Treatment patterns and outcomes in patients with metastatic synovial sarcoma in France, Germany, Italy, Spain and the UK

Synovial sarcoma (SS) is a rare and aggressive subtype of soft tissue sarcoma (STS), accounting for ≈5–10% of cases [1]. The estimated age-standardized STS incidence is 3.3–4.7 cases/100,000 individuals in Europe [2]; 993 incident SS cases were estimated in 2019 for European Union-28 countries [3]. A recent French study, based on expert pathology review of cases from 2013 to 2016, reported an incidence rate of 1,674/1,000,000/year [4]. Metastatic disease is present in ≈24% of patients at diagnosis and 50–70% of patients with localized disease will develop metastatic SS (mSS). Survival prospects for mSS are poor, with a 5-year survival rate of ≈10% [1].

The Clinical Practice Guidelines of the European Society for Medical Oncology–European Reference Network for rare adult solid cancers cover the treatment of STSs, but due to the rarity of the disease, there are no specific guidelines for SS alone [5]. In Europe, surgical resection with or without radiotherapy is the standard treatment for localized SS [5]. A joint prospective analysis showed that patients with adequately resected ≤5 cm SS, regardless of grade, can be safely treated with a surgery-only approach. Avoiding the use of adjuvant chemotherapy and radiotherapy in this low-risk patient population may decrease both short- and long-term morbidity and mortality [6]. For low-risk patients, surgery can be enough; however, for high-risk patients, neoadjuvant chemotherapy is standard first-line care. In adult patients, neoadjuvant chemotherapy is primarily offered to patients with high risk of mSS [7,8].

In patients with advanced (unresectable or mSS), standard first line of treatment (LOT) consists of anthracycline chemotherapy (monotherapy or combination) [5]. Given the high activity of ifosfamide (IFOS) in SS, most centers prefer IFOS/anthracycline combinations or IFOS as front-line therapy [8]. Clinical evidence guiding treatment options for patients with advanced/mSS who progress following anthracycline-based chemotherapy is lacking [5]. There is no standard-of-care chemotherapy in later-line settings [9], and while some studies have reported a degree of benefit in approved treatments for STS following progression on chemotherapy (such as pazopanib and trabectedin) [10–15], a recent meta-analysis indicates a lack of consistent evidence supporting a positive impact on overall survival (OS) [16].

Because patients with metastatic relapse of SS show significant morbidity and mortality with little opportunity for cure [7,17], it is essential to characterize treatment patterns, understand their real-world effectiveness, and examine the impact of disease on healthcare resource utilization (HCRU) [14,18,19]. In a chart review of patients with advanced STS in four European countries (the UK, France, Germany and Spain), the subpopulation of patients with SS comprised only 6% of the overall population, and the most common regimens used (and their duration), OS and overall response rates were only available for first LOT (1L) patients [18]. Typically, details on LOT 2 (2L)+ treatment patterns or HCRU associated with SS are lacking across published studies [9,18,19].

The aim of this study was to describe the real-world burden of illness in patients with mSS in five European countries (France, Germany, Italy, Spain and the UK), including patient characteristics, treatment patterns and outcomes by LOT, and HCRU.

Methods

Study design & objectives

This was a descriptive, non-interventional, retrospective cohort study of patients with mSS in France, Germany, Italy, Spain and the UK (Figure 1). Primary objectives were to describe the demographic and clinical characteristics of treated patients with mSS and to summarize their treatment patterns and outcomes. A secondary objective was to describe annualized HCRU in the period after diagnosis, stratified by country.

The study employed a customized data collection effort from Oncology Advantage (OA), an online physician survey reporting on a consecutive set of patients receiving therapy and their previous treatment history, by specific tumor cohort. OA targets large ‘center of excellence' hospitals and smaller local sites, to be representative of real-world treatment patterns and care settings in each country. The survey collected anonymous patient-level data from a specific time point (not designed to capture the same patients prospectively). Physicians reported on their most recent 2–5 consecutive patients treated for mSS in the past 12 months.

Eligibility criteria

Physicians eligible to complete the survey were: board certified ≥3 years; currently or previously initiated or recommended/prescribed treatment for patients with SS; seen ≥2 adult patients with SS in the past 12 months; and able to provide details of cytotoxic and/or targeted therapies prescribed for SS. Physicians were considered eligible if they spent >40% of their time seeing patients and >70% in a hospital setting (all other countries), or office-based setting (Germany only). The German healthcare system is organized differently compared with other European countries in that most patients, depending on disease, are treated at office-based sites rather than at main hospitals.

