CDK4/6 inhibitor selection

Patient and disease characteristics

Differences in patient populations include menopausal status, endocrine therapy sensitivity, prior treatments received and degree and location of metastatic spread (Table 1).

Table 1: Patient and disease characteristics included in the landmark clinical trials for CDK4/6 inhibitors. Subgroup analysis is available for the patient and disease characteristics highlighted in green (adapted^7-13,23). CDK4/6, cyclin dependent kinase 4/6; CNS, central nervous system; ECOG, Eastern Cooperative Oncology Group; NR, not reported.

To learn more about the efficacy and safety for the overall populations assessed in the landmark clinical trials, visit our section on CDK4/6 inhibitors in metastatic breast cancer.

Postmenopausal women

All three CDK4/6 inhibitors, abemaciclib, palbociclib and ribociclib, show efficacy in endocrine-sensitive and endocrine-resistant postmenopausal women²¹

There have been single-population focused clinical trials for all three CDK4/6 inhibitors in the sensitive setting (MONARCH 3, PALOMA-2 and MONALEESA-2), whereas only ribociclib has been assessed in this manner in the resistant setting (MONALEESA-3)^7,8,12,13. For palbociclib and abemaciclib, subgroup analysis data are available for clinical trials that included both pre- and postmenopausal women in the resistant setting (MONARCH 2, PALOMA-3)^11,23,24.

Endocrine sensitive population

Abemaciclib, palbociclib and ribociclib were all assessed in a postmenopausal and sensitive population in the first-line setting in the MONARCH 3, PALOMA-2 and MONALEESA-2 clinical trials (Table 2)^7,12,13. All three trials were phase 3, randomised and placebo controlled, and the CDK4/6 inhibitors were assessed in combination with a nonsteroidal aromatase inhibitor (NSAI) versus placebo plus NSAI. In all studies, the endocrine partner assessed was letrozole; however, the MONARCH 3 study also included patients who received anastrozole (20.9% of patients included in the trial). In this setting, abemaciclib, palbociclib and ribociclib all achieved a significant improvement in median progression-free survival (mPFS) with hazard ratios (HR) of 0.54, 0.58 and 0.57, respectively^7,12,13. Overall survival with ribociclib plus letrazone was more than 12 months longer than with ribociclib plus placebo (HR for death, 0.76; 95 confidence intervals 0.63-0.93; P=0.008)²⁸. However, the overall survival data for the MONARCH-3 and PALOMA-2 trials are yet to be published. However, results of an interim analysis of overall survival in MONARCH 3 were reported at the ESMO Congress 2022 showing that the difference in overall survival between abemaciclib + aromatase inhibitor (AI) and AI alone were not statistically significant²⁶. Similarly, final overall survival results for PALOMA-2 were reported at the 2022 ASCO Annual Meeting and demonstrated no significant survival benefit of treatment²⁷. Further, in a real word analysis of palbociclib + AI, compared withAI alone, a siginificant improvement in overall survival was observed in men and postmenopausal women treated with palbociclib²⁸.

Endocrine resistant population

The MONALEESA-3 clinical trial on ribociclib is the only trial that solely assessed postmenopausal women who were treatment-naive, or who received up to one line of prior endocrine therapy in the advanced setting⁸. This trial highlighted improved median progression-free survival for ribociclib plus fulvestrant (20.5 months; n = 484), compared with placebo plus fulvestrant (12.8 months; n = 242) (HR 0.593; P<0.001) in postmenopausal women²⁶. This improvement was consistent for patients who were treatment-naive (HR 0.577), and those who previously received endocrine therapy for advanced disease (HR 0.565)⁸.

Subgroup analysis in the MONARCH 2 and PALOMA-3 clinical trials also highlighted improved mPFS in postmenopausal women with resistant disease who were treated with fulvestrant plus abemaciclib or palbociclib, respectively^11,23,24. In MONARCH 2, 83.2% (371/446) of patients in the abemaciclib arm and 80.7% (180/223) of patients in the placebo arm were postmenopausal women¹¹. Sub-group analysis based on menopausal state indicated that postmenopausal women benefited from abemaciclib with an overall survival hazard ratio of 0.773 (95% CI, 0.609–0.980)²⁴. This study also found that all assessed patients with primary endocrine resistance benefited more from abemaciclib treatment than those with secondary resistance (HR 0.686 and 0.787, respectively)²⁴.

