Response to Standard of Care by Folate Receptor Status in Ovarian Cancer Patients Using an Observational Tumor-Linked Database

Authors

Mellissa Yong, PhD1*
Michael Nebozhyn, PhD1
Razvan Cristescu, PhD1
Reshma Rangwala, MD, PhD1
Kimberly Wilson, MS1**
Timothy Yeatman, MD2
Agnes Baffoe-Bonnie, MD, PhD1
Guochun Xie, PhD1
Andrey Loboda, PhD1
Theresa Zhang1, PhD
Daniel Sullivan, MD3
Johnathan Lancaster, MD, PhD3***
Robert Wenham, MD3
Mary E. Hanson, PhD1
James Hardwick, PhD1****

1Merck & Co., Inc., Kenilworth, NJ; 2Gibbs Cancer Center and Research Institute, Spartanburg, SC; 3H. Lee Moffitt Cancer Centerand Research Institute, Tampa, FL
Current affiliations: *Genentech, South San Francisco, CA; **Bristol-Myers Squibb, Princeton, NJ;
***Myriad Genetic Laboratories, Inc., Salt Lake City, UT;
****Pfizer Inc, San Diego, CA

Correspondence:Mellissa Yong, PhD,
Genentech,
1 DNA Way,
South San Francisco, CA 94080.
Tel: 650-225-4007
E-mail: yong.mellissa@gene.com

 

Abstract

Background:Despite high response rates to primary platinum-based therapy, the majority of patients with advanced ovarian cancer experience relapse and die from the disease.Understanding the genetic profile of ovarian cancers may provide a better understanding of chemotherapy resistance mechanisms and may lead to improved care and survival rates.Previous studies reported an association of folate receptor-α(FOLR1) expression with time to recurrence and poor prognosis in patients with serous ovarian tumors.

Methods: Treatment response of newly diagnosed patients with ovarian cancer receiving first-line platinum therapy (PLT) was evaluated, and associations between measures of progression-free survival(PFS) and FOLR1messenger RNA (mRNA) expression levels by treatment response were identified using data from the Merck-Moffitt Collaboration data warehouse.

Results: A total of 347 patients with ovarian cancer received PLT as first-line treatment and contributed tissue to this study; 6% of patients were refractory to PLT, 13% were partially sensitive, 49% were sensitive, and 19% were indeterminate.Patients (n=91) with high FOLR1mRNA expression, that is, ≥75th percentile on a molecular profiling scale of 0 to 4) had a borderline statistically significantly worse PFS on PLT compared with patients (n=238) with low FOLR1mRNA expression (≤75th percentile; P=0.06). The median PFS in patients with high FOLR1mRNA expression was 18.4 months versus 26 months in patients with low FOLR1 mRNA expression (hazard ratio=0.7188 [0.5129, 1.007]).

Conclusion:These observations and those of previous studies suggest that FOLR1 expression may be associated with poor clinical outcome for patients with ovarian cancer,and evaluation of FOLR1-targeted therapeutic strategies in women with ovarian cancer may be warranted.

Keywords:

disease progression; folate receptor 1; ovarian cancer; platinum therapy; mRNA expression

Introduction

Ovarian cancer is the ninth most diagnosed cancer and ranks fifth in mortality rate due to cancer in women.1,2In the United States, new cases of ovarian cancer were estimated to be diagnosed in 22 280women with 14 240deaths expected in 2016.3Until recently, the standard treatment for ovarian cancer has been limited to surgery and chemotherapy.In early disease, surgery alone may be sufficient;but in more advanced disease, cytoreductive surgery followed by chemotherapy is usually recommended.4Primary chemotherapy usually involves a combination of a platinum and taxane-based chemotherapy (commonly carboplatin and paclitaxel).5Despite successful initial chemotherapy, the majority of cases relapse, resulting in more than half of ovarian cancer patients eventually dying from the disease.6In 2011, bevacizumab, an antivascular endothelial growth factor monoclonal antibody,was approved by the European Union for first-line treatment of advanced ovarian cancer in combination with carboplatin and gemcitabine, as well as (in 2012) for the treatment of patients with first recurrence of platinum-sensitive disease.However, extended treatment with bevacizumab has been shown to result in altered patterns of recurrence, usually at distant sites, and recurrence rates as high as 82% inone study of patients with primary ovarian, fallopian tube, or peritoneal cancer treated with bevacizumab alone or in combination with other chemotherapy.7,8Because of the high recurrence rate for women with ovarian cancer, the overall 5-year survival rate is 44%.9.

