Following the demonstration of improvement in PFS with CDK4/6i and ET combination in ABC [6], the United States Food and Drug Administration (FDA) has approved all three CDK4/6i.
It is essential to identify patients who will benefit clinically from CDK4/6i treatments, evaluate their tolerability, and assess their impact on quality of life. Sensitivity and resistance mechanisms to CDK4/6i and ET combinations are highly complex. It is known that PgR-negatifve tumors in patients with ER-positive BC are more resistant to ET compared to PgR-positive tumors [9]. In recent years, very much evidence has shown that ER-positive/PgR-negative tumors exhibit more invasive clinical and pathological characteristics and have a worse prognosis compared to ER-positive/PgR-positive tumors [10]. In the 2013 St. Gallen Conference, emphasis was placed on the significant impact of PgR loss or low expression (≤ 20%) on the survival of BC patients [11].
A study evaluating the impact of PgR on metastasis and prognosis in HER2-negative BC patients showed that ER-positive/PgR-positive patients had a high incidence of bone metastasis. In contrast, ER-positive/PgR-negative patients had a higher incidence of visceral metastasis [12].
Several additional studies have demonstrated that PgR negativity is an independent risk factor for visceral metastasis [13]. Consistent with these findings, our study also revealed higher rates of visceral metastasis in patients with low PgR expression (47.0%) and non-visceral metastasis rates in patients with high PgR expression (66.3%) (p: 0.043).
A prospective study investigating the predictive role of PgR in tamoxifen response in ABC showed that high PgR levels are associated with better treatment response, prolonged time to treatment failure, and improvement in OS [14]. Similarly, in another study, PgR negativity was shown to be an independent predictive marker for tamoxifen resistance and breast cancer recurrence [15]. Roca et al. evaluated the impact of PgR and Ki-67 levels on the clinical benefit of first-line ET in ABC [16]. The cut-off value for PgR and Ki-67 was determined to be 20%. Patients with PgR > 20% demonstrated a longer median time to progression (mTTP) than those with PgR ≤ 20% (24 months vs. 12 months, P = 0.012). In the multivariate analysis, PgR was identified as a significant independent determinant of TTP (HR 2.45).
On the contrary, a meta-analysis by the Early Breast Cancer Trialists' Collaborative Group showed that tamoxifen improved survival independently of PgR status in ER-positive tumors [17].
Prat et al. found that PgR expression is prognostic in luminal A disease, with 20% being the most appropriate cut-off value [18]. In a study, the group with PgR < 20% and Ki-67 ≥ 20% was associated with a higher malignancy grade, and these patients were shown to benefit more from chemotherapy. Thus, it is believed that PgR and Ki-67 status will be beneficial in predicting prognosis and determining the most effective treatment strategy in ER-positive/HER2-negative BC [19]. In a different study, patients with ER-positive/PgR-negative tumors obtained similar poor outcomes as triple-negative tumors, and particularly in tumors with this tumor biology, chemotherapy has been shown to provide better survival benefits, especially in node-positive tumors [10].
For CDK4/6i based therapies, various parameters such as PIK3CA, ESR1, SMARCA4, PDK1, and many others have been investigated as predictive markers. However, no other biomarker besides ER expression has been identified to predict treatment for CDK4/6i [20, 21].
PALOMA-3 is a randomized study comparing the combination of palbociclib and fulvestrant with placebo and fulvestrant combination in patients with HR-positive/HER2-negative ABC who have previously not responded to prior ET [22].
Although the study demonstrated improved PFS and objective response rates with the combination of palbociclib and fulvestrant, along with improvements in quality of life and a favorable toxicity profile, no specific biomarker to predict response or benefit was identified in the final analysis [23]. A study using data from PALOMA-3 to identify biomarkers that could predict the long-term benefit of palbociclib and fulvestrant showed that the ER level had no impact on treatment duration. In patients with high PgR levels, it was shown that long-term responses occurred [24].
In a study evaluating predictive and prognostic factors in patients with HR-positive ABC receiving the combination of palbociclib and letrozole, the mPFS was found to be 12.99 months in PgR-negative tumors and 20.05 months in ER and PgR-positive tumors (p: 0.046) [25]. In a pooled analysis by the FDA (Makale 16), with CDK4/6i plus ET, mPFS was found to be 27.5 months (95% CI; 18.2–29.5) in PgR-negative patients and 29.1 months (95% CI; 26.2 to NE) in PgR-positive patients. In this study, positive ER status was considered the best biomarker for predicting the treatment benefit of CDK4/6i, and PgR was thought to have no prognostic value. In another study evaluating the clinical impact of CDK4/6i, subgroup analysis showed that in the palbociclib group, the 5-year PFS was 22.66% in PgR-positive patients and 21.07% in PgR-negative patients, demonstrating that PgR status did not affect survival [26].
