- Email: [email protected]
How accurate is frozen section histology of pelvic lymph nodes in patients with endometrial cancer?
Gynecologic Oncology 115 (2009) 12–17
Contents lists available at ScienceDirect
Gynecologic Oncology j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / y g y n o
How accurate is frozen section histology of pelvic lymph nodes in patients with endometrial cancer? Gunda Pristauz a,⁎, Arnim A. Bader a, Peter Regitnig b, Josef Haas a, Raimund Winter a, Karl Tamussino a a b
Department of Obstetrics and Gynecology, Medical University of Graz, Auenbruggerplatz 14, 8036 Graz, Austria Institute of Pathology, Medical University of Graz, Graz, Austria
a r t i c l e
i n f o
Article history: Received 3 June 2009 Available online 4 August 2009 Keywords: Endometrial cancer Frozen section Pelvic and para-aortic lymphadenectomy
a b s t r a c t Objective. Recent prospective data support the trend towards systematic retroperitoneal lymphadenectomy in patients with high-risk endometrial cancer. Because para-aortic node involvement in the absence of pelvic node involvement is uncommon, a reliable finding of negative pelvic lymph nodes (PLN) at intraoperative frozen section examination might allow omitting para-aortic dissection. We analyzed the diagnostic accuracy of frozen section examination of PLN in patients with endometrial cancer. Methods. We reviewed 131 patients with endometrial cancer who underwent surgery including systematic pelvic lymphadenectomy (n = 101) or pelvic and para-aortic lymphadenectomy (n = 27). Intraoperative frozen section examination of PLN was performed in 72 (55%) patients. Results of frozen section examination were compared with those of final histopathology and the diagnostic accuracy of frozen section examination of PLN was calculated. One pathologist measured the diameters of PLN metastases retrospectively. Results. A total of 1063 and 2666 PLN were analyzed by frozen section examination and by final histopathology, respectively. PLN metastases were found in 7 cases (10%) at frozen section examination, and in 17 cases (24%) at final histopathology (false negative rate, 59%). No false positive cases were noted. The mean diameter of all PLN metastases at final histopathology was 4.3 mm, as compared to 9.0 mm for the metastases detected at frozen section analyses. The mean diameter of PLN metastases missed at frozen section examination was 2.0 mm. Conclusion. In this review at a single institution, intraoperative frozen section histology missed nearly two of three endometrial cancer patients with positive nodes. These results do not support tailoring the extent of lymphadenectomy according to the results of frozen section examination. © 2009 Elsevier Inc. All rights reserved.
Introduction Endometrial cancer is the most common female genital malignancy in many Western countries [1]. The disease is staged surgically and this includes total hysterectomy, bilateral salpingo-oophorectomy, pelvic and para-aortic lymphadenectomy and peritoneal cytology. The status of the lymph nodes affects stage, prognosis and therapeutic decisions in patients with this disease [2,3]. Recent prospective data support the trend towards systematic pelvic and para-aortic lymphadenectomy in patients with high-risk endometrial carcinoma [4]. However, pelvic and para-aortic lymphadenectomy is associated with increased operating time and morbidity [5]. Because the likelihood of positive para-aortic lymph nodes in patients with negative pelvic nodes is low, a reliable finding of negative pelvic nodes at intraoperative frozen section examination might be used to decide against para-aortic dissection. We reviewed
⁎ Corresponding author. Fax: +43 316 385 4197. E-mail address: [email protected] (G. Pristauz). 0090-8258/$ – see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.ygyno.2009.07.007
the diagnostic accuracy of frozen section histology of pelvic lymph nodes in patients with endometrial cancer at our institution. Patients and methods Between January 1996 and July 2008 a total of 131 patients with endometrial cancer underwent surgery including lymphadenectomy at our institution. Of these, 101 patients had systematic pelvic lymphadenectomy. 27 patients had systematic pelvic and para-aortic lymphadenectomy, and three patients had resection of bulky nodes only. The latter three patients were not included in the present analysis, because frozen section histology and lymph node sampling was restricted to a limited number of bulky nodes. Whether lymphadenectomy was extended to the para-aortic region depended on intraoperative frozen section analyses and patient's medical status. The present study is based on 72 (55%) patients who had an intraoperative frozen section examination of unfixed grossly visible pelvic lymph nodes (Table 1). In all 72 patients pelvic lymphadenectomy was performed via laparotomy by one of five surgeons in a systematic manner from the femoral ring up to the bifurcation of the
G. Pristauz et al. / Gynecologic Oncology 115 (2009) 12–17 Table 1 Characteristics of 72 patients with endometrial cancer and primary surgery including systematic pelvic lymphadenectomy with intraoperative frozen section preparation. Variable
No. of patients/data
Total Mean age (years) Range (years) FIGO stage Ia Ib Ic IIa IIb IIIa IIIb IIIc IVb Histological type Endometrioid Clear cell Carcinosarcoma (MMMT) Serous papillary Grading G1 G2 G3
72 62 39–74
13
specificity, positive and negative predictive values and their 95%confidence intervals were calculated with StatsDirect 2.6.5 (StatsDirect Ltd., England). The accuracy of the test was evaluated with McNemar's test. P-values b 0.05 were considered significant. Results
7 15 17 2 9 3 1 14 4 53 8 7 4 15 20 37
aorta. The aim was to remove all accessible fat tissue containing lymph nodes around the entire circumference of major pelvic vessels and from the obturator fossa [6]. Tissue was removed as fifteen separate defined anatomic groups [7]. All lymphatic tissue was sent to the pathology laboratory for intraoperative pathologic evaluation. The pathologist palpated the unfixed fat tissue and dissected out all grossly suspicious lymph nodes. Nodes were selected for frozen section histology based on size, consistency and appearance of the cut surface. Larger lymph nodes were usually bisected, depending on size also cut in several pieces as described previously [7]. All intraoperative frozen section analyses and final histologies were performed by one out of three gynecopathologists. After frozen section the remaining tissue was formalin fixed and paraffin embedded for further histopathological analyses. One pathologist (P.R) measured the diameters of lymph node metastases retrospectively.
