Abstract
Breast cancer (BC) is the leading cause of cancer-related mortality among women, and hormone receptor (HR)-positive subtype makes up the majority of all cases. The standard of care in HR+/HER2− metastatic BC (MBC) is endocrine therapy (ET) plus a CDK4/6 inhibitor (CDK4/6i). ESR1 mutations could impair the clinical efficacy of the ETs. Similarly, PIK3CA mutations may serve as a negative prognostic marker. Furthermore, MBC is challenging to treat despite new drug approvals. Our patient received multiple lines of ET ± CDK4/6i and chemotherapy but persistently progressed after each or stopped the treatment due to adverse events. Here we showed for the first time that an all-oral combination of elacestrant plus alpelisib was feasible, tolerable, and clinically active in an ESR1 and PIK3CA co-mutated and heavily pretreated patient. We achieved a remarkable response in the metastatic lesions with minor toxicity issues. This case highlights the importance of utilizing up-to-date therapeutic agents and reactive decision-making during personalized cancer treatment.
Keywords
Introduction
Breast cancer (BC) is the most commonly diagnosed cancer type in the United States and worldwide according to estimates by ACS 1 and GLOBOCAN, 2 respectively. Hormone receptor (HR)-positive patients constitute 75%–80% of all BC cases. The standard of care (SOC) in HR+/HER2− metastatic patients is endocrine therapy (ET) in combination with CDK4/6 inhibitors.3,4 However, targeted treatment options after progression on the SOC are quite limited, and their sequencing for the optimum response is not clearly defined.5,6
ESR1 mutations can lead to acquired resistance to ETs, but are especially prevalent after progression on prior aromatase inhibitor (AI) treatment, even more prevalent after AI plus CDK4/6 inhibitor (CDK4/6i) and in the metastatic setting (de novo: 3.9% vs up to 55% posttreatment).7–10 As these mutations could be also associated with worse survival, new treatment strategies need to be explored. The recent approval of elacestrant, an orally available selective estrogen receptor degrader (SERD), for ER+/HER2−, ESR1-mutated advanced or metastatic breast cancer (A/MBC) patients exemplifies the success of such a strategy. PIK3CA is also frequently altered and is an actionable target in BC. Although some studies discovered PIK3CA mutation as a negative prognostic marker following the SOC, PI3K-alpha specific inhibitor alpelisib together with fulvestrant was not approved in the first line but following progression on or after an endocrine-based regimen in HR+/HER2−, PIK3CA-mutated, A/MBC patients. These alterations could limit the efficacy of endocrine and targeted therapies. However, there are currently no data regarding the safety and clinical activity for combination of elacestrant and inhibitors against PI3K/AKT/mTOR pathway and/or CDK4/6 in the estrogen receptor (ER)-positive, ESR1 and PIK3CA co-mutated BC patients, although ongoing phase Ib/II ELEVATE trial (NCT05563220 11 ) reported that the evaluated combinations (such as elacestrant plus alpelisib) were consistent with the known safety profiles of tested drugs, such as alpelisib.
Here, to our knowledge, we report the first case of safety and efficacy of all-oral elacestrant and alpelisib combination in a heavily pretreated, ESR1 and PIK3CA co-mutated HR+/HER2− MBC patient. There was a partial response following this combination with minor adverse events (AEs).
Results
Case presentation
The patient presented here is a 43-year-old Caucasian female diagnosed with MBC (cT2N3M1, stage IV) in January 2018. The PET/CT revealed a 28 mm mass in the left breast as well as FDG uptake in multiple lymph nodes (LN—right supraclavicular and right lower paratracheal, precarinal, and subcarinal) and mediastinal metastasis (proven by biopsy examination—see below). The patient had no history of chronic disease or continuous drug use (only oral contraceptive for 1 year), and no known allergies. She is a light smoker (10 packs/year) but a non-drinker. Although the patient’s cousin was also newly diagnosed with BC, there were no germline BRCA mutations.
