NEW YORK (GenomeWeb) – Survey results from a study funded by Castle Bioscience targeting specialists in the ophthalmology field who regularly treat patients with uveal melanoma have indicated that a substantial majority of these doctors regularly use molecular testing to guide treatment of their patients, including Castle's gene-expression assay, DecisionDx-UM.
According to Castle President and CEO Derek Maetzold, the data, published recently in the journal Clinical Ophthalmology, largely reflect what the company has seen in internal research on physician adoption. Castle estimates that 95 percent of specialists who treat uveal melanoma use its test and that at least 70 percent of diagnosed UM patients in the US receive the test at this point.
Interestingly, having solidified this hold on the uveal melanoma field, Maetzold said that Castle actually now sees its business focus shifting away from testing in niche or low-prevalence cancers like UM and toward more prevalent diseases, such as the cutaneous form of melanoma, as well as esophageal cancer, the target of its newest test, which it is rolling out in January of 2015.
The company's uveal melanoma test measures the expression of 15 genes and classifies patients into one of three risk categories: Class 1A, which is associated with about 2 percent chance of metastasis within five years; Class 1B, with a 21 percent chance of metastasis over five years; and Class 2, the highest risk, with 72 percent odds of metastasis over five years.
Whether uveal melanoma patients are treated initially with radiation, the most common treatment, or with eye removal surgery, about 50 percent see a reoccurrence of their disease in metastatic lesions. Castle has demonstrated that its gene expression test is superior to other clinical measures of disease aggressiveness and the risk of metastasis. This has allowed clinicians to better determine which patients can be monitored less aggressively post-surgery. It has also helped spur the first time clinical trials of adjuvant treatment for patients who are at a high risk of recurrence, according to Maetzold.
In the company's recent publication, Castle scientists collaborated with a team of researches led by Thomas Aarbeg of Michigan State University. The group conducted two sets of surveys to assess practice patterns among UM specialists in regard to molecular testing in general and to Castle's UM test in particular, as well as a study reviewing medical records to evaluate the impact of test results on treatment patterns.
In the medical record review study, the researchers examined the records of Medicare patients who received Castle's test in 2012. Among 191 total evaluable records, 88 patients had sufficient documentation of their treatment to evaluate how treatment and surveillance patterns differed among those with different DecisionDx-UM risk scores.
According to the study, 58 percent of patients in this sample set had a low DecisionDx-UM risk score while the other 42 percent were high-risk. For all of the low-risk cases, documentation indicated patients received a low-intensity observation plan including "liver function tests with or without abdominal ultrasound" once or twice per year. In contrast, the records indicated that all of the high-risk patients were followed with high intensity surveillance, defined as "liver function tests and abdominal ultrasound, CT, and/or positron emission tomography," two to four times per year.
In the two surveys also included in the group's Clinical Ophthalmology report, the researchers tracked both adoption of molecular testing by a subset of known UM-treating ophthalmologists, as well as the impact of test results on their post-surgery follow-up and clinical care of patients.
Respondents were asked both about their use of molecular testing in general, which could include either Castle's gene-expression assay or chromosomal testing for Chromosome 3 monosomy, which is associated with higher risk of metastasis, but which Maetzold said is not as sensitive as gene-expression testing, missing about one quarter of patients who later have disease recurrence.
One of the company's surveys, conducted in 2012, reported that 82 percent of the responding specialists said that they performed some type of tissue-based molecular analysis of UM cases, either Castle's test or a chromosomal analysis. The other survey in 2013 and 2014 found that 88 percent reported offering patients Castle's gene expression test specifically, while 24 percent said they offered chromosomal analysis.
The survey results also reflected the results of the team's medical records review study. Respondents said that most of the time molecular testing results impacted their clinical decision-making, prompting more active and intense surveillance for patients with high-risk results compared to those with low-risk results.
