21-Gene Assay and Breast Cancer Mortality in Ductal Carcinoma in Situ

Eileen Rakovitch, MD, MSc, FRCPC; Rinku Sutradhar, PhD; Sharon Nofech-Mozes, MD, FRCPC; Sumei Gu, MSc; Cindy Fong, BSc, CCRP; Wedad Hanna, MD, FRCPC; Lawrence Paszat, MD, MSc, FRCPC

Disclosures

J Natl Cancer Inst. 2021;113(5):572-579. 

In This Article

Results

Cohort Characteristics

The cohort includes 1362 individuals with pure DCIS treated with BCS; 787 received breast RT (conventional fractionation, n = 480; hypofractionation, n = 271; missing, n = 36). The median follow-up period was 16 years (range = 14–18). The median age was 56 years (interquartile range [IQR] = 49–64); 1118 (82.1%) had negative margins. The median value of the 21-gene RS of the index DCIS lesion was 15 (range = 0–86.7, IQR = 7.9–29.7); 29.6% (403 of 1362) had a high RS (>25). The median value of the 12-gene DS was 36.3 (IQR = 19.5–55). Baseline characteristics of the cohort are listed in Table 1.

After a median follow-up interval of 16 years, 236 of 1362 (17.3%) individuals in the cohort died: 36 (2.6%) from breast cancer and 200 (14.7%) from other causes. A higher proportion of women who died of breast cancer were aged 50 years or younger at diagnosis (27.8%) or had a high RS (38.9%) compared with those who died of other causes or were alive at last follow-up.

Recurrence Score and the Risk of Breast Cancer Mortality

Age at diagnosis (50 years or younger vs older than 50 years), pathological features of DCIS, treatment (BCS alone vs BCS plus RT) and RS risk group (>25 vs ≤25) were not individually associated with the risk of breast cancer mortality; however, there was a statistically significant interaction between the RS (>25 vs ≤25) and treatment (P = .02) and between RS and age at diagnosis (P = .04); therefore, subdistribution hazard ratios (HR) for the RS are presented stratified by age at diagnosis (50 years or younger, older than 50 years) and treatment. The presence of a high RS was associated with an 11-fold increased risk of breast cancer mortality (HR = 11.27, 95% CI = 3.00 to 42.33; P < .001) in women aged 50 years or younger at diagnosis treated with BCS alone; for those treated with BCS plus RT, there was no statistically significant association between high RS and breast cancer mortality (HR = 2.19, 95% CI = 0.50 to 9.58; P = .30). For women older than 50 years of age at diagnosis, the RS was not associated with breast cancer mortality irrespective of treatment by BCS alone (HR = 2.25, 95% CI = 0.64 to 7.88; P = .20) or BCS plus RT (HR = 0.44, 95% CI = 0.12 to 1.58; P = .21). Nuclear grade, tumor size, subtype, and necrosis were not associated with breast cancer death, and there were no statistically significant interactions between the RS and these features. The only factor associated with the risk of death from other causes was age at diagnosis (Table 2).

Recurrence Score and the Cumulative Risk of Breast Cancer Mortality

Overall, the 20-year risk of breast cancer mortality was low, as expected, and there was no difference in the risk of death from breast cancer by treatment, age at diagnosis, or RS. However, overall, women with a high RS who were treated with BCS alone had a statistically significantly higher incidence of death from breast cancer (6.4%, 95% CI = 3.1% to 11.3%) compared with those with a low RS (2.0%, 95% CI = 0.93% to 3.7%; P = .006) (Table 3, Figure 1, B). Additionally, for individuals who were aged 50 years or younger at diagnosis and were treated by BCS alone, a high RS was associated with an even greater risk of breast cancer death (9.4%, 95% CI = 2.3% to 22.5%) compared with 0.8% (95% CI = 0.07% to 4.1%) for those with a low RS (P = .006). For women aged older than 50 years at diagnosis treated with BCS alone, a high RS was associated with a small, absolute increased risk of breast cancer death compared with patients with a low RS, but it was not statistically significant (5.5%, 95% CI = 2.24% to 10.92% vs 2.5%, 95% CI = 1.1% to 4.82%; P = .10).

Figure 1.