Eligible patients were: diagnosed with mSS by case definition; ≥18 years of age at the time of data collection; positive for chromosomal translocation between SYT on the X chromosome and SSX1, SXX2, or SSX4 on chromosome 18, or had unknown chromosomal translocation status (i.e., the physician was uncertain whether the patient was tested); alive or died ≤12 months prior to the time of data collection, irrespective of last LOT; if living, currently receiving cytotoxic and/or targeted therapies, or received cytotoxic and/or targeted therapies ≤12 months prior to the time of data collection; or if deceased, received ≥2 cytotoxic and/or targeted therapies ≤12 months prior to the time of data collection.

Study dates & time periods

The data collection period for which physicians were surveyed began on 29 August 2019, and ended on 20 November 2019, or sooner if the desired patient quotas per country were fulfilled prior to this date. Physicians were instructed to evaluate all available medical history information when completing the survey and to include information on eligible patients treated most recently ≤12 months prior to the time of data collection.

The index date was the date of mSS diagnosis. The LOT start date was the date of the first dose of a specified LOT. The study inclusion date was the date of survey completion. Date of death was recorded at month and year level, so date of death was defined as the mid-point of the month at which patient death was recorded. For patients with a primary diagnosis of localized SS prior to the index date, the time from diagnosis to the index date, and prior LOTs received for localized SS, were also documented.

Follow-up & censoring

For living patients who had an end date recorded for their most recent LOT, follow-up time started at index date (inclusive) and at the end date (inclusive) of the most recent LOT (recorded closest in time to patient's study inclusion date). For living patients and whose most recent LOT was recorded as ‘ongoing’, or for patients without an end date of their most recent LOT prior to study inclusion date, follow-up time was censored at study inclusion date (inclusive). For deceased patients, follow-up time started at index date (inclusive) and ended at date of death (inclusive).

Study assessments

Physician & patient characteristics

Physician characteristics were described overall and by country, in terms of medical specialty, type of work setting, extent of clinical experience and number of patients treated. Patient baseline demographic and clinical characteristics were summarized overall, by country and by LOT.

Treatment patterns

Common treatments were assessed in terms of the proportion of patients in the five countries overall receiving each type of treatment regimen within their current/most recent LOT and current/historic LOT, defined as having a start date closest to the date survey completion or patient's death.

Treatment outcomes

Time-to-next-treatment (TTNT) was summarized for the five countries overall and by country. OS was described for the five countries overall. TTNT was defined as the time from the first dose of a given LOT (start date of given LOT; inclusive) until the first dose of the next distinct LOT (start date of next consecutive LOT; exclusive). Patients were considered to have experienced an event if they progressed to the next LOT from the given LOT.

Patients may not have progressed to their next LOT due to possible ‘censoring' events occurring prior to start date of the next LOT; for living patients (most recent LOT), TTNT was censored at study inclusion date if LOT was ongoing (exclusive: right censoring/end of follow-up). For deceased patients (most recent LOT), TTNT was censored at date of death (exclusive).

OS was derived within the combined living and deceased patient cohorts. Time to death (in days) was derived for the deceased cohort from the index date (inclusive) to date of death (exclusive). Time to death (OS) from index date (inclusive) for living patients was censored at study inclusion date (exclusive), as future information on date of death for these patients was not collected. Only month and year of death were recorded; therefore, the calculation of OS used date of death defined as the mid-point of the month at which patient death was recorded. Patients who died between the index date and study inclusion date were considered to have experienced an event.

HCRU

Annualized HCRU was investigated in the period after mSS diagnosis and stratified by country, including inpatient hospital admissions, total days of inpatient hospitalizations and outpatient visits.

Data analyses

Descriptive analyses were conducted for patient demographic and clinical characteristics, treatment patterns, and HCRU.

Each LOT was defined for chemotherapy/targeted treatments only (surgery and radiotherapy were excluded). Common treatment regimen patterns were reported for current/most recent and current and historic LOTs combined overall and by country. Percentages may not sum to 100% because some patients received no therapy while others received multiple lines of therapy. All 2L and 3L patients are also included in the 1L group.