A total of 1063 pelvic lymph nodes were examined by intraoperative frozen section, with a mean of 15 nodes (median, 13) examined per patient. A total of 2666 pelvic lymph nodes were analyzed in final paraffin sections for definitive histology, with a mean of 36 nodes (median, 35) per patient. Seventeen of the 72 patients (24%) showed pelvic lymph node metastases at final histology. These metastases were detected by intraoperative frozen section analysis in 7 patients (true positive rate, 41%) and by definitive paraffin histology in 10 patients (false negative rate, 59%). Those patients with intraoperative negative frozen section analyses and positive nodes in final histology showed 14 lymph node metastases (mean, 1; range, 1–3 per patient). Table 2 shows details of the 10 patients with false negative frozen section results. A total of 57 of the 2666 pelvic lymph nodes were positive at final histology. Thirteen of those were detected at frozen section (mean, 2; range 1–3 per patient) and 44 at final histology (mean, 2; range 1–10 per patient). No false positive cases by frozen section were noted. Overall 14 positive lymph nodes were missed at frozen section examination. Six positive nodes were not examined at frozen section examination; 5 positive nodes were examined at frozen section examination, but the metastases were only visible in deeper sections after paraffin embedding; and 3 positive nodes were examined at frozen section examination, but the metastases were not diagnosed. The mean diameter of all lymph node metastases at final histopathology was 4.3 mm (median, 3.5; range, 0.1–20 mm) as opposed to 9.0 mm (median, 5.0; range, 2–20 mm) for the metastases detected at frozen section analyses. The mean diameter of lymph node metastases missed at frozen section examination was 2.0 mm (median, 1.2; range, 0.1–6 mm) (Table 3). Table 4 shows the sensitivity, specificity, positive and negative predictive values of frozen section as a diagnostic test for predicting the final node status. Although the false negative rate was 59%, the negative predictive value was 85%.
Statistical analysis Discussion Frozen section histology was analyzed as a diagnostic test and its results were compared to the results of final histology. If the results from the paraffin sections did not differ from the results of the frozen sections the diagnosis was considered to be correct. Sensitivity,
This analysis of intraoperative frozen section examination of pelvic nodes in patients with endometrial cancer at a single institution shows a sensitivity of 41% and a false negative rate of 59%. Pathologists
Table 2 Details of 10 patients with false negative frozen section analyses. Patient
FIGO stage
Histology
Frozen section
Final histology
Number of nodes pos./examined
Number of nodes pos./examined
1
IIIb
Carcinosarcoma
0/25
3/42
2
IIIc
Endometrioid
0/21
2/35
3 4 5 6 7 8 9 10
IIb IIIc IIIc IIIc IIIc IIIc IIIc IIIC
Endometrioid Endometrioid Seröses Endometrioid Endometrioid Endometrioid Endometrioid Endometrioid
0/15 0/8 0/8 0/19 0/11 0/10 0/38 0/20
2/35 1/30 1/10 1/30 1/21 1/60 1/62 1/71
Node region with metastases
Diameter of lymph node metastases in mm
Reasons of discordance between frozen section and final analyses
Right external iliac Left external iliac Right external iliac Left common iliac Right external iliac lateral Left external iliac Left external iliac Right external iliac Left common iliac Right obturator (above nerve) Left obturator (above nerve) Right external iliac
4 1.5, 5 1.1 2.5 1, 0.1 1.2 1.5 1 0.5 1 1.2 6
C C+B A B C+A A A A C C A B
A = missed node not examined in frozen section examination. B = missed node examined in frozen section examination, but metastasis not diagnosed. C = missed node examined in frozen section examination, but metastasis only visible in deeper sections after paraffin embedding.