One way to improve survival rates is to increase knowledge of the biological basis of different cancer types using genomics, biomedical engineering, and information technology principles. Folate is an essential vitamin required by normal and tumor cells and plays an essential role in DNA synthesis and repair and DNA methylation.10Folate receptor has restricted expression in normal adult tissue, but is highly expressed in various nonmucinous tumors of epithelial origin, including ovarian carcinoma.11–13In a study focused on patients with early- or advanced-stage ovarian cancer, patients received 4 to 6cycles of adjuvant chemotherapy with platinum plus cyclophosphamide or paclitaxel regimens after surgery and were then monitored for recurrence.That study demonstrated that higher expression of folate receptor-α (FOLR1), 1 of 3 folate-specific receptors, in serous ovarian cancerous tissue was independently associated with poor prognosis.14Specifically, higher expression levels of FOLR1protein and messenger RNA (mRNA), asexamined by Western blot analysis and reverse transcriptase polymerase chain reaction,correlated with disease progression (stage I: 0.451; stage II: 0.415; stage III: 0.652; and stage IV: 0.768, P=0.004; the expression level of FOLR1 in each ovarian cancer tissue specimen was derived from the ratio of density of FOLR1 to density of glyceraldehyde-3-phosphate dehydrogenase).

An earlier study by Kalli et al13found no association of FOLR1protein expression (examined by immunohistochemistry) with time to recurrence in 186 patients with ovarian cancer; however, there was a modest trend toward earlier recurrence in patients with FOLR1-positive tumors. A subgroup analysis of high-grade serous tumors in the Kalli study did show an association with shorter time to recurrence and a higher proportion of FOLR1 expression.Further study is needed to determine whether there is an association between the presence of FOLR1 and disease progression and survival for patients with epithelial ovarian cancer.

Data from the Merck-Moffitt Collaboration (MMC) data warehouse15was utilized to conduct this current analysis with the following objectives: to obtain measures of treatment response of patients diagnosed with ovarian cancer who received first-line platinum therapy (PLT), and to identify associations between progression-free survival (PFS) and FOLR1 mRNA expression levels by treatment response.

Methods

The Merck-Moffitt Collaboration

The MMC is a public–private collaboration between the H. Lee Moffitt Cancer Center and Merck & Co., Inc. for translational and clinical oncology research, with a mission to build a large tumor tissue bank and patient registry linked to high-quality demographic, histopathology, clinical, and molecular data.15 The data warehouse included integrated clinical, molecular, and research data for basic, population, and clinical sciences. The patient population consisted of participants in the Total Cancer Care®protocol (TCCP), which is a prospective, observational study involving a consortium network composed of 10 Florida hospitals and 8 national sites.15 The study protocol for the MMC database was approved by the institutional review board at each consortium site (H. Lee Moffitt Cancer Center, Tampa, FL; Boca Raton Community Hospital, FL; Watson Clinic, Lakeland, FL; Martin Memorial Health System, Stuart, FL; Morton Plant Mease, Clearwater, FL; St. Joseph’s Hospital, Tampa, FL; Sarasota Memorial Hospital, Sarasota, FL; Tallahassee Memorial Hospital, Tallahassee, FL; Baptist Health South Florida, Miami, FL; Greenville Hospital System, Greenville, SC; Carolinas Medical Center, Charlotte, NC; Hartford Hospital, Hartford, CT; St. Vincent Hospital, Indianapolis, IN; SE Nebraska Cancer Center, Lincoln, NE; Our Lady of the Lake, Baton Rouge, LA; Marquette General Health System, Marquette, MI; and the University of Louisville, Louisville, KY). The patients signed a harmonized informed consent form that granted permission for the use of their data, to follow them for their lifetime, to access their tumor, and to contact them in the future for additional studies, such as a therapeutic clinical trial appropriate for the genetic profile of their tumor.

