advantage of endometrial cytology is that endometrial cells are scraped from the entire surface of the endometrial cavity with an Endocyte, which could affect results. Although only 12 cases used Pipet Curets, endometrial cytology tended to have a higher prediction rate for specific EC histologies than Pipet Curets. Endometrial cytology may also complement single-use endometrial suction curettes in predicting high-risk EC. The rate of predicting serous carcinoma and tumors of its related histology was low using endometrial cytology; only one case with psammoma bodies was precisely predicted in our study, though the cytotechnologist in our study was experienced with endometrial cytology and involved in a study using endometrial smears.11 A previous study reported important
© 2014 The Authors Journal of Obstetrics and Gynaecology Research © 2014 Japan Society of Obstetrics and Gynecology
Prediction of endometrial cancer types
(a)
(b)
(c)
(d)
Figure 2 (a,c) Hematoxylin–eosin and (b,d) immunohistochemical staining with monoclonal antibody p53. (a,b) Serous carcinoma with strong positive staining involving almost all epithelial cells but not the surrounding stroma. (c,d) Serous carcinoma with negative staining.
diagnostic findings of serous carcinoma without psammoma bodies, but showed clusters of tumor cells with papillary architecture, tumor diathesis, hyperchromatic nuclei, prominent nucleoli and anisonucleosis,5 which were also observed in some of our grade 2 or 3 endometrioid adenocarcinoma cases. If cytotechnologists did not suspect serous carcinoma on slides of endometrial carcinoma, those cases were diagnosed by the scoring system for endometrioid adenocarcinoma. Cases of serous carcinoma with cytological characteristics cited above were difficult to diagnose without psammoma bodies. Serous carcinoma misdiagnosed and treated as grade 3 endometrioid adenocarcinoma is not as potentially damaging to a patient as the reverse. Our study had three serous carcinoma cases that were misdiagnosed as grade 2 endometrioid adenocarcinoma and one case with serous carcinoma and clear cell carcinoma that was misdiagnosed as grade 2 endometrioid adenocarcinoma, all four by endometrial cytology. These four misdiagnoses could be potentially damaging to patients. Diagnosis of endometrial clear cell carcinomas by gynecologic pathologists is associated with a moderate level of interobserver variability.12 Clear cell carcinoma has core cytoarchitectural features that are of high diagnostic utility but also has a wide morphological spectrum.13 Pathologists may have trouble in diagnosing endometrial carcinoma when the carcinoma tissues include some clear cells. In cytology, carcinomas with
cells that show centrally located nuclei and translucent, clear cytoplasm may be diagnosed as clear cell carcinoma;6 such findings may be related to overdiagnosis. The characteristic feature of clear cell carcinoma of the genital tract is basement membrane-like substance. This hyaline extracellular material stains pinkish to purple-red in May–Grünwald–Giemsa stain preparations and is frequently observed within cancer cell clusters.14 Basement membrane material in scraped cytological specimens from patients with clear cell carcinoma showed metachromatic pink to purple staining in Diff-Quik-stained smears.15 Hyaline stroma was present in 93% of scraped cytology specimens of clear cell carcinoma.16 These stain methods may be useful to avoid overdiagnosis. Uterine serous carcinoma (USC) and its putative precursor, endometrial intraepithelial carcinoma (EIC), demonstrate strong p53 overexpression by immunohistochemistry, which suggests alteration of the p53 gene in their pathogenesis.17 When 21 cases of USC and nine cases of EIC were evaluated for mutations in the p53 gene using direct sequence analysis, 90% of USC and 78% of EIC were found to contain mutations.17 In our study, three of 12 tested cases of serous carcinoma were not diagnosed as such and were not diagnosed as type 2 EC by HE-stained sections of endometrial biopsy, but were diagnosed as type 2 EC by endometrial cytology. IHC for p53 could be useful in such cases, as they were p53+.
© 2014 The Authors Journal of Obstetrics and Gynaecology Research © 2014 Japan Society of Obstetrics and Gynecology
1937
M. Okadome et al.
Our study has some limitations. First, women who underwent endometrial biopsy and cytology were referred to our institution because they were already suspected of having EC by earlier biopsy and/or cytology; therefore, our findings do not necessarily indicate which modality could better screen for EC or predict EC type among women in the general population who may not be suspected of having EC. Second, most endometrial biopsies in this study were performed with thin metallic curettes; Pipet Curets were used in only 12 of the 110 cases. Third, the cytotechnologists in our study were aware that all 110 cases were EC, which obviously differs from the usual requirement to assess cells de novo. Finally, diagnostic qualities in the field of endometrial cytology are secured in most of the laboratories in Japan, but such qualities may be assured in a limited number of laboratories in other countries because endometrial cytology is not prevalently used to detect EC except Japan. Widespread acceptance of endometrial cytology may be needed in other countries. In conclusion, endometrial cytology has higher evaluable rate for type 1 and type 2 EC than does endometrial biopsy and could complement endometrial biopsy preoperatively. Endometrial cytology may have a fault in predicting serous carcinoma, though most misdiagnosed cases were diagnosed as type 2 EC. It may also occasionally overdiagnose type 1 EC as clear cell carcinoma. Endometrial biopsy may overdiagnose type 1 EC as serous carcinoma. We should carefully diagnose EC using endometrial biopsy and endometrial cytology in preoperative women, because both modalities have advantages in prediction. Endometrial cytology may also help verify endometrial biopsies performed with single-use endometrial suction curettes, such as the Pipelle or Pipet Curets; this merits further investigation.
