Trastuzumab in Esophagogastric Cancer: HER2-Testing and Treatment Reality outside Clinical Studies in Germany

We analysed trends over time in palliative first-line chemotherapy in patients with locally advanced or metastatic esophagogastric cancer. Special focus was on frequency and quality of HER2-testing and trends in drug use in combination with trastuzumab. Earlier published data about patients treated outside clinical studies showed a relatively low rate of HER2-testing and insufficient test quality. A total of 2,808 patients retrospectively documented in Therapiemonitor® from 2006 to 2013 were analysed regarding treatment intensity and trends in used drugs. Data on HER2-testing and therapies were analysed in two cohorts documented in 2010 and 2011 (1) compared to 2012 and 2013 (2). Treatment intensity increased: 49.3% of patients received at least a triplet in 2013 compared to 10.1% in 2006. In cohort 2 HER2 expression was tested in 79.1% of the cases. Still, in 26.9% testing was not done as requested by guidelines. Good performance status, multiple metastases, age ≤ 65 years, the objective “to prevent progression,” good cognitive capabilities, estimated good compliance, and social integration positively influenced the probability of HER2-testing; comorbidities negatively affected it. Usage of the combination of fluoropyrimidines and cisplatin with trastuzumab declined from 67% in cohort 1 to 50% in cohort 2.


Introduction
Gastric cancer is one of the most common malignancies worldwide, accounting for 950,000 new annual cases worldwide [1] and 15,840 newly diagnosed patients in Germany in 2010 [2].
In Western countries without specialized screening programs late diagnosis is common. Therefore the majority of patients present with locally advanced or metastatic disease and require palliative treatment during the course of their disease.
Several chemotherapy agents have been studied in the first-line therapy of advanced gastric cancer as single agents as well as a part of combination therapies. Among the "older" drugs 5-fluorouracil (5-FU), epirubicin, cisplatin, and mitomycin C are best known substances [3][4][5][6].
During past years, oxaliplatin, capecitabine, docetaxel, and irinotecan have been introduced in the treatment of gastric cancer. Oxaliplatin as a combination partner in multiagent therapies has been shown to be as effective as cisplatin [7] with fewer toxicities [8]. The oral fluoropyrimidine capecitabine has been investigated and proved to be a suitable substitute for 5-FU [7]. Taxanes such as paclitaxel and docetaxel have also shown to be effective [9,10] as single agent and in combinations [11].
Chemotherapy triplets are used more frequently, and even in elderly patients it could be shown that treatment 2 Gastroenterology Research and Practice intensification may be reasonable in certain clinical situations [12,13].
Trastuzumab in combination with chemotherapy proved to be effective in patients whose carcinomas show HER2 overexpression [14] and is thus recommended for these patients [15][16][17]. Market approval in Germany was granted in 2010.
Standardized immunohistochemical testing (IHC) and in situ hybridisation (ISH) techniques are required to test for HER2 overexpression [18,19]. HER2 overexpression is defined as either an IHC score of 3+ or an IHC score of 2+ and confirmation of gene amplification using in situ hybridisation (ISH) techniques.
In 2013, first data on the practical use of trastuzumab outside clinical studies in gastric cancer after market approval in Germany collected with Therapiemonitor were reported [20]. Main findings of this analysis were a relatively low testing rate for HER2 of only 49.1% of all documented patients in need for first-line palliative chemotherapy. In addition the recommended test algorithm was applied in only 52.2% of the tested patients.
Here we report data obtained with Therapiemonitor from 2006 to 2013 in patients with advanced esophagogastric cancer with a focus on patients undergoing first-line palliative chemotherapy outside of clinical studies reflecting "real-life" therapy strategies.
We were interested whether changes in drug use and treatment intensity in comparison to the previous reported data occurred. Moreover, we focused on HER2-testing strategies and variables influencing likelihood of testing as well as the quality of the used test algorithms.

Therapiemonitor.
Therapiemonitor is a method to collect real world data regarding treatment decisions and strategies in patients with malignant diseases. To achieve this, a retrospective documentation of anonymized clinical and epidemiological data of patients undergoing treatment decisions within a defined timespan is performed in a representative sample of institutions. A detailed description of the methodology has been reported earlier [20,21]. The selection of centers for documentation follows a two-step procedure. The apportioned and stratified random sample is based on an initial survey among all institutions (about = 800) dealing with the treatment of patients with advanced gastric cancer. According to this survey, the "treated prevalence" is ascertained and a collective of patients is apportioned according to treatment center and distributed regionally according to population density. In a second step, selected centers are asked to document their patients undergoing treatment decisions in the respective time period.
Treatment, demographic-and tumor-related data, former medical and surgical treatment, and socioeconomic data like insurance status available in a patient file are documented by the treating physicians. Response and outcome data such as progression-free or overall survival are not collected. Data are double checked: in a real time automatically and centrally for plausibility and completeness by clinical monitors.