Tumor Samples

As per the TCCP, tumor tissue samples from surgical resections were collected and snap-frozen in liquid nitrogen within 15 minutes of surgical extirpation. Snap frozen tissues were recorded, labeled, packaged, and placed in a quarantine status at each TCCP clinical consortium site for 14 days to allow a window for pathology to recall the specimen if needed for patient care. After the 2-week window, the frozen samples were shipped on dry ice to a TCCP biorepository at the Moffitt Cancer Center for further processing. Frozen sections were prepared for each tumor specimen and submitted for hematoxylin and eosin (H&E) staining. The H&E slide was submitted for a pathologist’s review in which each tumor specimen was scored for various characteristics, such as the percentages of stromal, malignant, necrotic, and normal tissue. This information was entered into the biorepository’s biospecimen inventory management system of record maintained at LabVantage Biobanking (Bridgewater, NJ).

As part of the pathology quality-control assessment, a pathologist marked each slide precisely for tumor composition, which served as a guide for the macrodissection of each tumor sample. The macrodissection was performed in liquid nitrogen to keep the tissue frozen. On completion of macrodissection, the enriched tumor was weighed and then transferred into a 2-dimensional bar-coded tube and stored in an automated –80°C freezer system. Finally, annotation of each tumor specimen was completed using an accompanying pathology report and was modeled based on College of American Pathologists standards.Specimen handling and pathology assessments with standardized quality-control methods were completed independently of the medical records.

The MMC database was interrogated to identify patients with a primary diagnosis of ovarian cancer who went on to receive a platinum-containing therapy as first-line treatment.Data from radiology and clinic visits were used to assess the end of progression-free time, which occurred on the date that one of the following factors was first noted after the first PLT: the first occurrence of a new lesion,the growth of an identified lesion, or a change of treatment due to progression.Treatment response categories were defined as follows:(1) refractory, in which progression occurred during the first 4 months of PLT; (2) resistant, in which progression occurred 0 to 6 months after the cessation of PLT; (3) partially sensitive, in which progression occurred 7 to 12 months after the cessation of PLT; and(4) sensitive, in which progression occurred >12 months after the cessation of PLT. Platinum resistance versus sensitivity was based on accepted criteria and a 6-month cutoff for disease recurrence after completion of platinum therapy.16,17

RNA Extraction and Processing

Total RNA was extracted from frozen tumors and shipped to the Gene Expression Laboratory at Rosetta Inpharmatics (a wholly owned subsidiary of Merck & Co., Inc., Seattle, WA) for profiling. The amplification and hybridization was performed for all patient samples (starting from total RNA using the NuGEN [San Carlos, CA] 50-ng amplification protocol and HRSTA-2.0 custom human Affymetrix [Santa Clara, CA] array GPL10379). Details of the custom array are available at the National Center for Biotechnology Information Gene Expression Omnibus repository as GPL10379 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GPL10379). The microarray data were processed by running a standard-reference robust multi-array average normalization method as implemented in Affymetrix Power Tools using the default settings, background correction, and quantile normalization with subsequent application of log10 to obtained probe intensities.

Three cutoff points for tumor expression of FOLR1 mRNA expression were tested by molecular profiling on a scale of 0 to 4:25th percentile computed on all Moffitt ovarian tumors (mRNA expression level=2.8); 50th percentile computed on all Moffitt ovarian tumors (mRNA expression level=3.3); and 75th percentile computed on all Moffitt ovarian tumors (mRNA expression level=3.6).For this analysis, FOLR1 mRNA expression level cutoff was prespecified as the 75th percentile across all Moffitt ovarian tumor profiles.

Statistical Analysis

Associations between PFS and overall survival (OS) were assessed using the Kaplan-Meier product-limit estimator, and comparisons between groups were evaluated using the log rank test. Associations between quartiles of FOLR1 mRNA expression levels and OS were evaluated using clinical data linked to molecular profiling data.

Results

Population Characteristics

A total of 347 ovarian patients received PLT as first-line treatment in the MMC database and contributed tissue to this study.Patient and tumor characteristics are shown in Table 1.The average (± standard deviation) age of the patients was 61 (±12) years,and the majority were Caucasian (86%). Approximately half of the patient population had carcinoma histology type, and the majority had composite stage 3 (60%) and grade 3 (66%) disease.

