|Home | About | Journals | Submit | Contact Us | Français|
Catherine Uzan: None; Aminata Kane: None; Annie Rey: None; Sebastien Gouy: None; Sophie Camatte: None; Patricia Pautier: None; Catherine Lhommé: None; Christine Haie-Meder: None; Pierre Duvillard: None; Philippe Morice: None.
Section Editor Dennis Chi discloses no financial relationships.
Section Editor Peter Harper discloses a consulting relationship with advisory boards for sanofi-aventis, Roche, ImClone, Pfizer, GlaxoSmithKline, Lilly, and Genentech; honoraria received from commercial symposia for Lilly, Novartis, sanofi-aventis, and Roche; and research funding received from Pfizer, Roche, and Novartis.
Reviewer “A” discloses honoraria received from Genzyme and Intuitive Surgical.
The content of this article has been reviewed by independent peer reviewers to ensure that it is balanced, objective, and free from commercial bias. On the basis of disclosed information, all conflicts of interest have been resolved.
After completing this course, the reader will be able to:
This article is available for continuing medical education credit at CME.TheOncologist.com
To determine the prognosis of a micropapillary (MP) pattern in patients with stage II and stage III serous borderline tumor of the ovary (SBOT).
Review of patients with stage II and stage III SBOT treated or referred to our institution with characterization of an MP pattern and its clinical impact.
In 1969–2006, 168 patients were reviewed. Fifty-six patients had SBOT-MP. The rate of conservative surgery was lower in the SBOT-MP group than in the typical SBOT group, but the rate of patients with more than three peritoneal sites with implants was higher in the SBOT-MP group. The rate of invasive implants was not statistically different between the two groups. Eighteen recurrences were observed (six of them in the form of invasive disease) in the SBOT-MP group. Only one death was observed. The overall survival times and recurrence-free intervals were similar in both groups. The only prognostic factor for recurrence in the SBOT-MP group was the use of conservative surgery.
In the present series, an MP pattern doesn't appear to signify a poor prognosis. The only prognostic factor for recurrence in SBOT-MP was the use of conservative surgery. Further studies on the MP pattern are needed to evaluate prognosis and the results of conservative surgery.
Serous borderline ovarian tumors (SBOTs) exhibiting micropapillary (MP) patterns (SBOT-MP) are a recent form of SBOT that was described in 1996 by Burks et al.  and Seidman and Kurman . These tumors are clinically characterized as more frequently associated with extraovarian implants (particularly invasive implants) [2, 3].
In patients with early-stage disease, the prognosis with an SBOT with and without an MP pattern seems to be similar, but recent publications reported a poor prognostic impact of an MP pattern in patients exhibiting peritoneal spread [4–6]. Nevertheless, many of the characteristics specific to this entity continue to fuel debate. To attempt to add new data on this topic, we decided to study the prognostic impact of an MP pattern in a continuous series of patients treated at or referred to our institution for an SBOT of the ovary with peritoneal implants.
The histological slides of 168 patients treated, from January 1969 to February 2006, at/or referred to our institution for an SBOT with peritoneal implants were reviewed by the same pathologist in our center. The MP pattern was evaluated according to the criteria of Burks et al.  and Seidman and Kurman . The MP pattern was retrospectively evaluated on slides from patients treated before 1996 (date of description of this entity). SBOT with a MP pattern was confirmed if this appearance continuously occupied >5 mm. Tumors were not considered as having an MP pattern if this pattern occupied a single focus or several foci <5 mm in greatest dimension. Ovarian tumors can exhibit stromal microinvasion defined as stromal infiltration <10 mm2.
The staging classification used was the 1987 Fédération Internationale de Gynécologie et d'Obstetrique (FIGO) classification (International Federation of Gynecology and Obstetrics ). Peritoneal implants were classified as noninvasive (epithelial or desmoplastic type) or invasive according to the criteria previously described by Bell and Scully  and Bell et al. .
Surgery consisted of either radical (bilateral salpingo-oophorectomy with or without a hysterectomy) or conservative (defined as conservation of the uterus and salvaging of at least a portion of one ovary) treatment. Additional staging surgical procedures were performed. This was dependent on the surgical team, on whether the SBOT was diagnosed during or after the surgical procedure, and on disease extension. Patients had undergone either a laparotomic or a laparoscopic approach performed either during a one-step surgical procedure, if the diagnosis of SBOT was made at the perioperative frozen section analysis, or during restaging surgery.
Adjuvant therapy (chemotherapy, external radiation therapy, and chemotherapy combined with external radiation therapy) was administered until 1985. The indications for and type of adjuvant therapy depended on the pathological results (the characteristics of the peritoneal implants: noninvasive versus invasive), on the existence or not of a persistent residual tumor, and on the date when treatment was administered.
