Carey K Anders, MD Medical Director of the Duke Center for Brain and Spine Metastases Duke Cancer Institute Durham, North Carolina

My practice has been to offer CDK4/6 inhibition in addition to endocrine therapy to patients with first-line, hormone receptor-positive breast cancer broadly, with the exception of frail patients or those with other significant, life-limiting comorbidities. This decision is based on the significant and consistent improvement in progression-free survival (PFS) across studies (PALOMA, MONARCH, MONALEESA) evaluating all 3 clinically available CDK4/6 inhibitors. In the de novo setting and after adjuvant therapy (at least 1 year since discontinuation of therapy), I tend to offer an aromatase inhibitor (AI) with CDK4/6 inhibition. If a patient relapses during adjuvant endocrine therapy, I will transition endocrine therapy from an AI to fulvestrant in combination with CDK4/6 inhibition as I am concerned about the emergence of ESR1 mutations. In this setting, I will also perform a circulating tumor DNA assay to help support this decision and to determine post-CDK4/6-inhibition options (ie, presence or absence of PIK3CA or BRCA mutations).

	Aditya Bardia, MD, MPH Director, Precision Medicine Center for Breast Cancer Attending Physician Massachusetts General Hospital Cancer Center Harvard Medical School Boston, Massachusetts

Given the improvement in overall survival (OS) with CDK4/6 inhibitors, my preference is to utilize endocrine therapy with a CDK4/6 inhibitor for all my patients unless there is a contraindication or the patient has bone metastases only and does not want frequent monitoring (refer to Figure 2 in Spring et al. Lancet 2020). The disease recurrence with adjuvant endocrine therapy (versus no adjuvant endocrine therapy) influences the choice of endocrine partner but not the choice of CDK4/6 inhibitor. For example, if a patient has disease recurrence on adjuvant AI therapy, I would consider fulvestrant with a CDK4/6 inhibitor, (usually ribociclib based on the MONALEESA-3 trial results) as first-line therapy.

	Erika Hamilton, MD Director Breast and Gynecologic Research Program Sarah Cannon Research Institute Nashville, Tennessee

For patients with de novo disease, I would treat with an AI and a CDK4/6 inhibitor. To me the efficacy of all 3 CDK4/6 inhibitors is equal, so I think you can take your pick. For premenopausal patients, I would use an AI and a CDK4/6 inhibitor with ovarian suppression. For patients who relapse while on adjuvant AIs or close to stopping them, I would use fulvestrant and a CDK4/6 inhibitor in the first line. For patients who completed their adjuvant endocrine therapy and relapsed much later, I would use an AI and a CDK4/6 inhibitor.

	Sara Hurvitz, MD Associate Professor of Medicine David Geffen School of Medicine at UCLA Director Breast Cancer Clinical Research Program Co-Director Santa Monica-UCLA Outpatient Oncology Practice Santa Monica, California

For patients with de novo disease after adjuvant endocrine therapy, I am compelled by data from the MONALEESA-3 and MONALEESA-7 trials, which demonstrated significant OS (Im et al. N Engl J Med 2019; Slamon et al. N Engl J Med 2020) and PFS benefits with the use of ribociclib and endocrine therapy (compared to endocrine therapy alone) in the first-line setting.

The MONALEESA-3 trial evaluated the use of fulvestrant with ribociclib or placebo in the first- or second-line setting, and MONALEESA-7 evaluated the use of ovarian suppression with an AI or tamoxifen in combination with ribociclib or placebo in premenopausal patients who had previously not received endocrine therapy for advanced disease. Approximately 20% of patients in MONALEESA-3 (Slamon et al. J Clin Oncol 2018) and 40% of patients in MONALEESA-7 (Tripathy et al. Lancet Oncol 2018) had de novo metastatic disease. Subgroup analysis in MONALEESA-7 demonstrates significant OS benefit in the subgroup of patients with de novo metastatic disease. Likewise, MONALEESA-3 demonstrated a significant OS benefit for first-line patients. To date, palbociclib (PALOMA-2) and abemaciclib (MONARCH 3) have not reported OS benefits in the first-line setting, thus my preference is to recommend ribociclib-based therapy for de novo metastatic disease.

During adjuvant endocrine therapy, both ribociclib (MONALEESA-3) and abemaciclib (MONARCH 2, Sledge et al. JAMA Oncol 2019) have been demonstrated to significantly improve OS when added to fulvestrant for patients whose disease progressed on or shortly after adjuvant endocrine therapy. To date, palbociclib has not demonstrated an OS benefit (PALOMA-3, Turner et al. N Engl J Med 2018) in this setting, although there is a suggestion that patients who demonstrated sensitivity to prior endocrine therapy may have an OS benefit with the use of palbociclib. Given that abemaciclib and ribociclib have each met the OS endpoint, either of these would be my favored approach.

