Pattern of endocrine treatment for epithelial ovarian cancer in the Southeast medical region of Sweden: a population-based study

Abstract

Aim of the study: Endocrine treatment (ET) is an alternative as salvage therapy in epithelial ovarian cancer (EOC) but the usage in routine care is unknown. We evaluated the treatment patterns and outcome of patients receiving ET for EOC in the Southeast medical region in Sweden.

Method: Patients were identified through the population-based Southeast Quality Registry for gynaecological cancer. Inclusion criteria were: age ≥18 years, histologically verified EOC diagnosed 2000–2013, ET for ≥4 weeks. Coverage compared with the Swedish National Cancer Registry was 100%. Data extracted from medical records was collected by means of a study-specific Case Report Form. Last date of follow-up was February 1st, 2018. All statistics were descriptive.

Results: Altogether 248 (18%) of 1414 patients were treated with ET. Most (49%) had received only one, and 34% two previous lines of chemotherapy. Time from last chemotherapy to ET was 4 months, range 0–55months. The reason for initiating ET was tumor progression (66%), chemotherapy related toxicity (29%) and maintenance (4%). Tamoxifen was prescribed in 94% of cases. Best response was partial (< 5%) and stable disease (50%). No patient had a complete response. 194 (78%) patients received subsequent chemotherapy, of these 27% had 3–7 lines of chemotherapy. Duration of ET was a median 4 months (range 1–80 months). Median time from ET to subsequent chemotherapy was 5 months (range 0–79). The median overall survival was 45 months (range 9–173).

Conclusion: In the Southeast region of Sweden, endocrine treatment for EOC was prescribed inconsistently and in various settings, usually initiated by a rising CA-125 level. Poorer documentation and irregular tumor response assessment were observed for endocrine treatment compared to chemotherapy.

Introduction

Epithelial ovarian cancer (EOC) is the seventh most common cancer and the eighth cause of death from cancer in women, with 239,000 new cases and 152,000 deaths annually worldwide [1]. In Sweden, approximately 700 new cases are diagnosed each year [2]. The majority of EOC are detected at advanced stage (FIGO III-IV) which, together with acquired treatment resistance, accounts for the poor prognosis. The relative 5-year survival rate is 46% [3]. Patients with EOC may be offered endocrine treatment (ET) as salvage therapy but the frequency and settings for prescribing ET in routine care is unknown.

Despite numerous published clinical studies on the effectiveness and safety of ET in EOC during the past decades the role of ET in EOC is not well defined [4–10]. We do not know which hormonal agent is most effective nor the optimal setting. The role of sex steroid receptor expression as biomarkers for predicting treatment response is also unclear [7,11–17]. In addition, there are few data on the effect of ET on symptom control and quality of life. The reason for this is multi-factorial but most studies in EOC are retrospective or phase II studies with limited number of patients, poorly defined tumor characteristics (e.g. histological subtypes) and includes mainly heavily pretreated patients. Despite the low evidence for efficacy, ET is an attractive treatment option, and included in many international guidelines for heavily pretreated patients, since it is associated with an often-favourable safety profile, easy intake and low cost [5,7,9–11,15]. Patient preference is important in the decision-making process of choosing best treatment, especially at late relapse, but must be based on thorough physician-patient communication including information regarding prognosis and treatment alternatives. It is important to clarify the prescription pattern of ET in routine care, at what point during the course of disease ET is prescribed and the decision-making process. If a treatment is considered as relatively nontoxic, cheap and easily administrated there may be a risk that the prescription pattern gradually extends without supporting evidence, risking more effective treatment to be withheld.

Personal identity numbers and official population-based registers in Sweden, covering all Swedish citizens, offer exceptionally good conditions for outcome studies. We performed a population-based study to elucidate the frequency, setting and outcome of endocrine treatment in patients diagnosed with epithelial ovarian, fallopian tube, and peritoneal cancer in the Southeast medical region of Sweden between 2000 and 2013.

Materials and method

Study population

We used the population-based Southeast Medical Region Quality Registry for Gynecological Cancer (SMRQRGC) to identify eligible patients. The Southeast medical region is one of six in Sweden, encompassing a population of 1.1 million inhabitants. Reporting to the Swedish National Cancer Registry, which started in 1958, is mandatory and the registry has over 95% coverage of all malignant tumors and 99% of the tumors registered are histologically verified [18]. The coverage between SMRQRGC and the national cancer registry is approximately 100% [19]. Inclusion criteria were: age at least 18 years, histologically verified epithelial ovarian, fallopian tube or primary peritoneal carcinoma diagnosed between January 1st, 2000 and December 31st, 2013 and ET for at least four weeks. Since we were interested in outcome we excluded patients (n = 2) who received ET for less than 4 weeks, see Flow chart (Figure 1). The ethical review board at Linköping University approved the study and deemed a separate study-specific consent as not needed.