14
G. Pristauz et al. / Gynecologic Oncology 115 (2009) 12–17
Table 3 Number of pelvic lymph nodes metastases and mean values of largest diameters as detected by intraoperative frozen section and by final histology of 72 patients with endometrial cancer. Lymph node region
Frozen section examination Patients with positive nodes
No. of pelvic node metastases
Right External iliac Common iliac Internal iliac Obturator (above nerve) Obturator (below nerve) External iliac lateral Common iliac lateral Subaortal
4 1 0 1 0 1 0 0
4 2 0 1 0 1 0 0
Left External iliac Common iliac Internal iliac Obturator (above nerve) Obturator (below nerve) External iliac lateral Common iliac lateral
2 0 0 1 0 0 0
4 0 0 1 0 0 0
Final histopathology Mean values of largest diameters of node metastases, mm (median) 8.8 (7.5) 3.0 (3.0) – 5.0 (5.0) – 20.0 (20.0) – –
8.5 (8.5) – – 4.0 (4.0) – – –
picked suspicious pelvic lymph nodes on the basis of size, consistency and overall appearance and detected 13 lymph node metastases out of 1063 lymph nodes examined at frozen section for a ratio of 1:82 node metastasis to nodes examined. This detection rate should be superior to a random selection rate of lymph nodes, but at paraffin embedding 57 nodes out of 2666 examined lymph nodes were detected with a ratio of 1:47. In contrast a similar study in patients with cervical cancer of our institution showed a 17% false negative rate for frozen section histology with the same pathologist as in our study detecting lymph node metastases at frozen section with a ratio of 1:18 node metastasis to nodes examined and a ratio of 1:35 at final histology [7,8]. The diagnostic accuracy of lymph node frozen section analyses depends on palpation of the lymphatic fatty tissue removed at surgery, i.e. on “who's at the sink”. Whether a lymph node or a lymph node metastasis is palpated by the pathologist depends on the presence of lymph node size and desmoplastic stromal reaction. Desmoplasia is defined as a newly formed fibrotic (collagenous) stroma surrounding tumor cells [9] and makes the lymph node rigid and palpable for the surgeon or pathologist. Positive pelvic lymph nodes in patients with squamous cell carcinomas of the cervix show an increased desmoplastic reaction in positive lymph nodes compared to positive lymph nodes in patients with adenocarcinoma of the uterus (Figs. 1 and 2). Para-aortic lymph node metastases occur in about 15% of patients with high-risk endometrial cancer and in about 60% of patients with pelvic lymph node metastases [4]. In contrast para-aortic metastases
Table 4 Sensitivity, specificity, positive predictive value and negative predictive value of frozen section as a diagnostic test for predicting the final node status. No. of patients True-positive True-negative False negative False positive
7 55 10 0
Sensitivity Specificity Positive predictive value Negative predictive value Test accuracy
Test characteristics with 95% confidence intervals 41.2% (18.4–67.1) 100.0% (93.5–100.0) 100.0% (59.0–100.0) 84.6% (73.5–92.4) 86.1%(75.9–93.1)
McNemar's test: P = 0.0016.
Patients with positive nodes
No. of pelvic node metastases
Mean maximum diameter of node metastases (median)
8 1 0 3 0 2 0 1
8 2 0 5 0 3 0 1
5.9 (5.0) 3.0 (3.0) – 2.3 (2.0) – 7.0 (1.0) – 2.0 (2.0)
6 3 1 4 0 1 1
12 3 2 6 0 1 1
5.7 (4.5) 1.7 (1.0) 4.5 (4.5) 2.7 (2.5) – 10.0 (10.0) 2.0 (2.0)
are uncommon in patients with negative pelvic nodes [4,10,11]. Imaging techniques such as computed tomography and magnetic resonance imaging are unreliable for evaluating pelvic or para-aortic lymph nodes and 2-[18F]fluoro-2-deoxy-D-glucose (FDG) positron emission tomography (PET) is only moderately sensitive for predicting lymph node metastasis preoperatively [12]. Therefore intraoperative frozen section analysis of pelvic lymph nodes was thought to be useful to identify patients for para-aortic lymphadenectomy. Lymphadenectomy may also have a therapeutic effect in patients with endometrial cancer, with an increase in 8-year disease specific survival from 60% to 85% when comparing patients who had extensive lymphadenectomy with those who did not have extensive lymphadenectomy [13–15]. However, pelvic and aortic lymphadenectomy is associated with longer operating times, longer hospital stays and higher estimated blood loss compared to patients undergoing pelvic lymphadenectomy only [15]. Why were positive nodes missed at frozen section? The first reason is that positive nodes were not palpated and thus not selected for frozen section. This was the case in 6 patients. Similarly McMeekin et al. [16] found that 36% of nodes with metastatic disease were abnormal on palpation. Also lymph node metastases are small. In this series 10 out of 13 missed metastases at frozen section were ≤2 mm. Girardi et al. [17] reported that 37% of node metastases in patients with endometrial cancer were ≤ to 2 mm in diameter. The most important factors for the failure of intraoperative frozen section examination seem to be the size of metastases [8]. Eight out of 14 positive lymph nodes missed at frozen section examination in our patients were picked correctly by the pathologist but metastases were not diagnosed due to the small diameters of the metastases. Therefore serial section at intraoperative frozen section examination is required to achieve a low false negative rate of frozen section examination but impossible due to the time limit, limited man power as well the large amount of nodes excised. Recent data suggest the possibility of identifying sentinel lymph node in patients with endometrial cancer [18,19] which would allow focusing intraoperative work on a few nodes. Finally, metastases from endometrial cancer may not induce a desmoplastic reaction like that seen with squamous cell carcinoma of the cervix. These results indicate that at our institution frozen section examination of pelvic lymph nodes in patients with endometrial cancer is not accurate enough to routinely guide intraoperative management decisions regarding the extent of lymphadenectomy. Although only 10 of the 72 overall patients (14%) were not correctly
G. Pristauz et al. / Gynecologic Oncology 115 (2009) 12–17
15
Fig. 1. Comparison of desmoplastic reaction in lymph node metastases of endometrial and cervical cancer. (A and B) Pelvic lymph node with metastatic endometrial cancer.
classified by frozen section examination, this would translate into approximately two of three patients with positive pelvic nodes not being identified intraoperatively. Selecting only patients with positive intraoperative pelvic nodes for para-aortic lymph node dissection would leave one out of ten patients with positive para-aortic nodes without para-aortic dissection (Fig. 3). It would be appropriate for other institutions using intraoperative frozen section histology of lymph nodes to plan the extent of surgery to analyze their results in this setting.