	Virginia Kaklamani, MD, DSc Professor of Medicine Ruth McLean Bowman Bowers Chair in Breast Cancer Research and Treatment AB Alexander Distinguished Chair in Oncology Associate Director for Clinical Research Leader of the Breast Cancer Program UT Health San Antonio The University of Texas MD Anderson Cancer Center San Antonio, Texas

In the first-line setting in de novo metastatic disease, based on 4 clinical trials (PALOMA-2, MONALEESA-2, MONALEESA-7 and MONARCH 3), I administer an AI in combination with a CDK4/6 inhibitor. All 4 of these trials show that the combination of an AI and a CDK4/6 inhibitor are superior to a single-agent AI. If a patient develops metastatic disease while on adjuvant therapy with an AI, then based on the PALOMA-3, MONALEESA-3 and MONARCH 2 trials I would administer the combination of fulvestrant and a CDK4/6 inhibitor. If a patient has completed adjuvant therapy with an AI and develops metastatic disease within 2 years of therapy, I would still use fulvestrant and a CDK4/6 inhibitor. Finally, if a patient develops metastatic disease more than 2 years after completing adjuvant therapy with an AI, then I would re-challenge with an AI and add a CDK4/6 inhibitor. There are a few cases in which single-agent endocrine therapy may be sufficient, such as for patients with minimal disease and a long disease-free interval or for elderly patients with indolent disease. However, data pointing toward an OS benefit with the addition of a CDK4/6 inhibitor suggest that these agents should be preferred in the first- or second-line metastatic setting.

	Ruth O’Regan, MD Endowed Professor of Medicine Chief Division of Hematology/Oncology and Palliative Medicine University of Wisconsin Carbone Cancer Center Madison, Wisconsin

In the de novo setting, for almost every postmenopausal patient I would use an AI with a CDK4/6 inhibitor, given the significant improvement in PFS in the first-line trials. For premenopausal women I would recommend ovarian suppression with an AI and a CDK4/6 inhibitor, with a preference for ribociclib given the survival advantage in MONALEESA-7, although retrospective data from first-line trials indicate that all CDK4/6 inhibitors appear effective in premenopausal patients who receive ovarian suppression and an AI. There are clearly patients who may not need a CDK4/6 inhibitor in the first-line setting, but no predictive biomarker is available at this time. For an elderly, frail patient, especially if the patient has bone-only disease, I would consider single-agent fulvestrant based on the FALCON trial (Robertson et al. Lancet 2016) or fulvestrant with anastrozole based on the SWOG-S0226 study (Mehta et al. N Engl J Med 2012).

During adjuvant endocrine therapy, if the patient is on an AI I would treat with fulvestrant with a CDK4/6 inhibitor. It is not clear that one CDK4/6 inhibitor is superior to another despite different PFS results in second-line studies due to patient populations with different numbers of endocrine-sensitive and endocrine-resistant cancers. For example, the patient population in PALOMA-3 was different to patients enrolled in MONALEESA-3. Patients should undergo a biopsy of a metastatic site when possible, and sending tissue for genomic testing is appropriate. The presence of an ESR1 mutation is associated with endocrine resistance, and these patients should definitely receive a CDK4/6 inhibitor. PI3-kinase mutations are not acquired and so can be checked on the same tissue, aiding with future treatment decisions. ER-positive cancers, especially invasive lobular cancers, can acquire mutations in HER2 and can ultimately be treated with neratinib. If the patient is on tamoxifen, an AI with a CDK4/6 inhibitor and ovarian suppression would be a reasonable option if the patient is premenopausal.

After adjuvant endocrine therapy, within a year or less of stopping adjuvant therapy, I would use fulvestrant and a CDK4/6 inhibitor as above. After 1 or more years I would consider restarting an AI with a CDK4/6 inhibitor. Metastatic biopsy should be performed. An important question is how we will manage the disease in patients who have received adjuvant CDK4/6 inhibitors as part of a clinical trial.

	Joyce O’Shaughnessy, MD Celebrating Women Chair in Breast Cancer Research Baylor University Medical Center Director, Breast Cancer Research Program Texas Oncology US Oncology Dallas, Texas

In the de novo setting, I would consider letrozole with palbociclib or ribociclib unless patients have high-grade, highly proliferative disease with liver metastases, in which case I would recommend letrozole with abemaciclib. I have been using more ribociclib in the first line, given the accumulating positive survival data with ribociclib.

During adjuvant endocrine therapy, I would use fulvestrant with palbociclib or ribociclib if there is no primary resistance to initial endocrine therapy, ie, if there is no recurrence on the first 2 to 3 years of endocrine therapy. If disease recurrence occurs within 3 years of adjuvant endocrine therapy, I would recommend fulvestrant with abemaciclib.

After adjuvant endocrine therapy, if the patient has recurrent disease that is aggressive with rapidly proliferative liver metastases, I would recommend fulvestrant and abemaciclib. Otherwise I would recommend fulvestrant with palbociclib or ribociclib. If a patient received adjuvant tamoxifen or adjuvant AI therapy and recurred having been off adjuvant AI for 3 years or longer, I would recommend letrozole and abemaciclib for aggressive liver metastases or otherwise letrozole with palbociclib or ribociclib.