The breast Tru-Cut biopsy was ER+ (70%), PR+ (20%), and HER2− (1+) with a Ki-67 proliferation index of 15%. Although the patient was initially considered T2N3 with suspected mediastinal metastasis, paratracheal plus subcarinal mediastinal LN biopsy (through mediastinoscopy with biopsy) was compatible with BC metastasis, revealing de novo metastatic disease. Accordingly, she received neoadjuvant chemotherapy consisting of dose-dense doxorubicin/cyclophosphamide (AC—4 cycles, doxorubicin 60 mg/m2 and cyclophosphamide 600 mg/m2 every 2 weeks) followed by weekly paclitaxel (T—12 cycles, 80 mg/m2) from February to July 2018 (Figure 1). Although there was a metabolic and pathologic complete response in August 2018 by PET/CT, breast-conserving surgery (lumpectomy) showed a 15 mm mass with histological grade 3 and diffuse Ductal Carcinoma In Stu (DCIS) (no invasive component) in September 2018. Although the patient was offered radiotherapy for the breast and lung after surgery, she refused. The patient was treated with adjuvant tamoxifen (20 mg/day) plus leuprorelin acetate (every 3 months) from September 2018 to January 2021. Due to increased CA 15-3 levels, a PET/CT was requested in January 2021, revealing disease progression in the mediastinal LNs and new bone metastases. She was administered letrozole (2.5 mg/day) plus palbociclib (125 mg/day, 3 weeks on/1 week off—plus leuprorelin acetate and zoledronic acid) from January to November 2021. A partial response was observed in the mediastinal LNs and bone lesions by PET/CT in April 2021 (RECIST v1.1 12 ). However, progressive disease was observed in the bone lesions, and three new liver lesions (a 3 cm one in segment 5 plus 1.8 and 1.3 cm lesions) emerged in November 2021, all of which were treated via liver radiofrequency ablation (RFA). Later, the treatment was changed to capecitabine (1250 mg/m2 twice daily) plus docetaxel (75 mg/m2 on day 1). This regimen was stopped after two cycles due to hand–foot syndrome caused by capecitabine in January 2022, and it was switched to gemcitabine (1200 mg/m2 on day 1 and 8 every 3 weeks) plus paclitaxel (75 mg/m2 on day 1, January to July 2022, 7 cycles). The PET/CT in March 2022 showed regression in the bone lesions, and the liver foci disappeared (potentially due to the RFA). The patient experienced grade 1 neuropathy, probably caused by paclitaxel. Accordingly, the treatment was changed to single-agent gemcitabine but could not be administered regularly.

Timeline of disease status and treatment response as well as comparison of the CGP results. (a) A 43-year-old female was diagnosed with metastatic breast cancer in 2018. She received multiple lines of chemotherapy or endocrine therapy-containing regimens from 2018 to 2024. Although partial responses had been achieved with some of these treatments, the disease eventually progressed or the treatment was halted due to toxicity issues. From July to August 2023, our team followed a precision oncology approach through CGP and achieved a partial response with an all-oral combination of elacestrant plus alpelisib as well as the ketogenic diet. (b) Comparison of the biomarkers and alterations detected and reported by three different CGP tests. (c) Co-occurrence pattern of ESR1 and PIK3CA alterations in a metastatic breast cancer patient cohort (INSERM, cBioPortal).
The patient was offered a comprehensive genomic profiling (CGP) test (OncoDEEP—April 2023, liver Tru-Cut biopsy—ER 10%, PR 60%, and HER2-negative, CGP #1 in Figure 1), which detected ESR1 Y537S, PIK3CA E545K, and CCND1 amplification with a tumor mutational burden (TMB) of 14.95 muts/Mb, microsatellite stable (MSS) and homologous recombination deficiency (HRD)-positive status. The test also reported the following immunohistochemistry (IHC) findings: AR 1%, HER2 0%, CD8 10%, and negative PD-L1 status (combined positive score (CPS): 0). The pathology examination confirmed HR-positive, HER2-negative BC with extensive metastasis to the liver. The patient received ET plus alpelisib (300 mg/day, April to May 2023) at an external clinic but experienced hyperglycemia (blood sugar > 300 mg/dL) that led to treatment discontinuation. She was later treated with carboplatin plus paclitaxel for two cycles (May to July 2023).