Conflict of interest statement None of the authors has financial conflicts of interests.
References [1] American Cancer Society. Available at: http://www.cancer.org. [2] Larson DM, Johnson K, Olson KA. Pelvic and para-aortic lymphadenectomy for surgical staging of endometrial cancer: morbidity and mortality. Obstet Gynecol 1992;79:998–1001.
16
G. Pristauz et al. / Gynecologic Oncology 115 (2009) 12–17
Fig. 2. Comparison of desmoplastic reaction in lymph node metastases of endometrial and cervical cancer. (A and B) Pelvic lymph node metastases of cervical cancer with increased desmoplastic reaction compared to lymph node metastases of endometrial cancer (Fig. 1). [3] Possover M, Krause N, Plaul K, Kühne-Heid R, Schneider A. Laparoscopic paraaortic and pelvic lymphadenectomy: experience with 150 patients and review of the literature. Gynecol Oncol 1998;71:19–28. [4] Mariani A, Dowdy SC, Cliby WA, Gostout BS, Jones MB, Wilson TO, et al. Prospective assessment of lymphatic dissemination in endometrial cancer: a paradigm shift in surgical staging. Gynecol Oncol 2008;109:11–8. [5] Benedetti-Panici P, Maneschi F, Cutillo G. Pelvic and aortic lymphadenectomy. Surg Clin North Am 2001;81:841–58. [6] Winter R. Pelvic and paraaortic lymphadenectomy. In: Burghardt E, Webb MJ, Monoghan JM, Kinderman G, editors. Surgical Gynecological Oncology. New York: Thieme; 1993. p. 281–93.
[7] Bader AA, Winter R, Moinfar F, Petru E, Pristauz G, Scholz H, et al. Is intraoperative frozen section analysis of pelvic lymph nodes accurate after neoadjuvant chemotherapy in patients with cervical cancer? Gynecol Oncol 2006;103:106–12. [8] Scholz HS, Lax SF, Benedicic C, Tamussino K, Winter R. Accuracy of frozen section examination of pelvic lymph nodes in patients with FIGO stage IB1 to IIB cervical cancer. Gynecol Oncol 2003;90:605–9. [9] Sternberg S, editor. Histology for Pathologists. 2nd. Lippincott-Raven; 1997. p. 143–59. [10] Mariani A, Keeney GL, Aletti G, Webb MJ, Haddock MG, Podratz KC. Endometrial carcinoma: paraaortic dissemination. Gynecol Oncol 2004;92:833–8.
G. Pristauz et al. / Gynecologic Oncology 115 (2009) 12–17
17
Fig. 3. Translation of our study results into clinical practice. [11] Goudge C, Bernhard S, Cloven NG, Morris P. The impact of complete surgical staging on adjuvant treatment decisions in endometrial cancer. Gynecol Oncol 2004;93:536–9. [12] Horowitz NS, Dehdashti F, Herzog TJ, Rader JS, Powell MA, et al. Prospective evaluation of FDG-PET for detecting pelvic and para-aortic lymph node metastasis in uterine corpus cancer. Gynecol Oncol 2004;95:546–51. [13] Kilgore LC, Patridge EE, Alvarez RD, et al. Adenocarcinoma of the endometrium: survival comparisons of patients with and without pelvic node sampling. Gynecol Oncol 1995;56:29–33. [14] Chan JK, Cheung MK, Huh WK, Osann K, Husain A, Teng NN, et al. Therapeutic role of lymph node resection in endometrioid corpus cancer: a study of 12,333 patients. Cancer 2006;107:1823–30. [15] Cragun JM, Havrilesky LJ, Calingaert B, Synan I, Secord AA, Soper JT, et al.
[16]
[17] [18]
[19]
Retrospective analysis of selective lymphadenectomy in apparent early-stage endometrial cancer. J Clin Oncol 2005;23:3653–5. McMeekin DS, Lashbrook D, Gold M, Scribner DR, Kamelle S, Tillmanns TD, et al. Nodal distribution and its significance in FIGO stage IIIc endometrial cancer. Gynecol Oncol 2001;82:375–9. Girardi F, Petru E, Heydarfadai M, Haas J, Winter R. Pelvic lymphadenectomy in the surgical treatment of endometrial cancer. Gynecol Oncol 1993;49:177–80. Perrone AM, Casadio P, Formelli G, Levorato M, Ghi T, Costa S, et al. Cervical and hysteroscopic injection for identification of sentinel lymph node in endometrial cancer. Gynecol Oncol 2008;111:62–7. Abu-Rustum NR, Khoury-Collado F, Pandit-Taskar N, Soslow RA, Dao F, Sonoda Y, et al. Sentinel lymph node mapping for grade 1 endometrial cancer: is it the answer to the surgical staging dilemma? Gynecol Oncol 2009;113:163–9.