Second-look surgery was defined in this study as a new surgical procedure (whatever the indication) performed within 6 months following the initial treatment of the ovarian malignancy in a clinically disease-free patient. The indication for second-look surgery was discussed or recommended at the discretion of the team treating the patient (when patients were treated outside our institution).
Patients were considered lost to follow-up whenever the duration of follow-up was <12 months following surgery (if no recurrence had occurred within that period). The prognosis for patients exhibiting an MP pattern (the SBOT-MP group) was compared with that of patients without this component (the typical SBOT group). Prognostic factors were also estimated in the SBOT-MP subgroup. For the statistical analysis, overall survival was not an endpoint because very few patients died as a result of their disease. The study endpoint was the recurrence rate. Two endpoints were retained for the analysis: (a) the rate of recurrence, whatever the histologic subtype, and (b) the rate of recurrence in the form of invasive disease or low-grade carcinoma (on the peritoneum and/or ovary, in the event of conservative surgery).
Recurrence-free intervals were calculated including patients who had relapsed and/or had died as a result of their disease. Patients who had died as a result of intercurrent disease were censored at the time of their death. The log-rank test was used to compare the curves and to estimate the p-value. The χ2 test was used to compare variables. A p-value <.05 was considered significant.
Among the 168 patients with stage II and stage III SBOT, the presence or absence of an MP pattern could not be accurately determined in 14 patients treated outside our institution. Therefore, 154 patients were included in the current series. The median age of these patients at the time of surgery was 36 years (range, 14–74 years). Patient characteristics are detailed in Table 1. Among these patients, 56 had an SBOT-MP.
Among patients with an SBOT-MP, 48 had undergone radical surgery. Eight patients had undergone conservative surgery (four had a unilateral salpingo-oophorectomy, three had a unilateral salpingo-oophorectomy with a contralateral cystectomy, and one had a unilateral cystectomy). The histological results are detailed in Table 1. In one case, the type of implant was indeterminate because the samples were too small.
Fifteen patients had received adjuvant treatment. This treatment was chemotherapy for 14 patients and external radiation therapy for one patient. Second-look surgery had been performed in six patients: borderline disease was found on the preserved ovary after conservative management in one patient, noninvasive peritoneal implants were found in two patients, and second-look surgery was negative in three patients.
Two factors were statistically different between the two groups: the rate of conservative treatment (which was significantly lower in the SBOT-MP group) and the rate of patients with more than three different sites harboring peritoneal implants (which was higher in the SBOT-MP group) (Table 1).
Seven patients were considered lost to follow-up. Among the 147 patients followed up, the median durations of follow-up in the two groups were, respectively, 49 months (range, 26–437 months) and 56 months (range, 12–357 months). In the SBOT-MP group, 13 patients had relapsed, six of them in the form of invasive peritoneal and/or ovarian disease. The detailed characteristics concerning these 13 patients are given in Table 2. Twelve of these patients had recurrent disease diagnosed within 5 years after initial treatment. Sites of recurrence were (some patients had disease at multiple sites): the remaining ovary after conservative surgery in five patients, the pelvic cavity in nine patients, the abdominopelvic peritoneum in five patients, nodes in three patients, and other sites in three patients. The histologic subtype of the recurrence was available in 12 cases. One patient (with noninvasive implants during initial management) had relapsed in the form of chest metastases. In that case, the histologic subtype of the recurrence was unknown.
In six patients, the recurrence was invasive disease (Table 3). Detailed characteristics of those patients are as follows. For two patients, this recurrent invasive disease had developed on the remaining ovary and peritoneum after conservative management (nodal involvement was also observed in one of these two cases). One of these patients had died as a result of progressive disease. That patient had initially presented with noninvasive implants. This is the only patient who succumbed to disease in the SBOT-MP group. The other patient is currently under treatment for progressive disease. For the third patient in this group with invasive recurrence, invasive disease had developed in the lymph nodes. For the three remaining patients in this group, the recurrence was peritoneal in the form of invasive implants. All these patients had received chemotherapy. In summary, of these six patients with recurrent invasive disease, one had died as a result of disease, three are alive with persistent or progressive disease, and two are alive without disease.
In six patients, recurrent disease was noninvasive. Three patients had borderline recurrent disease on the spared ovary after conservative management associated with peritoneal noninvasive implants. All of them are currently disease free. The other three patients in this group had peritoneal recurrences in the form of noninvasive implants and are also currently disease free. All these six patients were treated exclusively with surgery.
Among the eight patients who were treated conservatively, two had become pregnant at least once. Five patients had relapsed after conservative treatment: three had developed a borderline recurrence with noninvasive implants. Two had relapsed with invasive ovarian disease (one had died as a result of this disease and the other is under treatment for progressive disease). Two patients had been treated conservatively for recurrent disease.