In the PALOMA-3 trial, postmenopausal women made up 79% (275/347) of patients in the palbociclib plus fulvestrant arm and 138/174 (79%) of patients in the placebo plus fulvestrant arm²³. Median PFS was 9.9 months in the palbociclib arm, compared with 3.9 months in the placebo arm (HR 0.45; 95% CI, 0.34–0.59)²³. This was in line with the mPFS findings for all patients assessed in PALOMA-3 (HR 0.46; 95% CI, 0.36–0.59)²³.

Dr Alistair Ring, a consultant medical oncologist at The Royal Marsden NHS Foundation Trust discusses the benefit of CDK4/6 inhibitors in the endocrine therapy resistant metastatic breast cancer population.

Pre-/perimenopausal women

Ribociclib is the only CDK4/6 inhibitor investigated solely in pre-/perimenopausal women in the endocrine sensitive setting¹⁷

Of the landmark CDK4/6 clinical trials, only the MONALEESA-7 trial focused solely on the pre-/perimenopausal patient group. In the MONALEESA-7 study, ribociclib was compared to placebo, both given in combination with endocrine therapy (NSAI or tamoxifen) and goserelin in patients who had not received prior endocrine therapy in the advanced setting⁹. In this setting, the ribociclib arm demonstrated a median progression-free survival of 23.8 months compared to 13.0 months in the placebo arm (HR 0.55; 95% CI, 0.44–0.69; P<0.0001)⁹. The estimated overall survival at 42 months was 70.2% for the ribociclib arm and 46.0% for the placebo arm (HR for death 0.71; 95% CI, 0.54–0.95; P=0.00973)¹⁷.

Abemaciclib and palbociclib have not been assessed extensively in pre-/perimenopausal women. Furthermore, existing data include subgroup analysis on a relatively small number of patients from the MONARCH 2 and PALOMA-3 studies, both of which focused on the endocrine-resistant population (tested in combination with fulvestrant). In MONARCH 2, only 17% (114/669) of all patients enrolled were classed as pre-/perimenopausal¹¹. However, subgroup analysis on these patients indicated that those treated with abemaciclib and fulvestrant had a PFS hazard ratio of 0.415 (95% CI, 0.246–0.698) and overall survival hazard ratio of 0.689 (95% CI, 0.379–1.252) (note gonadotropin-releasing hormone agonist was added to both treatment arms for pre-/perimenopausal women)^11,24. In PALOMA-3, only 21% of patients in the palbociclib (72/347) and placebo (36/174) arms were pre-/perimenopausal women (goserelin was added to these treatment arms for pre-/perimenopausal women)²³. Subgroup analysis for this patient group found that those treated with palbociclib had a mPFS of 9.5 months compared to 5.6 months in the placebo arm (HR 0.50; 95% CI, 0.29–0.87)²³.

More robust data are needed for CDK4/6 inhibitors in pre-/perimenopausal women, and there are several ongoing studies that aim to assess the different settings for this patient group.

Elderly patients

Older patients with metastatic breast cancer are just as responsive to CDK4/6 inhibitors as younger patients, but they may experience greater toxicity²⁹

Adults aged 65 years or older account for the majority of breast cancer cases; however, only around 20% of patients included in clinical trials are in this age range, and most of those included are overall healthier patients within the 65–74 age range²⁹. Subgroup analyses in the landmark clinical trials for abemaciclib, palbociclib and ribociclib indicate that CDK4/6 inhibitors appear to be equally effective in older patients (≥65 years) and younger patients²⁹. In the treatment-naive older population, subgroup analysis highlighted similar mPFS hazard ratios of 0.57, 0.57 and 0.597 for abemaciclib (MONARCH 3), palbociclib (PALOMA-2) and ribociclib (MONALEESA-2), respectively^7,12,13. In the pretreated older population, hazard ratios were 0.620, 0.35 and 0.597 for abemaciclib (MONARCH 2), palbociclib (PALOMA-3) and ribociclib (MONALEESA-3), respectively^8,11,23,24.