Table_1_

The distribution of treatment response of the 347 patients receiving PLT is shown in Table 2; 6% of patients were refractory to PLT, 13% were partially sensitive, 49% were sensitive; and 19% were indeterminate (that is, patients who remained on therapy between 4 and 12 months and had not progressed).

Table_2_

FOLR1 Expression

The distribution of FOLR1 by treatment response group is shown in Figure 1 (N = 347). There was no significant difference between groups in the median mRNA distribution. The partial response group had lower variability in mRNA expressing tumors compared with all other response groups.

Fig1_fnl

Figure 1. Distribution of folate receptor-α (FOLR1) by treatment response group (N = 347): patients with ovarian cancer receiving platinum therapy as first-line treatment in the Merck-Moffitt Collaboration database. HR, hazard ratio; PFS, progression-free survival.

Based on the 75th percentile cutoff point of 3.6, there were 91 patients in the study with high FOLR1 expression (FOLR1 expression > 3.6). These patients with high FOLR1 expression had a borderline significantly worse PFS on PLT compared with the group of patients (n = 238) with FOLR1 expression < 3.6 (lower FOLR1) (P = 0.06; Figure 2). The median PFS in patients with high FOLR1 expression was 18.4 months compared with 26 months in patients with lower FOLR1 expression (hazard ratio = 0.7188 [0.5129, 1.007]).

Fig2_fnl

Discussion

The emergence of targeted therapies for treatment of ovarian cancer has become increasingly important in light of poor patient prognosis due in large part to chemotherapy resistance and advanced stage of disease at diagnosis.5 Identification of predictive biomarkers is necessary for targeted therapies to improve clinical results and minimize toxicities. Using a large tumor bank and patient registry linked to high-quality demographic, histopathology, clinical, and molecular data, we examined associations between the presence of tumor FOLR1 mRNA and disease progression and survival for patients with epithelial ovarian cancer. To date, this is the first study to use an observational tumor-linked database to assess these associations.

Our results showed that almost half of the patient population was classified as having “platinum-sensitive” disease (remained on therapy for at least a year without progression of their disease), and that there was a borderline statistically significantly worse PFS in patients with high FOLR1 mRNA expression compared with patients with lower FOLR1 mRNA expression. The study by Kalli et al13 observed that the majority of recurrent tumors were positive for FOLR1 protein expression, and FOLR1 protein expression was commonly maintained in recurrent disease, whereas the Chen et al14 study demonstrated that higher expression of FOLR1 protein in serous ovarian cancerous tissue was an independent predictor of a worse prognosis. The subgroup analysis of serous cancerous tissue by Kalli et al showed similar results to those of Chen et al.

In a recent evaluation of gynecological malignancies involving ovary, endometrium, and the fallopian tube, all serous ovarian cancers analyzed (n = 70) were positive for FOLR1 protein expression, and a significant difference in FOLR1 expression was observed between serous and mucinous ovarian carcinomas.18 The current analysis included a variety of ovarian cancer histologies, with a relatively low proportion of serous ovarian cancerous tissues. With half of the patients having a carcinoma histology type, and with a relatively high proportion of adenocarcinoma (23.6%) histology type, a borderline significantly worse PFS in patients with high FOLR1 mRNA expression was observed. Of note, FOLR1 protein expression has been identified in a majority of cases of lung adenocarcinoma in a separate analysis.19 Taken together, these results suggest that FOLR1 expression could be a prognostic factor in serous ovarian cancerous tissue, and potentially other ovarian cancer types rich in FOLR1 expression.

