Castle Biosciences Seeks to Bolster Skin Cancer Diagnostics Franchise With New Study

NEW YORK (360Dx) – After establishing its first two tests to predict the risk of metastasis in two melanoma indications, Castle Biosciences is laying the foundation to expand its technology for use in cutaneous squamous cell carcinoma (sCCC). 

The private, Texas-based company sponsored a study presented at the American College of Mohs Surgery meeting last week, which described a gene expression profile signature to predict recurrence risk in sCCC. Castle Bio is in the validation phase for the sCCC test and plans to commercialize it if it meets the company’s internal performance criteria. 

Castle Bio President and CEO Derek Maetzold said the study is part of the company's plan to establish Castle Bio as the leader in skin cancer diagnostics. Following a strategic review two years ago, the firm decided against applying its technology in esophageal and rectal cancer. Given its commitment to dermatologists and surgical oncologists in skin cancer, Maetzold said Castle Bio's resources would be spread too thin if it entered other oncology markets. 

In contrast, sCCC presents an opportunity for Castle Bio to home in on the skin cancer market. In the study looking at prediction of recurrence risk in sCCC, researchers led by Sarah T. Arron of the University of California, San Francisco, identified 73 candidate genes from scientific literature and pathway analysis. Six control genes were confirmed as showing consistent expression across all the samples tested. The six were then used to normalize expression values of the remaining genes. 

After analyzing mRNA expression, investigators identified 18 genes that were differentially expressed between recurrent and non-recurrent cases (p<0.05). 

The optimal model identified in the analysis was 71 percent sensitive and 90 percent specific, with a 50 percent positive predictive value, and a 96 percent negative predictive value (NPV) for recurrence. Based on a review of the scientific literature, Castle Bio and the researchers asserted that the data demonstrates an improvement over existing staging criteria like those developed at Brigham and Women's Hospital, which predict recurrence with only about 60 percent sensitivity. 

"For patients with cutaneous squamous cell carcinoma, accurate prediction of their individual risk of recurrence or metastasis remains a challenge," Arron said in a statement. "The development of a prognostic test that improves risk prediction could better inform important management decisions such as optimal surgical procedure, use of adjuvant radiation, and selection of patients for [sentinel lymph node biopsy] or adjuvant immunotherapy, and has the potential to drive improvements in patient management." 

Melanoma test gaining a foothold 

The sCCC results fit comfortably with Castle Bio's DecisionDx-Melanoma test for cutaneous melanoma, which is now being used to guide the treatment plans for a significant number of patients after they are diagnosed with stage I or stage II cancer, said Maetzold. 

Researchers, including some with financial ties to Castle Bio, evaluated the melanoma test and concluded in a study published recently in SKIN The Journal of Cutaneous Medicine that it "impacted clinical management decisions and led to changes in management that were aligned with guideline recommendations for the care of patients with melanoma."

In the SKIN study, 49 percent of patients overall experienced a change in clinical management recommendations after doctors received the DecisionDx-Melanoma results. Among stage I patients, 36 percent (66 of 181) experienced a post-test change in management plan after the test results, while 85 percent (56 of 66) of the stage II patients had a treatment plan, according to the study. 

"We are seeing strong adoption in cutaneous melanoma because of the implications for treatment because of the prospective study," said Maetzold. In a statement from Castle Bio, study coauthor Larry Dillon noted the study demonstrated that using DecisionDx-Melanoma resulted in "risk-appropriate changes in clinical management that matches the biological risk of the melanoma tumors. Importantly, the management changes are aligned with guideline recommendations to direct more frequent and intense surveillance to high-risk patients."

When Castle Bio launched the 31-gene DecisionDx-melanoma test three years ago, it was able to demonstrate that the test could distinguish between primary melanoma samples from patients who later metastasized and those who did not with an area under the receiver operating curve of more than 0.9.

Maetzold also noted that 8 percent of melanoma patients are diagnosed with stage III disease yet those patients comprise approximately 40 percent of the annual deaths in melanoma. The remaining 60 percent, stage I and II patients, are initially classified as low risk, but their melanoma metastasizes quickly despite the early diagnosis. 

"This is one of the problems we [in the melanoma community] have to solve — to use the biology of the tumor rather than just relying on the anatomical features to identify tumors that are aggressive," said Maetzold. "Physicians understand now if they classify patients as low risk through our tests, there is a low risk of [disease] recurrence following regular treatment protocols."

The acceptance of DecisionDx-Melanoma is increasing, according to Maetzold who said the test is now used by about one-third of the dermatologists and surgical oncologists who diagnose melanoma cases. Of that total, two-thirds are dermatologists and the rest are surgical oncologists. 

Last year, the test was used to identify about 12 percent of the roughly 75,000 cases of stage I and II melanoma cases, Maetzold said. This year, Castle Bio aims to boost that figure to 18 percent. 

If the company reaches that mark, Maetzold said Castle Bio's revenue should range between $40 million and $50 million, based on an average price of $3,500 per test. In 2017, the company expects to generate between $27 million and $32 million, depending on the company's collection of reimbursements from payors, a process that is still ongoing.

The cutaneous melanoma market has been estimated at $440 million per year, but that figure, Maetzold said, is based on epidemiology estimates of more than 100,000 patients, which differs from clinical estimates of the market of between 65,000 and 75,000 patients. 

The long-term goal for DecisionDx-Melanoma would be to reach 75 percent of the patients, as the company has achieved with its DecisionDx-UM test, launched in 2010, for uveal melanoma, according to Maetzold. Uveal melanoma is diagnosed in approximately 2,000 patients annually, a significantly smaller market than cutaneous melanoma. The DecisionDx-UM test generates approximately $4 million in annual revenue for Castle Bio. 

According to Maetzold, the DecisionDx-Melanoma test faces no commercial competition in the market regarding determination of risk for recurrence. While other firms have tests for diagnosing melanoma, he said none are focused on predicting recurrence. 

DermTech CMO Burkhard Jensen said that the market for melanoma gene expression tests is evolving to address three groups. The Castle Bio test addresses the care decisions of medical oncologists and surgical oncologists, as well as dermatologists, as a tool to replace or complement sentinel node biopsies, said Jansen. 

Dermtech's noninvasive melanoma gene expression test, guides biopsy decisions for clinicians following a visual assessment, he said. The company’s Pigmented Lesion Assay (PLA) helps rule out melanoma and the need for a surgical biopsy of atypical pigmented lesions identified by a clinician. 

Meantime, Myriad Genetics also offers a melanoma test, myPath Melanoma, which identifies whether a mole is benign or malignant, but does not differentiate the stages of melanoma, according to a company spokesperson.