Older patients are more likely to have decreased physiological functions and higher occurrence of comorbidities²⁹. Of the landmark trials that carried out subgroup analysis on efficacy, not all carried out subgroup analysis on toxicity in older patients (including both MONARCH 2 and 3, and the MONALEESA-3 trial)²⁹. Whilst the data are limited, the trials that did assess safety in older patients indicate that the safety profile is similar, or slightly increased, compared with younger patients. There are multiple ongoing clinical trials designed specifically to address CDK4/6 efficacy and safety in older patients²⁹.

Visceral and bone-only metastasis

All three CDK4/6 inhibitors are effective in patients with visceral and bone-only disease, with overall survival data for abemaciclib and ribociclib

In most of the landmark CDK4/6 inhibitor clinical trials for metastatic breast cancer, subgroup analyses of patients with either visceral (or lung and liver) metastasis or bone-only disease was carried out. For bone-only disease, the MONARCH 2 and 3 studies show equivalent mPFS for abemaciclib treated patients with bone-only disease (HR 0.543 and 0.58, respectively), compared with all patients in each study (HR 0.553 and 0.54, respectively)^11,12. Palbociclib also appears to be effective in combination with an NSAI for bone-only disease, with an mPFS hazard ratio of 0.41 compared to 0.62 in patients who had no bone-only disease³⁰. Ribociclib also showed improved mPFS in patients with bone-only disease in both the MONALEESA-2 (HR 0.642) and -3 trials (HR 0.379)^8,31. In the MONALEESA-7 study on pre-/perimenopausal women, mPFS hazard ratios varied more, with 0.70 for patients with bone-only disease, compared with 0.53 for those without⁹. Further, this patient subgroup did not derive overall survival benefit in the MONALEESA-7 trial, compared with patients without bone-only disease (HR 1.00 versus 0.65)¹⁷. Overall, all the CDK4/6 inhibitors show efficacy in patients with bone-only disease; however, more studies are needed to better understand if different patient groups with bone-only disease may derive better benefits from any one inhibitor over another.

Patients with visceral disease such as liver and lung metastasis have a very poor prognosis. All three CDK4/6 inhibitors found near equivalent efficacy in patients with and without visceral disease. In endocrine-resistant patients, PFS hazard ratios were 0.481, 0.47 and 0.645 for abemaciclib, palbociclib and ribociclib^8,11,23. Overall survival hazard ratio for abemaciclib in this setting was 0.675 versus 0.757 for all patients in the MONARCH 2 trial, indicating that abemaciclib may be particularly useful in this group of patients²⁴. Efficacy was also shown for ribociclib in the pre-/perimenopausal setting, and for all three CDK4/6 inhibitors in the endocrine sensitive setting^{9,12,17,30,31}.

CNS involvement

Abemaciclib can cross the blood brain barrier, and initial data indicate efficacy in patients with CNS involvement²¹

Breast cancer patients with central nervous system (CNS) involvement have a very poor prognosis³². Treatment options usually involve radiation therapy or surgery as most systemic treatments have limited penetration of the blood brain barrier. CDK4/6 inhibitors have been explored in this setting; however, more studies are needed. Indeed, most of the major phase 3 clinical trials for CDK4/6 inhibitors in metastatic breast cancer had limited inclusion of patients with CNS involvement³². This patient group was excluded from MONARCH 1, MONARCH 2, MONARCH 3, MONALEESA-2, MONALEESA-7, and the PALOMA-1 clinical trials. The remaining trials had very small numbers of patients with CNS involvement and subgroup analysis was not carried out³². This included MONALEESA-3 (8 patients total), PALOMA-2 (2 patients total from initial cohort, plus 9 patients who developed new brain lesions during the course of the study) and PALOMA-3 (5 patients total, plus 2 who developed new brain lesions)³².

Pre-clinical studies highlighted the ability of abemaciclib to cross the blood brain barrier, with the first study on human tissue indicating a comparable concentration in brain tissue and plasma³³. More recently, promising results were obtained in a phase 2 clinical trial that was specifically designed to address abemaciclib efficacy in 52 patients with CNS involvement³⁴. These patients had been heavily pretreated with a median of 4 prior systemic therapies, including chemotherapy (75%), endocrine therapy (71%), whole brain radiotherapy (50%), stereotactic radiosurgery (39%) and surgical resection of brain metastases (8%)³⁴. Treatment with abemaciclib resulted in 6% of patients achieving a confirmed cranial response, and 38% of patients showed a decrease in the sum of their intracranial target lesions. Intracranial benefit rate was 25%, mPFS was 4.4 months (95% CI, 2.6–5.5)³⁴. Further studies are needed to confirm these findings, and to better understand the utility of palbociclib and ribociclib in patients with CNS involvement.