Figure 1. Flow chart study population.

Data concerning baseline patient and tumor characteristics, treatment, and tumor responses were extracted from medical records by means of a study-specific Case Report Form. Staging was done according to the Federation Internationale de Gynecologie et d´Obstetrique, FIGO, classification from 1988 [20]. Grade of differentiation was performed according to WHO standards [21] and the Silverberg grading system [22]. The Malpica serous 2-tier criteria was not widely practiced in Sweden before 2014. We used available information on consecutive CA-125 levels, imaging findings and clinical investigations to evaluate tumor response. Response rate was defined as the proportion of patients who had achieved complete response or partial response.

Patients were followed up until February 1st, 2018 or to death, whichever came first.

In the Southeast region, treatment planning for gynaecological cancer is centralized to Linköping University Hospital but treatment is executed also at county hospitals. End-of-life treatment with best supportive care is provided at patients home with assistance from palliative care units or at palliative in-patient clinics. For some patients, data concerning their last months in life was not available in the medical records from the county hospitals. Patients who deceased within three months after the last note in records indicating on-going endocrine therapy, with missing data, were assumed not to have received any other oncological treatment before death.

Statistics

All statistics are descriptive. Date of diagnosis was defined as the date of tissue sampling providing the diagnosis. In most of the cases, this was synonymous with the date of the primary surgery. Overall survival was calculated from date of diagnosis until death or date of last follow-up.

Results

Patient characteristics and primary treatment

A total of 248 out of 1414 (18%) patients diagnosed with epithelial ovarian, fallopian tube or primary peritoneal cancer between the years 2000 and 2013, received endocrine treatment (ET) for more than four weeks at some time during the course of their disease (see Flow chart, Figure 1).

Patient and tumor characteristics are presented in Table 1. The median age was 63 years, with a range of 21–88 years. Seventy-seven percent of patients had FIGO stages IIIC-IV. The serous subtype was most common (76%), followed by the endometrioid subtype (10%). Few patients (4%) had clear cell or mucinous adenocarcinomas. The grade of differentiation was predominantly poor (63%), but not specified in almost one fifth of pathology reports. Details of primary treatment is provided in Supplementary Table S1.

Table 1. Patient and tumor characteristics of the study population.

Variable	Total N = 248 (%)
Age at diagnosis (years)
Median	63
Range	21–88
Diagnosis
Epithelial ovarian cancer	210 (85)
Fallopian tube cancer	18 (7)
Primary peritoneal cancer	14 (6)
Cancer of Undesignated primary tumor	6 (2)
FIGO stage
IA-IC	11 (4)
IIA-IIC	24 (10)
IIIA-IIIB	22 (9)
IIIC	140 (57)
IV	51 (21)
Histology
Serous	188 (79)
Endometrioid	24 (10)
Clear cell	5 (2)
Mucinous	4 (2)
Mixed tumor type	11 (4)
Other	16 (6)
Grade of differentiation^a
Well	16 (7)
Moderate	31 (13)
Poor	156 (63)
Not specified	45 (18)
Follow-up^b
Alive, no evidence of disease	6 (2)
Alive, with disease	13 (5)
Dead	229 (92)
Overall survival, months
Median	45
Range	9–173

Abbreviations: FIGO = Federation Internationale de Gynecologie et d´Obstetrique.

^aGrade of differentiation according to WHO international histological classification of tumors.

^bLast date of follow-up 1st February, 2018.

At the last date of follow-up, February 1st, 2018, 93% of the patients had died. The median overall survival was 45 months, with a range of 9–173 months. The prescription pattern for the endocrine treatment did not differ between the year cohorts of 2000–2007 and 2008–2013.

Chemotherapy preceding endocrine treatment

Most patients were not heavily pretreated with chemotherapy when ET was prescribed. Nearly 50% of patients had received only one prior line of chemotherapy regimen. Less than one fifth had received three or more previous chemotherapy lines (Table 2). Overall response rate to last line of chemotherapy prior ET was 57%. The time from last course of chemotherapy to start of ET was a median of 4 months, range 0–55 months. Administration dates and dose of chemotherapy, as well as tumor evaluation after primary and recurrent chemotherapy, was consistently performed and documented in the medical records, with no missing data (Supplementary Table S1 and Table 2).

