Exploratory Analysis of PTEN Deficiency by Immunohistochemistry from the Phase III CAPItello-291 Trial Authors: Komal Jhaveri, Hope S. Rugo, Javier Cortés, Mafalda Oliveira, Sacha J. Howell, Florence Dalenc, Henry L. Gomez, Xichun Hu, Petr Krivorotko, Sibylle Loibl, Meena Okera, Yeon Hee Park, Joohyuk Sohn, Masakazu Toi, Eriko Tokunaga, Lyudmila Zhukova, Agostina Nardone, Ian Wadsworth, Celina D’Cruz, Tanya Wantenaar, Dimple Das, Elza C. de Bruin, Nicholas C. Turner Published: 2026 ABSTRACT AbstractPurpose:The phosphoinositide 3-kinase (PI3K)/AKT serine/threonine kinase (AKT) pathway is frequently activated in hormone receptor (HR)-positive/human epidermal growth factor receptor 2 (HER2)-negative advanced breast cancer, with loss of phosphatase and tensin homolog (PTEN) activity contributing to this activation. Capivasertib is a potent pan-AKT inhibitor, approved in combination with fulvestrant for the treatment of HR-positive/HER2-negative locally advanced/metastatic breast cancer with one or more PIK3CA/AKT1/PTEN tumor alterations. Next-generation sequencing (NGS) is commonly used to identify patients with PIK3CA/AKT1/PTEN tumor alterations. However, the utility of immunohistochemistry (IHC) to identify patients with PTEN-deficient tumors has not been explored in this context.Experimental Design:This exploratory analysis was based on tumor samples collected from patients in the global phase III CAPItello-291 study.Results:PTEN IHC results were obtained for 367 tumor samples, with 70 (19.1%) identified as PTEN deficient by IHC. A total of 346 (94.3%) samples with a PTEN IHC test result also had NGS test results available for PIK3CA/AKT1/PTEN alteration status. When comparing PTEN deficiency by IHC with PTEN alteration status by NGS, the overall, positive, and negative percent agreements were 87.0% (301/346), 71.9% (23/32), and 88.5% (278/314), respectively. Exploratory analysis in patients with PTEN-deficient tumors by IHC (n = 70) showed improved progression-free survival in the capivasertib plus fulvestrant versus placebo plus fulvestrant treatment arm (median 9.3 vs. 3.7 months; hazard ratio: 0.52, 95% confidence interval, 0.28–0.90).Conclusions:These results suggest potential utility for IHC in determining tumor PTEN status in breast cancer and raise the possibility of IHC identifying additional patients who could benefit from treatment with capivasertib and fulvestrant. Translational RelevanceCapivasertib is a potent, selective inhibitor of all three AKT serine/threonine kinase isoforms and is recommended in combination with fulvestrant for patients with hormone receptor (HR)-positive/human epidermal growth factor receptor 2 (HER2)-negative advanced breast cancer with PIK3CA/AKT1/PTEN-altered tumors. Approximately half of all patients with HR-positive/HER2-negative advanced breast cancer harbor PIK3CA/AKT1/PTEN-altered tumors. Although next-generation sequencing (NGS) can readily detect activating mutations in PIK3CA and AKT1, more complex inactivating PTEN alterations are not always detectable by NGS. This study explored the ability of immunohistochemistry (IHC) to detect PTEN deficiency in tumor samples from the CAPItello-291 trial, in which patients received capivasertib plus fulvestrant and PIK3CA/AKT1/PTEN alteration status had been assessed by NGS. Patients with PTEN deficiency, as determined by IHC, showed improved progression-free survival with capivasertib plus fulvestrant versus placebo plus fulvestrant, regardless of PIK3CA/AKT1/PTEN alteration status by NGS, suggesting that IHC has potential utility in identifying some additional patients who can benefit from treatment with capivasertib plus fulvestrant. Translational Relevance Capivasertib is a potent, selective inhibitor of all three AKT serine/threonine kinase isoforms and is recommended in combination with fulvestrant for patients with hormone receptor (HR)-positive/human epidermal growth