Fax: (919) 966-2646
Uterine serous and grade 3 endometrioid carcinomas†
Is there a survival difference?
Article first published online: 27 SEP 2004
Copyright © 2004 American Cancer Society
Volume 101, Issue 10, pages 2214–2221, 15 November 2004
How to Cite
Boruta, D. M., Gehrig, P. A., Groben, P. A., Bae-Jump, V., Boggess, J. F., Fowler, W. C. and Van Le, L. (2004), Uterine serous and grade 3 endometrioid carcinomas. Cancer, 101: 2214–2221. doi: 10.1002/cncr.20645
See related editorial on pages 2152–4.
- Issue published online: 29 OCT 2004
- Article first published online: 27 SEP 2004
- Manuscript Accepted: 17 JUN 2004
- Manuscript Revised: 11 JUN 2004
- Manuscript Received: 29 JAN 2004
- uterine serous carcinoma (USC);
- papillary serous;
- endometrioid carcinoma;
Serous components within endometrial carcinoma are reportedly poor prognosticators. However, to the authors' knowledge the percentage of tumors which must be comprised of a serous component in order to affect outcome is unknown. The authors compared overall survival (OS) in women with endometrial carcinomas comprised of various percentages of uterine serous carcinoma (USC) with that of women with International Federation of Gynecology and Obstetrics (FIGO) Grade 3 endometrioid carcinoma (G3EC) to determine whether outcomes varied between these two poorly differentiated histologies.
Data concerning women with either G3EC or USC who were diagnosed between January 1990 and November 2000 were collected retrospectively. Cases were reviewed to confirm diagnosis and estimate the fraction of tumor comprised of USC. Variables assessed included patient age and race, tumor stage, and lymphovascular space invasion. Associations between variables were tested using the Fisher exact test. The Kaplan–Meier method was used to evaluate OS with comparisons performed using the log-rank test.
Fifty-two women with G3EC and 87 women with USC were identified. The OS of women with tumors comprised of > 50% USC was found to be significantly worse compared with women with G3EC (hazard ratio [HR] of 2.4; 95% confidence interval [95% CI], 1.2–5.2). Women with USC were more likely to present with extrauterine disease (odds ratio of 2.2; 95% CI, 1.1–4.5). The 5-year survival rate for women with G3EC was 75% compared with 41% for women with tumors that were > 50% USC (P = 0.01). There was a significant trend toward a worse OS in women with even 10% USC compared with women with G3EC.
USC involving > 50% of an endometrial carcinoma was found to be predictive of worse OS compared with the OS of women with G3EC. In patients with early-stage disease, a trend toward a worse prognosis was found to exist when USC comprised even 10% of a tumor. Investigation into the treatment of endometrial carcinoma should include and document tumors with any percentage comprised of USC. Cancer 2004. © 2004 American Cancer Society.
Endometrial cancer is the most common gynecologic malignancy in the U.S. It was estimated that 40,100 cases would be diagnosed in 2003 and that 6800 women would die from this disease.1 Whereas endometrioid carcinoma has relatively good cure rates, endometrial carcinomas containing a serous component have a much poorer prognosis, even in patients with early-stage disease.2–11 In a review of uterine serous carcinomas (USCs), women with FIGO Stage I disease had a 5-year survival rate of 60%, compared with a 5-year survival rate of 80–90% in patients with Stage I endometrioid carcinoma.2 Although the overall survival (OS) in women with USC is reported to be inferior to the OS in women with endometrioid carcinoma in general, to our knowledge no difference has specifically been established between USC and its poorly differentiated endometrioid counterpart.
Lauchlan12 and Hendrickson et al.13 first established USC as a distinct subtype of endometrial carcinoma. Since their reports, its presence has been repeatedly recognized as a poor prognostic factor. Surgical management with staging followed by tailored radiotherapy, although relatively effective in treating the endometrioid carcinoma, has to our knowledge failed to achieve similar rates of survival in women with USC. Compared with endometrioid carcinoma in general, USC is more frequently metastatic at the time of surgical staging and has a propensity for distant recurrence. Unfortunately, this is true even for women with apparently noninvasive endometrial lesions.14 Therefore, studies have begun to explore the role of systemic chemotherapy in the treatment of patients with USC.10, 15–18
Within any endometrial tumor, serous histologic components are frequently admixed with the more common endometrioid histology. Exactly what percentage of serous component is necessary before its presence becomes a negative prognostic indicator is to our knowledge unknown. Although the literature frequently includes women whose endometrial tumors were comprised of at least 25% serous histology, support for this is based on limited published data.11, 19 Just as current treatment recommendations for USC vary widely, so too do histologic inclusion criterion. As large, cooperative groups begin to consider the design of studies to evaluate treatment strategies for USC, specific criterion for entry, including the percentage of serous component required, will be critical.