Dr Gregory Vidal highlights that CDK4/6 inhibitor selection may be based on patient and disease features, including the aggressive nature of the tumour and CNS involvement.

Heavily pretreated populations

Abemaciclib is the only CDK4/6 inhibitor FDA approved as monotherapy for heavily pretreated HR+/HER2- metastatic breast cancer⁵

Abemaciclib is FDA approved as a monotherapy for patients with advanced or metastatic HR+/HER2-HER2− breast cancer that has progressed following endocrine therapy and prior chemotherapy in the metastatic setting⁵. This approval was influenced by the single-arm phase 2 MONARCH 1 trial that assessed the efficacy of abemaciclib in 132 patients with refractory metastatic breast cancer¹⁰. These patients were heavily pretreated with a median of 3 lines (range of 1–8) of prior systemic therapy including a median of 1 (range 1–3) line of chemotherapy and 2 (range of 1–6) lines of endocrine therapy in the metastatic setting¹⁰. An objective response rate (ORR) of 19.7% (95% CI, 13.3–27.5) was observed for this heavily pretreated group of patients, 90.2% of whom had visceral disease, and 50.8% that had three or more metastatic sites¹⁰. The clinical benefit rate (CBR) was 42.4%, the median progression-free survival (mPFS) was 6.0 months, and median overall survival was 17.7 months¹⁰. The studies which assess palbociclib and ribociclib as monotherapies are much smaller, and these agents are subsequently not approved as a monotherapy^35,36.

Palbociclib has been assessed in heavily pretreated patients. Hoste and colleagues treated 82 postmenopausal women who had at least 4 previous lines of therapy with at least one dose of palbociclib. mPFS in this group was 3.17 months and an overall clinical benefit rate was 41.5%³⁷. In another study on 118 heavily pretreated patients, median PFS was 4.5 months, and median overall survival was 15.8 months³⁸. These findings were similarSimilar findings were obtained in a small retrospective study on the efficacy of palbociclib in 24 heavily pretreated patients. which This study found that 58.3% of patients achieved stable disease, mPFS was 4.8 months, and median overall survival was 11 months³⁹. Whilst not heavily pretreated, subgroup analysis of patients in the PALOMA-3 trial found a higher hazard ratio for patients who had more lines of endocrine therapy or chemotherapy²³.

There is little data on the efficacy of ribociclib in heavily pretreated metastatic breast cancer patients. A recent analysis of real-world data on patients who received a median of 4 prior treatments (1–9) found that those treated with ribociclib had a median progression-free survival of 5 months and overall survival of 15.2 months⁴⁰.

Altogether, larger studies on heavily pretreated patients are needed to better understand the utility of each CDK4/6 inhibitor in this patient group.

Pharmacological features

There are notable differences in the pharmacokinetics and drug-drug interactions for the available CDK4/6 inhibitors.

Pharmacokinetics

Abemaciclib is more potent against CDK4 and CDK6 compared with palbociclib and ribociclib, and has additional potency against CDK9²¹

Abemaciclib, palbociclib and ribociclib also differ in their potency and affinity for CDKs. All three are orally active agents that bind to the ATP clefts of CDK4 and CDK6²¹. Palbociclib and ribociclib potencies are similar; however, palbociclib has greater affinity for CDK6, compared with ribociclib²¹. Abemaciclib is the most potent of all three inhibitors for both CDK4 and CDK6 (Table 2). Abemaciclib also has significant affinity for CDK9 and is therefore considered to be less specific than palbociclib or ribociclib⁴¹. It is thought that this affinity to CDK9 may partly explain the clinical efficacy of abemaciclib as a monotherapy, and the higher levels of gastrointestinal toxicity observed for abemaciclib, compared with ribociclib and palbociclib^10,21. Further, it may partially explain why some patients who progressed on palbociclib responded to subsequent abemaciclib treatment^42,43. Whilst this requires further investigation, the findings suggest that resistance mechanisms may at least partly differ between abemaciclib and palbociclib.