Cumulative incidence of invasive local recurrence and breast cancer mortality by recurrence score and treatment. A) Cumulative incidence function for development of invasive local recurrence as a first event with competing risks of ipsilateral ductal carcinoma in situ local recurrence, contralateral breast cancer, and death. B) Breast cancer mortality with competing risk of death from other causes. BCS = breast-conserving surgery; RS = recurrence score; RT = radiotherapy.

Among patients treated with BCS plus RT, the RS was not associated with a statistically significant increased risk of breast cancer mortality: the 20-year risk was 4.0% (95% CI = 1.1% to 10.2%) vs 2% (95% CI = 0.5% to 5.3%; P = .34) in patients aged 50 years or younger at diagnosis and 1.1% (95% CI = 0.2% to 3.5%) vs 2.8% (95% CI = 1.5% to 5.0%) for those older than 50 years (P = .16). For all subgroups of patients treated with BCS plus RT, the cumulative risk of dying of other causes exceeded the risk of death from breast cancer (Table 3).

The Impact of Breast Radiation

For women with a high (>25) 21-gene RS, treatment with breast RT was associated with a 71% (HR = 0.29, 95% CI = 0.10 to 0.89; P = .03) relative reduction and a 5% absolute reduction in the 20-year risk of death from breast cancer (absolute reduction = 5.4% for individuals aged 50 years or younger at diagnosis; absolute reduction = 4.4% for individuals aged 50–75 years) (Figure 2). For women with a low RS (≤25), RT was not associated with a statistically significant difference in the risk of breast cancer mortality (HR = 1.51, 95% CI = 0.63 to 3.61; P = .35) (P interaction = .02).

Figure 2.

Hazard ratios for the risk of breast cancer mortality. Hazard ratios (95% CI) for the overall cohort and subgroups by age at diagnosis, recurrence score, nuclear grade, multifocality, and treatment. BCS = breast-conserving surgery; CI = confidence interval; HR = hazard ratio; RS = recurrence score; RT = radiotherapy.

21-gene Recurrence Score is Associated With Ipsilateral Invasive Breast Recurrence

Among 36 women who died of breast cancer, 86.1% (n = 31) had developed an intervening invasive in-breast cancer between DCIS diagnosis and death, 52.7% (n = 19) had a prior ipsilateral invasive LR, 22.2% (n = 8) had a prior contralateral invasive breast cancer, and 11.1% (n = 4) had bilateral invasive breast cancer. By comparison, 15.0% of women who died of other causes and 20.6% of those who were alive at last follow-up had developed a prior in-breast invasive cancer (P < .001).

The presence of a high RS was associated with a higher risk of invasive LR as a first event, particularly in individuals treated by BCS alone. Women with a high RS treated by BCS alone had a 1.8-fold increased risk of invasive LR compared with individuals with a low RS (HR = 1.75, 95% CI = 1.09 to 2.82); the cumulative 20-year risk of invasive LR was 21.7% (95% CI = 15.2% to 29.0%), which was statistically significantly higher than the risk of invasive LR in individuals treated with RT (12.4%, 95% CI = 8.7% to 16.8%) or in those with a low RS whether treated by BCS alone (14.3%, 95% CI = 11.0% to 18.0%) or BCS plus RT (15.5%, 95% CI = 11.8% to 19.7%; P = .03) (Figure 1, A). A high RS was associated with a small increased risk of developing DCIS LR as a first event; the cumulative 20-year risk of DCIS LR as a first event was 12.9% (95% CI = 7.9% to 19.0%) vs 7.7% (95% CI = 5.3% to 10.7%) after BCS alone and 11.2% (95% CI = 7.7% to 15.4%) vs 5.1% (95% CI = 3.2% to 7.5%) after BCS plus RT (P = .007) (Supplementary Figure 2, available online).

The 12-gene DCIS Score

Most individuals (88.1%) with a high RS had an intermediate or high 12-gene DS (>39). In contrast, only about half (56.0%) of the 632 individuals with an intermediate or high-risk DCIS score had a high RS. When stratified by RS, there was no statistically significant difference in the cumulative risks of breast cancer mortality between women with a low or intermediate/high-risk DCIS score (P = .55), suggesting that the RS provides additional prognostic information beyond the DS.

processing....