	Mark D Pegram, MD Susy Yuan-Huey Hung Professor of Medicine Director of the Breast Oncology Program Associate Dean Clinical Research Quality Associate Director for Clinical Research Stanford Cancer Institute Stanford University School of Medicine Stanford, California

First-line CDK4/6 inhibition in combination with chemotherapy should be considered for all patients with ER-positive metastatic breast cancer, unless there is a medical comorbidity that precludes such consideration. This practice is based on many facts. First, in a subset analysis from the PALOMA-2 trial (Finn et al. N Engl J Med 2016), the addition of CDK4/6 inhibition was statistically significantly superior to endocrine therapy alone (for the PFS endpoint) in patients with de novo metastatic disease (HR 0.67, 95% CI 0.46-0.99), disease-free interval from prior endocrine therapy ≤12 months (HR 0.50, 95% CI 0.33-0.76), nonvisceral disease (HR 0.50, 95% CI 0.36-0.70), bone-only disease (HR 0.36, 95% CI 0.22-0.59), nonmeasurable disease (HR 0.35, 95% CI 0.22-0.57), prior chemotherapy or not (HR 0.53, 95% CI 0.40-0.72 and 0.61, 95% CI 0.44-0.84, respectively), a single site of metastatic disease (HR 0.51, 95% CI 0.34-0.77), ECOG PS 1 or 2 (HR 0.53, 95% CI 0.39-0.72) and age ≥65 (HR 0.57, 95% CI 0.39-0.84).

Second, the PFS Kaplan-Meier plots for the CDK4/6 inhibitors are virtually superimposable (Shah A et al. Clin Cancer Res 2018). Third, CDK4/6 inhibition has a survival benefit in ER-positive metastatic breast cancer, which is very likely to be a class effect (Slamon et al. ESMO 2019; Sledge GW Jr et al. JAMA Oncol 2019; Hurvitz SA et al. ASCO 2019).

	Sara M Tolaney, MD, MPH Associate Director Susan F Smith Center for Women’s Cancers Director of Clinical Trials Breast Oncology Director of Breast Immunotherapy Clinical Research Senior Physician Breast Oncology Program Dana-Farber Cancer Institute Assistant Professor of Medicine Harvard Medical School Boston, Massachusetts

As first-line therapy for ER-positive, HER2-negative metastatic breast cancer, I generally administer endocrine therapy with a CDK4/6 inhibitor. For a patient with de novo ER-positive, HER2-negative metastatic disease, I generally administer an AI with a CDK4/6 inhibitor. While data suggest that fulvestrant is superior to an AI in patients with de novo metastatic disease (ie, FALCON trial), we do not yet have data to know if there are differences in outcome in a randomized fashion between up-front AI with CDK4/6 inhibitor versus fulvestrant with CDK4/6 inhibitor. As for the choice of CDK4/6 inhibitor, I often use palbociclib but will use abemaciclib for patients with extensive visceral disease or those with CNS metastases.

For a patient with disease recurrence on adjuvant AI therapy, I would administer fulvestrant with a CDK4/6 inhibitor. For a patient with recurrence 2 or more years after an adjuvant AI, I would administer an AI and a CDK4/6 inhibitor. If the disease recurs within 2 years of an adjuvant AI, I would administer fulvestrant and a CDK4/6 inhibitor.

	Carey K Anders, MD Medical Director of the Duke Center for Brain and Spine Metastases Duke Cancer Institute Durham, North Carolina

While the efficacy of the 3 clinically available CDK4/6 inhibitors, palbociclib, ribociclib and abemaciclib, is highly consistent across landmark studies (PALOMA, MONALEESA and MONARCH series, respectively), the toxicity profiles are a bit different. As a class, CDK4/6 inhibitors all result in neutropenia, which requires careful monitoring with complete blood counts every 14 days during the first 2 cycles of therapy to establish the patient’s optimal dose. Many patients will require a dose reduction, and we have found that this is not dependent on age. In fact, many of our youngest patients have required dose reductions. Fatigue and nausea are also seen across each of the agents, as well as hair thinning. Monitoring both liver function tests and QTc by serial EKGs is unique to ribociclib, while gastrointestinal (GI) toxicity, particularly diarrhea, is unique to abemaciclib. We have found in our practice, through symptomatic care such as antiemetics, antidiarrhea mediation and appropriate dose reductions, that CKD4/6 inhibitors are well tolerated. It is very rare for a patient to be completely intolerant of CDK4/6 inhibition. We have also found that if one agent is not tolerated, switching to a second agent may result in better tolerability. In these cases, we have reduced the second CDK4/6 inhibitor one dose level upon initiation of therapy and escalated as tolerated.

With regard to the choice of CDK4/6 inhibition, my practice has been to start with palbociclib and endocrine therapy based on the efficacy and toxicity profile. The one caveat is for patients with brain metastases, for whom I start with abemaciclib with endocrine therapy. I have not used ribociclib as regularly, largely because of the toxicity monitoring requirement of this medication. The efficacy of ribociclib, however, is similar to abemaciclib and palbociclib, and this agent has been associated with improved outcomes, specifically among premenopausal patients.  

	Aditya Bardia, MD, MPH Director, Precision Medicine Center for Breast Cancer Attending Physician Massachusetts General Hospital Cancer Center Harvard Medical School Boston, Massachusetts

Ribociclib and palbociclib have similar toxicities, with myelosuppression, particularly neutropenia, being the most common adverse effect. The incidence of neutropenic fever is low and likely due to cytostatic rather than cytocidal effects. Ribociclib also has additional risk of LFT increase (5% to 10%) and a small but known risk of QTc prolongation.