We compared OS and progression-free survival (PFS) in women with endometrial carcinomas that were comprised of various percentages of USC with those of women with poorly differentiated (International Federation of Gynecology and Obstetrics [FIGO Grade 3]) endometrioid carcinoma (G3EC) to determine which histologic cell type portends a worse prognosis.
MATERIALS AND METHODS
Between January 1990 and November 2000, 243 women with USC or G3EC were treated at the University of North Carolina at Chapel Hill School of Medicine. The study group was comprised of 139 cases for which pathologic material was available for reexamination. A retrospective chart review was performed with extraction of clinical information including age at diagnosis, race, surgical procedure and staging, postoperative therapy, PFS, recurrence site, and length of follow-up. Surgical staging was based on the 1988 FIGO criteria20 and included peritoneal washings, total abdominal hysterectomy, bilateral salpingo-oophorectomy, and pelvic and paraaortic lymph node dissection. The study met the exemption criteria set forth by the Institutional Review Board at the University of North Carolina at Chapel Hill.
The diagnosis of USC was based on criteria described by Hendrickson et al.,13 which included a complex papillary architecture with tufted stratification of lining epithelium, marked nuclear pleomorphism, a high nuclear-to-cytoplasmic ratio, macronuclei, and a high mitotic rate. The presence of psammoma bodies was not necessary for the diagnosis of USC. A median number of 2 of slides per case (range: 1–23 slides per case) were available for review by the study pathologist (P.A.G.), who was blinded to all clinical data. Each was assigned to 1 of the following categories: 0%, < 10%, 10–24%, 25–50%, and > 50% USC.
Statistical analysis was performed using GraphPad Prism version 3.00 for Windows (GraphPad Software, San Diego CA). Associations between variables were tested using the Fisher exact test. OS and PFS, as well as 5-year survival probabilities, were obtained using the Kaplan–Meier product-limit method.21 Comparisons between survival curves were performed using the Mantel–Haenszel log-rank test and the log-rank test for trend.22 The log-rank test for trend calculates a P value testing the null hypothesis that there is no linear trend between the median survivals of different groups.
Study Population Characteristics
Of the 139 women who comprised our study group, 52 (37%) had G3EC with no serous component and 87 (63%) had a mixed tumor with a varying percentage of serous component. The distribution of clinical variables according to histologic category and early versus late stage is shown in Table 1. Women with G3EC did not appear to differ from those with USC with respect to age (≤ 65 years vs. > 65 years), race (white vs. other), weight, nulliparity, depth of myometrial invasion (≤ 50% vs. > 50%), or the presence of lymphovascular space invasion (LVSI) (all P values > 0.05). Women with USC were found to be more likely to present with cervical involvement (odds ratio [OR] of 3.3; 95% confidence interval [95% CI], 1.5–7.3) and extrauterine disease (OR of 2.2; 95% CI, 1.1–4.5). The administration of adjuvant therapy, including radiotherapy, chemotherapy, or combinations thereof, was balanced between women with G3EC and USC, both in the early and late stages of disease (Table 1).