Acknowledgments This study was supported by the National Cancer Center Research and Development Fund (23-A-17) and Grants-in-Aid for Cancer Research (no. 18-06, no. 10103749) from the Ministry of Health, Labor and Welfare of Japan.
Disclosure The authors have no conflicts of interest to declare.
1938
References 1. Huang GS, Gebb JS, Einstein MH, Shahabi S, Novetsky AP, Goldberg GL. Accuracy of preoperative endometrial sampling for the detection of high-grade endometrial tumors. Am J Obstet Gynecol 2007; 196: 243 e241–243 e245. 2. Nakagawa-Okamura C, Sato S, Tsuji I et al. Effectiveness of mass screening for endometrial cancer. Acta Cytol 2002; 46: 277–283. 3. Nishimura Y, Watanabe J, Jobo T, Hattori M, Arai T, Kuramoto H. Cytologic scoring of endometrioid adenocarcinoma of the endometrium. Cancer 2005; 105: 8–12. 4. AlHilli MM, Podratz KC, Dowdy SC et al. Preoperative biopsy and intraoperative tumor diameter predict lymph node dissemination in endometrial cancer. Gynecol Oncol 2013; 128: 294–299. 5. Hagiwara T, Kaku T, Kobayashi H, Hirakawa T, Nakano H. Clinico-cytological study of uterine papillary serous carcinoma. Cytopathology 2005; 16: 125–131. 6. Ohwada M, Suzuki M, Ohno T, Saito K, Sato I. Appearance of primary endometrial and ovarian clear cell adenocarcinoma 17 months postpartum. A case report. Acta Cytol 1998; 42: 765–768. 7. Voss MA, Ganesan R, Ludeman L et al. Should grade 3 endometrioid endometrial carcinoma be considered a type 2 cancer-a clinical and pathological evaluation. Gynecol Oncol 2012; 124: 15–20. 8. Dijkhuizen FP, Mol BW, Brolmann HA, Heintz AP. The accuracy of endometrial sampling in the diagnosis of patients with endometrial carcinoma and hyperplasia: A metaanalysis. Cancer 2000; 89: 1765–1772. 9. Jobo T, Arai T, Sato R, Kuramoto H. Clinicopathologic relevance of asymptomatic endometrial carcinoma. Acta Cytol 2003; 47: 611–615. 10. Todo Y, Okamoto K, Hayashi M et al. A validation study of a scoring system to estimate the risk of lymph node metastasis for patients with endometrial cancer for tailoring the indication of lymphadenectomy. Gynecol Oncol 2007; 104: 623– 628. 11. Okadome M, Saito T, Tsukamoto N, Nishi K, Nishiyama N, Nagata E. Endometrial scraping cytology in women with extragenital malignancies. Acta Cytol 2006; 50: 158– 163. 12. Fadare O, Parkash V, Dupont WD et al. The diagnosis of endometrial carcinomas with clear cells by gynecologic pathologists: An assessment of interobserver variability and associated morphologic features. Am J Surg Pathol 2012; 36: 1107–1118. 13. Fadare O, Zheng W, Crispens MA et al. Morphologic and other clinicopathologic features of endometrial clear cell carcinoma: A comprehensive analysis of 50 rigorously classified cases. Am J Cancer Res. 2013; 3: 70–95. 14. Atahan S, Ekinci C, Icli F, Erdogan N. Cytology of clear cell carcinoma of the female genital tract in fine needle aspirates and ascites. Acta Cytol 2000; 44: 1005–1009. 15. Jimenez-Heffernan JA, Vicandi B, Gonzalez-Peramato P et al. Cytologic features of clear cell carcinoma of the female genital tract. Diagnostic value of the ‘raspberry body’ in nonexfoliative cytologic specimens. Acta Cytol 2004; 48: 47–51.
© 2014 The Authors Journal of Obstetrics and Gynaecology Research © 2014 Japan Society of Obstetrics and Gynecology
Prediction of endometrial cancer types
16. Khunamornpong S, Thorner PS, Suprasert P, Siriaunkgul S. Clear-cell adenocarcinoma of the female genital tract: Presence of hyaline stroma and tigroid background in various types of cytologic specimens. Diagn Cytopathol 2005; 32: 336– 340.
17. Tashiro H, Isacson C, Levine R, Kurman RJ, Cho KR, Hedrick L. p53 gene mutations are common in uterine serous carcinoma and occur early in their pathogenesis. Am J Pathol 1997; 150: 177–185.