According to the study authors, 74 percent of ophthalmologists who participated in the 2012 survey said they used the information obtained from cytogenetic or gene-expression analyses to adjust the frequency of metastatic disease surveillance for their patients.
Similarly, 79 percent of respondents to the 2014 survey said they used the information from molecular analyses to change their clinical practice.
According to Maetzold, these data reflect what Castle has also seen in its internal tracking — that a majority of ophthalmologists who specialize in treating uveal melanoma have adopted molecular testing and use it to guide post-surgical clinical care for their patients.
About 95 percent of ophthalmologists and retina specialists who regularly treat these cancers use Castle's test, he said, and the company believes that currently 70 percent of diagnosed patients are tested with DecisionDx-UM.
Importantly, Maetzold said, the availability of a more accurate prognostic tool has led to significant changes in clinical practice, most notably, recruitment of patients whose DecisionDx-UM results place them in the high recurrence risk category to clinical trials of adjuvant or post-surgical treatment. These include a trial by Bristol Myers Squibb of its drug Yervoy (ipilimumab) in patients with a high-risk DecisionDx-UM score or the presence of Monosomy 3, and a trial of Pfizer's Xalkori (crizotinib) in patients with a Class II DecisionDx-UM result.
Looking to the future, Maetzold said that Castle is now anticipating a shift from its focus on rare, or orphan cancers to advancing newer tests targeting diseases that occur more frequently.
The company also offers tests for other less-common cancers, including mesothelioma, thymoma, and glioma, although Maetzold said its uveal melanoma assay has made up the bulk of its testing revenue prior to the launch this year of a cutaneous melanoma test.
Earlier this year Castle launched this DecisionDX-Melanoma assay, which measures the expression of 31 genes and is intended to assess the risk of recurrence of skin melanoma lesions. As in uveal melanoma, the test divides patients into risk groups, with the lowest risk subset having a very minimal risk of disease recurrence over five to 10 years, and the highest risk group having a very high risk of recurrence.
Unlike uveal melanoma, which strikes only as many as 2,000 people each year in the US, stage I or II cutaneous melanoma, which has not yet spread and thus may benefit from prognostic analysis using the company's test, occurs in 65,000 to 70,000 patients per year.
According to Maetzold, Castle expects its first validation study of this test to be published in January and its second validation to appear in print soon after. With a larger target population, and a larger group of treating physicians, the rapid, almost complete penetration Castlehas seen in the uveal melanoma space is less likely, Maetzold said. But even with lower market penetration, the company can expect to perform many more tests in this much larger group of patients.
In January 2015 the company also plans to more broadly launch its test for esophageal cancer. Standard of care for the disease is to treat patients with chemotherapy and radiation before surgery.
"Given neoadjuvant treatment, a certain group of esophageal adenocarcinoma patients, about a third, have a tremendous response and show no tumor in their resected tissue. Their tumor is essentially chemically ablated," Maetzold explained. "But, likewise, another 20 to 30 percent of patients look like they never even received chemo or radiation, so you end up delaying the only effective course, which is surgical resection."
Castle intends its test to help clinicians predict which patients are likely to respond to preoperative chemotherapy and radiation, and which will be unresponsive to this treatment and thus may benefit more from immediate surgery.
With a general shift to developing and marketing tests for more prevalent cancers, Maetzold said Castle's strategy has also moved away from licensing promising biomarkers and early-stage assays from the academic and research community to developing tests in house.
"There isn’t as much stuff available … as there was five years ago," Maetzold said. "Grants are tight and almost all funding is being directed toward identifying new druggable targets for targeted therapy, which doesn't necessarily lend toward tests to improve [clinical] management."
"We haven't actively looked for additional assays in last few months," he added. "We are focused now on the cutaneous melanoma test, validation of our esophageal cancer test, and have a test behind that for rectal cancer, which we are developing internally."
"We'll still look around," he said. "But I think more development will come internally, especially as we move out of rarer cancers into areas that are more common."