Impact of the Prosigna assay on neoadjuvant treatment decision-making in patients with early-stage HR-positive/HER2-negative breast cancer: a single-center prospective observational study

Abstract

Background: The 50-gene assay Prosigna is cleared by the United States Food and Drug Administration only in postoperative hormone receptor-positive (HR-positive) breast cancer (BC) specimens. Given studies showing that Prosigna may identify chemo-sensitive tumors, this decision-impact study evaluated whether the assay could influence physicians' choices and patients' confidence in the neoadjuvant treatment plan.

Patients and methods: This single-center prospective observational study included patients with early-stage HR-positive/HER2-negative BC measuring ≥0.5 cm, any nodal status, deemed as possible candidates for neoadjuvant treatment per physician's choice. Formalin-fixed, paraffin-embedded core biopsy specimens were centrally analyzed using Prosigna. Physicians and patients were surveyed before and after Prosigna testing. The primary endpoint was the proportion of patients whose treatment differed from the pre-Prosigna recommendation. Secondary endpoints included assessing (i) reasons for treatment changes, (ii) physicians' and patients' confidence in the treatment recommendation and (iii) the association of residual cancer burden (RCB) with risk of recurrence (ROR) risk groups (RG).

Results: Fifty-four patients were enrolled between March 2019 and April 2023. Tumors were luminal A, luminal B, HER2-enriched and basal-like in 27.8%, 63.0%, 5.6% and 3.7% of cases, respectively. ROR RG was low, intermediate, and high in 18.5%, 20.4% and 61.1% of cases, respectively. A change in the treatment plan occurred in 35.2% of cases [95% confidence interval (CI) 23.1%-50.2%]. Among 14 questionnaires regarding reasons for treatment changes, ROR RG alone and both ROR RG and intrinsic subtype were reported in 35.7% and 64.3% of cases, respectively. Of treating physicians, 83.7% felt their confidence in the appropriateness of the treatment plan remained stable or improved, and 62.5% of patients reported reduced anxiety. No association between RCB and ROR RG was found.

Conclusions: Performing Prosigna on upfront core biopsies may influence the neoadjuvant treatment decision-making in early-stage HR-positive/HER2-negative BC.

Clinicaltrials: gov number, NCT03749421.

Keywords: Prosigna; breast cancer; decision support; neoadjuvant; precision medicine.

Figure 1

Distribution of the intrinsic subtypes within each risk of recurrence risk group. To determine non-random associations between categorical variables, Fisher’s exact test was used at 0.05 level of statistical significance to reject the null hypothesis (P < 0.0001). HER2, human epidermal growth factor receptor 2; RG, risk group; ROR, risk of recurrence.

Figure 2

Neoadjuvant treatment recommendations before Prosigna assay and treatment received by the patients following Prosigna assay results (n = 53). Data are presented as frequencies (%); the percentages inside each pre-assay branch refer to the relative pre-assay treatment category. Information about the pre-Prosigna treatment recommendation was unavailable for one patient. One patient who received both NACT and NET was categorized as NACT for the purpose of this study, as the pre-Prosigna treatment intention was NET alone. NACT, neoadjuvant chemotherapy; NET, neoadjuvant endocrine therapy.

Figure 3

(A-C) Changes in confidence among physicians with pre- and post-Prosigna survey answers available (n = 49). Each panel corresponds to one of the domains investigated by the survey: physicians’ confidence in their diagnosis (Panel A); physicians’ confidence in the treatment plan being optimal for the patient (Panel B); physicians’ confidence in understanding how Prosigna works (Panel C). Improvement in physicians’ confidence was considered if the surveyed physician’s answer shifted from a lower level of agreement to a higher level of agreement. Worsening in physicians’ confidence was considered if the surveyed physician’s answer shifted from a higher level of agreement to a lower level of agreement. (D-F) Changes in confidence among patients with pre- and post-Prosigna survey answers available (n = 40). Each panel corresponds to one of the domains investigated by the survey: patients’ confidence in the treatment recommendation received (Panel D); patients’ confidence in understanding how Prosigna works (Panel E); and patients’ anxiety about treatment and disease outcomes (Panel F). For question number 2 (Figure 4E), improvement in patients’ confidence was considered if the surveyed patient’s answer shifted from a lower level of agreement to a higher level of agreement. Worsening in patients’ confidence was considered if the surveyed patient’s answer shifted from a higher level of agreement to a lower level of agreement. For questions number 1 and 3 (Figures 4D and F), only the post-Prosigna survey answer was considered, as the assessment of an increase in confidence about the treatment recommendation and a decrease in patients’ anxiety about treatment and disease outcomes were already included in the post-assay questionnaire. Data are presented as frequencies (%) calculated based on the number of patients represented in each row.

Figure 4

Pathologic response in patients who received neoadjuvant chemotherapy (NACT) and residual cancer burden (RCB) was evaluable. These were stratified by ROR RG (n = 25); dark blue, ROR-low (n = 1); orange, ROR-intermediate (n = 4); light blue, ROR-high (n = 20). Percentages listed at the end of each bar refer to the entire cohort, whereas percentages indicated above each bar segment pertain to each RCB category. To determine non-random associations between categorical variables, Fisher’s exact test was used at 0.05 level of statistical significance to reject the null hypothesis (P = 0.21). pCR, pathologic complete response; RG, risk group; ROR, risk of recurrence.