Table 2. Chemotherapy preceding endocrine treatment.

Variable	Total N = 248 (%)
Number of lines of chemotherapy preceding endocrine therapy
1	122 (49)
2	84 (34)
3	33 (13)
≥ 4	9 (4)
Range	1–7
Response to last line of chemotherapy preceding endocrine treatment
Complete response	56 (23)
Partial response	86 (35)
Stable disease	71 (29)
Progressive disease	34 (14)
Not evaluable	1 (0.4)
Time from last chemotherapy to start of endocrine treatment (months)
Median	4
Range	0–55

Endocrine treatment

Immunohistochemical analysis of estrogen receptor (ER) and/or progesterone receptor (PR) status had only been performed in 66 patients (27%) prior to ET. Staining of ER was positive in 54 out of 66 (82%) tumors and for PR staining was positive in 27 out of 51 (53%).

The most common reason for prescribing ET was tumor progression (Table 3). Monitoring of CA-125 levels had regularly been performed during follow-up in 202 (81%) patients, of which 90% had a rising value at the time when ET was prescribed. In 72% of 141 patients undergoing imaging, a radiological tumor progression was present. Almost one third of patients received ET due to chemotherapy-related toxicity. Eleven patients received ET as maintenance. Contributing factors for prescribing ET were patients request for additional treatment, and that some patients wanted to abstain further chemotherapy.

Table 3. Endocrine and subsequent treatments.

Variable	Total N = 248 (%)
Reason for prescribing endocrine treatment
Tumor progression	164 (66)
Chemotherapy related toxicity	73 (29)
Maintenance treatment	11 (4)
Drug used for endocrine treatment
Tamoxifen	234 (94)
Exemestan	9 (4)
Letrozole	3 (1)
Anastrozole	1 (0.4)
Progestins	1 (0.4)
Best response to endocrine treatment, n^a	198
Complete response	0
Partial response	9 (5)
Stable disease	98 (50)
Progressive disease	91 (46)
Duration of endocrine treatment, n	96
Median	4
Range	1–80
Chemotherapy following first endocrine treatment
Yes	194 (78)
No	54 (22)
Time from start of ET to start of first subsequent chemotherapy, n	194
0–3 months	53 (27)
4–6 months	59 (30)
7–12 months	43 (22)
>12 months	39 (20)
Median time (range) in months	5 (0–79)
Number of lines of chemotherapy following endocrine treatment, n	248
0	54 (22)
1	73 (29)
2	54 (22)
3	31 (13)
≥ 4	36 (15)
Median	2
Range	1–7
Re-challenge with endocrine therapy
Yes	45 (18)
No	203 (82)
Drug used for re-challenge with endocrine treatment, n	45
Tamoxifen	13 (30)
Exemestan	14 (32)
Letrozole	10 (22)
Gestagen	1 (2)
Two drugs used subsequentially^b	6 (14)

^aTumor evaluation was not performed in 39 of 237 (17%) patients with evaluable disease and in the 11 patients receiving endocrine maintenance treatment.

^bTamoxifen and letrozole; tamoxifen and exemestan; tamoxifen and gestagen.

Tamoxifen was prescribed to 94% of the patients, and for 81% in a dose of 40 mg per day. For 13 patients receiving tamoxifen, a switch to another endocrine drug (e.g. exemestane, letrozole) was made due to rising CA-125. ET was given as single-agent treatment in 90% of the patients. One patient received ET in combination with intraperitoneal chemotherapy, and one in combination with radiotherapy.

Of 237 patients with evaluable disease, tumor response assessment during ET was performed in 198 (84%) of the cases. Investigations used were CA-125 levels in 74% and imaging in 47%. Best response was partial (PR) in 5% and stable disease (SD) in 50% (Table 3). No patient had a complete response (CR). Median time to tumor progression was 4 months (range 0–79 months). Clinical assessment was performed in 207 patients and in 64% of the cases, symptom benefit was documented in the medical records.

For 152 patients (61%) the date and reason of ceasing ET was lacking in the medical records. For the 96 patients with adequate data, the median duration of ET was 4 months, with a range of 1–80 months. Adverse events (e.g. hot flushes or gastrointestinal effects) was reported as the cause of treatment discontinuation in 9% of patients. No patient had severe adverse events such as deep venous thrombosis or pulmonary embolism.