factor receptor 2 (HER2)-negative advanced breast cancer with PIK3CA/AKT1/PTEN-altered tumors. Approximately half of all patients with HR-positive/HER2-negative advanced breast cancer harbor PIK3CA/AKT1/PTEN-altered tumors. Although next-generation sequencing (NGS) can readily detect activating mutations in PIK3CA and AKT1, more complex inactivating PTEN alterations are not always detectable by NGS. This study explored the ability of immunohistochemistry (IHC) to detect PTEN deficiency in tumor samples from the CAPItello-291 trial, in which patients received capivasertib plus fulvestrant and PIK3CA/AKT1/PTEN alteration status had been assessed by NGS. Patients with PTEN deficiency, as determined by IHC, showed improved progression-free survival with capivasertib plus fulvestrant versus placebo plus fulvestrant, regardless of PIK3CA/AKT1/PTEN alteration status by NGS, suggesting that IHC has potential utility in identifying some additional patients who can benefit from treatment with capivasertib plus fulvestrant. Introduction Capivasertib is a potent, selective inhibitor of all three AKT serine/threonine protein kinase (AKT) isoforms (AKT1/2/3) used in the treatment of patients with hormone receptor (HR)-positive/human epidermal growth factor receptor 2 (HER2)-negative advanced breast cancer. Capivasertib targets the phosphoinositide 3-kinase (PI3K)/AKT pathway through pan-AKT inhibition (1). Phosphatase and tensin homolog (PTEN) is a negative regulator of the PI3K/AKT pathway (2). Based on results from the phase III CAPItello-291 trial (3), capivasertib in combination with fulvestrant was approved for patients with HR-positive/HER2-negative metastatic or locally advanced breast cancer with one or more PIK3CA/AKT1/PTEN tumor alterations whose disease relapsed or progressed during or after aromatase inhibitor treatment. Approximately 50% of patients with HR-positive/HER2-negative advanced breast cancer have tumors that harbor PIK3CA/AKT1/PTEN alterations (4). Inactivating PTEN alterations are present in 5% to 10% of HR-positive/HER2-negative metastatic breast cancer tumors and are associated with poorer overall survival for those patients versus those whose tumors do not have one or more PTEN alterations (5). Therefore, determining tumor PTEN status through the identification of PTEN genomic alterations or loss of PTEN protein is important for informing treatment selection and optimizing patient outcomes. Next-generation sequencing (NGS) is a powerful tool that uses high-depth sequencing of multiple genes to detect a range of clinically actionable alterations in tumor DNA with high sensitivity in a single assay (6). However, detection of complex loss-of-function alterations in PTEN can be more challenging than the simpler gain-of-function missense mutations at hotspot positions in AKT1 and PIK3CA (7). Immunohistochemistry (IHC) assays detect protein expression and/or location within the tumors and can be conducted with limited tissue sample. Although both NGS and IHC are well-established tools for informing treatment options, the comparison of the ability to classify PTEN status using NGS and IHC in HR-positive breast cancer is yet to be determined. Furthermore, the potential for IHC to identify patients with PTEN-deficient tumors that are not reported as PTEN altered by NGS warrants further investigation. Here, we report the concordance between PTEN deficiency by IHC and PTEN alterations by NGS using tumor samples from patients who participated in the CAPItello-291 study. In addition, a prespecified exploratory analysis was performed to evaluate progression-free survival (PFS) in patients with PTEN-deficient tumors identified by IHC. Materials and Methods Study population This study analyzed tumor samples collected from patients in the global phase III CAPItello-291 study (NCT04305496). A total of 708 women and men with HR-positive/HER2-negative advanced breast cancer