She was later admitted to our clinic for a second opinion, and we recommended a CGP test to design a patient-tailored treatment plan. We also requested a PET/CT (July 2023) to comprehend the current status of the disease. We first utilized a liquid biopsy to account for tumor heterogeneity of the extensive metastasis (FoundationOne®Liquid CDx—July 2023, CGP #2 in Figure 1) and observed the following alterations: ESR1 Y537S, PIK3CA E545K, FGFR2-TACC2 rearrangement, MAP3K1 Q1261*, SETD2 S470*, TP53 H193R, and equivocal amplifications of CCND1, MDM2, MYC, EMSY, FGF19, FGF3, and FGF4. The tumor fraction was elevated, the blood tumor mutational burden (bTMB) was 9 muts/Mb, and microsatellite instability-high (MSI-H) was not detected. We decided to use elacestrant plus alpelisib considering ESR1 and PIK3CA mutations. While waiting for a response from the off-label drug committee for the combination and access to elacestrant, we initially started with fulvestrant plus alpelisib (150 mg/day), complemented by the ketogenic diet. The patient tolerated the treatment without any blood sugar elevation but experienced grade 1 diarrhea, nausea, and asthenia. AEs were continuously assessed during the study and for 90 days after the last treatment. The AEs were evaluated according to the National Cancer Institute Common Terminology Criteria for Adverse Events (v5.0) through weekly biochemical and hematology tests as well as biweekly patient visits. We also offered a Tempus CGP test (August 2023, CGP #3 in Figure 1) due to the presence of an FGFR2 rearrangement at the DNA level, as Tempus could assess the presence of rearrangements/fusions at both DNA and RNA levels. RNA-based analysis is currently considered the gold standard for such alterations. While some alterations were missing in this test using liver biopsy, it reported many of the same alterations: ESR1 Y537S, PIK3CA E545K, MAP3K1 Q1261*, TP53 H193R as well as copy number gains in CCND1, EMSY, FGF3, and FGF4. The tTMB was 4.2 muts/Mb with an MSS and negative PD-L1 status (TPS < 1%, CPS < 1). This test additionally reported overexpression of ERBB3 and PGR, and a HRD score of 26.1%, where the cutoff to define HRD is 21% in BC. The biomarkers and alterations reported by each test are compared in Figure 1(b). The co-occurrence pattern of ESR1 with PIK3CA alterations was shown in an MBC cohort (INSERM, cBioPortal, Figure 1(c)), potentially offering our approach (outlined below) as a promising treatment option for a great number of BC patients. As a note, similar to our case, TMB predicted by different CGP tests and/or tissue versus liquid biopsy could differ dramatically. 13 Besides, before versus after treatment, TMBs could be highly variable. 14
As our patient received multiple lines of chemotherapy and hormonal therapy including short-term fulvestrant and due to the presence of ESR1 Y537S and PIK3CA E545K mutations, we recommended and switched to elacestrant (345 mg/day) plus alpelisib (100 mg/day, increased to 150 mg/day with a ketogenic diet) combination. Since our patient had previously experienced hyperglycemia with full-dose alpelisib (which is consistent with our own clinical experience), we started with a lower dose to avoid treatment discontinuation. The treatment was well tolerated with grade 1 asthenia, highlighting that alpelisib retreatment after hyperglycemia could be managed by dose reduction and ketogenic diet (increasing tolerability and efficacy). Furthermore, we have followed this approach in combination with ET and achieved remarkable responses in other A/MBC patients. The PET/CT (October 2023 vs July 2023) revealed a partial response in the metastatic lesions and improved clinical status (Figure 2(a) vs (c) and (b) vs (d)). Unfortunately, performance status deterioration and disease progression (time-to-progression: ~5.5 months) by ultrasonography were reported at the external clinic in January 2024. We added lenvatinib to the treatment to target FGF/R alterations but the patient succumbed to death after 3 months (April 2024). The cause of death was neither tumor progression as there was a stable disease response, nor any drug-related AEs. It was reported to be pneumonitis caused by an infection (pneumonia).

A partial response was observed in the metastatic lesions upon the elacestrant plus alpelisib combination. (a) July 2023 pet shows lesions in the liver parenchyma at the level of segments 5 and 6 with prominent FDG uptake, consistent with metastasis (arrows—see (a) vs (c)). (b) July 2023 PET shows lesions in the liver parenchyma in segments 8 and 4 with prominent FDG uptake, consistent with metastasis (arrows—see (b) vs (d)). (c) Three months later, a significant decrease in the FDG uptake of the lesions was observed, and only peripheral mild FDG uptake was present (arrows). (d) PET examination performed 3 months later shows only peripheral mild FDG uptake (arrows).