Statistical
Methods. All analyses presented herein are explorative. Depending on the size of the predicted numbers and the number of included categories, the Pearson Chisquare, likelihood-quota, or the exact Fisher-test were used.
For all comparisons a value of less than 0.05 was considered statistically significant. Analyses were performed using IBM SPSS Statistics for Windows, Version 19.0 (IBM Corp., Armonk, NY).

Patient and Tumor Characteristics.
Between 2006 and 2013 a total of 2,808 patients with locally advanced, unresectable, or metastatic adenocarcinoma of the stomach or esophagogastric junction and therefore in need for palliative first-line treatment were documented in Therapiemonitor.
Patient and tumor characteristics of all patients are shown in Table 1

Variables Influencing Treatment Intensity in 2012 and 2013.
Data about first-line chemotherapy outside clinical trials documented in 2012 and 2013 was available in 675 patients. The median age of these patients was 65 (range 19-88); sixtytwo percent of the patients ( = 419) were male. The majority (67.0%) had a good performance status with KPS ≥ 80%. Twenty-one percent ( = 141) were treated in university hospitals, fifty-one percent ( = 345) in other hospitals, and thirty-six percent ( = 246) in oncology practices. As mentioned above these patients characteristics remained stable compared to those previously in Therapiemonitor documented and published patient data. Note. § Information on gender is missing in one patient. § § Multiple answers were permitted. § § § Information is missing in one patient 2013. § § § § Other than statutory insurance, § § § § § in 2012 and 2012 the KPS was placed in the last form "Therapiestatus" and changed to KPS in the last therapy decision.
In these patients first-line chemotherapy consists of chemotherapy alone in 553 patients (81.7%) whereas 121 patients (17.9%) received chemotherapy in combination with trastuzumab and 3 patients (0.4%) received trastuzumab monotherapy. 43.6% ( = 295) were treated with single agents or doublet chemotherapy; 56.4% ( = 382) received at least triplets. Age <65 years correlated significantly with a more intense treatment ( < 0.001). Patients who received at least a triplet therapy were younger (median 63 years, range 21-88) than those who had monotherapy or doublet therapy (68 years, range 19-88). Performance status and presence of concomitant disease were also significant factors influencing treatment intensity: the overall amount of patients with KPS ≥ 80% was 67.0%. The group receiving triplets had a significantly higher amount of good-status patients (74.4%) than the lower intensity group (57.4%, < 0.001) and a significantly lower amount of patients with concomitant disease (51.1% versus 66.8/ < 0.001). In addition "achievement of resectability" as objectives of the systemic therapy led to the use of triplets ( < 0.001) while other objectives, namely, "prevention of progression," "improvement of tumour related symptoms and quality of life," had no impact on the decision to use a more intensive chemotherapy regimen. Also the presence of metastases, assumed compliance and cognitive  In addition, 20 patients were reported to have HER2 positive tumor and were treated with trastuzumab. However, the test results reported by the treating physicians of these patients in the case report forms did not indicate HER2positivity. In ten of these patients no IHC analyses as first step of the test algorithm were performed. Another eight patients with IHC 0 and 1+ underwent an ISH analysis indicating HER2 overexpression. In addition in two patients with IHC 2+ no additional ISH analysis was performed.

Factors Influencing the Probability for HER2-Testing.
Tumor specific as well as patient specific and institutional related variables were analysed for the likelihood of HER2testing. Variables affecting HER2-testing in the actually reported group in 2012 and 2013 compared to previous reported patients in 2010 and 2011 are depicted in Table 4. Significant correlations for a higher likelihood of applying a HER2 test were found for KPS ≥ 80% ( < 0.001), the presence of multiple metastases ( = 0.007), lower age ≤ 65 years ( = 0.048), the objective "to prevent progression" ( = 0.006) and patient specific factors assumed by the treating physicians as good cognitive capabilities ( < 0.001), good compliance ( < 0.001), and complete social integration ( = 0.047). Concomitant diseases negatively affected the decision to test for HER2 ( = 0.003). The percentage of patients tested for HER2-expression in relation to age compared in the actual  Other factors were also analysed: education, insurance status, the treating institution, gender, initial diagnosis with stage IV disease, tumor size, tumor localisation, localisation of metastatic site, or subtype of tumor histology had no impact on the decision to test for HER2 (data not shown).