Table 2: IC₅₀ of the three CDK4/6 inhibitors for CDK4 and CDK6 (adapted²¹). CDK4/6, cyclin dependent kinase 4/6; IC₅₀, half maximal inhibitory concentration

Drug-drug interactions

Ribociclib is unique in that it can prolong the QT interval and co-administration with other QT-prolonging drugs should be avoided²¹

All three CDK4/6 inhibitors are metabolised in the liver by the cytochrome P450 3A4 (CYP3A4) enzyme^2-4. Drug interactions have been observed with CYP3A inhibitors. Concomitant use of a CDK4/6 inhibitor with a strong CYP3A inhibitor, such as clarithromycin, itraconazole, ritonavir and grapefruit juice, should therefore be avoided^22-4,21. Likewise, CYP3A inducers should be avoided to ensure the correct plasma levels of the CDK4/6 inhibitors are maintained^22-4,21.

Ribociclib is unique in that it can prolong the QT interval, and co-administration with drugs known to prolong the QT interval should be avoided³. Of note, patients on ribociclib should avoid QT-prolonging agents, including fluoroquinolones, ketoconazole, itraconazole, some antidepressants (for example, citalopram) and some antipsychotic agents (for example, haloperidol)²¹.

Results of a recent retrospective observational study that enrolled 112 patients with oestrogen receptor positive (HR+), human epidermal growth factor 2 negative (HER2−) advanced breast cancer who were candidates for palbociclib as first-line treatment, indicated that progression-free survival (PFS) can be negatively affected when patients with advanced breast cancer treated with palbociclib are concomitantly treated with proton pump inhibitors. In such patients, it may be worth exercising caution when considering prescribing proton pump inhibitors, or to consider administering proton pump inhibitors or H2-antagonists for very short periods⁴⁴.

Toxicity, safety and tolerability

Abemaciclib, palbociclib and ribociclib are generally well tolerated, with some notable differences in side effects of the three CDK4/6 inhibitors²¹

Abemaciclib, palbociclib and ribociclib are now considered standard of care for HR+/HER2− metastatic breast cancer (mBC), with overall increases in quality of life. They are all generally well tolerated; however, there are notable differences in the side effects caused by each CDK4/6 inhibitor, and in the level and type of monitoring required. These differences provide an opportunity to select a CDK4/6 inhibitor that is optimal for a patient based on the presence of comorbidities. Here, we discuss differences in adverse events and monitoring for each of the CDK4/6 inhibitors.

Some of the noteworthy side effects of any frequency that may help to differentiate the available CDK4/6 inhibitors are neutropenia, diarrhoea, prolongation of QT intervals, venous thromboembolism, hepatotoxicity and creatinine levels (Figure 1)^21,22. More recently, reports of pneumonitis for all three CDK4/6 inhibitors indicate that this may be a rare but serious adverse event; however, details on how frequently this occurs for each CDK4/6 inhibitor is currently lacking⁴⁵.

Figure 1. Common grade 3–4 adverse events reported in the pivotal trials for the three CDK4/6 inhibitors (adapted²¹).

Most of the side effects of CDK4/6 inhibitors can be managed with dose reduction or supportive care measures, which have been shown to help restrict treatment discontinuation^21,22.

In the following clip, Dr Gregory Vidal from the West Cancer Centre and Research Institute discusses how he manages breast cancer patients who experience high levels of diarrhoea and neutropenia in response to CDK4/6 treatment.

To learn more about the frequency of side effects for the overall populations assessed in the landmark clinical trials, visit our section on CDK4/6 inhibitors in metastatic breast cancer.

Neutropenias

Neutropenia is a common side effect for all three CDK4/6 inhibitors; however it is less common for abemaciclib, compared with ribociclib and palbociclib⁴⁶

CDK4/6 inhibitors can cause neutropenia, leucopenia and, less frequently, anaemia and thrombocytopenia⁴⁶. This is probably due to the central role of CDK6 in the proliferation of haematological precursors²². Notably, neutropenia appears to be roughly twice as common with palbociclib and ribociclib, compared with abemaciclib (Figure 1)^7-13,23,47. Despite high rates of neutropenia, febrile neutropenia is not common, with ≤2% of patients affected in the palbociclib and ribociclib trials, and ≤0.9% in the abemaciclib trials^7-13,23.