Diarrhea is the most common adverse effect with abemaciclib, followed by myelosuppression. Increased risk of thromboembolism has also been reported with abemaciclib. Also, a rare risk of pneumonitis has been reported with all 3 of the CDK4/6 inhibitors. The toxicity is usually manageable with drug interruption and dose reduction as needed. The rate of drug discontinuation due to toxicity is low.

	Erika Hamilton, MD Director Breast and Gynecologic Research Program Sarah Cannon Research Institute Nashville, Tennessee

For me, across the first-line trials efficacy looks almost identical with the 3 agents, palbociclib, ribociclib and abemaciclib. The progression-free survival (PFS) hazard ratios are PALOMA-2 HR 0.58 (PFS 24.8 vs. 14.5 months), MONALEESA-2 HR 0.568 (PFS 25.3 vs 16 months), and MONARCH 3 HR 0.54 (PFS 28.18 vs 14.76 months). Even the trial dedicated to premenopausal patients in MONALEESA-7 showed a HR of 0.55 (PFS 23.8 vs 13.0 months).

Palbociclib and ribociclib are the CDK4/6 inhibitors with the higher rates of neutropenia (66.5% Grade 3 or 4 in PALOMA-2 and 62.0% Grade 3 or 4 in MONALEESA-2; Finn et al. N Engl J Med 2016; Hortobagyi et al. Ann Oncol 2018). Abemaciclib in MONARCH 3 had a 38.2% Grade 3 or 4 neutropenia rate (Johnston et al. NPJ Breast Cancer 2019). Febrile neutropenia, however, is extremely rare: <1% across all the studies. On the other hand, abemaciclib has higher rates of diarrhea with over 80% of patients experiencing diarrhea in MONARCH 2 (Sledge et al. J Clin Oncol 2017) and MONARCH 3 (Johnston et al. NPJ Breast Cancer 2019), with 10% to13% of this being Grade 3 or 4. The rates with palbociclib and ribociclib are 25% to 40% all grades and 0 to 2.4% Grade 3 or 4. Finally, ribociclib also comes with required liver function test monitoring due to an 11.7% increase in all-grade AST/ALT and 8.4% Grade 3 or 4 increased LFTs in MONALEESA-2 (Hortobagyi et al. Ann Oncol 2018). All of the CDK4/6 inhibitors also have the side effect of fatigue, with 25% to 40% of patients experiencing fatigue but Grade 3 or 4 fatigue rare at <3%.

	Sara Hurvitz, MD Associate Professor of Medicine David Geffen School of Medicine at UCLA Director Breast Cancer Clinical Research Program Co-Director Santa Monica-UCLA Outpatient Oncology Practice Santa Monica, California

In general, palbociclib and ribociclib have similar toxicity profiles. Both are associated with Grade 3 or 4 neutropenia in at least two thirds of patients, mandating a 3 week on, 1 week off dosing regimen to allow blood counts to recover. Unlike with the neutropenia associated with cytotoxic chemotherapy, it is rare for patients to develop fever or infection associated with CDK4/6 inhibition. In my experience, patients feel quite well on palbociclib and ribociclib and are able to maintain a good quality of life. Other less common side effects associated with ribociclib are transaminase elevation and QTc prolongation. Thus, liver transaminases and EKGs should be followed periodically. If a patient is on medication that could affect QT or if they have a history of significant cardiac arrhythmia, I would steer away from recommending ribociclib.

Abemaciclib is associated with a lower rate (~25%) of Grade 3 and 4 neutropenia but a higher risk of diarrhea. Patients tend to have more GI side effects in general. Patients should have loperamide on hand to use at the first sign of diarrhea. In general, this is a manageable side effect that tends to improve with time. Patients can also have an elevation in their liver transaminases, thus monitoring of liver tests is warranted monthly. Another uncommon but potentially significant side effect is the development of venous thromboembolic events.

All 3 agents carry a risk of interstitial lung disease. Patients should thus be advised to notify their physician if they develop shortness of breath, dry cough or fever.

	Virginia Kaklamani, MD, DSc Professor of Medicine Ruth McLean Bowman Bowers Chair in Breast Cancer Research and Treatment AB Alexander Distinguished Chair in Oncology Associate Director for Clinical Research Leader of the Breast Cancer Program UT Health San Antonio The University of Texas MD Anderson Cancer Center San Antonio, Texas

Palbociclib and ribociclib have similar toxicity profiles. However, with abemaciclib there is more diarrhea and a bit less neutropenia. I typically use palbociclib unless there is a contraindication for it, but that is mostly based on habit since the data we have seen are very similar among the 3 agents.

The toxicity profile of palbociclib includes neutropenia in approximately 80% of patients, infections in approximately 60% of patients, fatigue, nausea, alopecia and stomatitis in 30% to 40% of patients, and diarrhea in approximately 20% to 30% of patients (mostly Grade 1). Ribociclib has a similar toxicity profile with neutropenia in 70% to 80% of patients, infections in 30% to 40% of patients, fatigue in 20% to 40% of patients and diarrhea in 20% to 30% of patients (mostly Grade 1). A small number of patients on ribociclib were found to have QTc prolongation in their EKGs, which is why EKGs are required at baseline and during treatment with ribociclib. Finally, LFT abnormalities were also reported in 10% to 20% of patients, and therefore liver enzyme measurement is required during treatment with ribociclib.