who experienced disease recurrence or progression on or after aromatase inhibitor treatment, with or without a cyclin-dependent kinase 4/6 (CDK4/6) inhibitor, were eligible for inclusion. In the phase III study, patients were randomized 1:1 to receive fulvestrant (500 mg intramuscularly on days 1 and 15 of cycle 1 and day 1 of each subsequent 28-day cycle) with either placebo or capivasertib (400 mg twice daily; 4 days on, 3 days off). The data cutoff for PFS analysis was August 15, 2022. All patients required a recently collected formalin-fixed, paraffin-embedded (FFPE) tumor sample. An institutional review board and an independent ethics committee reviewed the CAPItello-291 trial protocol, amendments, and other relevant documents. All the patients provided written informed consent before enrollment. The trial was conducted in accordance with the applicable International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use and Good Clinical Practice guidelines and with the principles of the Declaration of Helsinki. NGS of tumor tissue samples PIK3CA/AKT1/PTEN alteration status was determined after randomization using NGS in tumor tissue collected prior to study enrollment, as described previously (3). Eligible PIK3CA/AKT1/PTEN alterations were detected using the FoundationOneCDx assay (Foundation Medicine, Inc.; ref. 8); eligible PTEN alterations are outlined in Supplementary Table S1. Of 708 patients that underwent randomization, 602 tumor samples passed quality control for NGS analysis of PIK3CA/AKT1/PTEN alteration status. Data from patients recruited in China who underwent NGS testing with OncoScreen Plus (Burning Rock Biotech, Ltd) were excluded from the analyses reported here (n = 8). This resulted in a total of 594 samples with PIK3CA/AKT1/PTEN alteration status determined by NGS included in this exploratory analysis, of which 381 (64.1%) were from primary tumor tissue, 156 (26.3%) were from recurrent tumor tissue, and 57 (9.6%) were unknown/not valid (9). IHC of tumor tissue samples Of 708 patients who underwent randomization, 472 (66.7%) had tumor samples available for central PTEN IHC testing at the VENTANA laboratory. PTEN IHC testing was not performed on patients enrolled in the study in China. FFPE tumor tissue samples were sectioned at approximately 4 to 5 μm in thickness and mounted on positively charged microscope slides. Slides were then stained with the VENTANA PTEN (SP218; RRID: AB_3675592) Rabbit Monoclonal Primary Antibody (Roche Diagnostics) at a dilution of 1:100 and left to incubate for 16 minutes at 37°C as per the manufacturer’s documentation. Following washes to remove unbound primary antibody, the slides were incubated with anti-rabbit secondary antibody bound by an enzyme-labeled tertiary antibody. The complex was then visualized with hydrogen peroxide substrate and 3, 3′-diaminobenzidine tetrahydrochloride chromogen. Slides were counterstained with hematoxylin, dehydrated, and mounted before observation by light microscopy. Evaluation of slides for PTEN IHC staining Slides were initially evaluated against clinical scoring criteria with a minimum of 100 viable tumor cells present and either no nonspecific background staining or low-level background staining that did not hinder interpretation of the specific PTEN staining pattern. Slides showing ambiguous PTEN staining in surrounding tissue (very faint or absent staining) were considered not evaluable (Supplementary Fig. S1). Interpretation of PTEN IHC staining Interpretation of PTEN IHC staining was performed by a pathologist blinded to PTEN alteration status by NGS. Slides were visualized at 20× magnification, and the percentage of cells with positive PTEN staining was determined through scoring of >100 viable tumor cells (defined as one of two buckets: 100–500 or >500 cells). PTEN proficiency was defined by a predefined threshold of >10% of viable malignant cells with cytoplasmic