Unlike neutropenia caused by chemotherapy (apoptosis-mediated), neutropenia caused by CDK4/6 inhibitors arises because of cell cycle arrest and can be effectively reversed within days through dose modification⁴⁶. This was highlighted in the PALOMA-2 and PALOMA-3 clinical trials, where neutropenia accounted for ≤1.6% of discontinuations, despite 66.5% and 62.0% of patients experiencing grade 3 or 4 neutropenia, respectively^13,23. For palbociclib, dose reduction in patients with persistent neutropenia is common; however, dose modification did not appear to negatively impact on progression-free survival^2,48.

Neutropenia occurs early on, within 4 weeks for ribociclib treatment and a median onset of 16 days for palbociclib^7,48. Indeed, palbociclib and ribociclib are given over 3 weeks followed by 1 week off in order to mitigate the severity of neutropenia^2,3,22. Given the lesser effect of abemaciclib on the occurrence of grade 3 or 4 neutropenia, compared with palbociclib and ribociclib, abemaciclib can be dosed continuously^4,22. Abemaciclib may therefore be considered for patients who present with baseline neutropenia, or for patients who find it difficult or confusing to adhere to treatment cycles.

Diarrhoea

Diarrhoea is the most common side effect for abemaciclib and is less common for palbociclib and ribociclib²¹

Diarrhoea is consistently the most common side effect associated with abemaciclib, and is more common in patients treated with abemaciclib, compared with palbociclib and ribociclib (Figure 1)^10–12,21. Grade 3 or 4 diarrhoea affected 19.7%, 13.4% and 9.5% of patients in the abemaciclib arm of MONARCH 1, 2 and 3 trials, respectively^10–12. The reduction observed in MONARCH 3 was achieved through a protocol-driven directive to manage diarrhoea with prompt commencement of anti-diarrhoeal agent at first onset, and subsequent dose reductions of abemaciclib for recurrent or high-grade episodes^12,49. Abemaciclib may therefore be less suited for patients who have pre-existing gastrointestinal comorbidities²¹.

At the European Society for Medical Oncology (ESMO) Congress in 2021, the results of a phase 2 randomised trial were presented in which 53 women with HR+/HER2– inoperable and/or recurrent breast cancers were enrolled. Fifty-one patients completed the study treatments, with results indicating that bifidobacterial formulation with or without trimebutine maleate reduced the duration of abemaciclib-induced diarrhoea and prevented diarrhoea of grade 3 or higher. This led to a reduced incidence of drug suspension/reduction⁵⁰.

Prolongation of QT interval

Ribociclib can lead to prolongation of the QT interval, an adverse event that is not shared with abemaciclib or palbociclib²¹

Ribociclib is unique in that it can lead to a dose-dependent prolongation of the QT interval by Fredericia (QTcF); however, this is a rare event that is reversible with dose interruption and reduction^3,7–9. In the MONALEESA-7 study, 16% of patients receiving ribociclib with tamoxifen had a QTcF interval increase by >60 ms from baseline versus 7% in the placebo plus tamoxifen group⁹. However, the incidence was lower in patients receiving an NSAI (7% with ribociclib vs 0% with placebo)⁹. Ribociclib should therefore not be initiated in patients with a corrected QTcF value >450 ms, and should also be avoided in patients who have, or who are at risk of, QT prolongation³. This may include patients who have a long QT syndrome, uncontrolled or significant cardiac disease, electrolyte abnormalities and those who take medications that may cause QT prolongation²². Prior to starting ribociclib treatment, an electrocardiogram (ECG) should be carried out to assess patient suitability and repeated throughout treatment³. Electrolytes should also be regularly monitored in patients receiving ribociclib³. Of note, patients on ribociclib should avoid QT-prolonging agents, including fluoroquinolones, ketoconazole, itraconazole, some antidepressants and antipsychotic agents²¹. They should also be made aware of this side effect and the need to report it if it occurs⁴⁶. Assessing concurrent medication that may prolong QTc prior to commencing therapy is also an important consideration.

