Treatment of hormone positive uterine leiomyosarcoma with aromatase inhibitors

We identified 29 patients with ULMS treated with an AI from January 2001 to July 2012. We excluded 8 cases that were treated in another institution and 5 patients that did not have evidence of measurable disease at the time of AI initiation. The demographics and tumour characteristics of the remaining 16 patients with advanced measurable disease are listed in Table 1 and their tumour characteristics are listed in Table 2. The median age at time of AI initiation as 1st line hormonal treatment was 55 years (range 39–72). Fourteen patients (87.5%) were postmenopausal at the time of treatment.

Tumours were classified according to FNCLCC criteria, as low-grade ULMS in 7 patients (44%), intermediate grade in 3 (19%) patients and high-grade ULMS in 6 patients (44%). We also classified our patients according to Stanford USMT criteria; 8 patients were found to have high-grade ULMS (50%) and 8 patients low-grade ULMS (50%; Table 2). Due to the small number of our patients, we eventually analysed them according to Stanford USMT criteria.

ER status was determined semi-quantitatively in all (n = 16) patients, and assigned as ‘weak’, ‘moderate’ or ‘strong’ in tumour nuclei. Three of these 16 patients (19%) had weak (1+) ER positive staining, while 13 patients (81%) had moderate to strong (2-3+) ER positive staining. With regards to PgR status: 10 (62.5%) had a moderate to strong PgR staining, and 6 (37.5%) had a weak PgR staining. Of the 10 patients with moderate to strong PgR positive disease, all were also moderately to strongly ER positive (Table 2). Moreover, in Table 2 is depicted the classification of tumours by the extent of ER/PgR as reported by the percentage of tumour expressing each receptor [17].

Prior surgical and systemic treatment details are provided in Table 3. All of the 16 evaluable patients had measurable disease at time of AI treatment. Six (37.5%) patients had low volume (defined as the absence of any tumour deposit >2 cm in longest diameter on radiographic imaging) measurable disease [16]. However, 10 (62.5%) had multiple metastases, with only one of these having low volume disease. Sites of disease included lung in 12 patients (75%), pelvis in 4 (25%) and peritoneum in 3 (19%).

Four patients (25%) received hormonal replacement therapy (HRT) with a median duration of treatment of 48 months (24–96; Table 3). Three patients progressed whilst on HRT; withdrawal of HRT led to regression of disease (equivalent to partial response with hormonal manipulation) in two patients that lasted for 10 months and 84 months respectively, while the other patient started treatment with letrozole. Three (19%) of the 16 patients received tamoxifen (Table 3); two of them were diagnosed with early breast cancer prior to their diagnosis of ULMS (and one of them also received HRT), while the third received tamoxifen as a therapeutic manoeuvre for ULMS and progressed one month later.

Letrozole was prescribed as 1st line therapy (2.5 mg daily) in all 16 patients (Table 1). Goserelin, a gonadotropin-releasing hormone agonist, was administered to suppress ovarian function in the two premenopausal women of our cohort, in order to permit treatment with an AI. One patient switched from letrozole to anastrozole due to side effects and eventually stopped treatment due to poor tolerance. Twelve (75%) patients received AI as 1st line treatment without receiving any anticancer agent prior to AI. Two patients had received gemcitabine plus docetaxel prior to 1st line AI, with one of them starting letrozole as maintenance treatment. The other two patients received two lines of chemotherapy prior to starting 1st line letrozole.

From these 16 patients, 6 of them received 2nd line endocrine treatment with exemestane (25 mg; 5 patients) or anastrozole (1 mg; 1 patient), when their disease progressed. The demographics and tumour characteristics of those 6 patients are listed in Table 4. The median age at time of 2nd line AI initiation was 56 years (range 40–74). One patient was premenopausal; hence goserelin was prescribed to suppress ovarian function. Five out of 6 tumours were classified as low-grade ULMS. With regards to ER and PgR status, five patients had moderate to strong (grade 2–3) ER with the four of them having also grade 2–3 PgR staining. Moreover, five patients had very high expression of ER (>90%; Table 4). All patients had measurable disease at time of 2nd line AI treatment; 4 patients had progressed in the lung and the other 2 patients to the pelvis and peritoneum (Table 4).