The toxicity profile of abemaciclib differs slightly, with diarrhea being more commonly seen in 80% to 90% of patients, neutropenia in approximately 50% of patients and fatigue in 40% to 50% of patients.

The neutropenia experienced with CDK4/6 inhibitors is rarely associated with neutropenic fever, and most infections seen are Grade 1. Therefore, I only dose reduce for neutropenia if a patient has neutropenic fever or a severe infection. Data have shown that patients whose dose is reduced have similar benefits from CDK4/6 inhibitors as do patients without any dose reductions. Diarrhea associated with abemaciclib is managed with common antidiarrheal agents such as loperamide.

	Ruth O’Regan, MD Endowed Professor of Medicine Chief Division of Hematology/Oncology and Palliative Medicine University of Wisconsin Carbone Cancer Center Madison, Wisconsin

Palbociclib is well tolerated and is associated with neutropenia but not febrile neutropenia. I have managed palbociclib-associated side effects with growth factors on rare occasions. Ribociclib is well tolerated and is associated with neutropenia but not febrile neutropenia. I have also managed ribociclib-associated side effects with growth factors on rare occasions. QT interval prolongation is not a huge issue with ribociclib. Abemaciclib is associated with diarrhea. I recommend antidiarrheals for my patients on abemaciclib.

I mainly use palbociclib, which is generally well tolerated. Myelosuppression, especially neutropenia, is common with palbociclib but is usually picked up incidentally as part of blood count tests because patients rarely experience febrile neutropenia or are symptomatic. I rarely use growth factors, but I delay therapy and reduce doses per the package insert. I have seen fatigue, but I rarely see rashes. Overall, palbociclib is generally very well tolerated. Ribociclib has a similar side-effect profile but requires EKG monitoring through the first few cycles. Additionally, drug-drug interactions need to be monitored. I have not used abemaciclib as much as the other 2 agents, but diarrhea is clearly an issue with abemaciclib and neutropenia is additionally seen.

	Joyce O’Shaughnessy, MD Celebrating Women Chair in Breast Cancer Research Baylor University Medical Center Director, Breast Cancer Research Program Texas Oncology US Oncology Dallas, Texas

Palbociclib and ribociclib are extremely well tolerated by patients and are optimal choices for patients with minimally symptomatic or asymptomatic disease. Small subsets of patients who receive palbociclib develop Grade 4 or prolonged Grade 3 neutropenia that is not improved with dose reduction, and I switch these patients to abemaciclib. I avoid ribociclib in patients with a history of serious cardiac disease, cardiac conduction dysfunction or those who require treatment with methadone or high-dose citalopram. A small subset of patients have severe Grade 3 diarrhea with abemaciclib, but this is almost always ameliorated with stopping therapy and then resuming at a reduced dose, with the use of loperamide after each loose stool and eating a bland diet. I try to avoid the use of ribociclib or abemaciclib in patients with preexisting hepatic dysfunction from autoimmune disease or cirrhosis.

	Mark D Pegram, MD Susy Yuan-Huey Hung Professor of Medicine Director of the Breast Oncology Program Associate Dean Clinical Research Quality Associate Director for Clinical Research Stanford Cancer Institute Stanford University School of Medicine Stanford, California

The major toxicities of CDK4/6 inhibitors are listed in the publication by Barroso-Sousa et al, Breast Care (Basel) 2016. Abemaciclib causes diarrhea (including Grade 3 and 4) and ribociclib is associated with QTc prolongation.

To “improve tolerability and manage toxicity,” palbociclib is the least toxic, and drug-related hematologic and nonhematologic toxicities are easily managed by dose modification, according to the FDA prescribing information.

	Sara M Tolaney, MD, MPH Associate Director Susan F Smith Center for Women’s Cancers Director of Clinical Trials Breast Oncology Director of Breast Immunotherapy Clinical Research Senior Physician Breast Oncology Program Dana-Farber Cancer Institute Assistant Professor of Medicine Harvard Medical School Boston, Massachusetts

Generally, palbociclib and ribociclib have several similarities — they both are administered 3 weeks on and 1 week off, they both cause Grade 3 or 4 neutropenia in about 60% of patients, and generally about 50% of patients require dose reduction. Ribociclib, however, does also cause prolongation of the QTc (requiring EKG monitoring on C1D1, C1D15, and C2D1) and can cause increase in LFTs.

Abemaciclib has more GI toxicity, with about 9% to 10% of patients having Grade 3 or 4 diarrhea. It does also lower white blood cells, but the rate of Grade 3 or 4 neutropenia is lower than with palbociclib or ribociclib, and about 20% of patients get Grade 3 or 4 neutropenia. Abemaciclib is dosed twice daily (150 mg BID) and continuously, as opposed to ribociclib and palbociclib, which have once-daily and intermittent dosing. Abemaciclib also has known CNS penetration and seems to provide even greater benefit in patients with higher-risk features, such as liver metastases or a short disease-free interval, than in patients with lower-risk features (bone-only disease). It is also the only CDK4/6 inhibitor with monotherapy activity.