© 2014 The Authors Journal of Obstetrics and Gynaecology Research © 2014 Japan Society of Obstetrics and Gynecology
1939
doi:10.1111/jog.12436
J. Obstet. Gynaecol. Res. Vol. 40, No. 7: 1931–1939, July 2014
Prediction of histological types of endometrial cancer by endometrial cytology Masao Okadome1, Toshiaki Saito1, Naoko Nishiyama2, Kazuya Ariyoshi1, Kumi Shimamoto1, Takako Shimada1, Keisuke Kodama1, Shogo Imamura2, Ken-ichi Nishiyama3 and Kenichi Taguchi2 1
Gynecology Service, National Kyushu Cancer Center, 2Department of Pathology, National Kyushu Cancer Center, and Department of Pathology, Japanese Red Cross Fukuoka Hospital, Fukuoka, Japan
3
Abstract Aim: Few studies have examined the accuracy of preoperative endometrial cytology in diagnosing low- and high-risk histology in women with endometrial cancer (EC). This single-institutional retrospective study compared the accuracy of endometrial cytology and biopsy in preoperatively predicting low-risk and high-risk histology of EC. Methods: Between January 2006 and March 2013, 198 women with EC were examined by endometrial cytology, endometrial biopsy and hysterectomy specimen in National Kyushu Cancer Center. Among these women, 110 had endometrial cytology samples available to compare with endometrial biopsy, and were enrolled in our study (mean age ± standard deviation: 59.57 ± 10.32 years). Single-use plastic endometrial suction curettes were used in 12 of the 110 cases and thin metallic curettes for the rest. Results: For type 2 EC, which includes grade 3 endometrioid adenocarcinoma and non-endometrioid histology, biopsy was 67.6% sensitive (25/37) and 84.9% specific (62/73); whereas cytology was 70.3% sensitive (26/37) and 91.8% specific (67/73). Cytology precisely diagnosed only one of 14 cases of serous carcinoma, but it diagnosed 11 of the 14 cases as type 2 EC, and its accuracy in distinguishing EC types was not inferior to endometrial biopsy (10/14). For EC, 9.1% (10/110) were unevaluable using biopsy, significantly more than the 0% (0/110) by cytology (P = 0.002). Conclusion: Although preoperative prediction of serous carcinoma was difficult, endometrial cytology had a higher evaluable rate for EC types. Endometrial cytology may complement endometrial biopsy in preoperative women with EC. Key words: endometrial cytology, endometrial neoplasms/pathology/surgery, endometrial sampling, neoplasm grading, preoperative care.
Introduction Preoperative endometrial sampling with a Pipelle is sensitive and accurate for diagnosis of high-grade endometrial tumors, including tumors with nonendometrioid histology.1 In Japan, endometrial cytology is prevalently used to detect endometrial cancer
(EC) as it is less painful and more cost-effective.2,3 Lymphadenectomy is proposed to be advantageous in patients with EC when preoperative histology indicates high-risk disease, for example, International Federation of Gynecology and Obstetrics (FIGO) grade 3 endometrioid or non-endometrioid histology if accurate frozen sections are unavailable.4
Received: October 17 2013. Accepted: February 23 2014. Reprint request to: Dr Masao Okadome, Gynecology Service, National Kyushu Cancer Center, 3-1-1 Notame, Minami-ku, Fukuoka 811-1395, Japan. Email: [email protected]
© 2014 The Authors Journal of Obstetrics and Gynaecology Research © 2014 Japan Society of Obstetrics and Gynecology
1931
M. Okadome et al.
Few studies have examined the accuracy of preoperative endometrial cytology in diagnosing low- and high-risk histology in patients with EC. A cytological scoring system for endometrioid adenocarcinoma was useful for predicting histological grade.3 Here, in a single-institution retrospective study, we compared endometrial cytology and endometrial biopsy in predicting low- and high-risk histology in preoperative women with EC.
Methods By reviewing the database, we found 270 women who underwent surgery for EC by the Gynecology Service of the National Kyushu Cancer Center from January 2006 to March 2013. A total of 198 women were examined by preoperative endometrial cytology, endometrial biopsy and hysterectomy specimen. Among these, we enrolled 110 women for whom endometrial cytology samples were available for review. We used thin metallic curettes when preoperative endometrial biopsies were performed from 2006 through March 2012 and began to use Pipet Curets (Cooper Surgical, Trumbull, CT, USA) which is similar to Pipelle, a single-use flexible plastic endometrial suction curette, from April 2012. Attending physicians collected biopsy specimens using thin metallic curettes in uterine cavities, with 2–4 passes according to the patient’s discomfort. Biopsy specimens were preoperatively diagnosed by two pathologists (K. N. and K. T.) using hematoxylin–eosin (HE)-stained sections. One gynecologic pathologist/ oncologist (K. A.) reviewed the biopsy specimens that were originally inspected by the two general pathologists (K. N. and K. T.). Endometrial cytology was performed by using Endocyte (Laboratoire CCD, Paris, France). The Endocyte was inserted into the uterine cavity and cells on the tips were spread onto glass slides. Endometrial cell slides were placed in 95% ethanol and then were processed by Papanicolaou stain. Cytological review was performed by two cytotechnologists (N. N. and S. I.) at National Kyushu Cancer Center, Department of Pathology, who were blinded to information about final histological diagnoses of hysterectomy specimens. When the results differed between the cytotechnologists, a final diagnosis was determined after discussion between them. Histological grade was assigned using published criteria. Grade 1–3 endometrioid adenocarcinoma was diagnosed by a cytological scoring system that used rates of cluster formation, nuclear pleomorphism, nuclear size,
1932
chromatin structure, chromatin distribution, presence of euchromatin, large nucleoli, histiocytes and necrosis, and absence of normal endometrial cells.3 Cytological diagnosis of uterine serous carcinoma was based on findings of tumor diathesis, psammoma bodies and papillary clusters composed of tumor cells with enlarged nuclei and numerous nucleoli.5 Cytological diagnosis of clear cell carcinoma is based on findings of adenocarcinoma cells with a translucent clear cytoplasm, centrally located nuclei, large and prominent nucleoli, and spherical aggregates of cells arranged in a single layer to a so-called mirror ball pattern.6 Histological types were divided into two groups: type 1 EC, including grade 1 and 2 endometrioid adenocarcinoma; and type 2, EC including grade 3 endometrioid adenocarcinoma, serous carcinoma and clear cell adenocarcinoma.7 Carcinosarcoma, squamous cell carcinoma and poorly differentiated adenocarcinoma of non-endometrioid type were classified as type 2 EC in the present study. Each endometrial cell slide was evaluated and categorized into type 1 EC or type 2 EC; specific histological types were also predicted for each slide.