The median duration of 1st line treatment with letrozole was 13 months (range 2–29). No patient achieved a CR, while PR was observed in 2 (12.5%). The first patient had low-grade ULMS strongly ER and PgR positive with low volume peritoneal disease. The other had high grade, moderately ER and PgR positive, oligometastatic disease and received letrozole as maintenance treatment after completing 6 cycles of chemotherapy, which may make difficult to assess accurately the therapeutic impact of letrozole alone. The duration of response in these patients was 10 and 11 months respectively.

Ten patients (62.5%) had SD as best response, with 8 of them maintaining it for more than 6 months. The duration of stable disease (median PFS) was 20.5 months (95% CI: 8.7 – 32.2 months), with 8 patients stopping letrozole because of disease progression, one because of poor tolerance, while one is still on letrozole (27+ months). Six of these patients had low-grade ULMS according to both grading systems and the remaining four had high-grade disease. ER status for all except one patient with SD was moderate to strong, while 7 patients were strongly PgR positive, 2 PgR negative and 1 patient had unknown PgR status. The CBR (CR + PR + SD >6 months) was 62.5% (10/16 patients).

Best response was progression of disease (PD) in 4 patients (25%). Three patients had high-grade disease; 2 patients had weak staining for ER and PgR, while the third strongly expressed both receptors. The fourth patient had low grade ULMS and ER and PgR strongly positive and she received letrozole after progressing on two previous lines of chemotherapy with symptomatic high volume disease.

The median duration of 2nd line AI therapy was 3 months (range 3–22 months). One patient achieved PR (after progression of peritoneal disease) but stopped treatment after 3 months due to toxicity. She had low grade, strongly ER and PgR positive disease.

Three patients had SD as best response and are continuing to date (duration of treatment 2, 6 and 22 months respectively). All of them had high volume disease with multiple lung metastases. Moreover all had low grade ULMS, with two of them expressing strongly both receptors and the third being weakly ER positive with unknown PgR status.

Two patients (on anastrozole and exemestane respectively) had PD as best response (duration of treatment 3 months); one of them with low-grade, strongly ER and PgR positive disease was switched to 3rd line treatment with exemestane, while the other patient had high grade, strongly ER/PgR positive disease and was eventually treated with 1st line chemotherapy.

At last follow-up, 4 (25%) of the 16 patients had died with disease progression, 5 (31%) were alive with disease progression, 2 (12.5%) were alive on 1st line letrozole without progression, 2 (12.5%) patients had stopped letrozole and exemestane respectively due to toxicity and were alive without progression and 3 (19%) patients were alive on 2nd line exemestane without progression (Figure 1). Median PFS for 1st line letrozole was 14 months (95% CI: 0–30 months). The 1-year progression-free rate was 63% (95% CI: 39%-87%).Figure 2 illustrates the difference in PFS in 1st line treatment with letrozole between patients with moderate and strong ER and/or PgR status and those with tumours with weak ER and PgR status. Patients with both hormone receptors moderately to strongly positive had superior median PFS [20 months (95% CI: 7– 34)] compared to those with weak ER and PgR expression [median PFS 12 months (95% CI: 0 – 29)]. However, due to the small number of patients it is not possible to drive safe conclusions due to lack of statistical significance in the overall cohort analysis.Moreover, patients with low-grade ULMS (n = 8) had superior median PFS [20 months (95% CI: 7 – 34)], compared to those with high-grade ULMS [n = 8; median PFS 11 months (95% CI: 0 – 30)], as demonstrated in Figure 2b. Accordingly, the 1-year progression-free rate for the low-grade ULMS was 75% (95% CI: 45% - 100%) compared to 50% (15% - 87%) in the high grade ULMS.

Median PFS for the 2nd line AI has not been reached after a median follow-up of 6 months. The 1-year progression-free rate for the 2nd line AI was 80% (95% CI: 21% - 97%).

Toxicities during the 1st line treatment with letrozole were noted among 31% of patients and all were mild (grade 1–2); specifically, hot flushes were recorded in 5 patients (31%), lethargy in 4 (25%), arthralgias/myalgias in 4 (25%), and weight gain in 1 patient (6%). One patient developed osteoporosis and started supplementation treatment. Another patient developed significant nausea and chest tightness and eventually discontinued anastrozole.

During the 2nd line treatment only one patient developed grade 1 arthralgias and another (premenopausal) developed grade 2 hot flushes, headaches and rashes and discontinued treatment. The rate of discontinuation of AIs secondary to toxicity among all patients was low (2/16 patients).