Predicting which patients with ductal carcinoma in situ (DCIS) will develop local disease recurrence: we have been trying to do that for 30 years. But now we can. The Holy Grail has been recovered in the form of Oncotype Dx, or so we have been told. A single genetic test capable of identifying the risk of local disease recurrence after breast conservation in patients with mammographically detected DCIS would obviate the need to factor in DCIS grade, size, and margin width, and patient age.
The Oncotype Dx DCIS gene signature assay was validated in a registration trial that strictly defined a permissible size of ≤25 mm for low-to-intermediate grade tumors and ≤10 mm for high-grade tumors and those with margin widths of ≥3 mm.
Unfortunately, most medical oncologists ordering the Oncotype Dx DCIS assay do not appreciate the limitations of the validated results. They order the $4000 test for patients with DCIS of all sizes and all margin widths. The test, if valid, is only valid if the strict protocol required of the Eastern Cooperative Oncology Group (ECOG) E5194 trial is adhered to. A transected DCIS is likely to recur in spite of a low score on the Oncotype Dx assay. A recent example of a 46-year-old woman with a low-risk Oncotype Dx score had 42 mm of disease and 3-mm margins. Her DCIS grade was intermediate with focal necrosis. A Van Nuys Prognostic Index calculation results in a total score of 9 (grade score of 2, size score of 3, margin score of 2, and age score of 2). Oncotype Dx provided a 10-year risk of disease recurrence of 15%, whereas the University of Southern California/Van Nuys Prognostic Index projects a local recurrence risk of 40% to 50% without irradiation. The Oncotype Dx assay is in essence a grading system based on gene expression, and is comparable to histologic grading of DCIS. Because it does not factor in a variety of important predictive factors such as age, margin width, extent of DCIS, nuclear grade, and necrosis, it is not likely to be accurate for the majority of patients. No single test, especially one in which other prognostic factors are ignored, can validly estimate the risk of local recurrence among patients with DCIS.