Diagnosis for any medical condition can be complex. For cancer it may be based on several factors, including the site of origin of the tumor, patient symptoms, and other variables. Although tumor pathology for the MMC was conducted in real time, it was handled separately and independently from medical records; therefore, the medical history and full treatment context were unknown to the pathologist at the time the specimens were processed and labeled. It was noted in a number of cases in this analysis that the tumor histology was not consistent with typical ovarian cancer histology (for example, dysgerminoma, granulosa, small cell carcinoma, and teratoma). Moreover, the histology of the patient with squamous cell carcinoma likely refers to a lung tumor. It is possible that this patient may have been diagnosed with ovarian cancer at some point prior to a diagnosis of lung cancer, but histology information for the ovarian tumor was not recorded. This is a limitation of the database, but it may also serve as further scientific information. This “noise” is not only what one may expect when analyzing a large observational database, it is also typical of what clinicians experience when treating patients. The advantage of using the MMC database for this study is the large quantity of specimens with high-quality, detailed clinical annotation. Because the MMC database is an observational database that is linked to centralized data repositories, it can be used to capture real-world clinical practice patterns and health outcomes captured under real-world conditions. It is reassuring that this type of observational database has results concerning outcomes for FOLR1 expression seen in smaller prior studies.

The use of PFS as a surrogate endpoint has come under scrutiny. Its original intended use was to aid in identifying signals of activity in early drug development rather than becoming an important endpoint in determining drug efficacy.20 In first-line therapy, PFS correlates with OS, particularly when a large effect on PFS is observed. The correlation diminishes with smaller effect sizes, so appropriate assessment is warranted. In this study, the use of PFS as the primary endpoint may be appropriate because there is evidence for the use of PFS as a surrogate for OS in the treatment of advanced ovarian cancers regardless of sample size.21

Limitations

There were some limitations to this study. The MMC database is a real-world observational database that was used to assess associations; thus the significant controls that would be present in a well-designed clinical trial were absent. In addition, treatment response to standard of care was evaluated using surrogate measures from abstracted clinical data and may not have been complete for all patients. Misclassification of treatment response may also have occurred to a minimal extent. However, these events are likely to be identified fairly completely for cancer patients because this type of information needs to be thorough and accurate to ensure that decisions about treatment regimens are appropriate for the patient. Finally, the results may not be generalizable to patients with cancers other than ovarian cancer and should be interpreted with that in mind.

Conclusion

Taken together, the observations in this study along with those of previous studies suggest that patients who are refractory to platinum therapy may have the poorest survival as well as poorest response to subsequent therapies. In addition, patients who are platinum sensitive may have the best survival and response to subsequent lines of therapy. The FOLR1 expression status in these subgroups of patients based on platinum sensitivity may serve as another measure of prognosis, and could identify a population that may respond to FOLR1 targeted therapies.

Conflict of Interest Statement

Mellissa Yong, PhD, Michael Nebozhyn, PhD, Razvan Cristescu, PhD, Reshma Rangwala, MD, PhD, Kimberly Wilson, MS, Agnes Baffoe-Bonnie, MD, PhD, Guochun Xie, PhD, Andrey Loboda, PhD, Theresa Zhang, PhD, Mary E. Hanson, PhD, and James Hardwick, PhD, are current or former employees of Merck & Co, Inc., and may own stock or hold stock options in the company. Robert Wenham, MD, has received research funding and honoraria from Merck. Timothy Yeatman, MD, Daniel Sullivan, MD, and Johnathan Lancaster, MD, PhD, have no conflict of interest to disclose.

Author Contributions

MY, RR, KW, TY ABB, and JH were involved in the conception, design, or planning of the study; MY, RR, AL, JL, RW, and MEH interpreted the results; MY, MN, RC, AL, TZ, RW, and JL analyzed the data; TY, DS, and JH provided study materials/patients; KW, GX, DS, and JH acquired data; and KW, MEH, and JH provided administrative, logistical, or technical support. All authors were involved in drafting, revising, or reviewing the manuscript for important intellectual content, and provided final approval of the version to be published. All authors agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved

Funding

Funding was provided by Merck & Co., Inc., Kenilworth, NJ.

Acknowledgements

Jennifer Rotonda, PhD, and Kristen Lewis of Merck & Co., Inc. (Kenilworth, NJ) provided editorial assistance.