PTEN staining at any staining intensity, including weak (1+), moderate (2+), and strong (3+) staining (Supplementary Fig. S2). PTEN deficiency was defined as ≥90% of viable malignant cells without any cytoplasmic PTEN staining in the presence of positive staining internal controls (Supplementary Fig. S3). Positive controls were based on PTEN staining, ranging from 1+ to 3+, in surrounding benign tissue (stromal cells, endothelial cells, and peripheral nerve fibers), as described previously (10). Concordance analysis Concordance analysis was performed to evaluate how often IHC detected samples as PTEN deficient/proficient versus the corresponding PTEN alteration status determined by NGS. Concordance was assessed using overall percent agreement (OPA), positive percent agreement (PPA), and negative percent agreement (NPA), and these metrics were calculated as follows:OPA= PTEN deficient by IHC and PTEN altered by NGS +PTEN proficient by IHC and PTEN non-altered by NGS Total samplesPPA= PTEN deficient by IHC and PTEN altered by NGSPTEN altered by NGSNPA=PTEN proficient by IHC and PTEN non-altered by NGSPTEN non-altered by NGS Statistics A prespecified exploratory analysis of PFS was conducted for the subset of patients with PTEN-deficient tumors by IHC stratified by treatment group (capivasertib plus fulvestrant vs. placebo plus fulvestrant), both with and without PIK3CA/AKT1/PTEN alterations identified by NGS. Hazard ratios were calculated using a Cox proportional hazards model and stratified by prior use of CDK4/6 inhibitor. Median PFS was estimated using Kaplan–Meier methods and was reported with 95% confidence intervals (CIs). PFS analysis was also performed for the subsets of these patients with confirmed PIK3CA/AKT1/PTEN alterations identified by NGS and without any PIK3CA/AKT1/PTEN alterations identified by NGS. For these analyses, hazard ratios were calculated using an unstratified Cox proportional hazards model due to the small number of events, in accordance with the predefined pooling strategy. Results Patient characteristics Baseline characteristics of the patients with a PTEN by IHC test result were generally similar to the overall study population and broadly comparable regardless of PTEN status (Table 1). Some differences between treatment arms were noted in the PTEN deficient by IHC subgroup; the capivasertib treatment group had a higher proportion of postmenopausal patients (85.3% vs. 72.2%), a lower incidence of liver (32.4% vs. 58.3%) and visceral disease (58.8% vs. 72.2%), and a lower proportion of patients who had undergone adjuvant or neoadjuvant chemotherapy (44.1% vs. 63.9%) compared with the placebo group (Table 1). Table 1.Baseline patient characteristics and prior treatments.CharacteristicOverall population (n = 708)PTEN proficient by IHC (n = 297)PTEN deficient by IHCTotal (n = 70)Capivasertib–fulvestrant (n = 34)Placebo–fulvestrant (n = 36)Median age, years (range)58 (26–90)59 (26–83)62 (33–86)61.5 (33–83)63 (35–86)Female, n (%)701 (99)294 (99)70 (100)34 (100)36 (100)Postmenopausal, n (%)547 (77.3)227 (76.4)55 (78.6)29 (85.3)26 (72.2)Race, n (%)​​​​​ White407 (57.5)175 (58.9)42 (60)21 (61.8)21 (58.3) Asian189 (26.7)75 (25.3)13 (18.6)7 (20.6)6 (16.7) Black/African American8 (1.1)3 (1)2 (2.9)1 (2.9)1 (2.8) Other104 (14.7)44 (14.8)13 (18.6)5 (14.7)8 (22.2)Region,an (%)​​​​​ 1395 (55.8)164 (55.2)29 (41.4)14 (41.2)15 (41.7) 2136 (19.2)65 (21.9)28 (40)13 (38.2)15 (41.7) 3177 (25)68 (22.9)13 (18.6)7 (20.6)6 (16.7)Metastatic sites, n (%)​​​​​ Bone only103 (14.5)48 (16.2)9 (12.9)4 (11.8)5 (13.9) Livera306 (43.2)117 (39.4)32 (45.7)11 (32.4)21 (58.3) Visceral478 (67.5)194 (65.3)46 (65.7)20 (58.8)26 (72.2)HR status,bn (%)​​​​​ ER+/PgR+501 (70.8)213 (71.7)55 (78.6)27 (79.4)28 (77.8) ER+/PgR−197 (27.8)78 (26.3)15 (21.4)7 (20.6)8 (22.2) ER+/PgR unknown9 (1.3)6 (2)000Endocrine resistance, n (%)​​​​​ Primary262 (37)112 (37.7)29 (41.4)13 (38.2)16 (44.4) Secondary446 (63)185 (62.3)41 (58.6)21 (61.8)20 (55.6)Prior lines of endocrine therapy, n (%)​​​​​ 093 (13.1)46 (15.5)14 (20)6 (17.6)8 (22.2) 1539 (76.1)222 (74.7)49 (70)25 (73.5)24 (66.7) 276 (10.7)29 (9.8)7 (10)3 (8.8)4 (11.1)Previous CDK4/6 inhibitor,an (%)489 (69.1)188 (63.3)45 (64.3)21 (61.8)24 (66.7)Previous chemotherapy, n (%)​​​​​ Neoadjuvant/adjuvant350 (49.4)158 (53.2)38 (54.3)15 (44.1)23 (63.9) Advanced disease129 (18.2)50 (16.8)12 (17.1)7 (20.6)5 (13.9)Abbreviations: ER+/−, estrogen receptor–positive/negative; ESO-ESMO, European School of Oncology and the European Society for Medical Oncology; PgR+/−, progesterone receptor–positive/negative.Primary and secondary resistances were defined using the 4th ESO-ESMO International Consensus Guidelines for advanced breast cancer.aBaseline stratification factors include the presence of liver metastases, prior use of CDK4/6 inhibitor, and region. Region 1: United States, Canada, Western Europe, Australia, and Israel. Region 2: Latin America, Eastern Europe, and Russia. Region 3: Asia.bOne patient in the capivasertib–fulvestrant group was ER−. Baseline patient characteristics and prior treatments. Abbreviations: ER+/−, estrogen receptor–positive/negative; ESO-ESMO, European School of Oncology and the European Society for Medical Oncology; PgR+/−, progesterone receptor–positive/negative. Primary and secondary resistances were defined using the 4th ESO-ESMO International Consensus Guidelines for advanced breast cancer. Baseline stratification factors include the presence of liver metastases, prior use of CDK4/6 inhibitor, and region. Region 1: United States, Canada, Western Europe, Australia, and Israel. Region 2: Latin America, Eastern Europe, and Russia. Region 3: Asia. One patient in the capivasertib–fulvestrant group was ER−. Concordance between IHC and NGS in assaying for PTEN deficiency Of the 708 patients in the CAPItello-291 study, 472 underwent IHC testing, of which PTEN status was determined for 367 samples (77.8%; Fig. 1). Seventy of 367 (19.1%) patient tumor samples were identified as PTEN deficient by IHC (Fig. 1). IHC testing failed for 105 of 472 (22.2%) samples, predominantly due to failed internal controls (51/105, 48.6%) or insufficient tumor tissue (47/105, 44.8%; Fig. 1). Of the 708 patients in the CAPItello-291 study, 594 had PIK3CA/AKT1/PTEN tumor alteration status determined by NGS, with PTEN alterations identified in 51 of 594 (8.6%) of these samples. Identification of tumor PTEN IHC status from tumor samples collected from patients in the CAPItello-291 trial. Analysis of concordance between PTEN status by IHC and PTEN alteration status by NGS was performed in the subset of patients for whom both NGS and IHC results were available (n = 346; Fig. 2A). The OPA, PPA, and NPA were 87.0%, 71.9%, and 88.5%, respectively. Of the 314 samples that were PTEN non-altered by NGS, 36 (11.5%) were found to be PTEN deficient by IHC (Fig. 2A). Some samples (n = 9) with PTEN alterations identified by NGS were PTEN proficient by IHC (Fig. 2A). Of these nine samples, two contained missense mutations, and seven contained truncating PTEN mutations (Fig. 2B). Truncating mutations affecting only one PTEN allele or occurring in a subclonal population of tumor cells may result in partial loss of PTEN protein without abolishing PTEN protein levels below the threshold necessary to be scored as PTEN deficient by IHC. Notably, three samples with truncating mutations were within the 80% to 89% of cells with PTEN-negative staining category and so were close to the threshold that defines PTEN deficiency by IHC (Fig. 2B). The majority of PTEN alterations identified through NGS were truncating mutations or homozygous deletions (Fig. 2B). Notably, all tumor samples with homozygous deletions or structural variants of PTEN by NGS, for which IHC results were also available, were PTEN deficient by IHC (Fig. 2B). Concordance between