The PARSIFAL study presented at ASCO 2020 demonstrated that PgR and Ki-67 levels significantly influenced the benefit of palbociclib and aromatase inhibitor (AI) therapy. Patients with low PgR and high Ki-67 levels were shown to benefit less from the combination of palbociclib and AI. Palleschi et al. found that PFS was inversely related to Ki-67 levels but not to PgR status in patients receiving palbociclib and ET [27].
In a retrospective analysis performed by Shao et al., a cohort treated with the combination of palbociclib and AI achieved a longer PFS in patients with PgR values ≥ 20% compared to those with < 20% (not reached vs 5.8 months; P = 0.012) [28].
In our study, the mPFS was 23.85 months (95% CI; 15.47–32.23) for patients with low PgR expression and 34.66 months (95% CI; 24.30-45.02) for patients with high PgR expression (p:0.008). In patients with Ki-67 levels ≤ 20%, mPFS was not reached, while for those with Ki-67 levels > 20%, the mPFS was 24.08 months (95% CI; 18.89–29.28) (p:0.005). Thus, we obtained results supporting studies suggesting that PgR and Ki-67 levels predict response to CDK4/6i.
Tang et al. investigated whether PgR expression affected survival outcomes in patients receiving CDK4/6i [26]. mPFS was found to be 38 months in ER-positive/PgR-positive tumors and 19.2 months in ER-positive/PgR-negative tumors (p = 0.0038). In the ribociclib group, the mPFS was 44 months in ER-positive/PgR-positive tumors and 10.1 months in ER-positive/PgR-negative tumors (p: 0.0014). In the palbociclib group, the PgR status did not affect survival, and the 5-year PFS rates were 22.66% in PgR-positive tumors and 21.07% in PgR-negative tumors.
Unlike this study, our study showed no significant difference in mPFS between patients treated with ribociclib with high PgR expression (32.07 months, 95% CI; 23.48–40.65) and those with low PgR expression (30.85 months, 95% CI; 20.71–40.99) (p: 0.488). However, in patients treated with palbociclib, the mPFS was 40.51 months (95% CI; 17.83–63.19) for those with high PgR expression, whereas it was 16.72 months (95% CI; 5.71–27.73) for those with low PgR expression (p: 0.001).
It is known that patients with recurrent breast cancer have a worse prognosis compared to those with de novo disease [29]. In the study by Tang et al., significantly prolonged mPFS was found in de novo disease compared to recurrent disease (47.1 months vs. 20.3 months, p = 0.0002). Furthermore, the rate of PgR positivity was higher in de novo disease than in recurrent disease (p: 0.0018) [26]. In the pooled analysis conducted by the FDA [30], patients receiving CDK4/6i plus AI treatment had an mPFS of 30.8 months in denovo disease and 27.1 months in recurrent disease.
Our study observed no difference between low or high PgR expression in denovo and recurrent disease. Independent of PgR expression, mPFS was significantly higher in denovo disease compared to recurrent disease (respectively, NR vs. 25.66 months (95% CI; 20.51–30.81)) (p: 0.021). Additionally, whether the disease was denovo or recurrent was an independent predictive factor for CDK4/6i treatments.
The limitations of our study include the PgR expressions of patients being recorded from pathology reports and not re-evaluated by a single pathologist, the study being single-center and retrospective, and the inability to reach mOS due to the short follow-up period.
Approaches are being investigated to identify patients who are likely to benefit from single-agent ET in the first line for ER-positive/HER2-negative advanced breast cancer and thus avoid exposure to the toxicities of CDK4/6i. However, ER positivity is currently the only established biomarker for identifying breast cancer patients who may be candidates for CDK4/6i therapy. The use of existing biomarkers and the development of new ones are crucial for identifying these patients.
Our study obtained a statistically significant difference in mPFS, with 34.66 months in high PgR expression patients and 23.85 months in low PgR expression patients given CDK4/6i (p: 0.008). Additionally, the multivariate analysis demonstrated that PgR expression is an independent predictive factor for CDK4/6i treatments.
Our study is clinically significant due to the lack of known biomarkers predicting which patients will benefit most from CDK4/6i treatments and the demonstrated impact of PgR expression levels on PFS in patients receiving these treatments. However, further research is needed to support these positive findings.