Castle Biosciences, Inc., a provider of molecular diagnostics to improve cancer treatment decisions, today announced the publication of results from a study comparing and combining risk prediction using DecisionDx®-Melanoma, the Company’s gene expression profile (GEP) test used to predict risk of metastasis in cutaneous melanoma (CM) patients, and the American Joint Committee on Cancer’s (AJCC’s) clinical staging system using the Individualized Melanoma Patient Outcome Prediction Tool. The AJCC online prognostic tool is used by clinicians to estimate survival rates for patients with Stage I and II melanoma based on tumor depth, ulceration status, location and patient age to determine appropriate patient management.

The paper, “Identification of High Risk Cutaneous Melanoma Tumors is Improved When Combining the Online AJCC Melanoma Patient Outcome Prediction Tool with a 31-Gene Expression Profile-Based Classification” was published today in the Journal of the American Academy of Dermatology (JAAD). In the study, researchers found that when used in combination with the AJCC clinical staging system, DecisionDx-Melanoma can significantly improve the identification of those Stage I and II melanoma patients who have a high risk for recurrence and/or developing metastatic disease.

“For physicians and patients, the risk assessment provided by the GEP test may ultimately lead to earlier identification of metastatic disease when tumor burden is lower and the disease is potentially more treatable,” said Laura Ferris, M.D., Ph.D., Associate Professor of Dermatology, University of Pittsburgh, and lead author of the study. “This information is vital in determining the best possible follow-up care plans for patients to ensure that their intensity of surveillance is appropriate for their individual risk of melanoma recurrence.”

Study Details

Study investigators determined survival predictions by examining records from 205 Stage I and Stage II melanoma patients collected from six U.S. centers. AJCC clinical staging factors (Breslow thickness, ulceration, site of lesion, age and sentinel lymph node biopsy status) were entered into the AJCC Individualized Melanoma Patient Outcome Prediction Tool to attain predicted 5-year survival rates for each patient. Patients also received the DecisionDx-Melanoma GEP test, which provides a binary prediction of risk (Class 1 or Class 2) for melanoma recurrence, with refinement into subclasses (1A, 1B, 2A and 2B) that reflect proximity of risk to the crossover point between Class 1 and Class 2.

The cohort included 109 Stage I and 96 Stage II cases. To compare risk prediction from the AJCC clinical staging system to the binary GEP prediction of risk, five-year survival rates of 79 percent and 68 percent were used as cutoff scores to separate patients into low risk and high risk groups. These cutoff scores reflect the 5-year survival rates for Stage IIA or IIB patients, respectively, and are used to identify cohorts of patients who can receive significantly different surveillance and therapeutic opportunities based upon national treatment guidelines and center-specific management strategies for Stage II patients.

In comparing the GEP test with the AJCC clinical staging system, researchers found that the GEP test accurately identified early stage, high risk disease at both cutoff scores, when analyzed for regional or distant recurrence, distant metastasis, and death (from all causes).

Multivariate Cox regression analysis to compare the binary GEP classification and AJCC predictions of risk (n=205)

    Variable   HR   95% CI   p-value
Regional or distant recurrence   GEP   5.9   2.9-11.9   <0.0001
  AJCC 79%   3.6   2.0-6.6   <0.0001
 
GEP 8.9 4.6-17.1 <0.0001
  AJCC 68%   2.3   1.4-3.8   0.002
Distant metastasis GEP 5.3 2.3-12.2 <0.0001
AJCC 79% 3.0 1.5-6.1 0.002
 
GEP 7.2 3.4-15.5 <0.0001
  AJCC 68%   2.7   1.5-4.9   0.001
Death from all causes GEP 5.3 2.4-11.4 <0.0001
AJCC 79% 2.2 1.2-4.3 0.02
 
GEP 7.1 3.5-14.3 <0.0001
  AJCC 68%   1.5   0.8-2.9   0.2
HR = Hazard Ratio
 

Further analysis was conducted to compare outcomes from the GEP test and AJCC tool when risk prediction was not concordant. One in five results (21%) were not concordant. As seen in the table below, 13 patients with documented evidence of regional or distant metastasis were predicted as low risk (survival risk above 79%) by the AJCC clinical staging system but high risk (Class 2) by DecisionDx-Melanoma. Eleven had Stage I or IIA disease at diagnosis. For those patients, national management guidelines recommend low intensity surveillance. However this approach would not allow these patients to benefit from early detection of metastatic disease that can be achieved by appropriate high-intensity surveillance in high-risk patients.