All these drugs have been associated with a risk of interstitial lung disease, so that is something to be aware of. Also, there seems to be an increased risk of thromboembolic events with abemaciclib, although overall the rate is quite low. I generally monitor complete blood counts with differential and comprehensive metabolic panel blood tests every 2 weeks for the first 2 months in all patients on any CDK4/6 inhibitor. I also make sure to prescribe loperamide for patients starting abemaciclib, and instruct them on how to use it at the onset of loose stool. I tend to check cystatin C in patients on abemaciclib because it can elevate the creatinine levels, making the assessment of renal function challenging. I generally hold CDK4/6 inhibitors prior to surgery and during radiation therapy.

In choosing among the agents, I generally use palbociclib for most patients, and I use abemaciclib for patients with significant visceral metastases or CNS metastases.

	Carey K Anders, MD Medical Director of the Duke Center for Brain and Spine Metastases Duke Cancer Institute Durham, North Carolina

Therapy after CDK4/6 inhibition can be quite complex. I use next-generation sequencing (NGS), either tissue-based or via circulating tumor DNA (ctDNA), to help guide my decisions, as well as the patient’s overall health/comorbidities, overall disease burden and patient preference. If a mutation is found in PIK3CA, I have prescribed alpelisib with fulvestrant. Management of alpelisib toxicity requires intensive monitoring and input from our PharmD colleagues. We have routinely been starting antihistamines prior to alpelisib treatment to prevent rash and, in most cases, metformin to prevent significant hyperglycemia. In one case, dose reduction was also required to maintain glycemic control despite the use of 2 antidiabetic agents. Consultation with endocrinology and engaging with them early and often is recommended. From an efficacy perspective, our experience has been similar to the PFS results published in the SOLAR-1 study, around 6 to 7 months of benefit.

In the absence of PIK3CA mutation, I have routinely offered everolimus with endocrine therapy based on the BOLERO-2 data, with a steroid mouth rinse to prevent stomatitis and careful monitoring of toxicities (ie, glucose, LFTs and complete blood counts). Beyond CDK4/6 inhibition, I am also utilizing BRCA mutations to help guide the use of PARP inhibitors, with BRCA2 mutations more commonly seen in our patients with hormone receptor (HR)-positive breast cancer.

Once targeted therapy and/or endocrine therapy have been exhausted, I then move to chemotherapy, preferably capecitabine as an oral, alopecia-sparing option, before intravenous chemotherapy. I also recommend clinical trials at this time, with an emphasis on oral SERDs and the NCI-MATCH study.

	Aditya Bardia, MD, MPH Director, Precision Medicine Center for Breast Cancer Attending Physician Massachusetts General Hospital Cancer Center Harvard Medical School Boston, Massachusetts

There are a number of treatment options in the second-line setting for HR-positive metastatic breast cancer, and this is somewhat guided by the first-line regimen and therapeutic response. The potential options include fulvestrant with or without a CDK4/6 inhibitor, fulvestrant with alpelisib (for PIK3CA-mutated breast cancer), exemestane with everolimus and chemotherapy such as oral capecitabine. In addition, a number of clinical trials are asking various questions related to sequencing as well as the role of new drugs. Broadly, they can be categorized into (a) the role of continued CDK4/6 blockade, eg, the PACE and MAINTAIN trials; (b) the role of CDK4/6 inhibitor combination therapy with other drugs such as everolimus, eg, the TRINITI-1 trial and (c) novel drugs in the post-CDK4/6 inhibition setting such as oral SERDs (the EMERALD trial) and antibody-drug conjugates (TROPICS-02 and DESTINY-Breast04 trials).

In the past, tumor genotyping was a research tool, but now with the approval of alpelisib for PIK3CA-mutated metastatic breast cancer, we have level 1 evidence for the use of tumor genotyping to guide clinical decision-making in HR-positive metastatic breast cancer.

Alpelisib tends to cause on-target toxicities such as hyperglycemia. The reason for this is that insulin signaling is dependent on the PI3K axis and blocking PI3K results in interference with insulin-mediated effects, including lowering of blood sugars. Potential strategies to mitigate this side effect include the use of antidiabetes medications such as metformin or lifestyle changes such as a ketogenic diet. In my experience, early recognition and management is the key, and I often send a prescription of a glucometer along with alpelisib. Other side effects of alpelisib include diarrhea and rash.

	Erika Hamilton, MD Director Breast and Gynecologic Research Program Sarah Cannon Research Institute Nashville, Tennessee

I universally obtain a molecular profile with a broad-based panel after disease progression on an aromatase inhibitor (AI) with a CDK4/6 inhibitor. A broad-based panel can tell you not only about somatic BRCA mutations but also about PI3K mutations and ESR1 mutations. With 2 approved PARP inhibitors, I use those after CDK4/6 inhibitors for patients with ER-positive disease if needed, because unlike CDK4/6 inhibitors, they have not shown improvement in overall survival. With the approval of alpelisib, PI3K mutations are now actionable outside a clinical trial, and I administer fulvestrant with alpelisib per SOLAR-1 data to patients with a tumor mutation. Finally, ESR1 mutations predict resistance to AIs. In the absence of BRCA or PI3K mutations, we have multiple second-line options, including everolimus with exemestane, fulvestrant, and chemotherapy. A novel class of drugs is also readily available in multiple clinical trials, namely oral SERDs. Knowing about the presence of ESR1 mutations helps me to select my next line of therapy as exemestane would not be appropriate in the presence of one, and I would go to fulvestrant or an oral SERD on a clinical trial at that point.