References

  1. Ferlay J, Shin H, Bray F, Forman D, Mathers C, Parkin D. Cancer incidence and mortality worldwide: IARC CancerBase No. 10. GLOBOCAN 2008. 2010:2.0. http://globocan.iarc.fr.
  2. American Cancer Society. Survival rates of ovarian cancer, by stage. http://www.cancer.org/cancer/ovariancancer/detailedguide/ovarian-cancer-survival-rates.
  3. Siegel R, Miller K, Jemal A. Cancer statistics, 2016. CA Cancer J Clin. 2016;66:7–30.
  4. Gubbels JA, Claussen N, Kapur AK, Connor JP, Patankar MS. The detection, treatment, and biology of epithelial ovarian cancer. J Ovarian Res. 2010;3:8.
  5. Jelovac D, Armstrong DK. Recent progress in the diagnosis and treatment of ovarian cancer. CA Cancer J Clin. 2011;61:183–203.
  6. Heintz AP, Odicino F, Maisonneuve P, et al. Carcinoma of the ovary. International Federation of Gynecology and Obstetrics (FIGO) 26th Annual Report on the Results of Treatment in Gynecological Cancer. Int J Gynaecol Obstet. 2006;95(Suppl 1):S161–S192.
  7. Dao MD, Alwan LM, Gray HJ, Tamimi HK, Goff BA, Liao JB. Recurrence patterns after extended treatment with bevacizumab for ovarian, fallopian tube, and primary peritoneal cancers. Gynecol Oncol. 2013;130:295–299.
  8. Rauh-Hain JA, Guseh SH, Esselen KM, et al. Patterns of recurrence in patients treated with bevacizumab in the primary treatment of advanced epithelial ovarian cancer. Int J Gynecol Cancer. 2013;23:1219–1225.
  9. American Cancer Society. Cancer Facts and Figures, 2013. Atlanta, GA: American Cancer Society; 2013.
  10. Higdon J, Drake V, Shane B. Folic Acid. Cornvallis, OR: Micronutrient Information Center, Linus Pauling Institute, Oregon State University; 2011. http://lpi.oregonstate.edu/infocenter/vitamins/fa/.
  11. Parker N, Turk MJ, Westrick E, Lewis JD, Low PS, Leamon CP. Folate receptor expression in carcinomas and normal tissues determined by a quantitative radioligand binding assay. Anal Biochem. 2005;338:284–293.
  12. Elnakat H, Ratnam M. Role of folate receptor genes in reproduction and related cancers. Front Biosci. 2006;11:506–519.
  13. Kalli KR, Oberg AL, Keeney GL, et al. Folate receptor alpha as a tumor target in epithelial ovarian cancer. Gynecol Oncol. 2008;108:619–626.
  14. Chen YL, Chang MC, Huang CY, et al. Serous ovarian carcinoma patients with high alpha-folate receptor had reducing survival and cytotoxic chemo-response. Mol Oncol. 2012;6:360–369.
  15. Fenstermacher DA, Wenham RM, Rollison DE, Dalton WS. Implementing personalized medicine in a cancer center. Cancer J. 2011;17:528–536.
  16. Blackledge G, Lawton F, Redman C, Kelly K. Response of patients in phase II studies of chemotherapy in ovarian cancer: implications for patient treatment and the design of phase II trials. Br J Cancer. 1989;59:650–653.
  17. Eisenhauer EA, Vermorken JB, van Glabbeke M. Predictors of response to subsequent chemotherapy in platinum pretreated ovarian cancer: a multivariate analysis of 704 patients. Ann Oncol. 1997;8:963–968.
  18. O’Shannessy DJ, Somers EB, Smale R, Fu YS. Expression of folate receptor-alpha (FRA) in gynecologic malignancies and its relationship to the tumor type. Int J Gynecol Pathol. 2013;32:258–268.
  19. Bremer RE, Scoggin TS, Somers EB, O’Shannessy DJ, Tacha DE. Interobserver agreement and assay reproducibility of folate receptor alpha expression in lung adenocarcinoma: a prognostic marker and potential therapeutic target. Arch Pathol Lab Med. 2013;137:1747–1752.
  20. Booth CM, Eisenhauer EA. Progression-free survival: meaningful or simply measurable? J Clin Oncol. 2012;30:1030–1033.
  21. Bast RC, Thigpen JT, Arbuck SG, et al. Clinical trial endpoints in ovarian cancer: report of an FDA/ASCO/AACR Public Workshop. Gynecol Oncol. 2007;107:173–176.
Tagged with: , , , ,