The number of HCRU events was reported by the surveyed physician for each patient over their follow-up time. Annualized HCRU for each variable was described continuously and reported for the five countries overall and by country. Annualized rate of HCRU was derived via the equation:

Results

Patient & physician characteristics

Overall, 296 patients with mSS were included from France (14.9%), Germany (18.6%), Italy (22.3%), Spain (22.3%) and the UK (22.0%; Table 1). At the time of data abstraction, 68.6% of patients were living (median age: 54 years [range: 18–89]). The median number of metastases/patient was 4 (range: 1–30), with lung/respiratory system as the most common metastatic sites (77.7%). Most patients (76.4%) had a primary diagnosis of mSS. Among patients with a primary diagnosis of localized SS (23.6%), median time from diagnosis to index date overall was 502 days (range: 0–6760; Supplementary Table 1).

The majority (86.1%) received 1 LOT; 38.9% received 2 LOTs and 8.4% received 3+ LOTs (Supplementary Table 2). A total of 70.2% 1L, 51.3% 2L and 48.0% of LOT 3 (3L)+ patients had an Eastern Cooperative Oncology Group performance status (ECOG PS) of 0–1. A larger proportion of 2L (47.8%) and 3L+ (52.0%) patients had an ECOG PS ≥2 compared with 1L patients (27.5%).

Seventy-one physicians fulfilled eligibility criteria in France (14.1%), Germany (21.1%), Italy (22.5%), Spain (21.1%) and the UK (21.1%; Supplementary Table 3). Most physicians (77.5%) were medical oncologists, were from teaching or research hospitals (43.7%) and had 15–34 years of experience (57.7%). The median number of patients with mSS who were seen by physicians within the past 12 months was 12 (interquartile range [IQR] 6–30). Physicians in the UK saw the highest median (IQR) number of patients (34 [8–50]), followed by physicians in France (15.5 [10–30]), Spain (12 [6–20]), Italy (10 [3–29]) and Germany (8 [5–10]).

Treatment patterns

The most common 1L regimens overall were doxorubicin/IFOS-based regimens (37.4%), doxorubicin monotherapy (27.0%) and IFOS-based regimens (excluding combination with doxorubicin; 14.9%; Figure 2A). Doxorubicin/IFOS-based regimens were the most common 1L regimens in France (62.2%), Germany (40.7%) and the UK (45.0%). Doxorubicin monotherapy was most common in Spain (47.1%) and IFOS-based regimens most common in Italy (33.3%). In 2L, the most common treatments overall were trabectedin-based regimens (29.7%), followed by IFOS-based (16.8%), pazopanib- or trabectedin-based (15.8%) and doxorubicin-based (13.9%) regimens. Trabectedin-based regimens were most common in the UK (50.0%), Germany (34.8%) and France (33.3%), whereas IFOS-based regimens were most common in Italy (37.0%; Figure 2B). In 3L+, trabectedin-based regimens were reported in 28.6% of patients, whereas 33.3% received ‘other' regimens (Figure 2C). Trabectedin-based regimens were most common in Germany (40.0%) and Italy (66.7%).

These observations by country were consistent when current and historic treatment patterns were combined for 1L and 3L+ (Figure 2). The most common reasons for stopping treatment were treatment course completion at 1L (62/122, 50.8%) and distant progression/relapse at 2L (10/21, 47.6%) and 3L+ (5/9, 55.6%). Concurrent surgery was reported in 23.6, 11.9 and 19.0% of 1L, 2L and 3L+ patients, respectively (Supplementary Table 4). Twenty-three percent of 1L, 21.8% of 2L and 14.3% of 3L+ patients had radiotherapy. Most patients starting 1L (81.0%), 2L (77.2%) or 3L+ (90.5%) received supportive care.

Treatment outcomes

TTNT

For 1L, median (IQR) TTNT overall was 13.1 (9.1–18.0) months for living patients and 6.0 (2.8–11.0) months for deceased patients (Table 2). Among all patients (living and deceased), TTNT was 12.2 (10.3–20.0) months in 2L, 13.2 (4.2–13.2) months in 3L, 15.6 (15.6–15.6) months in 4L, and was not reached in 5L. For TTNT by country level, data are limited due to the small number of patients.