Among the nine patients with invasive implants, only one had relapsed in the form of borderline disease on the ovary, after conservative management, and she is currently disease free. The other patients were disease free after a median follow-up of 40 months.
The 5-year recurrence-free interval rates in the SBOT-MP and typical SBOT groups were, respectively, 65% (95% confidence interval [CI], 49%–89%) and 77% (95% CI, 68%–86%) (p = .8). The 10-year recurrence-free interval rates in the SBOT-MP and typical SBOT groups were, respectively, 61% (95% CI, 43%–79%) and 63% (95% CI, 49%–77%) (not significant). The 5-year rates of recurrent invasive disease in the SBOT-MP and typical SBOT groups were, respectively, 18% (95% CI, 4%–32%) and 4% (95% CI, 0%–7%) (p = .03). Overall survival curves for both groups are shown in Figure 1. Recurrence-free interval curves are shown in Figure 2. A single prognostic factor for recurrence in the SBOT-MP group was identified in the univariate analysis: the type of surgery (conservative) (Table 2). The median times to recurrence in patients with an SBOT-MP or a typical SBOT were, respectively, 31 months (range, 4–56 months) and 44 months (range, 4–264 months) (not significant).
SBOT-MP was described nearly 10 years ago. Few specific papers have been devoted to the study of the prognosis of patients with such tumors and many questions remain unanswered about this disease. Different reports suggest that the prognosis for patients with early-stage disease (without peritoneal implants) seems to be similar in SBOTs with and without an MP pattern [3, 10, 11]. Nevertheless, this entity is considered as having a higher risk for extraovarian disease. Interest in complete peritoneal staging surgery is currently debated in BOTs but, given the higher rate of extraovarian disease from lesions displaying an MP pattern, the use of complete staging surgery remains the rule in MP disease . Very few series have specifically been devoted to evaluating the survival of patients with SBOT-MP and peritoneal implants. Those that have are summarized in Table 4. The present study includes the largest number of patients with stage II and stage III SBOT-MP (Table 4). Several papers describing the clinical behavior of SBOT-MP claimed that these tumors have a higher rate of invasive implants than with typical SBOT . These rates were similar in our study (16% versus 10%; p = .3). Other series have reported similar findings. In a study by Slomovitz et al. , including six patients with stage II/III SBOT-MP, none of them had invasive implants. In a series by Deavers et al. , the difference in the percentages of patients with invasive implants did not reach statistical significance (three of 18 [18%] SBOT-MP patients versus five of 81 [6%] typical SBOT patients). One explanation for such results in the study by Deavers et al. , as well as in ours, is that the rates were different but failed to reach statistical significance because the number of patients was too small. Nevertheless, in other series in which the type of implant (invasive versus noninvasive) had a significant prognostic value, the number of patients studied was smaller than in ours. Another explanation is perhaps that, finally, no statistical difference can be demonstrated between SBOT-MP and typical SBOT in terms of the type of implant. In contrast, the rate of invasive implants was lower in the SBOT-MP group (8% [one of 13 patients] versus 17% [four of 24 patients]) in the series reported by Prat and de Nictolis .
In our study, two characteristics were statistically different between the two groups. The first concerned disease extension, which seemed to be greater in the SBOT-MP group—the percentage of patients with more than three different involved peritoneal sites was significantly higher and the rate of stage III disease (versus stage II) was also higher and almost of borderline statistical significance in the SBOT-MP group. Nevertheless, although this difference could have a potential impact on prognosis, the overall survival and recurrence-free survival times were similar in the two groups of patients.
This is a major point because it suggests that the overall prognosis for patients with SBOT, with or without a MP pattern, may be similar. It is essential to review the literature regarding this question because the crucial question behind the description of a new pathological entity is whether this new lesion will exert a different impact on survival (and, if so, whether it should be treated in a different way). Having examined different series that reported on SBOT-MP with peritoneal implants, we found two different “groups” of papers in terms of prognosis (Table 4). In three series that (taken together) reported on 71 patients, 21 deaths or persistent/progressive disease were observed [4–6]. In the series reported by Smith Sehdev et al. , among 12 patients with persistent/progressive disease, 11 initially had invasive implants. Similarly, in the series by Longacre et al. , three of four patients who died as a result of the disease also had invasive implants. Such poor results would logically lead to a modification of the current classification of serous tumors of the ovary, as suggested by Kurman et al. . In his new classification, SBOT-MP is considered as a “non-invasive micropapillary (or intraepithelial) low-grade serous carcinoma,” an intermediate form between an atypical serous tumor (typical SBOT for other authors) and low-grade carcinoma, so-called “invasive micropapillary serous carcinoma” (which exhibits stromal infiltration). Recent molecular genetic studies (particularly BRAF and KRAS mutations) provide cogent evidence to support this new classification. Such a classification is logical if the prognosis for patients with SBOT-MP is considered to be clearly different from that of patients with non-MP forms.