|Histology No. (%)||All stages||Early stage (I/II)||Late stage (III/IV)|
|G3EC 52 (37)||USC 87 (63)||P value||G3EC 26 (46)||USC 31 (54)||P value||G3EC 21 (28)||USC 53 (72)||P value|
|≤ 65||22 (42)||29 (33)||0.36||11 (42)||12 (39)||0.79||9 (43)||17 (32)||0.43|
|> 65||30 (58)||58 (67)||15 (58)||19 (61)||12 (57)||36 (68)|
|White||37 (71)||49 (56)||0.10||19 (73)||16 (52)||0.11||15 (71)||31 (58)||0.43|
|Other||15 (29)||38 (44)||7 (27)||15 (48)||6 (29)||22 (42)|
|Myometrial invasion (%)|
|≤ 50||25 (53)||41 (48)||0.72||18 (69)||19 (61)||0.59||7 (33)||22 (42)||0.60|
|> 50||22 (47)||44 (52)||8 (31)||12 (39)||14 (67)||31 (58)|
|Yes||29 (62)||55 (66)||0.70||11 (42)||14 (47)||0.79||18 (86)||40 (77)||0.53|
|No||18 (38)||28 (34)||15 (58)||16 (53)||3 (14)||12 (23)|
|Yes||12 (26)||41 (49)||0.01||4 (15)||12 (39)||0.08||8 (40)||29 (57)||0.29|
|No||34 (74)||42 (51)||22 (85)||19 (61)||12 (60)||22 (43)|
|Percent of tumor comprised of USC|
|0 (100% G3EC)||52 (100)||0 (0)||26 (100)||0 (0)||21 (100)||0 (0)|
|< 10||0 (0)||8 (9)||0 (0)||5 (16)||0 (0)||3 (6)|
|10–24||0 (0)||16 (18)||0 (0)||7 (23)||0 (0)||9 (17)|
|25–50||0 (0)||12 (14)||0 (0)||6 (19)||0 (0)||5 (9)|
|> 50||0 (0)||51 (59)||0 (0)||13 (42)||0 (0)||36 (68)|
|Yes||43 (83)||70 (80)||0.82||18 (69)||24 (77)||0.55||20 (95)||43 (81)||0.16|
|RT only||27 (52)||37 (42)||16 (62)||21 (68)||6 (29)||14 (26)|
|Chemotherapy only||10 (19)||21 (24)||2 (7)||2 (6)||8 (37)||18 (34)|
|Both RT and chemotherapy||6 (12)||12 (14)||0 (0)||1 (3)||6 (29)||11 (21)|
|None||9 (17)||17 (20)||8 (31)||7 (23)||1 (5)||10 (19)|
Peritoneal or omental sampling was performed as part of surgical staging in 73 of the 139 women in the current study group (53%). In women with early-stage disease (FIGO Stages I and II), a peritoneal or omental specimen was obtained in 11 of 26 women (42%) with G3EC and 11 of 31 women (36%) with USC. Of the 21 women with advanced stage G3EC (FIGO Stages III and IV) and the 53 women with advanced stage USC, only 1 woman with a G3EC was classified as such solely secondary to metastatic disease within an omental or peritoneal biopsy.
Survival Comparisons Examining Variables Other than Histology
The survival analysis of women with either G3EC or USC when variables other than histology were examined is detailed in Table 2. In women with G3EC, stage of disease, LVSI, and depth of myometrial invasion were found to have a significant effect on OS whereas age, race, and adjuvant therapy did not. In women with USC, only race was found to significantly affect OS, whereas the potential effect of age, stage of disease, LVSI, and adjuvant therapy did not achieve statistical significance. Both race and stage were found to significantly impact PFS in women with either G3EC or USC.