Written informed consent was obtained from the patient for both treatment and publication of her anonymized data in this scientific publication. We also submitted an off-label drug use application to the Ministry of Health and received approval.
Methods
CGP tests
The first CGP test (OncoDEEP) was carried out using a liver biopsy collected on April 5, 2023. Tumor percentage was 70%. This test analyzes both DNA (638 genes) and RNA (22 genes) as well as genomic biomarkers (MSI, TMB, HRD) plus TERT promoter status. Additional tests, such as PD-L1 IHC and MGMT methylation status, could be requested. The second test (FoundationOne Liquid CDx) was performed using liquid biopsy blood collected on July 10, 2023. This test provides alteration data for over 300 genes and bTMB, MSI-H, and tumor fraction values. The tumor fraction was 69% (elevated), implying that all types of alterations including copy number variants (CNVs) could be reliably detected. The third and final test (Tempus) was done using a liver biopsy collected on August 5, 2023. The tumor percentage was 80% post microdissection. Tempus xT test using solid tumor plus a normal control is a 648-gene panel that reports clinically relevant alterations, immunotherapy biomarkers such as TMB and MSI, HLA genotyping, as well as potential germline findings. With the add-ons, Tempus could provide PD-L1 IHC, HRD, and RNA-seq for structural variants and transcriptomic analysis (present in our case).
The reporting of this study conforms to the CARE guidelines. 15
Discussion
BC, as the most diagnosed cancer, is a global health problem with a 5-year survival rate of less than 30% in distant metastatic cases. 16 The first-line treatment in the HR+/HER2− MBC without visceral metastasis and germline BRCA1/2 mutation is ET (a nonsteroidal AI or fulvestrant) plus a CDK4/6 inhibitor.3,4 However, de novo or especially acquired resistance to hormone therapy is observed in many patients, rendering current approaches ineffective. Some resistance mechanisms include acquisition of ESR1 mutations, loss of ER and high cyclin E1 expression, dysregulation in ER co-regulatory proteins, amplifications in MYC and FGFR1 as well as TP53 mutations.17–20
ESR1 mutations are frequently detected in metastatic tumors as compared to the primary tumors. 21 Analysis through liquid biopsy tools (ctDNA + circulating tumor cells (CTCs) revealed the relationship of different mutations with organ tropism and reported that ESR1 mutations were the only significant parameter associated with liver metastasis. Consistently, all CGP tests from liver biopsies or liquid biopsies in our case detected an ESR1 mutation. 22 Many of the ESR1 mutations are activating and can drive ligand-independent ER activity but may differentially affect the pathway activity or the response to endocrine therapies (Y537S vs D538G, E380Q, S463P). 23 For example, 537 mutations were usually accompanied by single nucleotide variants (SNVs) in the estrogen receptor (ER) and RAF pathways, CNVs in the MYC pathway and bone metastases, while ESR1 538 with SNVs in the cell cycle pathway and liver metastases. 24 They are mostly found after AI treatment (or AI plus CDK4/6i) and are associated with worse patient outcomes.9,25 Unlike AIs, some clinical studies showed that fulvestrant-based treatment could be efficacious in the ESR1-mutant patients similar to the ESR1 wild-type (WT) patients. 7 However, ESR1 Y537S mutation was found to be statistically enriched in the posttreatment (fulvestrant plus palbociclib) samples through the genomic landscape and clonal evolution analysis in the PALOMA-3 trial. 26 Similarly, patients harboring baseline Y537S mutation had a shorter progression-free survival (PFS—1.8 vs 3.5 months) than those WT for this residue following fulvestrant treatment (cohort A) in the plasmaMATCH phase II trial. 27 One strategy to overcome the resistance could be chemotherapy and fulvestrant combination; however, a study demonstrated that the combination was synergistic in the Y537S-mutant models but dependent on p53 activity. 28 Considering the previous failures with multiple lines of chemotherapy and the patient’s TP53 mutation, this approach was not prioritized in our case. Moreover, PIK3CA mutations or activation have been previously associated with resistance to chemotherapy in many cancer types.29–31 As the next generation of anti-estrogen therapies active against ESR1 mutations has been developed, they could form the backbone for the treatment of endocrine-refractory patients. Following the phase III EMERALD trial, elacestrant has received regulatory approval as monotherapy for postmenopausal women or adult men with ER+/HER2−, ESR1-mutated A/MBC with disease progression following at least one line of ET. 32 The duration of prior ET + CDK4/6i in metastatic setting was associated with elacestrant efficacy (⩾12 vs ⩾6 months). 