	Sara Hurvitz, MD Associate Professor of Medicine David Geffen School of Medicine at UCLA Director Breast Cancer Clinical Research Program Co-Director Santa Monica-UCLA Outpatient Oncology Practice Santa Monica, California

When a patient experiences disease progression on a CDK4/6 inhibitor, I check for tumor PIK3CA mutation, first by liquid biopsy and if negative on tumor tissue. If there is no PIK3CA mutation, I would recommend everolimus with endocrine therapy (AI if the patient was on fulvestrant previously or fulvestrant if the patient’s disease progressed on an AI with a CDK4/6 inhibitor). If there is a PIK3CA mutation, I recommend alpelisib and endocrine therapy.

Clinically the use of alpelisib can be quite challenging. However, I recall that when we began to use everolimus, it was also quite challenging with many patients experiencing significant stomatitis. As clinicians have learned to manage the side effects of everolimus with the use of steroid mouthwash and careful patient monitoring, its use has become much easier. The optimal use of alpelisib requires intense patient education prior to the initiation of therapy and very close monitoring, especially in the first few months of therapy. If a patient has a history of noncompliance with treatment recommendations, I am less enthusiastic about trying alpelisib, as the side effects such as severe rash, diarrhea and hyperglycemia can be significant and in some cases dangerous. It is important to ensure that patients have antidiarrheal medication on hand at home. They should also start an oral antihistamine concurrently with alpelisib initiation to reduce the risk of rash. In addition to checking hemoglobin A1c and fasting plasma glucose levels at baseline, I have also begun to advise patients to initiate a low carbohydrate diet. With careful patient selection, thorough patient education and close monitoring, the side-effect profile of alpelisib is manageable and roughly reflects what was reported in SOLAR-1.

	Virginia Kaklamani, MD, DSc Professor of Medicine Ruth McLean Bowman Bowers Chair in Breast Cancer Research and Treatment AB Alexander Distinguished Chair in Oncology Associate Director for Clinical Research Leader of the Breast Cancer Program UT Health San Antonio The University of Texas MD Anderson Cancer Center San Antonio, Texas

When a patient is first diagnosed with metastatic ER-positive breast cancer, I perform testing for PIK3CA mutation and germline BRCA1/2 testing. These tests help in treatment selection. The frequency of PIK3CA mutation in ER-positive metastatic breast cancer is approximately 40%, and these mutations are seen both in primary and metastatic tissue. Testing can be performed on tissue (other than bone) or plasma and results are highly concordant. If a patient is found to have a PIK3CA mutation, then I will commonly prescribe alpelisib in combination with fulvestrant as second-line therapy. A caution with this medication is the high rate of hyperglycemia. If a patient’s tumor does not harbor a mutation in PIK3CA, then second-line therapies will include single-agent fulvestrant, everolimus and exemestane combination or other antiestrogens such as tamoxifen. The efficacy of alpelisib as seen in the SOLAR-1 trial is 26.6% overall response rate with an 11-month PFS in the cohort of patients with PIK3CA-mutated disease. In my clinical experience, the efficacy has been similar to that of clinical trial data. The major side effect of alpelisib is hyperglycemia, which is an on-target effect. Hyperglycemia is seen in 64% of patients and Grade 3 or 4 hyperglycemia seen in about 40% of patients. It is important to remember that patients on the SOLAR-1 trial had to have a baseline hemoglobin A1c level of ≤6.4%. Oral antidiabetic agents such as metformin are used to manage hyperglycemia. Insulin is not helpful in treating hyperglycemia. Patients must be counseled extensively on monitoring their blood sugars and maintaining a diet low in sugars.

	Ruth O’Regan, MD Endowed Professor of Medicine Chief Division of Hematology/Oncology and Palliative Medicine University of Wisconsin Carbone Cancer Center Madison, Wisconsin

Outside of a clinical trial, I have rarely continued a CDK4/6 inhibitor or switched to an alternative agent. We check genomics on archived tissue or do another metastatic biopsy. If the patient has a PI3-kinase mutation, we administer fulvestrant with alpelisib. If they had experienced disease progression on fulvestrant, I would consider an alternative endocrine agent such as exemestane or tamoxifen, although data on these combinations are minimal. If the tumor does not have a PI3-kinase mutation, I would give everolimus with tamoxifen or exemestane. Fulvestrant with everolimus is also an option. There are data that support revisiting a prior endocrine agent and starting everolimus (Paplomata et al. Clin Breast Cancer 2019). Capecitabine is another reasonable option for these patients. I have used alpelisib for a few patients with some efficacy. It is a toxic agent and is associated with hyperglycemia and rashes.