|Variable||Overall survival||Progression-free survival|
|HR||95% CI||P value||HR||95% CI||P value|
|Stage, early vs. late||0.17||0.06–0.70||0.01||0.24||0.08–0.68||0.01|
|Age, ≤ 65 yrs vs. > 65 yrs||0.81||0.25–2.64||0.73||0.53||0.21–1.40||0.21|
|Race, white vs. other||0.48||0.12–1.58||0.20||0.39||0.11–0.90||0.03|
|LVSI, present vs. absent||6.54||1.07–13.26||0.04||2.81||0.87–6.98||0.09|
|Myometrial invasion, ≤ 50% vs. > 50%||0.18||0.06–0.71||0.01||0.26||0.09–0.73||0.01|
|Stage, early vs. late||0.57||0.26–1.28||0.17||0.45||0.25–0.88||0.02|
|Age, ≤ 65 yrs vs. > 65 yrs||1.01||0.44–2.35||0.98||1.07||0.55–2.09||0.84|
|Race, white vs. other||0.32||0.14–0.70||<0.01||0.55||0.29–0.98||0.04|
|LVSI, present vs. absent||1.98||0.82–4.28||0.14||1.47||0.76–2.78||0.26|
|Myometrial invasion, ≤ 50% vs. > 50%||0.84||0.38–1.84||0.65||0.82||0.44–1.52||0.52|
Survival in Women with G3EC and Women with USC
|Variable||All stages||Early stage (I/II)||Late stage (III/IV)|
|HR||95% CI||P value||HR||95% CI||P value||HR||95% CI||P value|
|G3EC vs. USC (any %)||0.53||0.28–1.02||0.06||0.24||0.08–0.97||0.04||0.89||0.39–2.05||0.78|
|G3EC vs. < 10% USC||1.31||0.20–8.08||0.79||0.20||0.001–2.74||0.14||0.26|
|G3EC vs. 10–24% USC||0.45||0.09–1.30||0.12||0.23||0.01–1.61||0.23||0.58||0.10–2.49||0.40|
|G3EC vs. 25–50% USC||2.42||0.45–8.30||0.38||0.49||0.02–7.26||0.55||0.20|
|G3EC vs. < 50% USC||0.86||0.33–2.24||0.75||0.30||0.06–1.52||0.14||1.56||0.44–5.30||0.45|
|G3EC vs. > 50% USC||0.42||0.19–0.83||0.01||0.17||0.02–0.68||0.02||0.75||0.32–1.76||0.51|
|< 50% vs. > 50% USC||2.23||0.97–4.64||0.06||0.56||0.12–2.32||0.40||0.47||0.18–1.45||0.21|
|≤ 65 vs. > 65||0.77||0.40–1.49||0.45||0.25||0.08–1.01||0.05||1.39||0.61–3.32||0.41|
|White vs. other||0.33||0.15–0.59||< 0.01||0.24||0.06–0.83||0.03||0.35||0.12–0.69||0.01|
|Present vs. absent||2.70||1.19–4.73||0.01||3.23||0.87–13.11||0.08||1.44||0.53–3.65||0.50|
|≤ 50% vs. > 50%||0.55||0.28–1.06||0.07||1.07||0.28–4.12||0.92||1.87||0.81–4.02||0.15|
|Yes vs. no||0.74||0.30–1.69||0.44||0.55||0.12–2.08||0.35||0.42||0.06–1.26||0.10|
|Variable||All stages||Early stage (I/II)||Late stage (III/IV)|
|HR||95% CI||P value||HR||95% CI||P value||HR||95% CI||P value|
|G3EC vs. USC (any %)||0.56||0.34–0.95||0.03||0.36||0.14–1.06||0.06||0.73||0.38–1.42||0.35|
|G3EC vs. < 10% USC||1.03||0.24–4.44||0.97||0.49||0.02–6.57||0.51||1.47||0.24–8.30||0.71|
|G3EC vs. 10–24% USC||0.41||0.10–0.80||0.02||0.28||0.02–1.14||0.07||0.47||0.11–1.27||0.11|
|G3EC vs. 25–50% USC||4.27||0.79–7.80||0.12||1.13||0.13–9.53||0.91||0.11|
|G3EC vs. < 50% USC||0.83||0.39–1.76||0.61||0.48||0.12–1.79||0.26||1.01||0.38–2.66||0.98|
|G3EC vs. > 50% USC||0.43||0.23–0.73||< 0.01||0.27||0.05–0.83||0.03||0.64||0.32–1.29||0.21|
|< 50% vs. > 50% USC||2.07||1.08–3.66||0.03||0.55||0.16–1.81||0.31||0.63||0.29–1.40||0.26|
|≤ 65 vs. > 65||0.91||0.54–1.53||0.72||0.37||0.14–1.10||0.08||1.59||0.86–3.31||0.13|
|White vs. other||0.45||0.24–0.70||< 0.01||0.28||0.09–0.74||0.01||0.47||0.21–0.82||0.01|
|Present vs. absent||1.79||1.00–3.00||< 0.05||1.41||0.49–4.18||0.52||1.14||0.53–2.44||0.73|
|≤ 50% vs. > 50%||0.59||0.34–0.99||< 0.05||1.05||0.36–3.09||0.92||0.61||0.33–1.16||0.13|
|Yes vs. no||0.85||0.42–1.68||0.62||0.62||0.17–1.92||0.37||0.55||0.15–1.37||0.16|
Early-stage (Stage I/II) disease
In women with early-stage disease whose tumors were comprised of > 50% USC, the OS and PFS were found to be significantly worse than in similarly staged women with G3EC (hazards ratio [HR] of 5.9, 95% CI, 1.5–61.4; and HR of 3.7, 95% CI, 1.2–19.3, respectively). The log-rank test for trend was significant when comparing the PFS and OS of women with G3EC and < 10%, 10–50%, and > 50% USC (P = 0.04 and P = 0.03, respectively). The PFS and OS at 5 years in women with G3EC was found to be 82% and 89%, respectively, compared with 35% and 43%, respectively, in women with > 50% USC (P = 0.03 and P = 0.02, respectively).