Immunostaining for p53 Immunohistochemistry (IHC) for p53 was analyzed in 12 of 14 specimens of serous carcinoma and one specimen of serous carcinoma + clear cell carcinoma. The IHC used mouse monoclonal anti-p53 antibody (DO-7; Novocastra, Newcastle, UK), and was processed by Leica BOND-III automated system (Leica Microsystems, Tokyo, Japan) according to manufacturer’s instructions. Positive and negative controls used for IHC were histological sections that already had been established as either positive or negative. Statistical analyses Fisher’s exact test was used to compare endometrial biopsies with endometrial cytology on agreement of EC types and specific histologies, and on rates of overdiagnosis, underdiagnosis and unevaluable cases of EC types. P < 0.05 was considered significant. Analyses used Dr SPSS II software, version 11.0.1 (SPSS Japan, Tokyo, Japan).
Results The age distribution of 198 women of the study population was 27–85 years (mean age ± standard deviation [SD], 57.33 ± 10.77 years). Of the 198 women who provided preoperative endometrial cytology, endometrial
© 2014 The Authors Journal of Obstetrics and Gynaecology Research © 2014 Japan Society of Obstetrics and Gynecology
Prediction of endometrial cancer types
biopsy and hysterectomy specimens, 10 women showed no atypical cells and no atypical tissues in their cytology or biopsy specimens. For nine women, no atypical cells were found in their cytology specimens but atypical tissues were found in their biopsy specimens. For four women, no atypical tissues were found in their biopsy specimens but atypical cells were found in their cytology specimens. In all, no atypical cells were found in 19 of 198 (9.60%) cytology specimens and no atypical tissues were found in 14 of 198 (7.07%) biopsy specimens. The difference was not significant (P = 0.468). Of the remaining 175 women, we were able to review both cytology specimens and biopsy specimens from 110 women (mean age ± SD, 59.57 ± 10.32 years). The other specimens were not available for review because they had been returned to the hospitals or clinics where patients were initially examined. Figure 1 shows one of 13 cases of grade 3 endometrioid adenocarcinoma. In this case, preoperative diagnosis by endometrial biopsy was grade 1 endometrioid adenocarcinoma (Fig. 1a) but the diagnosis for the resected uterus was grade 3 endometrioid adenocarcinoma (Fig. 1b). The resected specimen had some well-differentiated adenocarcinoma but more
(a)
(c)
than 50% of its cells indicated solid proliferation of carcinoma cells (Fig. 1c). A review of the preoperative endometrial cytology of this case using a cytological scoring system3 was grade 3 endometrioid adenocarcinoma (Fig. 1d,e). Table 1 shows detailed histological types predicted by biopsy and cytology. Final diagnoses were histological diagnoses of resected uterus specimens. Endometrial biopsies provided 10 cases that could not be classified for detailed histological type, including five for adenocarcinoma not otherwise specified (AC NOS), one for carcinoma not otherwise specified (Ca NOS), and three for atypical cells or glands. ‘AC NOS’ means that adenocarcinoma tissue was seen but the collected tissue was scant and pathologists could not diagnose it as endometrioid adenocarcinoma or non-endometrioid adenocarcinoma. ‘Ca NOS’ means that carcinoma tissue was seen but the collected tissue was too small to be specified. ‘Atypical cells or glands’ mean that atypical cells or glands were seen but most of the specimens were mucous materials or necrosis. Endometrial cytology showed no case to be unevaluable (Table 1). Of 14 patients with serous carcinoma ± endometrioid adenocarcinoma, one case was precisely
(b)
(d)
(e)
Figure 1 Histological and cytological appearance of one case with grade 3 endometrioid adenocarcinoma. (a) Histological specimen preoperatively obtained by a thin metallic curette. (b) Histological specimen from hysterectomized uterus. (c) Specimen from hysterectomized uterus: low-powered appearance. More than 50% of the tumor showed solid proliferation, as seen in the right side of the figure. (d,e) Cytological appearances of the endometrial smear for the same case, which was preoperatively obtained by Endocyte. Nuclear pleomorphism is mild to moderate, but tumor diathesis is remarkable and euchromatin is noticeable. The cytological score of this case was >20; cytological diagnosis was grade 3 endometrioid adenocarcinoma (Papanicolaou stain, original magnification ×400).
