QHow can the Decision-Dx-SCC test contribute to understanding and managing local recurrence risk in CSCC, particularly for patients with traditionally ambiguous risk factors?

The DecisionDx-SCC test has already been shown to predict metastatic risk for cSCC patients with one or more high risk factors by quantifying expression of 40 genes from the primary tumor with RT-PCR, applying a validated neural network algorithm, and classifying their tumors as at low, higher or highest biological risk of metastasis.1,2  To complicate matters, even before they develop metastases, some patients experience local recurrence (LR), even after excision of their tumors with clear margins. These patients may require more invasive surgeries, radiation therapy, or other adjuvant treatment to minimize their subsequent risk of metastasis and prolong their survival. It would therefore be advantageous to have a tool to predict LR risk to identify at risk patients early and manage them more aggressively.

Recent analysis of the DecisionDx-SCC data has focused on predicting LR risk by developing and validating an algorithm to correlate genetic testing results with outcomes. To accomplish this, a cohort of 1094 patients with primary, high-risk cSCC with available 40-GEP test results and pathologically confirmed negative margins after definitive surgery were randomized into a training cohort of 367 patients for algorithm development and two separate cohorts of 360 and 367 patients for validation. The algorithm was trained on patients with a history of LR less than 1 year after diagnosis. One-year LR-free survival was analyzed and serial algorithms evaluated, which included the 40-GEP test and clinical and pathologic factors available at the time of biopsy and diagnosis. A cut-point for low vs. high LR risk was set based on highest model performance and confirmed with the second validation set.

The best LR predictive model included genes within the 40-GEP test along with immune status, tumor diameter, degree of histologic differentiation, and presence of any perineural invasion (together termed LR-GEP). 8.5% of patients (93/1094) experienced LR. Over half of these LR (54%) were concurrent with or followed by regional/distant metastasis, while only 7% of patients without LR experienced metastasis. LRs followed by metastasis occurred sooner after diagnosis (median 0.5 years) than LRs not followed by metastasis (median 1.2 years; P<0.0001). In the combined validation cohorts (727 patients), the LR-GEP significantly stratified one-year LR-free survival between patients at low LR risk (98,8%; 95%CI, 97.7-99.8%, n=411) and those at high LR risk (89,5%; 95%CI, 86.2-93.0%; P<0.0001, n=316).

These results indicate that DecisionDx-SCC testing has the potential to classify both LR risk and non-local metastasis risk simultaneously via a single high-risk cSCC biopsy specimen. This information would be particularly helpful for dermatologists and Mohs surgeons, who can use their patient’s predicted level of risk to help guide the width and depth of excisional margins as well as the frequency of follow up. GEP testing could therefore play an instrumental role in guiding short-term procedural planning and patient counseling for patients with cSCC, while also providing an assessment of their longer-term metastatic risk.

References:

1. Ibrahim S et al. Future Oncology 2021
2. Wysong et al. Dermatology & Therapy 2024