By Turna Ray
An advisory agency to UK's National Health Service has found that four molecular diagnostics lack sufficient clinical and cost-effectiveness evidence to support their use in assessing whether early-stage breast cancer patients should receive adjuvant chemotherapy treatment.
The UK's National Institute for Health and Clinical Excellence last month issued a diagnostics consultation document that recommended against the use of Genomic Health's Oncotype DX, Agendia's MammaPrint, and General Electric Healthcare/Clarient's Mammostrat to guide adjuvant chemotherapy treatment decisions for women with estrogen receptor-positive, lymph node-negative, HER 2-negative early-stage breast cancer.
NICE's Diagnostics Advisory Committee "could not recommend these tests because of uncertainty in the evidence of clinical effectiveness leading to uncertainty about the cost effectiveness," NICE said in its provisional recommendations.
Additionally, NICE limited its recommendation for the use of IHC4, an immunohistochemistry test developed by the Royal Marsden Hospital and Queen Mary University London, to research settings involving early-stage breast cancer patients. "Although the modeling suggested that IHC4 could be both beneficial and cost saving, weaknesses in the evidence base meant that it could currently only be recommended for use in research studies to collect data needed to reduce the uncertainty on clinical and cost effectiveness," NICE said.
Comments on the draft recommendations are due by Feb. 24. After consultations with stakeholders, NICE may change its initial recommendation for the final guidance. Representatives from the three commercial test providers confirmed this week that they plan to hold discussions with NICE during this period.
For the analysis, NICE consulted various independent experts and an external assessment group regarding the clinical effectiveness and cost effectiveness of the four tests. Upon reviewing the analytical validity, clinical validity, and clinical utility of the tests, NICE found the data wanting for all of them.
While the advisory committee said that the clinical evidence for Oncotype DX and MammaPrint was the most mature, with the "highest quality evidence" reported for Oncotype DX, "most studies, for all tests, were retrospective in design, analyzing archived tumor samples from a cohort of patients with documented information on patient characteristics and outcomes." According to NICE, such study designs are more likely to be influenced by bias compared to prospective randomized-controlled trials.
Several studies submitted to NICE included heterogeneous breast cancer populations, although the institute's review was focused on ER-positive, node-negative, HER2-negative, early-stage breast cancer patients. Further weaknesses of the data included pooled analysis, small patient cohorts in certain studies, and short follow-on periods. In addition, a large proportion of studies enrolled non-UK populations.
Currently, UK doctors use the Nottingham Prognostic Index and Adjuvant! Online to assess whether patients are likely to experience disease recurrence and therefore should be treated with adjuvant chemotherapy. NPI determines disease prognosis by factoring in tumor size, grade, and spread. Adjuvant! Online takes into account patients' age, tumor size, histological grade, comorbidities, as well as whether the cancer has spread to the lymph nodes and involves hormone receptors. NICE recognized that there is variability in how these standard clinical tools are used at local healthcare facilities and may not provide the clearest indication to doctors regarding whether their patients should receive chemotherapy.
The economic analysis conducted by the external assessment group assessed cost-effectiveness of the four tests assuming they were being used in addition to current practice. First, the assessment group considered test cost-effectiveness for all women, 75 years or younger, with ER-positive, node-negative, HER2-negative early-stage breast cancer. Second, the expert group considered cost-effectiveness when the tests were given to women with these same characteristics, but who also had a Nottingham Prognostic Index score of above 3.4 to account for women who were at intermediate risk of recurrence by the molecular diagnostics.
NICE considers a medical intervention to be cost effective when its incremental cost-effectiveness ratio falls below £20,000 ($31,778) per quality-adjusted life year gained.
Cost-Effectiveness Analysis
Genomic Health's Oncotype DX is a RT-PCR based test that assesses the expression of 21 genes in women with ER-positive, node-negative, HER2-negative early-stage breast cancer, and yields a score between one and 100 indicating their recurrence risk over 10 years. The recurrence risk score has been shown in studies to be able to predict which women will benefit from chemotherapy in addition to the hormonal therapy tamoxifen and which patients can be treated with just tamoxifen.
The incremental cost-effectiveness ratio for Oncotype DX when compared to current practice was £26,940 ($42,833) per QALY gained. At a threshold of £20,000 per QALY gained, NICE found that, compared to current practice, the probability of Oncotype DX being cost effective in this population was 12.44 percent.
When the test was offered to women with a Nottingham Prognostic Index score above 3.4 the ICER was £9,007 ($14,320) per QALY gained. At a threshold of £20,000 per QALY gained, the probability of Oncotype DX being cost effective in this population was 91.56 percent.
NICE considered the cost-effectiveness ratio to be "too high to recommend Oncotype DX for use in the NHS for all women with ER+, LN-, HER2- early breast cancer," it said. Furthermore, with regard to women with an NPI score over 3.4, the committee noted that more than 38.5 percent of patients needed to benefit from chemotherapy across all Oncotype DX risk groups in order for the test to be cost effective.