© 2014 The Authors Journal of Obstetrics and Gynaecology Research © 2014 Japan Society of Obstetrics and Gynecology
1933
1934
1 1
1
1
1
1 2 2
1 2 1
Numbers in boldface agree with final histological diagnoses of hysterectomized specimens. Underscored numbers were within same endometrial types (e.g. type 1 endometrial cancer and type 2 endometrial cancer). AC, adenocarcinoma; AEH, atypical endometrial hyperplasia; C, clear cell carcinoma; Ca, carcinoma; CS, carcinosarcoma; EM, endometrioid adenocarcinoma; G1, grade 1 EM; G2, grade 2 EM; G3, grade 3 EM; NOS, not otherwise specified; S, serous carcinoma; Sm, small cell carcinoma.
1
7 7
1
1
4 1
1 1
2
3 2 1
5 15 5 3 1 36 10 1 1 1 1 1 2 1
2 2 2 7 1 1 1 8 1 1 7 1 38 13 2 1 3 G1 G2 G3 S ± EM S+C C ± EM CS
46 27 13 14 1 4 5
AEH
Diagnostic measures Endometrial biopsy diagnosis (reviewed diagnosis) Endometrial cytology diagnosis (reviewed diagnosis) G1 G2 G3 S ± EM S + C C ± AC Ca Atypical AEH G1 G2 G3 G3 + Sm S ± EM C ± EM CS EM NOS NOS cells or glands Final diagnosis
Table 1 Detailed histological types predicted by endometrial biopsy and endometrial cytology
M. Okadome et al.
diagnosed with serous carcinoma using endometrial cytology (Table 1); in that case, psammoma bodies were seen on the slide. Other cytological diagnoses were seven cases of grade 3 endometrioid adenocarcinoma, two of clear cell carcinoma ± endometrioid adenocarcinoma, one of carcinosarcoma and three of grade 2 endometrioid adenocarcinoma. A case of clear cell carcinoma + serous carcinoma was also misdiagnosed as grade 2 endometrioid adenocarcinoma on cytology. All misdiagnosed serous carcinoma cases lacked psammoma bodies on the slides. Seven of 14 cases of serous carcinoma were correctly diagnosed by the gynecological pathologist/oncologist (K. A.). Only one of the 14 cases was correctly diagnosed as serous carcinoma by the cytotechnologists but 11 of 14 cases were diagnosed as type 2 EC and their accuracy on distinguishing type 2 EC was not inferior to endometrial biopsy (10/14). Three of the 14 cases of serous carcinoma ± endometrioid adenocarcinoma were not evaluated for detailed histological types on biopsy; they were two cases of AC NOS and one case of atypical cells. Of four cases of clear cell carcinoma ± endometrioid adenocarcinoma, two cases were diagnosed as having clear cell carcinoma components by endometrial biopsy, whereas all cases were correctly diagnosed by endometrial cytology. None of five cases of carcinosarcoma were correctly diagnosed by either endometrial biopsy or endometrial cytology (Table 1). Five cases of type 1 EC were overdiagnosed as clear cell carcinoma ± endometrioid adenocarcinoma by endometrial cytology, and two of the five cases were unevaluable by biopsy. One case of type 1 EC was overdiagnosed as clear cell carcinoma ± endometrioid adenocarcinoma by endometrial biopsy. Four cases of type 1 EC were overdiagnosed as serous carcinoma by endometrial biopsy (Table 1). For type 1 EC, endometrial biopsy was 80.8% (59/73) sensitive and 67.6% (25/37) specific; whereas cytology was 90.4% (66/73) sensitive and 70.3% (26/37) specific. For type 2 EC, endometrial biopsy was 67.6% sensitive (25/37) and 84.9% (62/73) specific; whereas cytology was 70.3% (26/37) sensitive and 91.8% (67/73) specific (Table 2). Agreement between biopsy results and final diagnosis for type 2 EC was 67.6% (25/37); agreement between cytology results and final diagnosis was 70.3% (26/37; P = 1.0; Tables 2,3). For EC types, the rate of unevaluable cases by biopsy was 9.1% (10/110)
© 2014 The Authors Journal of Obstetrics and Gynaecology Research © 2014 Japan Society of Obstetrics and Gynecology
Prediction of endometrial cancer types
Table 2 Type 1 and type 2 endometrial cancer predicted by endometrial biopsy and endometrial cytology Final histological type
n
Type 1 Type 2
73 37
Diagnostic measures Endometrial biopsy AEH Type 1 3
59 7
Type 2
NE
AEH
6 25
5 5
1
Endometrial cytology Type 1 Type 2 66 11
NE
6 26
Numbers in boldface indicate agreement between diagnosis and final histological type. AEH, atypical endometrial hyperplasia; NE, not evaluable.
Table 3 Agreement rates for endometrial cancer types, detailed histological types, overdiagnosis rate, underdiagnosis rate and unevaluable rate
Agreement of endometrial cancer types† Agreement of type 2 endometrial cancer Agreement of detailed histological types Overdiagnosis of endometrial cancer types Underdiagnosis of endometrial cancer types Unevaluable cases on endometrial cancer types
Endometrial biopsy
Endometrial cytology
P
84/110 (76.4%) 25/37 (67.6%) 61/110 (55.5%) 6/110 (5.5%) 10/110 (9.1%) 10/110 (9.1%)
92/110 (83.6%) 26/37 (70.3%) 63/110 (57.3%) 6/110 (5.5%) 12/110 (10.9%) 0/110 (0%)
0.238 1.0 0.892 1.0 0.823 0.002*
*Significant difference. †Endometrial cancer types are type 1 endometrial cancer and type 2 endometrial cancer. Type 1 endometrial cancer includes G1 endometrioid adenocarcinoma and G2 endometrioid adenocarcinoma. Type 2 endometrial cancer includes G3 endometrioid adenocarcinoma and non-endometrioid type cancer.