On the other hand, five other series published—reporting, respectively, six, 12 (in two series), 15 and 53 cases with follow-up—reported an overall “good” prognosis in SBOT-MP patients. Of a total of 98 cases reported in those different publications, only nine deaths or persistent/progressive disease were reported [3, 10, 13, 15]. It is interesting to note that, in three of these series, patients with persistent/progressive disease or who had died initially had invasive implants. Such was the case for both deaths in the series reported by Eichhorn et al.  and the only death in the series reported by Prat et al.  and Gilks et al. [3, 13, 15]. In our series, among the nine patients with invasive implants during initial management in the SBOT-MP group, only one relapsed in the form of noninvasive disease and is currently disease free. In the present series, the 5-year rates of recurrent invasive disease in the SBOT-MP and typical SBOT groups were, respectively, 18% and 4%. This difference was not statistically significant, perhaps because the number of patients with invasive disease was too small throughout the series. Nevertheless, this (nonsignificant) difference had no effect on survival.
How can one explain such differences among the results in the literature? The first explanation concerns the histological interpretation of an MP pattern. It is not always easy to characterize SBOT in terms of the pathological analysis. Nonetheless, the results in Table 4 were reported by experienced teams in expert centers who were well-versed in the histological analysis of ovarian tumors in their country. This is therefore unlikely to be the explanation. As pointed out by Prat , a plausible explanation for such differences is the way in which the analyzed cases were recruited in these different series. Some of the published series were obtained from consultation or referral material (inevitably increasing the rate of patients with “ambiguous lesions” akin to carcinoma, i.e., patients with invasive implants). Other series were based on the analysis of records of cases treated in a single institution .
Another potential explanation is the follow-up duration. Two recent series highlighted the time dependence of the recurrence rate in SBOT [4, 17]. Recurrences are not uncommon after 5 years for patients with this tumor [4, 17]. Nevertheless, the duration of follow-up reported in the current series is similar to, or longer than, that of other series that reported a clear difference in terms of prognosis between SBOT-MP and typical SBOT patients (Table 4).
The only factor identified as a risk factor for recurrence in the SBOT-MP group in this series was the use of conservative surgery (Table 2). However, the rate of conservative treatment was lower in the SBOT-MP group than in their typical SBOT counterparts. Less recourse to conservative surgery in the SBOT-MP group could have been a result of the fact that, in most published series on SBOT-MP (including early- and advanced-stage disease), the tumors were more frequently bilateral than in typical SBOT [3, 4]. If tumors with an MP pattern are more frequently bilateral, the use of conservative surgery decreases. In our review of the literature concerning the conservative management of SBOT-MP patients, only one series was published by our group on this topic (15 cases including early- and advanced-stage disease). New cases of advanced-stage disease have since been included in the current series. In other series, conservative treatments were described in case reports and were more frequently used in early-stage disease . These different data (concerning a few cases) suggest that conservative surgery could be safely proposed, at least for early-stage SBOT-MP patients, but a high rate of recurrence is expected . Among the eight patients treated conservatively in the current series, five relapsed, three in the form of noninvasive disease and two in the form of invasive carcinoma at least on the remaining ovary. One of these latter patients died as a result of the disease. Here, we have probably reached one of the limits of conservative surgery in borderline ovarian disease. In patients with SBOT-MP and peritoneal implants, the results of conservative surgery (five of eight patients relapsed, two of them in the form of invasive carcinoma, and one of them succumbed to the disease) were not good, and suggest that, in cases of bilateral ovarian involvement during initial management, bilateral salpingo-oophorectomy should be preferred to conservative surgery in such patients.
In the present series, the rates of invasive implants were similar in patients with an SBOT with peritoneal implants, with and without an MP pattern. The presence of an MP pattern does not appear to portend a worse prognosis than in patients whose tumor is devoid of this feature. The only prognostic factor for recurrence in such patients was the use of conservative surgery. Further studies are needed to evaluate the results of conservative surgery in this context and to evaluate the prognosis of patients with an MP pattern.
We thank Lorna Saint Ange for editing.
Conception/Design: Catherine Uzan, Aminata Kane, Sebastien Gouy, Philippe Morice
Provision of study material or patients: Sophie Camatte
Collection and/or assembly of data: Aminata Kane, Sophie Camatte, Pierre Duvillard
Data analysis and interpretation: Catherine Uzan, Aminata Kane, Annie Rey, Pierre Duvillard, Philippe Morice
Manuscript writing: Catherine Uzan, Philippe Morice
Final approval of manuscript: Sebastien Gouy, Patricia Pautier, Catherine Lhommé, Christine Haie-Meder, Pierre Duvillard, Philippe Morice