No significant difference in recurrence rate was observed when women with early-stage G3EC were compared with women with early-stage USC (P > 0.05) (Table 5). Although none of the women with locally recurrent G3EC had been treated with adjuvant therapy, two of the three women with locally recurrent USC had received adjuvant radiotherapy (one to the whole pelvis and one to the whole abdomin). Distant, intraabdominal recurrence was noted in three women with early-stage USC who had undergone whole abdominal radiotherapy and in one other woman who received pelvic radiotherapy alone. Of the six women who developed a pulmonary recurrence of early-stage disease, four patients had received adjuvant whole pelvic radiotherapy (two with USC and two with G3EC) and the remaining two women with USC received no adjuvant therapy.
|Histology (no.)||All stages||Early stage (I/II)||Late stage (III/IV)|
|G3EC (52)||USC (87)||P value||G3EC (26)||USC (31)||P value||G3EC (21)||USC (53)||P value|
|Recurrent disease, no. (%)|
|Yes||13 (25)||26 (30)||0.32||4 (15)||11 (35)||0.14||5 (24)||14 (26)||1.0|
|No||33 (63)||44 (51)||22 (85)||20 (65)||10 (48)||23 (44)|
|Persistent disease, no. (%)||6 (12)||17 (19)||0 (0)||0 (0)||6 (28)||16 (30)|
|Site of recurrence, no. (%)|
|Local||3 (23)||8 (31)||0.72||2 (50)||3 (27)||0.56||0 (0)||5 (36)||0.26|
|Vagina||3 (23)||6 (23)||2 (50)||3 (27)||0 (0)||3 (21)|
|Pelvis||0 (0)||2 (8)||0 (0)||0 (0)||0 (0)||2 (14)|
|Distant||10 (77)||18 (69)||2 (50)||8 (73)||5 (100)||9 (64)|
|Abdomen||5 (39)||9 (35)||0 (0)||4 (36)||3 (60)||5 (36)|
|Lung||3 (23)||6 (23)||2 (50)||4 (36)||0 (0)||1 (7)|
|Other||2 (15)||3 (11)||0 (0)||0 (0)||2 (40)||3 (21)|
In the current study, three women with early-stage USC received adjuvant chemotherapy (carboplatin and paclitaxel in two patients and cisplatin and doxorubicin in one patient), one of whom also received radiotherapy to the vaginal cuff. At the time of last follow-up, none of the patients had experienced a disease recurrence after a median follow-up of 31 months (range, 3–65 months). Two women with early-stage G3EC received adjuvant chemotherapy alone. At the time of last follow-up, neither woman had experienced recurrence after 30 months and 60 months, respectively, of follow-up.
Late-stage (Stage III/IV) disease
The PFS and OS in women with late-stage disease whose tumors were comprised of USC of any percentage, including > 50%, did not appear to differ from the PFS and OS reported in women with G3EC (P > 0.05). In addition, in women with late-stage disease, the log-rank test for trend was not found to be significant when comparing the PFS and OS of those women with G3EC and < 10%, 10–50%, and > 50% USC. The PFS and OS at 5 years in women with G3EC was reported to be 50% and 58%, respectively, compared with 34% and 53%, respectively, in women with USC at any percentage (P > 0.05).