Table 4 Detailed histological types predicted by endometrial biopsy using Pipet Curets and endometrial cytology Final histological type G1 G2 G3 CS
G1 1 6 3 2
1 4 2
Pipet Curets G2 G3
Diagnostic measures Endometrial cytology C G1 G2 G3 G3 + Sm 1
1 1 2
1 1
5 1
2 1
1
Numbers in boldface indicate agreement between diagnosis and final histological type. C, clear cell carcinoma; G1, grade 1 endometrioid adenocarcinoma; G2, grade 2 endometrioid adenocarcinoma; G3, grade 3 endometrioid adenocarcinoma; Sm, small cell carcinoma.
and 0% (0/110) by cytology (Tables 2,3), which was significant (P = 0.002, Table 3). Agreements for EC types and detailed histological types, overdiagnosis and underdiagnosis of EC types were not significant (P = 0.238, 0.892, 1.0 and 0.823, respectively, Tables 1–3). The 10 cases unevaluable by biopsy included five cases of type 1 EC and five cases of type 2 EC. Three of the five type 1 EC cases and four of the five type 2 EC cases were correctly diagnosed by cytological reviews. The two misdiagnosed cases of type 1 EC were both grade 2 endometrioid adenocarcinoma; they were overdiagnosed as clear cell carcinoma. The misdiagnosed case of type 2 EC was serous carcinoma, which was underdiagnosed as grade 2 endometrioid adenocarcinoma.
There were 12 cases of EC for which Pipet Curets were used in endometrial biopsies among the cohort. Agreement between biopsy results using Pipet Curets and final diagnosis for specific histologies was 25.0% (3/12) and 66.7% (8/12) between endometrial cytology and final diagnosis (Table 4), which was not significant (P = 0.099).
p53 immunostaining Eleven of 12 tested specimens of serous carcinoma, and one tested specimen of serous carcinoma + clear cell carcinoma were p53+ (Table 5). The negative case (S6) was diagnosed as serous carcinoma by the gynecologic pathologist/oncologist (K. A.) who reviewed the HE-stained section and was diagnosed as grade 2
© 2014 The Authors Journal of Obstetrics and Gynaecology Research © 2014 Japan Society of Obstetrics and Gynecology
1935
M. Okadome et al.
Table 5 Immunocytochemical expression of p53 in preoperative endometrial biopsy specimens of serous carcinoma and serous carcinoma with clear cell carcinoma (reviewed diagnosis for both types)
S1 S2 S3 S4 S5 S6 S7 S8 S9 S 10 S 11 S 12 S 13 S 14 S+C
p53
Endometrial biopsy
Endometrial cytology
+ + + + + − + + + + + + ND ND +
AC NOS G1 G3 Serous Clear Serous Serous Serous Serous Clear AC NOS Serous Atypical cells Serous Serous
Clear + EM G3 Carcinosarcoma G2 G3 G2 Serous G3 Clear + EM G3 Clear + EM G3 G3 G3 G2
AC, adenocarcinoma; C, clear cell carcinoma; EM, endometrioid adenocarcinoma; G1, grade 1 EM; G2, grade 2 EM; G3, grade 3 EM; ND, not done; NOS, not otherwise specified; S, serous carcinoma.
endometrioid adenocarcinoma by the cytotechnologists (N. N. and S. I., Table 5). The three p53+ cases (S1, S2 and S11) were not diagnosed as serous carcinoma by review of HE-stained sections, but were diagnosed as type 2 EC by cytological review (Table 5). A p53+ serous carcinoma case and the p53– case are shown in Figure 2.
Discussion The Pipelle is generally considered to be the best device in detecting endometrial carcinoma, with detection rates of 99.6% and 91%, respectively, in postmenopausal and premenopausal women;8 it is 81% sensitive and more than 98% specific in detecting EC.8 Preoperative endometrial sampling with Pipelle is sensitive and accurate for the diagnosis of high-grade endometrial tumors, including tumors with non-endometrioid histology.1 In Japan, an attempt has been made to detect asymptomatic EC using endometrial cytology, and to clarify whether detection of asymptomatic EC correlates with reduced mortality.9 Another study evaluated the usefulness of cytological scoring in assessing morphological differentiation of endometrioid adenocarcinomas
1936
using endometrial smears, and concluded that the cytological scoring system applied for endometrioid adenocarcinoma was useful for predicting histological grade and tumor malignant potential.3 When no macroscopic extrauterine disease is present, tumor diameter is 2 cm or less, and preoperative findings show FIGO grade 1 or 2 endometrioid histology or complex and/or atypical hyperplasia (low-risk patients), the risk of positive lymph nodes or lymph node recurrence is less than 1% and the 3-year recurrence-free survival rate is 98.7%. In contrast, highrisk patients with grade 3 or non-endometrioid histology or mixed Müllerian tumors or intraoperative macroscopic extrauterine disease have significantly higher risk of positive lymph nodes or lymph node recurrence (27%) than low-risk patients and a lower 3-year recurrence-free survival rate (71.2%). In these high-risk patients, lymphadenectomy may be advantageous.4 In addition, multivariate analysis confirmed that tumor volume, high serum CA125 level and grade 3/serous adenocarcinoma were independent risk factors for lymph node metastasis among 211 patients with endometrial carcinoma. This validation study was carried out using only preoperative data.10 Consequently, preoperative diagnosis of type 2 EC disease is important. In the present study, we investigated whether endometrial cytology could predict grades of endometrioid adenocarcinoma and tumors of non-endometrioid histology or complement preoperative endometrial biopsy in diagnosing type 1 and type 2 EC. For type 1 and type 2 EC, endometrial cytology had a significantly lower rate of unevaluable cases than did endometrial biopsy when biopsy was done by a metallic curette. An advantage of endometrial cytology is that endometrial cells are scraped from the entire surface of the endometrial cavity with an Endocyte, which could affect results. Although only 12 cases used Pipet Curets, endometrial cytology tended to have a higher prediction rate for specific EC histologies than Pipet Curets. Endometrial cytology may also complement single-use endometrial suction curettes in predicting high-risk EC. The rate of predicting serous carcinoma and tumors of its related histology was low using endometrial cytology; only one case with psammoma bodies was precisely predicted in our study, though the cytotechnologist in our study was experienced with endometrial cytology and involved in a study using endometrial smears.11 A previous study reported important