Substage analysis of cases with discordant prediction of risk with GEP test and AJCC* clinical staging system

Class 2/AJCC low risk (n = 30)

  

Class 1/AJCC high risk (n = 13)

    All       All   1A**   1B***
Regional or distant met   13 (43%) Regional or distant met   5 (38%)   1 (8%)   4 (31%)
Stage IA   2 Stage IA   0   0   0
Stage IB   2 Stage IB   0   0   0
Stage IIA   7 Stage IIA   1   0   1
Stage IIB   2 Stage IIB   4   1   3
Stage IIC   0 Stage IIC   0   0   0
Cases without event   17 Cases without event   8   5   3
Distant met   9 (30%) Distant met   3 (23%)   0 (0%)   3 (23%)
Stage IA   1 Stage IA   0   0   0
Stage IB   2 Stage IB   0   0   0
Stage IIA   5 Stage IIA   0   0   0
Stage IIB   1 Stage IIB   3   0   3
Stage IIC   0 Stage IIC   0   0   0
Cases without event   21 Cases without event   10   6   4
Death (any cause)   11 (37%) Death   2 (15%)   1 (8%)   1 (8%)
Stage IA   1 Stage IA   0   0   0
Stage IB   2 Stage IB   0   0   0
Stage IIA   7 Stage IIA   0   0   0
Stage IIB   1 Stage IIB   2   1   1
Stage IIC   0 Stage IIC   0   0   0
Cases without event   19   Cases without event   11   5   6
*AJCC risk based on 79% cutoff
**1A – Class 1A represents a Class 1 low risk test result that falls within the normal confidence range
***1B – Class 1B represents Class 1 test result that is near the crossover point between Class 1 and Class 2.
 

The study was designed to compare independent prediction of risk by the GEP test or the AJCC Individualized Melanoma Patient Outcome Prediction Tool, and the potential utility of combining the results of the 2 prognosticators. The researchers identified the GEP test as a strong independent predictor of risk, and combining results from the two tools enhanced sensitivity, reflecting a more accurate identification of patients with high-risk Stage I and II melanoma.

Derek Maetzold, President and CEO of Castle Biosciences commented, “Traditionally, the patients in this study whose tumor was determined to be more aggressive than AJCC staging had predicted would have received less intensive surveillance plans than appropriate for their individual risk of recurrence. Importantly they may have missed the opportunity for enhanced surveillance to identify recurrence of disease earlier. Incorporation of the DecisionDx-Melanoma test with current staging methods may offer the ability to identify the majority of patients at risk for metastasis and death who could potentially benefit from early therapeutic intervention due to enhanced surveillance efforts.”

About DecisionDx-Melanoma

The DecisionDx-Melanoma test uses tumor biology to provide a prediction of individual risk of melanoma recurrence beyond traditional factors. Using tissue from the primary melanoma, the test measures the expression of 31 genes. The test has been validated in three multi-center studies demonstrating consistent results in over 800 patients. Performance has been confirmed in two independent, prospective studies of over 400 patients. The consistent, high rate of accuracy demonstrated in these studies provides confidence in disease management plans that incorporate DecisionDx-Melanoma test results. Clinical impact has been demonstrated in a multi-center study showing that test results add additional patient-specific prognostic information to complement traditional staging tools. More information about the test and disease can be found at www.SkinMelanoma.com.

About Castle Biosciences

Castle Biosciences is a molecular diagnostics company dedicated to helping patients and their physicians make the best possible treatment and follow-up care decisions based on the individual molecular signature of their tumor. The Company currently offers tests for patients with cutaneous melanoma (DecisionDx®-Melanoma; www.SkinMelanoma.com) and uveal melanoma (DecisionDx®-UM and DecisionDx®-PRAME; www.MyUvealMelanoma.com), with development programs in other underserved cancers. Castle Biosciences is based in Friendswood, TX (Houston), and has laboratory operations in Phoenix, AZ. More information can be found at www.CastleBiosciences.com.

DecisionDx-Melanoma, DecisionDx-UM and DecisionDx-PRAME are the trademarks of Castle Biosciences, Inc. Any other trademarks are the property of their respective owners.