	Joyce O’Shaughnessy, MD Celebrating Women Chair in Breast Cancer Research Baylor University Medical Center Director, Breast Cancer Research Program Texas Oncology US Oncology Dallas, Texas

I obtain NGS reports on initial metastatic biopsy as well as ctDNA at initial metastatic diagnosis and after disease progression on a CDK4/6 inhibitor. If the tumor has a PIK3CA mutation on NGS or ctDNA, and hemoglobin A1c is less than 6.5%, I would recommend treatment with fulvestrant and alpelisib, regardless of metastatic sites, as long as the patient does not urgently need chemotherapy because of impending organ failure. Otherwise, I would recommend exemestane with everolimus if patients do not have liver metastases at progression on a CDK4/6 inhibitor. If I cannot offer alpelisib to patients with liver metastases after progression on a CDK4/6 inhibitor because the tumor harbors PIK3CA wild type or if the patient has uncontrolled hyperglycemia, I would recommend capecitabine. I have not seen any clinical benefit with everolimus in patients with liver metastases after progression on a CDK4/6 inhibitor, so I would recommend capecitabine for patients with PIK3CA wild-type disease.

The main things I have learned about alpelisib by using it in practice:

1. Most patients’ cancer responds with improved serum tumor markers and reduced tumor burden and symptoms.
2. Hyperglycemia is very challenging to control in patients whose baseline hemoglobin A1c is above 6.4%, and these patients should not receive alpelisib unless their hemoglobin A1c level can be brought down.
3. Most normo-glycemic patients require only metformin to control hyperglycemia on alpelisib. Patients already on metformin when they begin alpelisib often need the addition of an SGLT2 inhibitor to treat hyperglycemia.
4. Twice-a-day nondrowsy antihistamine use is effective in preventing treatment-limiting rash.
5. Most patients lose weight on alpelisib, sometimes a concerning amount of weight.
6. Most of my patients have required dose reduction to 200 mg po daily within 4 to 6 weeks of initiating therapy in order to continue on therapy/manage toxicity, and none of these patients have lost their developing responses to therapy.

What I do not yet know about alpelisib:

1. Will patients with PIK3CA-mutated liver metastases benefit from alpelisib after progression on a CDK4/6 inhibitor?
2. Will some patients experience very prolonged benefit from alpelisib after progression on a CDK4/6 inhibitor? I have several patients who have been on alpelisib for more than 8 months and are continuing to benefit, but I do not know yet if their response will persist for 12 or 18+ months.

	Mark D Pegram, MD Susy Yuan-Huey Hung Professor of Medicine Director of the Breast Oncology Program Associate Dean Clinical Research Quality Associate Director for Clinical Research Stanford Cancer Institute Stanford University School of Medicine Stanford, California

For patients with a PIK3CA mutation (in the primary tumor, metastatic biopsy or ctDNA), alpelisib in combination with fulvestrant is a logical choice after disease progression on a CDK4/6 inhibitor in combination with an AI (André et al. N Engl J Med 2019). For patients whose tumors have wild-type PIK3CA gene sequence, fulvestrant alone or in combination with everolimus are considerations, the latter based on randomized Phase II data from the PrE0102 trial (Kornblum et al. J Clin Oncol 2018).

Published clinical experience on the SOLAR-1 trial with alpelisib indicates PFS of 11.0 months at a median follow-up of 20 months (95% CI, 7.5 to 14.5) in the alpelisib/fulvestrant group, as compared to 5.7 months (95% CI, 3.7 to 7.4) in the placebo/fulvestrant group (HR for progression or death, 0.65; 95% CI, 0.50 to 0.85; p < 0.001). Whether these results can be recapitulated in a prior CDK4/6 inhibitor-treated population is unknown. In a subset analysis of patients enrolled in the SOLAR-1 trial, the PFS hazard ratio for the 6% of patients with prior CDK4/6 inhibitor treatment was promising (HR 0.48) but with a wide confidence interval (95%CI 0.17-1.36) due to the low number (n = 20) in this subset .

Important adverse events requiring close attention include hyperglycemia, cutaneous rash frequently requiring intervention, fatigue, anorexia/weight loss and diarrhea.

	Sara M Tolaney, MD, MPH Associate Director Susan F Smith Center for Women’s Cancers Director of Clinical Trials Breast Oncology Director of Breast Immunotherapy Clinical Research Senior Physician Breast Oncology Program Dana-Farber Cancer Institute Assistant Professor of Medicine Harvard Medical School Boston, Massachusetts

For patients whose disease progresses on first-line endocrine therapy and a CDK4/6 inhibitor, I will often obtain a ctDNA panel to see if the patient has a tumor with a PIK3CA mutation and/or ESR1 mutation. If the patient has a PIK3CA mutation, then I generally offer endocrine therapy with alpelisib. For a patient without a PIK3CA mutation, I generally administer endocrine therapy with everolimus. For choice of endocrine therapy, I will use fulvestrant for patients who did not receive it in the first-line setting, or reuse fulvestrant if a patient has an ESR1 mutation. If a patient received first-line fulvestrant and does not have an ESR1 mutation, then I usually choose an AI.

Tolerability of alpelisib has proven to be similar to what was published from the SOLAR-1 trial, and hyperglycemia is an issue. I check the baseline hemoglobin A1c level and then repeat tests every 3 months, and I monitor fasting glucose level at baseline, weekly intervals for the first 2 weeks and then approximately every 4 weeks. I have had to start several patients on metformin (usually if the fasting glucose level is >150 mg/dL). Rash is also an issue, so I start all patients on a prophylactic antihistamine and ask them to take it daily with therapy.