33 Patients with ESR1- and PIK3CA E542X/E545X-mutated tumors and ET + CDK4/6i ⩾12 months had PFS of 5.5 versus 1.9 months (HR: 0.80) with elacestrant and the SOC, respectively. Those with concurrent ESR1 and TP53 mutations (8.61 vs 1.87 months), or ESR1 mutations with liver and/or lung metastases (7.26 vs 1.87 months) exhibited remarkably longer PFS upon elacestrant versus placebo. 33 There are also ELAINE 2 phase II trial 34 (with abemaciclib) and a case report (monotherapy) 35 showing the clinical efficacy of lasofoxifene (LAS), a nonsteroidal selective ER modulator, in a similar group of patients after disease progression on prior therapies. Lastly, early results from a randomized phase II trial (SERENA-2) demonstrated improved and promising clinical outcomes with a next-generation oral SERD, camizestrant versus fulvestrant in postmenopausal women with advanced ER+/HER2− BC with disease recurrence or progression after ⩽1 ET and a baseline ESR1 mutation, including Y537S.36,37 Since our patient did not obtain clinical benefit from the ET-containing regimens and potential resistance to fulvestrant through Y537S mutation, we switched to elacestrant to enhance therapeutic efficacy.
PIK3CA is the most commonly mutated gene in TCGA PanCancer Atlas 38 and METABRIC BC cohorts. 39 A meta-analysis in advanced HR+, HER2− BC patients not receiving PI3K-targeted drugs revealed that PIK3CA mutation was a negative prognostic factor for both progression-free and overall survival. 40 For example, PFS was shorter in patients with concomitant ESR1 and PIK3CA mutations compared to all ESR1-mutated patients (5.45 vs 8.61 months) although elacestrant still led to significantly higher PFS compared to placebo (5.45 vs 1.94 months).33,41 While ESR1- and PIK3CA E542X/E545X-mutated tumors and ET + CDK4/6i ⩾12 months exhibited PFS of 5.5 versus 1.9 months with elacestrant versus the SOC, it was 4.6 versus 3.3 months for ESR1- and PIK3CA H1047X-mutated patients. In line, LAS plus abemaciclib yielded a 73% CBR and 12.9 months PFS in the ESR1-mutant patients, whereas they were 63% and 7.8 months in the ESR1 and PIK3CA co-mutant patients. 42 Just five hotspot mutations cover approximately 75% of all PIK3CA mutations in BC, where E545K accounts for more than 15%. 43 PIK3CA variant classes show differential patterns of associations in terms of concomitant mutations, pathway activity, or clinical phenotype. 24 For example, 545 mutations were associated with the SNVs in the P53 pathway while 542 mutations were correlated with the SNVs in receptor tyrosine kinase (RTK)/RAS and CNVs in the PI3K pathway. Furthermore, PIK3CA mutations were linked to lung tropism in MBC while TP53 alterations were associated with LN localization. 22 These alterations could be currently targeted by alpelisib, capivasertib, or mTOR inhibitor, everolimus.44,45 While alpelisib with fulvestrant is indicated for postmenopausal women and men with HR+/HER2−, PIK3CA-mutated, A/MBC as following progression on or after an endocrine-based regimen, the everolimus with exemestane is for patients with metastatic HR-positive BC who progressed on nonsteroidal aromatase inhibitor therapy. 46 Prior exposure to CDK4/6 inhibitors did not lead to differential survival outcomes with this combination. More recently, pan-AKT inhibitor capivasertib in combination with fulvestrant has been approved in HR+/HER2− locally A/MBC harboring one or more AKT1, PIK3CA, or PTEN alterations following the positive results from the phase III CAPItello-291 trial. 45 However, the currently available data regarding the tolerability and efficacy of elacestrant with these agents are limited even though there are ongoing early-phase clinical trials (NCT05563220 and NCT05386108, n = 3 for elacestrant 300 mg/day plus alpelisib 250 mg/day) evaluating the combinations of elacestrant with alpelisib, capivasertib, everolimus, or CDK4/6 inhibitors. 47 One study reported that elacestrant plus alpelisib was tolerable and could impair tumor growth in letrozole and letrozole plus palbociclib-pretreated PIK3CA-mutant (ESR1 WT) patient derived xenograft (PDX) model. 48 Here our precision oncology center, for the first time, showed that elacestrant and alpelisib at doses selected through molecular tumor board discussions were well tolerated with minor toxicities such as grade 1 asthenia. In our clinical experience, alpelisib at 150 mg/day is clinically active and more tolerable than the regular dose in combination with ET in BC patients. The ketogenic diet was reported to potentially improve the tolerability and efficacy of PI3K inhibitors. Similarly, prophylactic metformin could diminish the incidence and severity of hyperglycemia. 49 Overall, we achieved a partial response in the liver and bone metastatic lesions upon combinatorial treatment (Figure 2).