Contents lists available at ScienceDirect
Gynecologic Oncology j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / y g y n o
How accurate is frozen section histology of pelvic lymph nodes in patients with endometrial cancer? Gunda Pristauz a,⁎, Arnim A. Bader a, Peter Regitnig b, Josef Haas a, Raimund Winter a, Karl Tamussino a a b
Department of Obstetrics and Gynecology, Medical University of Graz, Auenbruggerplatz 14, 8036 Graz, Austria Institute of Pathology, Medical University of Graz, Graz, Austria
a r t i c l e
i n f o
Article history: Received 3 June 2009 Available online 4 August 2009 Keywords: Endometrial cancer Frozen section Pelvic and para-aortic lymphadenectomy
a b s t r a c t Objective. Recent prospective data support the trend towards systematic retroperitoneal lymphadenectomy in patients with high-risk endometrial cancer. Because para-aortic node involvement in the absence of pelvic node involvement is uncommon, a reliable finding of negative pelvic lymph nodes (PLN) at intraoperative frozen section examination might allow omitting para-aortic dissection. We analyzed the diagnostic accuracy of frozen section examination of PLN in patients with endometrial cancer. Methods. We reviewed 131 patients with endometrial cancer who underwent surgery including systematic pelvic lymphadenectomy (n = 101) or pelvic and para-aortic lymphadenectomy (n = 27). Intraoperative frozen section examination of PLN was performed in 72 (55%) patients. Results of frozen section examination were compared with those of final histopathology and the diagnostic accuracy of frozen section examination of PLN was calculated. One pathologist measured the diameters of PLN metastases retrospectively. Results. A total of 1063 and 2666 PLN were analyzed by frozen section examination and by final histopathology, respectively. PLN metastases were found in 7 cases (10%) at frozen section examination, and in 17 cases (24%) at final histopathology (false negative rate, 59%). No false positive cases were noted. The mean diameter of all PLN metastases at final histopathology was 4.3 mm, as compared to 9.0 mm for the metastases detected at frozen section analyses. The mean diameter of PLN metastases missed at frozen section examination was 2.0 mm. Conclusion. In this review at a single institution, intraoperative frozen section histology missed nearly two of three endometrial cancer patients with positive nodes. These results do not support tailoring the extent of lymphadenectomy according to the results of frozen section examination. © 2009 Elsevier Inc. All rights reserved.
Introduction Endometrial cancer is the most common female genital malignancy in many Western countries [1]. The disease is staged surgically and this includes total hysterectomy, bilateral salpingo-oophorectomy, pelvic and para-aortic lymphadenectomy and peritoneal cytology. The status of the lymph nodes affects stage, prognosis and therapeutic decisions in patients with this disease [2,3]. Recent prospective data support the trend towards systematic pelvic and para-aortic lymphadenectomy in patients with high-risk endometrial carcinoma [4]. However, pelvic and para-aortic lymphadenectomy is associated with increased operating time and morbidity [5]. Because the likelihood of positive para-aortic lymph nodes in patients with negative pelvic nodes is low, a reliable finding of negative pelvic nodes at intraoperative frozen section examination might be used to decide against para-aortic dissection. We reviewed
⁎ Corresponding author. Fax: +43 316 385 4197. E-mail address: [email protected] (G. Pristauz). 0090-8258/$ – see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.ygyno.2009.07.007
the diagnostic accuracy of frozen section histology of pelvic lymph nodes in patients with endometrial cancer at our institution. Patients and methods Between January 1996 and July 2008 a total of 131 patients with endometrial cancer underwent surgery including lymphadenectomy at our institution. Of these, 101 patients had systematic pelvic lymphadenectomy. 27 patients had systematic pelvic and para-aortic lymphadenectomy, and three patients had resection of bulky nodes only. The latter three patients were not included in the present analysis, because frozen section histology and lymph node sampling was restricted to a limited number of bulky nodes. Whether lymphadenectomy was extended to the para-aortic region depended on intraoperative frozen section analyses and patient's medical status. The present study is based on 72 (55%) patients who had an intraoperative frozen section examination of unfixed grossly visible pelvic lymph nodes (Table 1). In all 72 patients pelvic lymphadenectomy was performed via laparotomy by one of five surgeons in a systematic manner from the femoral ring up to the bifurcation of the
G. Pristauz et al. / Gynecologic Oncology 115 (2009) 12–17 Table 1 Characteristics of 72 patients with endometrial cancer and primary surgery including systematic pelvic lymphadenectomy with intraoperative frozen section preparation. Variable
No. of patients/data
Total Mean age (years) Range (years) FIGO stage Ia Ib Ic IIa IIb IIIa IIIb IIIc IVb Histological type Endometrioid Clear cell Carcinosarcoma (MMMT) Serous papillary Grading G1 G2 G3
72 62 39–74
13
specificity, positive and negative predictive values and their 95%confidence intervals were calculated with StatsDirect 2.6.5 (StatsDirect Ltd., England). The accuracy of the test was evaluated with McNemar's test. P-values b 0.05 were considered significant. Results
7 15 17 2 9 3 1 14 4 53 8 7 4 15 20 37
aorta. The aim was to remove all accessible fat tissue containing lymph nodes around the entire circumference of major pelvic vessels and from the obturator fossa [6]. Tissue was removed as fifteen separate defined anatomic groups [7]. All lymphatic tissue was sent to the pathology laboratory for intraoperative pathologic evaluation. The pathologist palpated the unfixed fat tissue and dissected out all grossly suspicious lymph nodes. Nodes were selected for frozen section histology based on size, consistency and appearance of the cut surface. Larger lymph nodes were usually bisected, depending on size also cut in several pieces as described previously [7]. All intraoperative frozen section analyses and final histologies were performed by one out of three gynecopathologists. After frozen section the remaining tissue was formalin fixed and paraffin embedded for further histopathological analyses. One pathologist (P.R) measured the diameters of lymph node metastases retrospectively.