PTEN status by IHC and NGS. A, Concordance table between PTEN status by IHC and PTEN alteration status by NGS (n = 346). B, Distribution of the proportion of tumor cells lacking PTEN staining by IHC (binned by %) with corresponding PTEN alteration status by NGS. C, Analysis of overlap between samples that were PTEN deficient by IHC and PIK3CA/AKT1 or PTEN altered by NGS (n = 185). D, Schematic showing types of PIK3CA/AKT1/PTEN alterations by NGS (n = 594) with PTEN status by IHC indicated for the samples with available IHC results (n = 367). When the NGS analysis was broadened beyond PTEN-only to also include PIK3CA/AKT1 tumor alterations, 38 (54.3%) of the 70 patients with tumors identified as PTEN deficient by IHC had PIK3CA/AKT1/PTEN-altered tumors identified by NGS, and 32 (45.7%) were not identified as PIK3CA/AKT1/PTEN altered by NGS, with 21 (30.0%) having confirmed PIK3CA/AKT1/PTEN non-altered tumors identified by NGS and 11 (15.7%) with unknown NGS results (Fig. 2C and D). Exploratory PFS analyses in patients with IHC-classified PTEN-deficient tumors Exploratory efficacy analyses in patients with PTEN-deficient tumors by IHC (n = 70) showed improved PFS in the capivasertib plus fulvestrant versus the placebo plus fulvestrant treatment arm (median 9.3 vs. 3.7 months; hazard ratio: 0.52, 95% CI, 0.28–0.93; Fig. 3A), stratified according to prior CDK4/6 inhibitor treatment. This population includes patients both with and without PIK3CA/AKT1/PTEN tumor alterations identified by NGS in the original CAPItello-291 study. Assessing those subpopulations separately using unstratified analyses, a similar PFS benefit was observed in the subset of patients with PTEN-deficient tumors by IHC and also PIK3CA/AKT1/PTEN tumor alterations identified by NGS (n = 38) in the capivasertib plus fulvestrant versus the placebo plus fulvestrant treatment arm (9.3 vs. 3.7 months; hazard ratio: 0.43, 95% CI, 0.20–0.92; Fig. 3B). A similar PFS benefit was also observed in the subset of patients with PTEN-deficient tumors by IHC but without PIK3CA/AKT1/PTEN alterations detected by NGS (n = 32; 16.2 vs. 4.6 months; hazard ratio: 0.43, 95% CI, 0.15–1.05; Fig. 3C). However, this subpopulation includes 11 patients with unknown NGS results, which therefore could include some patients with PIK3CA/AKT1/PTEN-altered tumors. Nevertheless, IHC did allow detection of PTEN deficiency in cases in which NGS did not provide a test result, often due to insufficient sample availability or low tumor cell content. PFS analysis for patients with PTEN-deficient tumors. A, PFS in patients with PTEN-deficient tumors by IHC (n = 70), stratified according to prior CDK4/6 inhibitor treatment. B, PTEN deficient by IHC and PIK3CA/AKT1/PTEN altered by NGS (n = 38) with unstratified PFS analysis. C, PTEN deficient by IHC and PIK3CA/AKT1/PTEN non-altered by NGS (n = 32) with unstratified analysis. The PIK3CA/AKT1/PTEN non-altered group includes patients with confirmed PIK3CA/AKT1/PTEN non-altered and unknown results. NC, noncalculable. Discussion The prevalence of PTEN deficiency by IHC in this study was 19%, which is higher than the reported prevalence of PTEN alterations identified by NGS (∼7%) (3). IHC also detected tumor PTEN deficiency in patients whose tumors were classified as PIK3CA/AKT1/PTEN non-altered in the original CAPItello-291 study. Analogous observations have been described in endometrial and prostate cancers, in which IHC identified additional patients with PTEN-deficient tumors but without PTEN genomic alterations (11, 12). Various mechanisms have been suggested to account for this observation, including limited ability to detect complex rearrangements by NGS (13), miRNAs (14, 15), and epigenetic gene silencing (16, 17). Conversely, some samples with PTEN alterations identified by NGS were not scored as PTEN deficient by IHC. However, this was only observed in a small number of patients (n = 9); it