A cell-cycle gene array test demonstrated independent value for predicting metastatic progression after surgery for organ-confined renal cell carcinoma (RCC) of clear cell histology.
James A. Brown, MD, a professor of Urology and Biomedical Engineering at the University of Iowa Carver College of Medicine in Iowa City
James A. Brown, MD
A cell-cycle gene array test demonstrated independent value for predicting metastatic progression after surgery for organ-confined renal cell carcinoma (RCC) of clear cell histology, a retrospective study of patient data showed.
Univariate and multivariate analyses identified the cell-cycle progression (CCP) score as a significant predictor of metastatic progression. When the score was excluded from a model that included other significant predictors of progression, the overall performance characteristics decreased from area under the curve (AUC) 0.84 to AUC 0.78, further emphasizing the independent prognostic value of the gene test.
“The test appears useful, on the basis of the results from this proof-of-principle study,” said James A. Brown, MD, a professor of Urology and Biomedical Engineering at the University of Iowa Carver College of Medicine in Iowa City. “Clearly, the test needs to be validated in larger studies, and there are plans for that.
“This test was aimed at a group of patients that had a 40% rate of metastasis at 5 years, making it a high-risk cohort,” Brown continued. “There is potential utility in the high-risk arena, as patients who have a concerning test score might be better served by entering a clinical trial or by starting treatment with a targeted agent sooner.”
Rationale for Test Approach
Alterations in expression of CCP genes have demonstrated prognostic value in other malignancies, including prostate and breast cancer. Clear cell RCC, somewhat analogous to prostate cancer, poses a clinical conundrum in that many kidney cancers are low risk and might never threaten survival. However, some cases of clear cell RCC are extremely aggressive and make cure unlikely with current approaches to treatment, Brown noted.
A biomarker that could help distinguish aggressive from nonthreatening clear cell RCC could help focus intensive therapy on the most difficult tumors and avoid unnecessary treatment of low-risk tumors.
The 46-gene CCP test previously demonstrated the ability to identify prostate cancers likely to recur following definitive treatment. Because some tumors have genes in common, the test could have clinical applicability across several types of cancers, including clear cell RCC.
To examine the prognostic potential of the CCP test in kidney cancer, Brown and colleagues retrospectively reviewed records of patients with clear cell RCC treated with surgical resection. They excluded patients with distant metastasis or lymph node involvement at surgery and included patients with T2A-T3B cancers. Patients who had no evidence of metastasis after ≥4.5 years of follow-up served as a control group.
Parameters Used in Analyses
The 46-gene test consists of 31 cell-cycle genes and 15 housekeeping or control genes. The cell-cycle genes were normalized, and a CCP score was calculated by the test manufacturer, Myriad Genetics.
Parameters available for analysis were CCP score, patient sex, age at surgery, smoking status, tumor size and stage, Fuhrman nuclear grade (FNG), lymphovascular invasion (LVI), follow-up, and time to metastasis. The analysis included 64 patients with good-quality cell-cycle progression scores.
The study group comprised 26 patients who subsequently developed metastatic disease and 38 who remained metastasis free. Patients with metastatic disease were older (aged 61 years vs 58-59 years), had slightly larger tumors (9-10 cm vs 8 cm), were more likely to have T3 disease, were more likely to have FNG III/IV (high-grade) as opposed to I/II, and more likely to have LVI.
The patients who developed metastatic disease had consistently higher CCP scores, including mean, median, quartile 1, and quartile 3.
Univariate analysis identified three significant predictors of progression to metastasis: LVI (odds ratio [OR] 4.64,P= .0050), CCP score (OR 2.65,P= .0091), and FNG (OR 4.16,P= .0099). Multivariate logistic regression modeling that included CCP and all clinical variables revealed age (OR 1.22,P= .0045), tumor size (OR 1.34,P= .022), and CCP score (OR 3.40,P= .026) as the only statistically significant predictors of metastatic disease.
Multivariate logistic regression model using a stepwise variable selection of age, CCP score, tumor size, and LVI produced an AUC of 0.84. Excluding the CCP score lowered the AUC to 0.78.
Other analyses involving the same cohort have been planned, said Brown. An example is an ongoing study to evaluate the influence of tumor size on CCP score.