New Studies Question Utility of Gene Expression Testing for Cancers With Unknown Primary Site, Low-Risk Breast Cancer

Gene expression profiling has been making inroads into clinical cancer care, with application in cancer classification, most notably with breast cancer. Two recent studies are calling into question the clinical utility and the cost-effectiveness of such tests.

Cancers of Unknown Origin
For patients with cancer of unknown primary site (CUP), use of microarray profiling to guide site-specific cancer treatment does not result in a significant improvement in survival compared to empirical treatment, according to a phase II Japanese study published Jan. 17 in the Journal of Clinical Oncology.

“With the increasing number of effective molecularly targeted therapies available as well as the recent success of immunotherapies for several types of cancer, the gap in treatment advances between CUP and known primary cancer is widening,” write the authors led by Hidetoshi Hayashi, MD, PhD, from Kindai University in Japan.

Gene expression profiling that has the potential to predict the origin of tumor tissue for patients with metastatic disease of unknown origin, based on the basis of site-specific expression profiles.

In this clinical trial, comprehensive gene expression profiling was performed on frozen biopsy tissue using microarray analysis. An established algorithm was applied to results to predict tumor origin. Tissue of origin was predicted for all 130 patients across 16 different anatomical sites—pancreatic and gastric cancer and malignant lymphoma were the three most common predicted primary sites, accounting for approximately 60 percent of all patients. From consent to results took 3 weeks. Based on these results, patients were randomized (1:1) to receive either site-specific therapy or empirical chemotherapy based on test results between October 2008 and February 2015.

The researchers found that the 1-year survival rate was similar between the site-specific and empirical treatment groups, as were median overall and progression-free survival. No subgroup of patients showed a significant benefit in terms of overall survival from site-specific treatment.

“Despite an accuracy rate of 78.6 percent for site prediction in silico, our results suggest that comprehensive genome-wide profiling of gene expression by microarray analysis might not yet be suitable for clinical application inpatients with CUP,” the authors conclude. “However, the observation that patients predicted to have more-responsive tumor types had a better OS and PFS than those with less-responsive tumor types suggests that pre-diction of original tumor site has prognostic value.”

Low-Risk Breast Cancer
OncotypeDx (ODX; Genomic Health; Redwood City, Calif.) is the most commonly used tumor profiling test and is included in practice guidelines worldwide. However, a new study, published in the January issue of the Journal of the National Comprehensive Cancer Network, is questioning the tests’ cost-effectiveness in women with a low clinical risk of breast cancer recurrence.

Unlike previous analysis, this study examined cost-effectiveness by a patients’ risk category, as determined by clinical and pathologic features. This analysis did confirm the tests’ cost effectiveness in intermediate- and high-risk estrogen receptor (ER)–positive, HER2-negative, lymph node–negative breast cancer.

ODX, a 21-gene assay, is widely used in clinical practice to accurately identify patients who would benefit from chemotherapy. Using gene-expression profiling, the test provides a recurrence score (RS) between 0 and 100, with scores categorized as low (RS < 18), intermediate (18 ≤ RS ≤ 30), or high (RS ≥ 31) risk.

The authors note that the previous analyses that concluded ODX is cost-effective for all women with ER-positive, node-negative breast cancer had “important methodologic limitations” in that they combined all patients into one group, regardless of clinical and pathologic features.

“Such an approach, ignoring tumor characteristics when evaluating ODX, is not consistent with actual clinical practice because it assumes that all patients would be treated without consideration of their clinical characteristics,” write the authors led by Shi-Yi Wang, M.D., Ph.D., from Yale University in New Haven, Conn.

The new analyses used data from the Connecticut Tumor Registry to identify 4,281 women diagnosed with ER-positive, HER2-negative, node-negative breast cancer between 2011 and 2013. Just over half of these women (54.6 percent) received ODX testing.  The researchers classified the 2,245 patients into 3 clinical risk groups according to the PREDICT model, a risk calculator developed by the National Health Service in the United Kingdom.

The researchers found that the PREDICT risk calculator categorized the sample as 82.5 percent low-, 11.9 percent intermediate-, and 5.6 percent high-risk. When the three groups were combined, ODX had an incremental cost-effectiveness ratio (ICER) of $62,200 per QALY for patients aged 60 years. However, ICERs, differed substantially across clinical risk groups, ranging from $124,600 per quality-adjusted life years (QALY) in the low-risk group, to $28,700 per QALY in the intermediate-risk group, and $15,700 per QALY in the high-risk group.

There was also substantial variability in the cost-effectiveness of ODX when considering variation in life expectancy and age-based utility weights. ICERs ranged from $77,100 per QALY for patients aged 45 years to $344,600 per QALY for patients aged 75 years in the PREDICT low-risk group. In the intermediate- and high-risk groups QALYs remained lower than the $100,000 threshold for women aged 45 to 75 years.

“Given that most women in our population-based sample were classified as low risk, our study suggests that clinicopathologic information needs to be incorporated in ODX testing decision-making,” writes Wang and colleagues.

Despite the cost concerns raised, some experts are wary to change clinical practice immediately

“I applaud any efforts to reduce unnecessary testing and chemotherapy in patients where there is no benefit,” said Lori Goldstein, M.D., a member of the National Comprehensive Cancer Network (NCCN) Guidelines Panel for Breast Cancer. “While the present analysis raises some thoughtful questions, until future prospective randomized trials are able to further help confirm subsets of patients with no clear benefit from additional chemotherapy, the NCCN recommendations remain appropriate. The current costs associated with the ODX test seem justifiable based on patient benefit.”

Takeaway: While the long-term utility of gene expression profiling is not in question, these studies raise some questions about optimizing testing to most cost-effectively inform patient care.