Although the prescription and administration of postoperative adjuvant therapy in women with late-stage disease was heterogeneous, no significant difference was observed between women with G3EC compared with women with USC (Table 1). The most common treatments included whole abdominal radiotherapy alone (29% and 21%, respectively, of women with late-stage G3EC and USC) or whole pelvic radiotherapy sequenced with chemotherapy (29% and 17%, respectively, of women with late-stage G3EC and USC). Chemotherapy agents included combinations of doxorubicin, cyclophosphamide, cisplatin, carboplatin, paclitaxel, etoposide, altretamine, and megestrol acetate. One woman with late-stage USC who received whole abdominal radiotherapy and another who received vaginal cuff brachytherapy also received sequential chemotherapy. Three women (6%) with late-stage USC received pelvic radiotherapy alone. Chemotherapy alone was administered to 8 women with late-stage G3EC (38%) and 18 women with late-stage USC (34%).
Fifteen women with advanced stage G3EC (71%) and 35 women with advanced stage USC (66%) were found to be free of macroscopic disease after surgical staging and cytoreduction. Both the OS and PFS were found to be significantly worse in women with residual macroscopic tumor after surgical staging and cytoreduction (all P values < 0.001). The OS at 5 years was reported to be 14% and 83%, respectively, in women with and without residual macroscopic G3EC. Similarly, the OS at 5 years was reported to be 21%and 66%, respectively, in women with and without residual macroscopic USC.
Stages I-IV disease
When women with all stages of disease were included in survival analysis (Tables 3 and 4), both the OS and PFS were found to be significantly worse in those women whose tumors were comprised of > 50% USC compared with those with G3EC (HR of 2.4, 95% CI, 1.2–5.2; and HR of 2.3, 95% CI, 1.4–4.4, respectively). In addition, the PFS in women whose tumors were comprised of 10–24% USC also was found to be significantly worse compared with that of women with G3EC (HR of 2.5, 95% CI, 1.2–9.9). The log-rank test for trend was significant when comparing the PFS and OS of women with G3EC and < 10%, 10–50%, and > 50% USC (P < 0.01 and P < 0.02, respectively). The PFS and OS at 5 years in women with G3EC was 64% and 75%, respectively, compared with 28% and 41%, respectively, in women with > 50% USC (P < 0.01 for both).
In comparison with women with tumors comprised of < 50% USC, a tumor comprised of > 50% USC was found to be significantly associated with late-stage disease (FIGO Stage III/IV) at the time of diagnosis (OR of 2.9; 95% CI, 1.7–7.3). The presence of > 50% USC was not found to be significantly associated with an increased overall risk of recurrence when compared with < 50% USC (OR of 1.92; 95% CI, 0.71–5.1). Furthermore, in women experiencing recurrent disease, a tumor comprised of > 50% USC did not appear to be predictive of an increased risk of extrapelvic recurrence when compared with a tumor with < 50% USC (OR of 0.66; 95% CI, 0.08–3.6).
Although a large number of studies have established a worse prognosis for women with USC compared with women with uterine endometrioid carcinoma in general, to our knowledge the current study is the first to establish a survival difference between women with USC and those with G3EC, both of which are considered poorly differentiated histologic types. Both the PFS and OS were found to be significantly worse in women diagnosed with early-stage (FIGO Stages I and II) USC compared with women with similar stage G3EC. This difference was not found to be present in women with late-stage (FIGO Stages III and IV) disease. The OS in women with G3EC was reported to be negatively impacted by stage, LVSI, and the depth of myometrial invasion (P < 0.05). These variables were not found to be significant prognosticators of OS in women with USC, in whom race alone was found to be of statistical significance. Regardless, there were no significant differences noted with regard to the distribution of these or other variables between the population of women with early-stage or late-stage G3EC or USC.
The current study also attempted to correlate potential survival differences based on the percentage of serous component that comprises the overall uterine tumor specimen. Although criterions vary between institutions, uterine carcinomas frequently have been characterized as USC when their serous histologic component comprises ≥ 25% of the entire specimen. In the current study, a significant difference in OS between women with G3EC and women with > 50% USC was observed, whereas there was no difference in OS noted when the specimens contained < 50% USC. Admittedly, a limited sample size restricted the development of conclusions regarding the potential impact of < 50% USC on survival compared with G3EC. The log-rank test for trend was significant when applied to OS in women with < 10%, 10–50%, and > 50% USC. This suggests the presence of a linear trend between the median survivals of these different groups. Furthermore, there was a significant difference in PFS noted between women with G3EC and either 10–24% or > 50% USC. This suggests the presence of an underlying difference in the behavior of uterine carcinomas with as little as a 10% serous component.