"The committee considered that the benefit of chemotherapy was unlikely to be as great as 38.5 percent across all Oncotype DX risk groups," NICE noted. Therefore, it "was not able to recommend Oncotype DX for use in the NHS for women with ER+, LN-, HER2- early breast cancer and a Nottingham Prognostic Index score above 3.4."
NICE added, however, that data from the National Cancer Institute-led Trial Assigning Individualized Options for Treatment, or TAILORx, study may provide further clinical evidence on the utility of Oncotype DX. The TAILORx trial has enrolled more than 10,000 breast cancer patients who received an intermediate recurrence score from Oncotype DX and researchers are investigating how the test can be used to make treatment decisions for this patient group. Study results are slated for 2015.
When it came to MammaPrint, Agendia's 70-gene microarray-based breast cancer recurrence test, NICE found the data on the test's predictive capabilities lacking. As such, the ICER for MammaPrint was between £12,240 ($19,448) and £53,058 ($84,303) per QALY gained when the test was used for all women with ER-positive, node-negative, HER2-negative, early-stage disease. When women with a Nottingham Prognostic Index score above 3.4 were considered, the ICER for the test ranged from between £6,053 ($9,617) and £29,569 ($46,982) per QALY gained.
NICE said it provided the ICERs as a range "because there was uncertainty around the data for the predicted benefit of chemotherapy."
GE/Clarient's Mammostrat, meantime, would designate more women to receive chemotherapy than would standard techniques, according to NICE's analysis. Mammostrat gauges five immunohistochemical markers (SLC7A5, HTF9C, P53, NDRG1 and CEACAM5) and stratifies patients into risk groups associated with chemotherapy benefit.
NICE estimated the ICER for Mammostrat to be £26,598 ($42,261) per QALY gained when the test is used for all women with early stage breast cancer. The Committee reported that the test had a 36 percent chance of being cost effective in the ER-positive, node-negative, HER2-negative population.
When used for women with ER+, LN-, HER2- early breast cancer and an NPI score above 3.4, Mammostrat was "dominated by current practice" and had an 18.04 percent chance of being cost-effective, NICE said.
Although NICE said it needed more validation data to make definitive recommendations for IHC4, preliminary economic modeling showed that it saved £498 ($791) compared to standard procedures.
"The cost-effectiveness acceptability curve showed that the probability of IHC4 being cost effective, when compared to current practice only and at a threshold of £20,000 per QALY gained, was almost 100 percent," NICE said.
IHC4 measures levels of the ER, progesterone receptor, HER2, and Ki-67 proteins, as well as clinical and pathological variables and calculates a distant recurrence risk score. An online calculator for the test is slated for launch this year. NICE noted that data for IHC4 were particularly lacking on the reliability and reproducibility of the Ki-67 marker.
Next Steps
In response to NICE's preliminary recommendations, test sponsors plan to hold discussions with NICE and possibly submit additional data to the committee during the consultation period.
"NICE's draft guidance is not unexpected given that they have had a conservative approach to emerging technologies in oncology. It is surprising to the extent that there is cost-effectiveness data available, only in the UK, but also in the rest of the world," Steve Shak, Genomic Health's chief medical officer, said during a call with analysts this week to discuss 2011 revenues.
Genomic Health plans to discuss with NICE clinical evidence regarding Oncotype DX, and to point out the test is recommended in breast cancer guidelines issued by the St. Gallen Breast Cancer Conference, the European Society for Medical Oncology, the National Comprehensive Cancer Network, and the American Society of Clinical Oncology. The company has previously presented data from multiple countries, including the UK, showing that with the aid of the Oncotype DX risk score doctors change their treatment recommendations approximately 30 percent of the time.
"[We] will continue to work with relevant stakeholders to ensure UK patient access to Oncotype DX through studies, selected trusts that have established contracts, and private payers who currently provide broad access for patients with private health insurance," Genomic Health said in a statement.
The company estimates that as many as 700 women in the UK have already used Oncotype DX to guide cancer treatment decisions. According to Genomic Health, the UK women tested on Oncotype DX include those who have had testing covered by private UK insurers, such as AXAPPP, Bupa, Aviva, PruHealth, SimplyHealth, Cigna, and WPA.
"We will work to provide [NICE] with more clinical information to support the inclusion of Oncotype DX," Shak said. "And we … have a long-term goal of having all patients in the UK being provided access to the test."
Meanwhile, a spokesperson for Agendia told PGx Reporter that the company will provide comments to NICE during the consultation period and submit data from prospective studies once they are available. Currently, absent a coverage decision from NHS, "private payors occasionally pay [for MammaPrint] on a case-by-case basis," the spokesperson added.
GE Healthcare expressed disappointment in NICE's preliminary recommendations regarding its test. "We firmly believe in the value of Mammostrat and what it offers to patients, doctors and payors around the world," a company spokesperson told PGx Reporter. "We remain committed to further demonstrating the benefit of Mammostrat to patients. We will continue to work with payors and agencies around the world to expand the clinical evidence of performance."
Have topics you'd like to see covered in Pharmacogenomics Reporter? Contact the editor at tray [at] genomeweb [.] com.