Australian researchers have discovered that circulating antibodies against tumor antigens could serve as accurate biomarkers for the early detection of melanoma. The findings, presented at the European Academy of Dermatology and Venereology Congress 2024, suggest that a test incorporating these antibodies could lead to earlier and more accurate diagnoses, which is essential for determining the most appropriate treatment, especially with the rise of immunotherapy for melanoma.
Identifying Key Tumor Antigens
Cristina Vico-Alonso, MD, lead researcher from the Victorian Melanoma Service, explained that melanoma's high mutation rate leads to the production of tumor antigens, specifically cancer-testis antigens (CTAgs), which trigger an immune response. The research team utilized a novel cancer-specific array to determine if these circulating antibodies could reliably aid in early melanoma detection. The array, developed in collaboration with the Olivia Newton John Cancer Research Institute, can detect and quantify anti-CTAg IgG antibodies against over 100 tumor antigens.
Validation and Accuracy
The array was tested on plasma samples from 199 patients with stage I and stage II melanoma at diagnosis and within 30 days after curative-intent surgery. These results were compared with those from 38 healthy individuals without melanoma. The analysis identified three melanoma-specific tumor antigens showing promise as diagnostic biomarkers for early-stage melanomas. The area under the curve (AUC) values for these antigens ranged from 0.857 to 0.981 in the discovery cohort and from 0.824 to 0.985 in the internal validation cohort.
For the biomarker with an AUC value of 0.981, the corresponding sensitivity for melanoma diagnosis was 98% in the discovery cohort and 99% in the validation cohort. Specificity was 82% in both cohorts. According to Vico-Alonso, these results indicate that 99% of melanoma patients in the validation cohort tested positive for this marker, while 82% of healthy individuals were correctly identified as negative using the recommended threshold.
Improving Accuracy with Multiparameter Signatures
While 18% of healthy individuals were incorrectly identified as positive for this marker, combining it with the other two markers into a multiparameter signature helps improve accuracy. Based on the validation data, only one percent of melanoma patients would have a negative test for this top marker, suggesting that a negative result is highly indicative of the absence of melanoma.
Data from an external validation cohort revealed a signature of seven antigens with an AUC of 0.903, a sensitivity of 83%, and specificity of 84%. Vico-Alonso noted that while this isn’t as high as in the previous cohort, it is still considered a good AUC for a diagnostic test. The team is working on identifying the main confounding factors that might have contributed to these differences with the internal validation cohort.
The Need for Early Detection
Early melanoma detection is crucial because earlier diagnosis significantly improves patient outcomes, either by facilitating timely surgical intervention or by initiating systemic therapy when necessary. However, melanoma diagnosis remains challenging due to difficulties in distinguishing between benign and malignant skin lesions and the complexities of histopathological diagnosis.
Vico-Alonso emphasized that additional diagnostic tools offering insights into the biological significance of skin lesions would be helpful and could reduce the number of unnecessary procedures like biopsies, particularly for high-risk patients.
Future Directions
The researchers plan to investigate the prognostic value of CTAgs in melanoma and identify candidates for a pan-cancer diagnostic test, initially focusing on melanoma, lung, bowel, and pancreatic cancers. Vico-Alonso highlighted the tumor-agnostic nature of the cancer array, noting that CTAgs are expressed in many solid tumors, making the diagnostic signature applicable beyond melanoma. However, the specific cognate antigen combination remains unique to melanoma compared to other solid tumors.
In addition, the team is collaborating with Professor Zongyuan Ge from Monash University on a multimodal AI-based approach for early detection of melanoma. This approach aims to reduce diagnostic variability and overdiagnosis by integrating features from macroscopic and dermoscopic images, associated histopathology, antibody profiling, and proteomic data through a multi-modal analysis.