OS

There were 79 (26.7%) death events overall, and median OS was 22.0 months (95% CI: 18.0–31.2) from the date of first mSS diagnosis for all LOTs. The percentage of living patients at 6, 12 and 24 months of the post-index date was 63.2, 34.1 and 11.1%, respectively. The OS probabilities were 0.94 (95% CI: 0.91–0.97) at 6 months, 0.79 (95% CI: 0.73–0.85) at 12 months and 0.49 (95% CI: 0.40–0.59) at 24 months. Median OS calculated from the date of respective LOT initiation was 6.0 (95% CI: 5.2–9.0) months in 2L (n = 122), 4.9 (4.4–N/A) months in 3L (n = 21), and 7.5 (5.9–N/A) months in 4L (n = 7).

HCRU

The median (IQR) number of hospital admissions/year overall was 1 (0–3) (Table 3). Higher median (IQR) numbers of hospital admissions/year were observed in France (2 [1–4]) and the UK (1.5 [1–6]) compared with the other countries. Median (IQR) total days in hospital was 3 (0–12)/year overall; estimates for France (4 [2–10]), Germany (5 [0–17]) and the UK (5.5 [0–16]) were similar and were higher than in Italy and Spain (0 [0–8]).

The median (IQR) number of outpatient visits/year was 4 (2–9) overall, with the highest number of visits in Germany (8 [4–11]), followed by the UK (7 [3–12]), Spain (6 [2–14]), Italy (3 [2–6]) and France (2 [1–4]; Table 3). The median (IQR) number of visits to the medical/clinical oncologist was 3.5 (1–7), being highest in Spain at 6 (1–10) visits/year and lowest in France (2 [0–4]). The median (IQR) number of visits to the surgeon was 0 (0–1), being highest in the UK at 1 (0–2) visits/year and lowest in France (no visits recorded). The median (IQR) number of visits/year to the radiotherapist was 0 (0–1) overall, and in all countries except for the UK, where the median (IQR) was 1 (0–1) visits/year.

Discussion

To our knowledge, this retrospective assessment of the demographic and clinical characteristics, treatment patterns and HCRU among 296 patients diagnosed with mSS in five European countries (France, Germany, Italy, Spain and the UK) is the largest European real-world evidence (RWE) study of patients with mSS, and the first to characterize HCRU in the management of mSS.

The number of patients seen per physician varied between countries, which may reflect country-specific centralization of care for patients with SS; indeed, a median of 34 patients were seen by UK physicians compared with eight for Germany, which operates a more decentralized healthcare system. Most patients had a primary diagnosis of mSS, and, of these, lung/respiratory metastases were the most common. The median age of patients with mSS was 54 years, which is higher than previously reported in a real-world study of mSS in Italy (37 years) [20] and a review of patients with mSS in 15 European clinical trials (40 years) [21]. It is possible the higher median age observed in this study was a result of the patient selection criteria (including those most recently seen by their treating physician), leading to a degree of survival bias. Owing to the paucity of data for mSS, other like-for-like comparisons to published data are difficult, as most previous studies have provided results for STS overall, with small populations of SS cases included.

The treatment pattern findings suggest broad adherence to current treatment recommendations in Europe, where doxorubicin (monotherapy or with IFOS) is recommended in 1L; IFOS monotherapy provides an alternative if doxorubicin is contraindicated [5,22,23]. This may particularly be the case if patients have received doxorubicin for localized disease before and have reached the maximum cumulative dose. A small minority of patients (6.3%) were receiving trabectedin or pazopanib as 1L. Trabectedin is approved after progression following anthracyclines and IFOS and for patients unsuited to receive these agents [11], while pazopanib is only approved for mSS following prior chemotherapy, or in patients who have progressed within 12 months of (neo)adjuvant therapy [10]. It is possible that some patients received pazopanib following doxorubicin as perioperative treatment of localized disease and then metachronous metastases. Similar to the results reported in this study, a real-world retrospective review of medical records of 807 patients treated with systemic therapy for advanced histologically diagnosed STS in the UK, Spain, Germany and France by Nagar et al. found that doxorubicin alone, or in combination with ifosfamide, was frequently used at 1L in all four countries [18]. Similarly, Leahy et al. analyzed data from a retrospective medical record review of 213 patients with metastatic STS across nine countries, including the five European countries of our study, and found that the most common 1L regimens were doxorubicin monotherapy (34%) and anthracycline plus ifosfamide (30%) [24]. In the US specifically, Pokras et al. reported that patients with mSS treated in community practices most commonly received doxorubicin and ifosfamide (35.9%) followed by either monotherapy or combination treatment with pazopanib (17.9%) at 1L [25]. Similarly, the Pollack et al. multi-center retrospective study of 249 patients with advanced or mSS from nine sarcoma centers in the US reported that the most common treatment in 1L was doxorubicin and ifosfamide combination (37%) followed by ifosfamide monotherapy (20%) [26].