A total of 1063 pelvic lymph nodes were examined by intraoperative frozen section, with a mean of 15 nodes (median, 13) examined per patient. A total of 2666 pelvic lymph nodes were analyzed in final paraffin sections for definitive histology, with a mean of 36 nodes (median, 35) per patient. Seventeen of the 72 patients (24%) showed pelvic lymph node metastases at final histology. These metastases were detected by intraoperative frozen section analysis in 7 patients (true positive rate, 41%) and by definitive paraffin histology in 10 patients (false negative rate, 59%). Those patients with intraoperative negative frozen section analyses and positive nodes in final histology showed 14 lymph node metastases (mean, 1; range, 1–3 per patient). Table 2 shows details of the 10 patients with false negative frozen section results. A total of 57 of the 2666 pelvic lymph nodes were positive at final histology. Thirteen of those were detected at frozen section (mean, 2; range 1–3 per patient) and 44 at final histology (mean, 2; range 1–10 per patient). No false positive cases by frozen section were noted. Overall 14 positive lymph nodes were missed at frozen section examination. Six positive nodes were not examined at frozen section examination; 5 positive nodes were examined at frozen section examination, but the metastases were only visible in deeper sections after paraffin embedding; and 3 positive nodes were examined at frozen section examination, but the metastases were not diagnosed. The mean diameter of all lymph node metastases at final histopathology was 4.3 mm (median, 3.5; range, 0.1–20 mm) as opposed to 9.0 mm (median, 5.0; range, 2–20 mm) for the metastases detected at frozen section analyses. The mean diameter of lymph node metastases missed at frozen section examination was 2.0 mm (median, 1.2; range, 0.1–6 mm) (Table 3). Table 4 shows the sensitivity, specificity, positive and negative predictive values of frozen section as a diagnostic test for predicting the final node status. Although the false negative rate was 59%, the negative predictive value was 85%.
Statistical analysis Discussion Frozen section histology was analyzed as a diagnostic test and its results were compared to the results of final histology. If the results from the paraffin sections did not differ from the results of the frozen sections the diagnosis was considered to be correct. Sensitivity,
This analysis of intraoperative frozen section examination of pelvic nodes in patients with endometrial cancer at a single institution shows a sensitivity of 41% and a false negative rate of 59%. Pathologists
Table 2 Details of 10 patients with false negative frozen section analyses. Patient
FIGO stage
Histology
Frozen section
Final histology
Number of nodes pos./examined
Number of nodes pos./examined
1
IIIb
Carcinosarcoma
0/25
3/42
2
IIIc
Endometrioid
0/21
2/35
3 4 5 6 7 8 9 10
IIb IIIc IIIc IIIc IIIc IIIc IIIc IIIC
Endometrioid Endometrioid Seröses Endometrioid Endometrioid Endometrioid Endometrioid Endometrioid
0/15 0/8 0/8 0/19 0/11 0/10 0/38 0/20
2/35 1/30 1/10 1/30 1/21 1/60 1/62 1/71
Node region with metastases
Diameter of lymph node metastases in mm
Reasons of discordance between frozen section and final analyses
Right external iliac Left external iliac Right external iliac Left common iliac Right external iliac lateral Left external iliac Left external iliac Right external iliac Left common iliac Right obturator (above nerve) Left obturator (above nerve) Right external iliac
4 1.5, 5 1.1 2.5 1, 0.1 1.2 1.5 1 0.5 1 1.2 6
C C+B A B C+A A A A C C A B
A = missed node not examined in frozen section examination. B = missed node examined in frozen section examination, but metastasis not diagnosed. C = missed node examined in frozen section examination, but metastasis only visible in deeper sections after paraffin embedding.