Neither abemaciclib nor palbociclib affect QTcF intervals or require regular ECG monitoring and should be considered for patients who are not suited to ribociclib treatment. Additionally, the need for regular ECG assessments may not be well suited to some patients who may be given ribociclib treatment, such as those who lack mobility or who live far away from their healthcare providers.

Dr Gregory Vidal talks about the CDK4/6 inhibitors that he uses for metastatic breast cancer patients at his clinic, and how he chooses the right one for his patient.

Venous thromboembolism

Venous thromboembolism was reported consistently for abemaciclib, less consistently for palbociclib and rarely for ribociclib⁵¹

Thromboembolism was reported in the PALOMA-1 study (5% in the palbociclib arm versus 0% in the control arm), and in the MONARCH 2 and MONARCH 3 clinical trials (4.9–5.0% in the abemaciclib arms versus 0.6–0.9% in the control arms) (Table 3)^10-12,47,51. While consistent in the clinical trials for abemaciclib, the occurrence of venous thromboembolism from palbociclib was less evident in the PALOMA-2 and PALOMA-3 trials (≤2% in the palbociclib arm)^13,23,51. Interestingly, thromboembolic events are much less common for ribociclib (<1%), which may be partly related to differences in the patient inclusion criteria⁵¹. It is therefore recommended that patients who are given abemaciclib should be monitored for signs and symptoms of venous thrombosis and pulmonary embolism, and treated as medically appropriate^4,5. Further studies are needed to better understand the mechanisms that lead to increased venous thromboembolism risk, and comparative studies would help to better understand the differences in risk between the inhibitors.

For future consideration, all three CDK4/6 inhibitors are currently being explored in the pre-/perimenopausal setting in combination with tamoxifen⁵¹. Patients who use tamoxifen are at a greater risk of venous thromboembolism and the impact of combination therapy with abemaciclib or palbociclib in this setting remains to be understood^51,52.

Table 3. Incidence of thromboembolic events in major trials for CDK4/6 inhibitors (adapted⁵¹). CDK4/6, cyclin dependent kinase 4/6

Hepatotoxicity

Ribociclib is more likely to cause hepatotoxicity compared with palbociclib and abemaciclib²¹

In the MONALEESA-2 and MONALEESA-3 trials, 5–10% of patients treated with ribociclib presented with Grade 3/4 elevations in alanine aminotransferase (ALT) and aspartate aminotransferase (AST)^7,8,21. Elevated Grade 3/4 levels of these enzymes were also observed in patients treated with abemaciclib, albeit at a lower rate of 2–7%^11,12. By contrast, abnormal hepatic function was observed in <1% of patients in the palbociclib treatment arm of the PALOMA-3 trial²³. Liver function tests (LFT) should therefore be performed prior to the start of ribociclib and abemaciclib treatment, with LFT monitoring every 2 weeks for the first 2 cycles and at the beginning of each of the subsequent 4 cycles^3,4. Patients who experience persistent or recurrent elevation may be managed with dose modification and patients who initially present with moderate or severe hepatic impairment (Child-Pugh class B and C) should start on a reduced dose for abemaciclib, palbociclib and ribociclib^2-4.

Renal function

Abemaciclib can increase creatinine levels, which is much less evident for ribociclib and palbociclib⁵³

High levels of creatinine (15–40% over baseline) were noted for 98.3% of patients who received abemaciclib, with a Grade3/4 occurrence of 1.9%⁵³. This increase usually occurs within the first month of treatment initiation and remains stable throughout treatment. Of note, abemaciclib treatment did not significantly alter other markers of kidney function including glomerular filtration rate⁵³. Following treatment, the serum creatinine levels return to normal, with no lasting effect on other markers of renal function.

CDK4/6 inhibitor cross-resistance

Clinical trials have not yet identified any predictive markers for CDK4/6 inhibitor response in patients with HR+/HER2− metastatic breast cancer; however, it is an active area of research and interest. An understanding of predictive markers for CDK4/6 inhibitor response may one day enable better patient selection for treatment. A biomarker specific to any one CDK4/6 inhibitor may additionally inform on choosing the best CDK4/6 inhibitors for a patient.