© 2014 The Authors Journal of Obstetrics and Gynaecology Research © 2014 Japan Society of Obstetrics and Gynecology
Prediction of endometrial cancer types
(a)
(b)
(c)
(d)
Figure 2 (a,c) Hematoxylin–eosin and (b,d) immunohistochemical staining with monoclonal antibody p53. (a,b) Serous carcinoma with strong positive staining involving almost all epithelial cells but not the surrounding stroma. (c,d) Serous carcinoma with negative staining.
diagnostic findings of serous carcinoma without psammoma bodies, but showed clusters of tumor cells with papillary architecture, tumor diathesis, hyperchromatic nuclei, prominent nucleoli and anisonucleosis,5 which were also observed in some of our grade 2 or 3 endometrioid adenocarcinoma cases. If cytotechnologists did not suspect serous carcinoma on slides of endometrial carcinoma, those cases were diagnosed by the scoring system for endometrioid adenocarcinoma. Cases of serous carcinoma with cytological characteristics cited above were difficult to diagnose without psammoma bodies. Serous carcinoma misdiagnosed and treated as grade 3 endometrioid adenocarcinoma is not as potentially damaging to a patient as the reverse. Our study had three serous carcinoma cases that were misdiagnosed as grade 2 endometrioid adenocarcinoma and one case with serous carcinoma and clear cell carcinoma that was misdiagnosed as grade 2 endometrioid adenocarcinoma, all four by endometrial cytology. These four misdiagnoses could be potentially damaging to patients. Diagnosis of endometrial clear cell carcinomas by gynecologic pathologists is associated with a moderate level of interobserver variability.12 Clear cell carcinoma has core cytoarchitectural features that are of high diagnostic utility but also has a wide morphological spectrum.13 Pathologists may have trouble in diagnosing endometrial carcinoma when the carcinoma tissues include some clear cells. In cytology, carcinomas with
cells that show centrally located nuclei and translucent, clear cytoplasm may be diagnosed as clear cell carcinoma;6 such findings may be related to overdiagnosis. The characteristic feature of clear cell carcinoma of the genital tract is basement membrane-like substance. This hyaline extracellular material stains pinkish to purple-red in May–Grünwald–Giemsa stain preparations and is frequently observed within cancer cell clusters.14 Basement membrane material in scraped cytological specimens from patients with clear cell carcinoma showed metachromatic pink to purple staining in Diff-Quik-stained smears.15 Hyaline stroma was present in 93% of scraped cytology specimens of clear cell carcinoma.16 These stain methods may be useful to avoid overdiagnosis. Uterine serous carcinoma (USC) and its putative precursor, endometrial intraepithelial carcinoma (EIC), demonstrate strong p53 overexpression by immunohistochemistry, which suggests alteration of the p53 gene in their pathogenesis.17 When 21 cases of USC and nine cases of EIC were evaluated for mutations in the p53 gene using direct sequence analysis, 90% of USC and 78% of EIC were found to contain mutations.17 In our study, three of 12 tested cases of serous carcinoma were not diagnosed as such and were not diagnosed as type 2 EC by HE-stained sections of endometrial biopsy, but were diagnosed as type 2 EC by endometrial cytology. IHC for p53 could be useful in such cases, as they were p53+.
© 2014 The Authors Journal of Obstetrics and Gynaecology Research © 2014 Japan Society of Obstetrics and Gynecology
1937
M. Okadome et al.
Our study has some limitations. First, women who underwent endometrial biopsy and cytology were referred to our institution because they were already suspected of having EC by earlier biopsy and/or cytology; therefore, our findings do not necessarily indicate which modality could better screen for EC or predict EC type among women in the general population who may not be suspected of having EC. Second, most endometrial biopsies in this study were performed with thin metallic curettes; Pipet Curets were used in only 12 of the 110 cases. Third, the cytotechnologists in our study were aware that all 110 cases were EC, which obviously differs from the usual requirement to assess cells de novo. Finally, diagnostic qualities in the field of endometrial cytology are secured in most of the laboratories in Japan, but such qualities may be assured in a limited number of laboratories in other countries because endometrial cytology is not prevalently used to detect EC except Japan. Widespread acceptance of endometrial cytology may be needed in other countries. In conclusion, endometrial cytology has higher evaluable rate for type 1 and type 2 EC than does endometrial biopsy and could complement endometrial biopsy preoperatively. Endometrial cytology may have a fault in predicting serous carcinoma, though most misdiagnosed cases were diagnosed as type 2 EC. It may also occasionally overdiagnose type 1 EC as clear cell carcinoma. Endometrial biopsy may overdiagnose type 1 EC as serous carcinoma. We should carefully diagnose EC using endometrial biopsy and endometrial cytology in preoperative women, because both modalities have advantages in prediction. Endometrial cytology may also help verify endometrial biopsies performed with single-use endometrial suction curettes, such as the Pipelle or Pipet Curets; this merits further investigation.