14
G. Pristauz et al. / Gynecologic Oncology 115 (2009) 12–17
Table 3 Number of pelvic lymph nodes metastases and mean values of largest diameters as detected by intraoperative frozen section and by final histology of 72 patients with endometrial cancer. Lymph node region
Frozen section examination Patients with positive nodes
No. of pelvic node metastases
Right External iliac Common iliac Internal iliac Obturator (above nerve) Obturator (below nerve) External iliac lateral Common iliac lateral Subaortal
4 1 0 1 0 1 0 0
4 2 0 1 0 1 0 0
Left External iliac Common iliac Internal iliac Obturator (above nerve) Obturator (below nerve) External iliac lateral Common iliac lateral
2 0 0 1 0 0 0
4 0 0 1 0 0 0
Final histopathology Mean values of largest diameters of node metastases, mm (median) 8.8 (7.5) 3.0 (3.0) – 5.0 (5.0) – 20.0 (20.0) – –
8.5 (8.5) – – 4.0 (4.0) – – –
picked suspicious pelvic lymph nodes on the basis of size, consistency and overall appearance and detected 13 lymph node metastases out of 1063 lymph nodes examined at frozen section for a ratio of 1:82 node metastasis to nodes examined. This detection rate should be superior to a random selection rate of lymph nodes, but at paraffin embedding 57 nodes out of 2666 examined lymph nodes were detected with a ratio of 1:47. In contrast a similar study in patients with cervical cancer of our institution showed a 17% false negative rate for frozen section histology with the same pathologist as in our study detecting lymph node metastases at frozen section with a ratio of 1:18 node metastasis to nodes examined and a ratio of 1:35 at final histology [7,8]. The diagnostic accuracy of lymph node frozen section analyses depends on palpation of the lymphatic fatty tissue removed at surgery, i.e. on “who's at the sink”. Whether a lymph node or a lymph node metastasis is palpated by the pathologist depends on the presence of lymph node size and desmoplastic stromal reaction. Desmoplasia is defined as a newly formed fibrotic (collagenous) stroma surrounding tumor cells [9] and makes the lymph node rigid and palpable for the surgeon or pathologist. Positive pelvic lymph nodes in patients with squamous cell carcinomas of the cervix show an increased desmoplastic reaction in positive lymph nodes compared to positive lymph nodes in patients with adenocarcinoma of the uterus (Figs. 1 and 2). Para-aortic lymph node metastases occur in about 15% of patients with high-risk endometrial cancer and in about 60% of patients with pelvic lymph node metastases [4]. In contrast para-aortic metastases
Table 4 Sensitivity, specificity, positive predictive value and negative predictive value of frozen section as a diagnostic test for predicting the final node status. No. of patients True-positive True-negative False negative False positive
7 55 10 0
Sensitivity Specificity Positive predictive value Negative predictive value Test accuracy
Test characteristics with 95% confidence intervals 41.2% (18.4–67.1) 100.0% (93.5–100.0) 100.0% (59.0–100.0) 84.6% (73.5–92.4) 86.1%(75.9–93.1)
McNemar's test: P = 0.0016.
Patients with positive nodes
No. of pelvic node metastases
Mean maximum diameter of node metastases (median)
8 1 0 3 0 2 0 1
8 2 0 5 0 3 0 1
5.9 (5.0) 3.0 (3.0) – 2.3 (2.0) – 7.0 (1.0) – 2.0 (2.0)
6 3 1 4 0 1 1
12 3 2 6 0 1 1
5.7 (4.5) 1.7 (1.0) 4.5 (4.5) 2.7 (2.5) – 10.0 (10.0) 2.0 (2.0)
are uncommon in patients with negative pelvic nodes [4,10,11]. Imaging techniques such as computed tomography and magnetic resonance imaging are unreliable for evaluating pelvic or para-aortic lymph nodes and 2-[18F]fluoro-2-deoxy-D-glucose (FDG) positron emission tomography (PET) is only moderately sensitive for predicting lymph node metastasis preoperatively [12]. Therefore intraoperative frozen section analysis of pelvic lymph nodes was thought to be useful to identify patients for para-aortic lymphadenectomy. Lymphadenectomy may also have a therapeutic effect in patients with endometrial cancer, with an increase in 8-year disease specific survival from 60% to 85% when comparing patients who had extensive lymphadenectomy with those who did not have extensive lymphadenectomy [13–15]. However, pelvic and aortic lymphadenectomy is associated with longer operating times, longer hospital stays and higher estimated blood loss compared to patients undergoing pelvic lymphadenectomy only [15]. Why were positive nodes missed at frozen section? The first reason is that positive nodes were not palpated and thus not selected for frozen section. This was the case in 6 patients. Similarly McMeekin et al. [16] found that 36% of nodes with metastatic disease were abnormal on palpation. Also lymph node metastases are small. In this series 10 out of 13 missed metastases at frozen section were ≤2 mm. Girardi et al. [17] reported that 37% of node metastases in patients with endometrial cancer were ≤ to 2 mm in diameter. The most important factors for the failure of intraoperative frozen section examination seem to be the size of metastases [8]. Eight out of 14 positive lymph nodes missed at frozen section examination in our patients were picked correctly by the pathologist but metastases were not diagnosed due to the small diameters of the metastases. Therefore serial section at intraoperative frozen section examination is required to achieve a low false negative rate of frozen section examination but impossible due to the time limit, limited man power as well the large amount of nodes excised. Recent data suggest the possibility of identifying sentinel lymph node in patients with endometrial cancer [18,19] which would allow focusing intraoperative work on a few nodes. Finally, metastases from endometrial cancer may not induce a desmoplastic reaction like that seen with squamous cell carcinoma of the cervix. These results indicate that at our institution frozen section examination of pelvic lymph nodes in patients with endometrial cancer is not accurate enough to routinely guide intraoperative management decisions regarding the extent of lymphadenectomy. Although only 10 of the 72 overall patients (14%) were not correctly
G. Pristauz et al. / Gynecologic Oncology 115 (2009) 12–17
15
Fig. 1. Comparison of desmoplastic reaction in lymph node metastases of endometrial and cervical cancer. (A and B) Pelvic lymph node with metastatic endometrial cancer.
classified by frozen section examination, this would translate into approximately two of three patients with positive pelvic nodes not being identified intraoperatively. Selecting only patients with positive intraoperative pelvic nodes for para-aortic lymph node dissection would leave one out of ten patients with positive para-aortic nodes without para-aortic dissection (Fig. 3). It would be appropriate for other institutions using intraoperative frozen section histology of lymph nodes to plan the extent of surgery to analyze their results in this setting.