Dr Laura Spring also highlights that an understanding of the mechanisms of resistance to CDK4/6 inhibitors may one day help to guide the sequencing of treatment for metastatic breast cancer.

Possible markers of CDK4/6 inhibitor response

Recent clinical and laboratory research indicate that the molecular landscape of resistance to CDK4/6 inhibitors is diverse. Investigative markers of interest include factors involved in the cell cycle and oncogenic signal transduction pathways²². The RB1 tumour suppressor that plays a signalling role downstream of CDK4/6 is a key marker of interest (Figure 2)⁵⁴. Two separate studies identified RB1 gene alterations in 2.6% and 4.7% of patients receiving CDK4/6 treatments^55,56. Despite the low prevalence, a small study found that RB1 gene alterations were acquired in 3 patients who ultimately progressed on either ribociclib or palbociclib treatment⁵⁷. More recently, analysis of serial biopsies from 5 patients treated with ribociclib and letrozole further revealed that loss of RB and PTEN of the PI3K signalling pathway coincided with the onset of treatment resistance⁵⁸.

Figure 2. Retinoblastoma (RB) lies downstream of CDK4/6 signalling (adapted⁵⁸). AKT, protein kinase B; CDK4/6, cyclin dependent kinase 4/6; E2F, E2 transcription factor; ER, oestrogen receptor; mTOR, mammalian target of rapamycin; PI3K, phosphoinositide 3-kinases.

Cell cycle regulation factors other than RB1 are also being explored as resistance markers. These include alterations in the FAT1 gene and the overexpression of specific micro-RNAs that regulate the levels of CDK6, overexpression of which promotes in vitro resistance to CDK4/6 inhibitors^55,59,60. Further, amplification of cyclin E1 was associated with shorter time to progression in the PALOMA-3 study on palbociclib, whilst aurora kinase A (AURKA) amplification was associated with CDK4/6 inhibitor resistance in tumour biopsies^61,62.

Other than cell cycle regulators, factors such as PTEN, AKT and PDK1 of the PI3K oncogenic signalling pathway have also been implicated in CDK4/6 inhibitor resistance^58,62,63. Further, analysis of patient samples and translational models highlight an involvement of the ERBB2 (HER2) and FGFR signalling pathways^64–66. Altogether, multiple cellular factors may impact on CDK4/6 inhibitor resistance, with no clear dominant alteration.

Cross-resistance

Dr Laura Spring of the Massachusetts General Hospital Cancer Centre shares her analysis of a retrospective study by Wander and colleagues which looked at the utility of abemaciclib following progression on palbociclib⁴³.

Currently, there is not enough evidence to separate the three CDK4/6 inhibitors in terms of unique biomarkers for CDK4/6 inhibitor resistance, or treatment response. While only suggestive, an investigational, retrospective study on the sequential use of CDK4/6 inhibitors found that some patients who progressed on palbociclib responded to subsequent abemaciclib treatment⁴². In this small study, the median progression-free survival (PFS) was 7.0 months (range of 1.8–12.1 months) for patients treated with abemaciclib who progressed following prior palbociclib treatment⁴². This finding was mirrored in a second study by Wander and colleagues, which found that patients treated with abemaciclib following progression on ribociclib or palbociclib had a median PFS of 5.8 months (95% CI, 3.4–8.0)⁴³. Although the mechanisms of resistance to palbociclib or ribociclib were not determined in these studies, they indicate that abemaciclib may be able to overcome at least one mechanism of resistance to palbociclib and ribociclib. However, it is important to note that for both studies, patient numbers were small (n = 23 and n = 58, respectively), and there was wide variation in the type of patient population included^42,43. For instance, patients differed in number of prior treatments, intervening treatments and endocrine partners. Whilst these observations are promising, particularly since heavily pretreated patients who previously received multiple lines of chemotherapy were included, larger prospective studies are needed, and it is important to note that this sequential use of CDK4/6 inhibitor is not yet approved or recommended.

In general, most of the identified alterations that may contribute to CDK4/6 inhibitor resistance have demonstrated cross-resistance between the three CDK4/6 inhibitors in a laboratory setting. Further studies are therefore needed to identify all the alterations relevant to resistance, and whether any of these alterations are specific to any one CDK4/6 inhibitor.