was therefore not possible to conduct meaningful analyses of clinical outcomes in this subset. NGS may detect subclonal PTEN alterations that result in PTEN deficiency in only a subset of cells (insufficient to meet the PTEN-deficient IHC threshold). In addition, the fact that IHC detects PTEN protein expression, irrespective of whether the protein is functional, is an inherent limitation of IHC (5, 7). Indeed, all of these samples contained missense or truncating PTEN mutations, which may not deplete PTEN protein levels to the same extent as homozygous deletions. Notably, all samples with homozygous PTEN deletions were PTEN deficient by IHC. In addition to tissue-based NGS testing, there is growing interest in using blood-based liquid biopsies to inform treatment options for patients with HR-positive/HER2-negative advanced breast cancer (18, 19). FoundationOneLiquid CDx is an FDA-approved companion diagnostic test for the PI3K inhibitors alpelisib, in combination with fulvestrant, and inavolisib, in combination with palbociclib and fulvestrant, for patients with activating PIK3CA tumor alterations (20). Although activating mutations in PIK3CA and AKT1 can be readily detected in liquid biopsies (with tumor fractions ≥1%), the detection of PTEN homozygous deletions, copy number alterations, and rearrangements in liquid biopsies is more variable (13, 21). Guidelines recommend confirmatory tissue-based testing for the detection of certain biomarkers if no alterations are detected by liquid biopsy NGS (19, 22). IHC could provide a practical alternative for the identification of patients with tumor PTEN deficiency, with faster turnaround times than tissue-based NGS. The retrospective IHC testing in this study failed to provide conclusive PTEN results for approximately one-fifth of patients. This was largely due to either failure of internal controls, which highlights a limitation of IHC, or insufficient tumor content, which underscores the importance of ensuring sufficient tumor tissue when using IHC to inform treatment options. The availability of adequate tissue with sufficient tumor cells for additional exploratory biomarker analyses was limited in the context of the CAPItello-291 trial with NGS testing as part of the preplanned primary endpoint. A recent publication suggests best practices for minimizing analytic variability of PTEN IHC testing and standardization of scoring and reporting of PTEN IHC results (23). Preplanned exploratory analyses of PFS showed a consistent benefit of capivasertib plus fulvestrant in patients with PTEN-deficient tumors by IHC, regardless of PIK3CA/AKT1/PTEN alteration status determined by NGS. Although interpretation is limited by the small number of patients in certain subgroups and the inclusion of patients with unknown PIK3CA/AKT1/PTEN alteration status (n = 11), the observed hazard ratio seems to be similar to that reported for the PIK3CA/AKT1/PTEN-altered population in the global CAPItello-291 study (0.50, 95% CI, 0.38–0.65; P < 0.001; ref. 3). This suggests that addition of PTEN IHC testing alongside PIK3CA/AKT1/PTEN testing by NGS may identify additional patients who could benefit from treatment with capivasertib and fulvestrant, which warrants further exploration. Conclusion Although these analyses were exploratory in nature and based on a small number of patients, they suggest that PTEN IHC testing can identify patients who benefit from treatment with capivasertib and fulvestrant, some of whom were not identified by NGS. This suggests detection of PTEN deficiency by IHC may have clinical relevance as a biomarker for optimal treatment selection in HR-positive/HER2-negative advanced breast cancer. Further exploration of PTEN IHC testing in breast cancer and comparisons with NGS are warranted to better inform the clinical utility of IHC in this context. Supplementary Material Figure S1, Figure S2, Figure S3, Table 1 References.