Other studies exploring survival in women with early-stage USC have been unable to detect a difference in the OS or PFS between women with G3EC or USC.23–25 Cirisano et al.24 reported a significant difference in OS in women with either clinically or surgically staged early-stage clear cell carcinoma or USC (n = 71 and n = 39, respectively) compared with similarly staged women with all grades of endometrioid uterine carcinoma (n = 503 and n = 441) (P < 0.001). When the endometrioid group was restricted to those patients with poorly differentiated histology (G3EC) (n = 59 and n = 42), the difference was no longer present (P = 0.11). Although the 5-year survival in women with early-stage disease appeared to differ between women with clear cell carcinoma or USC and those with G3EC (56% vs. 71%), it was not found to be statistically significant. The inclusion of clear cell histology into the group with USC limits our conclusions because to our knowledge the specific impact of clear cell histology on survival is not well understood. Furthermore, the inclusion of women with < 25% USC in the endometrioid group may have led to the appearance of a worse survival in women with G3EC.
Alektiar et al.23 compared women with early-stage clear cell carcinoma or USC with women with G3EC and failed to detect a difference in either the PFS or OS. Again, conclusions regarding differences in survival between women with G3EC and those with USC are limited by the inclusion of clear cell histology. Furthermore, although the tumors of 40% of the women with clear cell carcinoma or USC did not contain an endometrioid component, the percentage of endometrioid component in the mixed tumors was not specified. Halperin et al.25 found no significant difference in the OS between women with 100% USC (n = 10) and women with mixed tumors containing ≥ 25% USC (n = 24). In addition, no significant difference in survival was noted between women with ≥ 25% USC (n = 34) and women with moderately or poorly differentiated endometrioid uterine carcinoma (n = 30). However, there was an imbalance with regard to stage of disease between the 2 groups, with 65% of the women with USC having either Stage III or Stage IV disease compared with only 23% of the women with moderately or poorly differentiated endometrioid histology. This imbalance between the two groups may have resulted in a favorable OS within the endometrioid group.
The current study, along with all other similar attempts to compare survival in women with G3EC versus those with USC, is limited by its retrospective nature. The retrospective study design may have led to an apparent difference in survival, or lack thereof, between groups. For example, although all the early-stage patients in the current study underwent complete surgical staging, the extent of lymph node sampling may have varied between surgeons. This could lead to the presence of a group of “understaged” women in whom survival appears falsely reduced. Routine peritoneal sampling and omentectomy, which have been advocated by some authors as an important addition to the surgical staging of women with USC,26 was not performed in our study group. Although this may have led to “understaging” a percentage of the women with early-stage disease, only one woman with advanced stage disease would have been mistakenly classified as having early-stage disease had either omental or peritoneal sampling been omitted from the surgical staging procedures. The heterogeneous nature of the prescription and administration of postoperative adjuvant therapy in the group of patients in the current study also may have served as a confounding variable with respect to survival analysis. In addition, although all available material for each case was examined thoroughly by the study pathologist (who was unaware of clinical history), retrospective inspection of tissue specimens is limited in its accuracy.
Investigation into effective treatments for women with USC is warranted and sorely needed. Based on the results of the current study, future cooperative group studies should include women whose tumors contain any fraction of USC. The PFS and OS of women with early-stage USC differ when compared with those of women with G3EC. In the current study, USC comprising as little as 10% of the entire tumor was found to influence disease behavior and the outcome for women based on its impact on disease-free survival. A statistically significant trend toward a worsening OS was established between women with G3EC and women with increasing percentages of USC (< 10%, 10–50%, and > 50%). Formally established criteria for assigning a percentage component of USC for each tumor and determining the number of slides to be evaluated must be developed and applied to future prospective trials. Only then will a more detailed understanding of the significance of the percentage of USC comprising a uterine carcinoma be elucidated. Hopefully, this will allow for the development of more effective treatment for women with this disease.