Skip to main content
Premium Trial:

Request an Annual Quote

BioTheranostics Modeling Study Demonstrates Cost Effectiveness of CancerType ID Test


This story has been updated to clarify the current state of the Tissue of Origin test formerly of Pathwork Diagnostics, now owned and marketed by Response Genetics.

NEW YORK (GenomeWeb) — Modeling conducted by the Partnership for Health Analytic Research and funded by BioTheranostics has found that the company's CancerType ID test, a 92-gene assay to determine the origin of cancers of unknown primary site, is cost-effective and improves patient care.

The study used four separate models covering a range of possibilities of how the test might be used in the clinical treatment of metastatic cancer patients, and these simulations showed that overall the test increased the number of patients treated according to their cancer type, and also increased quality-adjusted survival with an incremental cost-effectiveness ratio of about $50,000 per quality adjusted life year — significantly below the commonly used threshold of $100,000.

BioTheranostics has conducted a number of other studies to support the clinical utility of its CUP test. The new modeling report, which appeared in the Journal of Medical Economics, is the first published evidence that the test is not only clinically impactful, but also cost-effective.

Brock Schroeder, director of medical and scientific affairs at BioTheranostics and an author of the modeling study told PGx Reporter that while the company is also interested in collecting real-world confirmation of the test's cost effectiveness, models are important evidence of the test's practical value.

Schroeder said the modeling was undertaken carefully and scrutinized for sensitivity to any one data input. Overall, the study used four different models, one attempting to closely represent real-world use of the test, one looking at the impact of integrating the test earlier, one using it later, and a final model simplifying the clinical process to compare immediate treatment based on the test results versus immediate empiric or nonspecific chemotherapy.

In creating such models, the researchers attempted to use hard data for as many inputs as possible, Schroeder explained. However, for certain costs or other model inputs, sufficient data was not available. In those cases — such as estimates of the percentage of patients who should be expected to be treated correctly under different clinical and diagnostic scenarios — the CancerTypeID models also relied on expert opinion, which the study authors reported was sourced from consultations with multiple practicing clinical oncologists.

As a failsafe against any one of these inputs skewing the models, Schroeder said the research included sensitivity analyses that evaluated how changes in any particular input might affect the result.

BioTheranostics' collaborators at the Partnership for Health Analytic Research "were really the primary designers of the study and they are really industry experts," he said, adding that PHAR relies on "hard published data" for its modeling wherever possible.

"Where it becomes a little difficult is when there is not high-quality data published, so they do a series of interviews with experts [followed by] the sensitivity analyses that were run as part of this paper. Those look at each individual variable … in either direction to see how that skews the study results," Schroeder said.

According to the authors, inputs based on expert opinion rather than published data were not key drivers of the model results according to these sensitivity tests, nor were other suspect inputs such as information used to calculate patient quality of life.

The models included patients with eight common cancers, including breast, colon and rectum, kidney and renal pelvis, liver and intrahepatic bile duct, lung and bronchus, ovarian, pancreatic, and prostate.

For patients receiving empiric therapy, the models estimated survival of 13.13 months, based on published literature. The group calculated cancer type-specific treatment survival for each cancer subtype using surveillance epidemiology and end results data, and adjusted this using published survival estimates associated with site-specific versus empiric therapy, the authors reported.

Costs for different testing, including immunohistochemistry, and for oncology consultations and other clinical costs were obtained from published estimates and expert opinion, and the commercial list price for CancerType ID was used as the test's cost input to the model, the authors wrote. Quality of life measures were estimated from published literature as well.

Overall, the study found that in the primary model, which was designed to reflect expected clinical use of the test, CancerType ID increased the proportion of patients treated "correctly," or according to their tumor-type: 81 percent versus 58 percent. The test also decreased the proportion of patients treated incorrectly and those treated with nonspecific or empiric therapy

According to the model, use of the assay also increased quality-adjusted survival by 1.15 months compared to patients who did not receive the test, with an incremental cost-effectiveness ratio —the ratio of the change in costs to the incremental benefits of a therapeutic intervention or treatment — of $50,273 per quality-adjusted life year. According to BioTheranostics, this compares favorably to the societal willingness-to-pay threshold in oncology of at least $100,000 per quality-adjusted life year.

In the secondary modeling scenario where CancerType ID was used earlier in the diagnostic process, even more patients were diagnosed and treated correctly, 83 percent versus 63 percent, and both life expectancy and total cost increased resulting in a ratio of $63,972 per QALY, the authors reported.

The model that pushed testing later in the clinical process also resulted in a similarly increased cost-effectiveness ratio, as did the fourth, simplified model, in which survival increased by 3.4 quality-adjusted months, and cost-effectiveness ratio increased to $85,584 per QALY.

According to the study authors, the results suggest overall that even when a range of assumptions regarding standard practice patterns are included, the test should remain under the acceptable cost-effectiveness threshold of $100,000 per QALY in increasing overall and quality-adjusted survival.

This also held true when varying parameters to explore uncertainty in model inputs, as Shcroeder explained.

One potential confounder of the models was the inclusion and distribution of only eight cancer sites or types. According to the authors, the distribution of patients in the model among these eight sites may not accurately reflect reality. Additionally, they wrote, survival and treatment cost data, which were culled from retrospective data, may underestimate both costs and survival benefit of newer therapies not included in that dataset.

However, the authors wrote, the models suggested that using CancerType ID can improve treatment accuracy and clinical outcomes cost effectively.

In the market for cancer of unknown primary testing, BioTheranostics is also joined by Rosetta Genomics with its microRNA-based Rosetta Cancer Origin Test. Before it went out of business last year, Pathwork Diagnostics also marketed a test to determine the origin of unknown primary metastatic cancer. Response Genetics later acquired the IP related to that assay and currently offers the Tissue of Origin test clinically.