Gene fusions are strong oncogenic drivers that can drive tumorigenesis, cancer progression, and therapeutic failure. However, FGFR2-TACC2 rearrangement reported in the FoundationOne Liquid CDx was later predicted to be “not in-frame” upon additional information request from the test provider. It was a result of chromosome 10 duplication fragment with breakpoints at chr10:123241583 and chr10:123788371 (5′-FGFR2 (x1-17)-3′:3′-TACC2 (x2-1)-5′). Consistently, the Tempus test from the liver biopsy did not report any FGFR2-TACC2 rearrangement although the alteration was detected at the DNA level, it was absent at the RNA level. Besides, out-of-frame fusions are often assumed to be inactivating mutations and labeled as passenger mutations.50–52 Considering that the RNA-based approach is the gold standard for the detection of structural variants, we initially did not include any FGFR inhibitor in the combination therapy. Similarly, the specimen harbors CCND1 and FGF amplifications (albeit equivocal in the F1 Liquid CDx) that can undermine therapeutic benefit. Pan-FGFR inhibitors such as pemigatinib have limited activity against these amplifications. 53 Multikinase FGFR inhibitors, unlike specific FGFR inhibitors, have also activity against other RTKs, such as VEGFR, which could be upregulated due to the presence of the TP53 mutation.54,55 For example, lenvatinib in combination with palbociclib was highly active (50% PR) in the FGF/FGFR and accompanying alterations. 56 Therefore, these alterations should be monitored carefully during treatment.
Our study shows the clinical efficacy and tolerability of the all-oral elacestrant plus alpelisib combination for the first time but it has some shortcomings. Although our patient was heavily pretreated (and prior ET + CDK4/6i ~10 months), the elacestrant plus alpelisib combination yielded a short time to progression (TTP) of approximately 5.5 months, which is quite similar to PFS in patients with ESR1- and PIK3CA E542X/E545X-mutated tumors (prior ET + CDK4/6i ⩾12 months) with elacestrant monotherapy. Furthermore, the potential effect of alterations other than ESR1, PIK3CA, and TP53 toward elacestrant efficacy was not reported in the subgroup analysis in the EMERALD trial. Therefore, we cannot clearly pinpoint potential mechanisms of adaptive and/or acquired resistance to elacestrant ± alpelisib. Finally, a single case would preclude making generalized assumptions about the clinical activity of the combination in ESR1/PIK3CA co-mutated patients.
Conclusion
Overall, HR+/HER2− MBC is a challenging clinical entity to treat, especially after the failure of the SOC and in heavily pretreated patients. Fortunately, we demonstrated for the first time that an all-oral combination of elacestrant plus alpelisib with dose modifications and a ketogenic diet was feasible and clinically active in an ESR1 and PIK3CA co-mutated patient. Accordingly, utilizing up-to-date therapeutic agents and reactive decision-making during personalized cancer treatment may be a promising approach. It is also important to note that we should strive to make precision oncology the SOC, not the currently organ-based therapeutic approach. The timely evaluation and approval of such off-label drug combinations by health authorities could speed up the paradigm shift toward precision oncology.