Further oncological treatment

After discontinuation of ET, 194 of 248 patients (78%) were treated with at least one line of chemotherapy while almost one third of the patients received at least three lines of chemotherapy regimens (Table 3). The time from start of ET to start of subsequent chemotherapy was at a median of 5 months, with a range of 0–79 months.

Forty-five patients (18%), were prescribed an additional line of endocrine regimen later in their course of disease. Twelve patients were prescribed tamoxifen, 13 exemestane, 11 letrozole, and one progestin. Six women were treated with two additional lines of ET with two different agents.

Discussion

In a population-based cohort study of patients with epithelial ovarian cancer in the Southeast medical region in Sweden, we found that almost one fifth of patients received endocrine treatment for at least 4 weeks at some point during the course of their disease. Endocrine treatment was mainly initiated due to a rising level of CA-125 and prescribed to a heterogeneous group of patients. The majority was not heavily pretreated and most received several lines of chemotherapy regimens after endocrine treatment had ended. We also found differences in pattern of tumor evaluation and documentation in medical records when comparing chemotherapy to endocrine treatment in our patient cohort. Tumor evaluation was irregularly performed during endocrine treatment and the documentation of e.g., date and reason for treatment discontinuation, was poor. A patient benefit of endocrine treatment i.e., symptom relief was documented in medical records for the majority of patients.

To our knowledge this is the first population-based study on the usage of endocrine treatment in routine care of invasive EOC. We can therefore not compare our findings of one fifth of the patients receiving ET, with those of others. However, since most advanced-staged ovarian cancer patients will recur and ultimately succumb of their disease one might expect that the frequency of ET prescription would be higher, despite the low evidence of efficacy, due to the low risk for toxicity. From a patient perspective, the acceptance of treatment side effects seems to change over the disease continuum. In a recent online survey, ovarian cancer patients were asked about goals of care and priorities [23]. Responders (328/434; 76%) reported a high tolerance for side effects if the goal of care was cure (e.g. 86% for fatigue, 86% tolerance for alopecia). The tolerance for most symptoms however declined significantly when the goal was remission and even less for stable disease. A recent systematic review, including ten studies, of patient preferences and expectations for the treatment of recurrent ovarian cancer showed that patients value both overall and progression-free survival and are willing to accept side effects if survival may be increased [24]. These, and other studies, underscores the importance of physician-patient communication and the need to provide accurate prognostic information together with realistic goal of treatment, treatment alternatives, side effects and possible impact on quality of life.

Our results, based on the whole cohort of patients diagnosed with EOC in the Southeast region of Sweden between 2000 and 2013, demonstrates the lack of consistency concerning prescription patterns of ET in EOC. To our surprise, ET was not only prescribed to heavily pretreated ovarian cancer patients. In some cases, ET was given as maintenance treatment after successful re-challenge of platinum-based chemotherapy for recurrent EOC. Patient’s desire for further oncological treatment contributed to prescribing ET.

Most patients were prescribed ET for tumor progression, where a rising CA-125 level was the main reason for initiating treatment. The value of routine surveillance of CA-125 in the follow-up of patients with ovarian cancer has been questioned since no survival benefit has been demonstrated with early treatment of recurrences based on a raised CA-125 level alone [25]. Still it may be difficult to abstain from CA-125 measurements, and not to react when the tumor marker is rising, which is evident by the findings of our study indicating an early initiation of ET. After the results from the CA-125 surveillance study [25] had been presented a survey was sent out to ovarian cancer patient asking how these results might influence their opinion on the utility of CA-125 monitoring during follow-up [26]. Despite the limitation of the study, including low response rate and small sample size (n = 416), it is interesting to note that one fifth of patients would want to start treatment as soon as possible if there was evidence that their cancer had recurred. Perhaps more important 25% of the survey population reported they wanted to know their CA-125 results because it provided them to have more control suggesting that not all patients request action to a rising CA-125. It may be difficult for physicians to resist patient desire for treatment due to a rising CA-125 without any other proof of tumor progression. In these cases, a choice of a relative nontoxic oncological treatment for heavily pretreated patients is understandable. However, most patients in our study were not heavily pretreated as demonstrated by the frequent use of chemotherapy following ET, with nearly one third of patients receiving at least 3 lines of chemotherapy regimens. In addition, not all patients with rising CA-125 levels were undergoing imaging before treatment onset. The choice of ET for these patients is less clear nor understandable.