Conflict of interest statement None of the authors has financial conflicts of interests.
References [1] American Cancer Society. Available at: http://www.cancer.org. [2] Larson DM, Johnson K, Olson KA. Pelvic and para-aortic lymphadenectomy for surgical staging of endometrial cancer: morbidity and mortality. Obstet Gynecol 1992;79:998–1001.
16
G. Pristauz et al. / Gynecologic Oncology 115 (2009) 12–17
Fig. 2. Comparison of desmoplastic reaction in lymph node metastases of endometrial and cervical cancer. (A and B) Pelvic lymph node metastases of cervical cancer with increased desmoplastic reaction compared to lymph node metastases of endometrial cancer (Fig. 1). [3] Possover M, Krause N, Plaul K, Kühne-Heid R, Schneider A. Laparoscopic paraaortic and pelvic lymphadenectomy: experience with 150 patients and review of the literature. Gynecol Oncol 1998;71:19–28. [4] Mariani A, Dowdy SC, Cliby WA, Gostout BS, Jones MB, Wilson TO, et al. Prospective assessment of lymphatic dissemination in endometrial cancer: a paradigm shift in surgical staging. Gynecol Oncol 2008;109:11–8. [5] Benedetti-Panici P, Maneschi F, Cutillo G. Pelvic and aortic lymphadenectomy. Surg Clin North Am 2001;81:841–58. [6] Winter R. Pelvic and paraaortic lymphadenectomy. In: Burghardt E, Webb MJ, Monoghan JM, Kinderman G, editors. Surgical Gynecological Oncology. New York: Thieme; 1993. p. 281–93.
[7] Bader AA, Winter R, Moinfar F, Petru E, Pristauz G, Scholz H, et al. Is intraoperative frozen section analysis of pelvic lymph nodes accurate after neoadjuvant chemotherapy in patients with cervical cancer? Gynecol Oncol 2006;103:106–12. [8] Scholz HS, Lax SF, Benedicic C, Tamussino K, Winter R. Accuracy of frozen section examination of pelvic lymph nodes in patients with FIGO stage IB1 to IIB cervical cancer. Gynecol Oncol 2003;90:605–9. [9] Sternberg S, editor. Histology for Pathologists. 2nd. Lippincott-Raven; 1997. p. 143–59. [10] Mariani A, Keeney GL, Aletti G, Webb MJ, Haddock MG, Podratz KC. Endometrial carcinoma: paraaortic dissemination. Gynecol Oncol 2004;92:833–8.
G. Pristauz et al. / Gynecologic Oncology 115 (2009) 12–17
17
Fig. 3. Translation of our study results into clinical practice. [11] Goudge C, Bernhard S, Cloven NG, Morris P. The impact of complete surgical staging on adjuvant treatment decisions in endometrial cancer. Gynecol Oncol 2004;93:536–9. [12] Horowitz NS, Dehdashti F, Herzog TJ, Rader JS, Powell MA, et al. Prospective evaluation of FDG-PET for detecting pelvic and para-aortic lymph node metastasis in uterine corpus cancer. Gynecol Oncol 2004;95:546–51. [13] Kilgore LC, Patridge EE, Alvarez RD, et al. Adenocarcinoma of the endometrium: survival comparisons of patients with and without pelvic node sampling. Gynecol Oncol 1995;56:29–33. [14] Chan JK, Cheung MK, Huh WK, Osann K, Husain A, Teng NN, et al. Therapeutic role of lymph node resection in endometrioid corpus cancer: a study of 12,333 patients. Cancer 2006;107:1823–30. [15] Cragun JM, Havrilesky LJ, Calingaert B, Synan I, Secord AA, Soper JT, et al.
[16]
[17] [18]
[19]
Retrospective analysis of selective lymphadenectomy in apparent early-stage endometrial cancer. J Clin Oncol 2005;23:3653–5. McMeekin DS, Lashbrook D, Gold M, Scribner DR, Kamelle S, Tillmanns TD, et al. Nodal distribution and its significance in FIGO stage IIIc endometrial cancer. Gynecol Oncol 2001;82:375–9. Girardi F, Petru E, Heydarfadai M, Haas J, Winter R. Pelvic lymphadenectomy in the surgical treatment of endometrial cancer. Gynecol Oncol 1993;49:177–80. Perrone AM, Casadio P, Formelli G, Levorato M, Ghi T, Costa S, et al. Cervical and hysteroscopic injection for identification of sentinel lymph node in endometrial cancer. Gynecol Oncol 2008;111:62–7. Abu-Rustum NR, Khoury-Collado F, Pandit-Taskar N, Soslow RA, Dao F, Sonoda Y, et al. Sentinel lymph node mapping for grade 1 endometrial cancer: is it the answer to the surgical staging dilemma? Gynecol Oncol 2009;113:163–9.