Acknowledgments This study was supported by the National Cancer Center Research and Development Fund (23-A-17) and Grants-in-Aid for Cancer Research (no. 18-06, no. 10103749) from the Ministry of Health, Labor and Welfare of Japan.
Disclosure The authors have no conflicts of interest to declare.
1938
References 1. Huang GS, Gebb JS, Einstein MH, Shahabi S, Novetsky AP, Goldberg GL. Accuracy of preoperative endometrial sampling for the detection of high-grade endometrial tumors. Am J Obstet Gynecol 2007; 196: 243 e241–243 e245. 2. Nakagawa-Okamura C, Sato S, Tsuji I et al. Effectiveness of mass screening for endometrial cancer. Acta Cytol 2002; 46: 277–283. 3. Nishimura Y, Watanabe J, Jobo T, Hattori M, Arai T, Kuramoto H. Cytologic scoring of endometrioid adenocarcinoma of the endometrium. Cancer 2005; 105: 8–12. 4. AlHilli MM, Podratz KC, Dowdy SC et al. Preoperative biopsy and intraoperative tumor diameter predict lymph node dissemination in endometrial cancer. Gynecol Oncol 2013; 128: 294–299. 5. Hagiwara T, Kaku T, Kobayashi H, Hirakawa T, Nakano H. Clinico-cytological study of uterine papillary serous carcinoma. Cytopathology 2005; 16: 125–131. 6. Ohwada M, Suzuki M, Ohno T, Saito K, Sato I. Appearance of primary endometrial and ovarian clear cell adenocarcinoma 17 months postpartum. A case report. Acta Cytol 1998; 42: 765–768. 7. Voss MA, Ganesan R, Ludeman L et al. Should grade 3 endometrioid endometrial carcinoma be considered a type 2 cancer-a clinical and pathological evaluation. Gynecol Oncol 2012; 124: 15–20. 8. Dijkhuizen FP, Mol BW, Brolmann HA, Heintz AP. The accuracy of endometrial sampling in the diagnosis of patients with endometrial carcinoma and hyperplasia: A metaanalysis. Cancer 2000; 89: 1765–1772. 9. Jobo T, Arai T, Sato R, Kuramoto H. Clinicopathologic relevance of asymptomatic endometrial carcinoma. Acta Cytol 2003; 47: 611–615. 10. Todo Y, Okamoto K, Hayashi M et al. A validation study of a scoring system to estimate the risk of lymph node metastasis for patients with endometrial cancer for tailoring the indication of lymphadenectomy. Gynecol Oncol 2007; 104: 623– 628. 11. Okadome M, Saito T, Tsukamoto N, Nishi K, Nishiyama N, Nagata E. Endometrial scraping cytology in women with extragenital malignancies. Acta Cytol 2006; 50: 158– 163. 12. Fadare O, Parkash V, Dupont WD et al. The diagnosis of endometrial carcinomas with clear cells by gynecologic pathologists: An assessment of interobserver variability and associated morphologic features. Am J Surg Pathol 2012; 36: 1107–1118. 13. Fadare O, Zheng W, Crispens MA et al. Morphologic and other clinicopathologic features of endometrial clear cell carcinoma: A comprehensive analysis of 50 rigorously classified cases. Am J Cancer Res. 2013; 3: 70–95. 14. Atahan S, Ekinci C, Icli F, Erdogan N. Cytology of clear cell carcinoma of the female genital tract in fine needle aspirates and ascites. Acta Cytol 2000; 44: 1005–1009. 15. Jimenez-Heffernan JA, Vicandi B, Gonzalez-Peramato P et al. Cytologic features of clear cell carcinoma of the female genital tract. Diagnostic value of the ‘raspberry body’ in nonexfoliative cytologic specimens. Acta Cytol 2004; 48: 47–51.
© 2014 The Authors Journal of Obstetrics and Gynaecology Research © 2014 Japan Society of Obstetrics and Gynecology
Prediction of endometrial cancer types
16. Khunamornpong S, Thorner PS, Suprasert P, Siriaunkgul S. Clear-cell adenocarcinoma of the female genital tract: Presence of hyaline stroma and tigroid background in various types of cytologic specimens. Diagn Cytopathol 2005; 32: 336– 340.
17. Tashiro H, Isacson C, Levine R, Kurman RJ, Cho KR, Hedrick L. p53 gene mutations are common in uterine serous carcinoma and occur early in their pathogenesis. Am J Pathol 1997; 150: 177–185.
© 2014 The Authors Journal of Obstetrics and Gynaecology Research © 2014 Japan Society of Obstetrics and Gynecology
1939