Doxorubicin and IFOS-based regimens were also being utilized in later LOTs, which may reflect some degree of ‘rechallenge' with these agents. This is not possible to confirm from the survey data, but research suggests that this approach has clinical activity in STS and may be a viable option for the treatment of advanced/mSS [27,28]. The most common 2L+ regimens were trabectedin-based, which is consistent with product labeling and guideline recommendations [5,11,22,23]. Pazopanib monotherapy and combinations were used in 15.8% and 9.5% of patients in 2L and 3L+, respectively. ECOG PS decreased as patients progressed to later LOTs. Our results were consistent with the findings of a RWE study of 5298 patients with advanced STS across France, Germany, Spain and Italy that trabectedin was the most commonly used choices of treatment in 2L [29]. In contrast to the results of our study, Pollack et al. reported that, in the US, trabectedin was the 5th-most common choice of treatment (12%) for patients with advanced or mSS in 2L [26]. Pokras et al. reported the most common regimens for patients with mSS in the USA involved monotherapy or combination of either ifosfamide (17.9%) or pazopanib (17.9%) in 2L, and monotherapy or combination of pazopanib (20.5%) in 3L [25]. Overall, our findings for treatment patterns across LOTs are similar to published STS RWE in Europe. However, all the RWE so far has been in a broader group of advanced STS, in which SS forms a small population. This is the first RWE study to focus predominately on a larger SS population. Moreover, our study shows differences in treatment pattens at 2L and beyond compared with US studies in patients with SS.

The median TTNT for the five countries overall was similar for 1L, 2L and 3L (13.1, 12.2 and 13.2 months for living patients, respectively). Differences were observed among individual countries in 1L and 2L, with the highest median TTNT observed in the UK (1L: 16.5 months; 2L: 20.0 months). The TTNT increase observed here may be due to small patient numbers in later lines.

The estimated median OS (from the date of mSS diagnosis across the five countries for all LOTs) was 22 months, similar to that reported in the recent METAstatic SYNovial sarcoma (METASYN) database study of 417 patients with mSS in France (median OS: 22.3 months [95% CI: 19.7–24.1]) [30]. Another study of patients with mSS treated at a single institution in Italy had an approximate median OS of <18 months (estimated value from a Kaplan–Meier plot showing 5-year OS) [31]. In the USA, Pokras et al. reported that patients with mSS who received systemic treatment at 1L in community practices had a median OS of 24.5 months (95% CI: 17.1–35.0) [25], which is similar to the median OS of 24.7 months (95% CI: 20.9–29.4) reported in Pollack et al. [26]. The median OS for 2L (6.0 months; 95% CI: 5.2–9.0) is slightly lower than other reports, which range from 10.4 months (95% CI: 8.5–12.7) in a meta-analysis of trials and RWE for pazopanib and trabectedin efficacy in 2L+ [16] to 16 months (95% CI: 13.5–18.9) in an analysis of patients with mSS receiving various anticancer therapies at nine major US academic centers [26]. Our estimates may be lower than those reported in the meta-analysis [16] due to a variety of regimens being used, and lower than in the other analysis [26], as that study included patients treated at major academic medical centers. Overall, these findings confirm the need for more effective treatment options in previously treated patients.

The HCRU data showed that the UK, Germany, and Spain had higher levels of HCRU than France and Italy; the latter had particularly low levels of inpatient and outpatient hospital visits. These findings likely reflect between-country variances in healthcare practices, including utilization of various treatments with different administration routes and/or drug-related adverse event management, rather than being an indication of the severity of disease, which seemed comparable between countries. With the differences noted across countries, it is important for future researchers to ensure HCRU can differ by country for the same condition and disease severity, and this should be considered in cost-effectiveness and economic analyses.