We found that chemotherapy-induced toxicity was a contributing reason for choosing ET in almost one third of the patients. This is lower than reported in a retrospective cohort study of 99 ovarian cancer patients where chemotherapy intolerance was the most common (64%) reason for shifting to ET [11]. Unlike the case in our study, all patients were heavily pretreated with a mean of four prior lines of chemotherapy.

Yet there is no validated predictive marker of endocrine sensitivity in ovarian cancer. ERα expression, the perhaps most promising biomarker, has been associated with increased probability of response to anti-estrogen treatment in some studies but not in others [5–7,10]. The lack of validated biomarkers probably explains the low rate (26%) of hormone receptor analysis in our study. A more detailed review of predictive biomarkers of endocrine sensitivity in EOC has recently been published by Langdon et al.[10].

Most patients (94%) in our study received tamoxifen. The response rate was <5%, with no patient in complete response. Approximately 50% of the patients demonstrated disease stabilization. Median time to tumor progression was 4 months. The group of patients in our study was however heterogeneous and in various setting of disease therefore our results on outcome needs to be interpreted with caution. Interestingly, a symptom benefit of ET was observed for some patients in the study. The magnitude of benefit is however difficult to interpret due to insufficient documentation in the medical records and the lack of patient reported outcome. In the Ovaresist trial, 238 patients were randomized to chemotherapy versus tamoxifen for platinum-resistant recurrent EOC [27]. Median progression-free survival on tamoxifen was 8 weeks (95% CI, 8.0–10.4) compared to 13 weeks (95% CI, 9.0–16.3) on chemotherapy (HR, 1.54; 95% CI, 1.16–2.05; log-rank p = .003). No difference in overall survival was found between the treatment arms. Patients on chemotherapy experienced more toxicity and poorer health-related quality of life, compared with tamoxifen. Control over gastrointestinal symptoms was not better for chemotherapy.

In a recent meta-analysis by Paleari et al. of 53 trials including 2490 patients the overall clinical benefit rate (CR, PR or SD) was 41% (95% CI, 0.34–0.48) for any endocrine treatment [5]. There was also a trend for a higher benefit in platinum-sensitive versus platinum-resistant tumors. The odds ratio (OR) for death, showed a reduced mortality with ET (Summary OR 0.69; 95% CI, 0.50–0.97), with a possible greater effect in first-line and low grade tumors. We agree with the authors that the activity of ET support large properly designed randomized trials in the first treatment of hormone sensitive EOC. Although the results of this systematic review are encouraging, the overall quality of the randomized clinical trials included was considered low and hence should not be implemented into clinical care until the evidence of efficacy is stronger.

Some strengths of our study are that it is population-based with high coverage, minimizing selection bias, and no patient was lost to follow-up. Since this is to our knowledge the first population-based study on the pattern of endocrine treatment for EOC we cannot compare our results with those from other regions in Sweden or other countries. We observed that the documentation in medical records was poorer for ET compared to treatment with chemotherapy. We can only speculate on possible reasons for this but the general view of ET as an easily administered and cheap treatment without major side-effects may contribute to a less stringent as well as poorer quality of routine care. In some cases, ET was given late in the course of disease, at end of life, where palliative care units had shortly thereafter taken over patient responsibility and treatment with best supportive care. The medical records from these units were not available for the study. Patient compliance is another aspect that needs to be considered. We cannot rule out that some women have discontinued the treatment due to e.g. adverse events without informing their physician.

In conclusion, our study demonstrates an inconsistent prescription of ET in routine care of EOC. The setting varied from maintenance in early relapse to heavily pretreated recurrences, most treatments were initiated by a rising CA-125 level. A poorer documentation in medical records and an irregular tumor response assessment was observed for endocrine treatment compared to chemotherapy. Our data may contribute to improve awareness and quality of clinical care of women with epithelial ovarian cancer considered for endocrine treatment. From the results of our study the need for guidelines based on strong level of evidence is apparent.

Authors contribution

All authors approved the final manuscript and agree to be accountable for all aspects of the work.

Acknowledgments

We would like to thank Lena Wigren for her exceptional administrative assistance.

Disclosure statement

The authors declare that there are no conflicts of interest. The granting institutions had no roles in the design of the trial, data collection, analysis, interpretation or writing of the manuscript.

Additional information

Funding

This study was supported by funding from the Medical Research Council of Southeast Sweden and by ALF Grants, Region Östergötland.