Discussion
Here we report on a large cohort of 2,808 patients with metastatic esophagogastric adenocarcinoma who were treated between 2006 and 2013 and documented with a healthcare research tool to collect and analyse data on treatment reality in cancer patients. We focused on trends over time in treatment intensity, used drugs, HER2-neu testing frequency, and quality as well as anti-HER directed treatment patterns. Patient and tumour characteristics of these patients remained stable through the time span from 2006 to 2013. Noteworthily the amount of patients with poor KPS < 80% receiving chemotherapy increases over the years from 27.1% in 2006 to 33.4% in 2013 meaning even frail patients were treated. A possible explanation of this "real-life" trend are the encouraging results in studies as FLOT 65+ performed especially in elderly patients [12,13].
An ongoing trend to use oxaliplatin instead of cisplatin as well as an increase of the use of docetaxel as part of the 1st-line treatment was observed [20,21]. Most probably these findings reflect the results of several studies demonstrating the efficacy and feasibility of docetaxel and oxaliplatin containing regimens [8,11,22] and as mentioned above in terms of FLOT-65 even in elderly patients [12,13]. In contrast, epirubicin as part of chemotherapy triplets in the 1st-line setting is less used since 2010. One could speculate this is due to several disadvantages in the application of the most common used epirubicin containing regimens in Germany such as ECF and ECX or EOX like the need for continuous infusion of 5-FU through the whole treatment period or difficulties in swallowing while using an oral application form as capecitabine in patients with esophagogastric cancer. As reported before [20,21] most of the patients received combination treatment as first-line chemotherapy with half of the patients receiving at least a chemotherapy triplet since 2010 reflecting the actual German and international guidelines for treatment of locally advanced or metastatic esophagogastric cancer [16].
We analysed factors influencing treatment intensity in 675 patients documented in the last cohorts in 2012 and 2013. Age < 65 years, good KPS > 80, lower presence of concomitant disease, and the objective "achievement of resectability" lead to the use of at least a chemotherapy triplet in these patients. This compares adequately with what was observed in previous documented cohorts [20,21].
Another focus of our present analyses was frequency and quality of HER2-testing in a cohort of 683 patients documented in 2012 and 2013. Almost 80% of the patients were tested for HER2 expression. In 26.9% of the patients the suggested test algorithm was not applied in the appropriate manner. Compared to earlier published data documented in 2010 [20] with only 49.1% of patients tested for HER2expression and 52.2% of patients with inappropriate test algorithms this represents almost doubling in test rate and "quality" of testing. The most probable reasons for this efforts, especially in test quality, are the widespread availability of testing and commonly adoption of the proposed test algorithm in qualified laboratories as published in 2011 [23,24]. This may be underlined by the finding that in contrast to earlier published data [20] the treating institution clinics versus office based physicians are no longer a variable affecting the frequency of HER2-testing.
Regarding other variables affecting the likelihood for testing for HER2 expression in the actual cohort (2012 and 2013) the following factors were significant positive predictors in bivariate analyses: higher KPS, good cognitive capabilities, assumed good compliance, the objective "to prevent progression," the presence of multiple metastases, lower age ≤ 65 years, and complete social integration. Concomitant diseases negatively affected the decision to test for HER2. Again, this compares adequately with what was observed in previous documented cohorts for most of the analysed factors [20]. Noteworthily the factor age ≤ 65 years seems to be a less strong predictor for testing for HER2-expression only 50% of the patients were treated with this combination. Accordingly the use of trastuzumab in combination with a 5-FU or docetaxel monotherapy or the use of trastuzumab monotherapy rose from 4% to 10% reflecting the fact the treating physicians consider a less toxic monochemotherapy regimen in combination with trastuzumab or even a trastuzumab monotherapy as an effective treatment regimen even in the elderly and therefore especially these patients showed higher test rates than in previous years. Regarding the ongoing trend for treatment intensification in the good performance status patients the use of a chemotherapy triplet in combination with trastuzumab rose from 15% in 2010 and 2011 to 22% in 2012 and 2013. The most commonly used chemotherapy triplets in combination with trastuzumab in the patients documented in 2012 and 2013 were 5-FU or capecitabine/oxaliplatin/docetaxel in 19% of the patients. Another 18% of the patients received trastuzumab in combination with 5-FU or capecitabine and oxaliplatin confirming previously reported data [20,21] and the ongoing trend of the use of less toxic oxaliplatin instead of cisplatin as well in combination with trastuzumab. The efficacy and feasibility of the use of oxaliplatin in combination with trastuzumab in the treatment of locally advanced or metastatic gastric cancer were confirmed in a recently published phase II study [25] and a retrospective analyses [26].

Conclusion
In summary, in the present analysis we found an ongoing trend toward the use of more intensive treatment in first-line chemotherapy in patients with esophagogastric adenocarcinoma. In 2012 and 2013, more than two years after market approval of trastuzumab in the treatment of esophagogastric cancer, HER2-testing is widely used according to the suggested and standardized test algorithms and independent of the treating institutions, reflecting the feasibility of testbased treatment strategies. Half of the patients documented in 2012 and 2013 and treated with trastuzumab outside clinical studies were not treated according to label with cisplatin and 5-FU as chemotherapy combination partners leading to uprising test rates and treatment especially in the elderly.

Disclaimer
Roche Pharma AG had no influence on data acquisition and analysis nor on the writing of the paper and on the decision to submit it for publication.