Strengths of our study include the use of real-world, standardized survey data collected by experienced physicians across specialties and five countries. The study included large center of excellence hospitals and collected patient-level data based on physician-abstracted charts rather than claims data. The paper contains details on patient characteristics, treatment patterns and outcomes. The study had broad patient eligibility criteria and accrued a large sample of patients with mSS, despite the rarity of this disease. These factors help ensure wider applicability and generalizability of the data, and provision of detailed information on mSS treatment patterns by LOT and HCRU addresses some of the gaps in existing literature in these five countries.

Treatment patterns for 2L reported here differ from those reported for patients with mSS in regions outside of Europe. For example, in a single-center, retrospective analysis in Japan, the most common 2L therapies were doxorubicin with or without IFOS and IFOS plus etoposide [32]. In a retrospective cohort study of patients with mSS treated in US community practices, IFOS- and pazopanib-containing regimens were most common in 2L [25]. However, these studies also highlight the urgent need for new approaches to improve outcomes in patients with mSS in the 2L setting, irrespective of the treatment received [25,32]. In the Japanese study, the 1-year OS rate was 67.1% in patients receiving 2L therapy (median OS was not reported) [32], while in the US community study, median OS was 24.5 months (range: 17.1–35.0) in patients receiving 1L therapy [25].

The study had the limitation of being a retrospective analysis, with possible selection bias for participating physicians. As the cross-sectional study design did not allow for follow-up after the study inclusion date, the outcomes of ongoing treatment or future LOTs could not be determined. Restricting eligibility to patients who received treatment or died in the past 12 months may have introduced selection bias by excluding patients in remission (likely to have better outcomes), and patients receiving palliative care (likely to have worse outcomes). In the living cohort, inadvertent selection of patients at earlier stages of treatment may have occurred. There may also be an under-representation of living patients in later LOTs compared with the deceased cohort, where eligibility was restricted to patients with ≥2 LOTs. Another limitation of the design is that a certain number of living and deceased patients were preselected, so the OS is not based on natural distribution of death in this population. In addition, LOT was defined by use of chemotherapy/targeted treatments only (information on surgery and radiotherapy was only descriptive), which could potentially affect our understanding of the breadth of treatment used by patients and consequent impact on outcomes. Patients initially diagnosed with localized disease and treated with cytotoxic/targeted regimens who subsequently went on to develop metastatic disease and were unable to take the same regimen were not accounted for in this analysis.

The study assumed that physicians reported data based on medical record abstraction, which may be subject to errors arising from the abstraction process and comprehension of the data collection tools. While major events such as patient hospitalizations were captured, patients could have received healthcare services in other settings (e.g., pharmacy visits or other specialty clinics outside of the physicians' institutions), which may not have been adequately recorded in patients' medical records. This is likely to be country-specific, influenced by variations in the healthcare systems of the countries sampled and, therefore, represents a potential confounding factor for regional associations observed. In addition, reasons for each healthcare visit and costs were not available. While this is the largest sample of patients with mSS investigated to date, there remain limitations associated with interpreting data in subgroups, in which event numbers are relatively small. For example, further stratification of analyses (e.g., outcomes for current vs historic LOTs) could not be fully explored, and the survey only captured 21 patients being treated in 3L+. However, this might be expected for mSS, where relapsing patients are increasingly likely to be managed in supportive care settings, rather than having regular contact with a specialist. Finally, as the German healthcare system is structured to have more office-based physicians compared with the other four counties, the selection criteria for physician eligibility could not be applied. Consequently, prescribed treatments and outcomes may have been affected.

Taken together, the variation in treatments across LOTs, the documented worsening of ECOG PS with subsequent LOTs, and poor clinical outcomes highlight the continued unmet need for effective therapies to improve outcomes in patients with mSS. Therapeutic progress for mSS will require mSS-specific clinical trials, and trial designs in Europe should consider regional patient characteristics, treatment patterns and outcomes.

Conclusion

The use of real-world, standardized survey data that have broad patient eligibility criteria and accrue a large sample of patients is key for exploring unmet needs and burden of illness, especially in rare populations in which only small sample sizes exist at single centers. In the future, owing to cell and gene therapies in development for metastatic synovial sarcoma, there are expected to be more treatment options available for patients, with possible changes in treatment patterns and patient outcomes.