Radiol Oncol 2023; 57(2): 220-228. doi: 10.2478/raon-2023-0027 220 research article Association between PIK3CA activating mutations and outcomes in early-stage invasive lobular breast carcinoma treated with adjuvant systemic therapy Domen Ribnikar 1,2 , Valentina Jeric Horvat 1 , Ivica Ratosa 2,3 , Zachary W Veitch 4 , Biljana Grcar Kuzmanov 5 , Srdjan Novakovic 6 , Erik Langerholc 7 , Eitan Amir 8 , Bostjan Seruga 1,2 1 Department of Medical Oncology, Institute of Oncology Ljubljana, Ljubljana, Slovenia 2 Faculty of Medicine Ljubljana, University of Ljubljana, Ljubljana, Slovenia 3 Department of Radiation Oncology, Institute of Oncology Ljubljana, Ljubljana, Slovenia 4 Division of Medical Oncology and Hematology, Royal Victoria Hospital, Bariie, Ontario, USA 5 Department of Pathology, Institute of Oncology Ljubljana, Ljubljana, Slovenia 6 Department of Molecular Diagnostics, Institute of Oncology Ljubljana, Ljubljana, Slovenia 7 Institute of Biostatistics and informatics, Faculty of Medicine Ljubljana, University of Ljubljana, Ljubljana, Slovenia 8 Division of Medical Oncology and Hematology, University of Toronto and Princess Margaret Cancer Centre, Toronto, Canada Radiol Oncol 2023; 57(2): 220-228. Received 13 February 2023 Accepted 18 May 2023 Correspondence to: Assoc. Prof. Boštjan Šeruga, M.D., Ph.D., Department of Medical Oncology, Institute of Oncology Ljubljana, Zaloška cesta 2, SI-1000 Ljubljana, Slovenia. E-mail: bseruga@onko-i.si Disclosure: No potential conflicts of interest were disclosed. This is an open access article distributed under the terms of the CC-BY license (https://creativecommons.org/licenses/by/4.0/). Background. The aim of the study was to evaluate the independent prognostic role of PIK3CA activating mutations and an association between PIK3CA activating mutations and efficacy of adjuvant endocrine therapy (ET) in patients with operable invasive lobular carcinoma (ILC). Patients and methods. A single institution study of patients with early-stage ILC treated between 2003 and 2008 was performed. Clinicopathological parameters, systemic therapy exposure and outcomes (distant metastasis-free sur- vival [DMFS] and overall survival [OS]) were collected based on presence or absence of PIK3CA activating mutation in the primary tumor determined using a quantitative polymerase chain reaction (PCR)-based assay. An association between PIK3CA mutation status and prognosis in all patient cohort was analyzed by Kaplan-Meier survival analysis, whereas an association between PIK3CA mutation and ET was analyzed in estrogen receptors (ER) and/or progester- one receptors (PR)-positive group of our patients by the Cox proportional hazards model. Results. Median age at diagnosis of all patients was 62.8 years and median follow-up time was 10.8 years. Among 365 patients, PIK3CA activating mutations were identified in 45%. PIK3CA activating mutations were not associated with differential DMFS and OS (p = 0.36 and p = 0.42, respectively). In patients with PIK3CA mutation each year of tamoxifen (TAM) or aromatase inhibitor (AI) decreased the risk of death by 27% and 21% in comparison to no ET, re- spectively. The type and duration of ET did not have significant impact on DMFS, however longer duration of ET had a favourable impact on OS. Conclusions. PIK3CA activating mutations are not associated with an impact on DMFS and OS in early-stage ILC. Patients with PIK3CA mutation had a statistically significantly decreased risk of death irrespective of whether they received TAM or an AI. Key words: invasive lobular carcinoma; PIK3CA mutation; endocrine therapy; genomics of invasive lobular carcinoma Radiol Oncol 2023; 57(2): 220-228. Ribnikar D et al. / PIK3CA activating mutations and outcomes in early-stage invasive lobular breast carcinoma 221 Introduction Breast cancer represents a spectrum of heterogene- ous diseases with different clinical behaviour and response to specific systemic therapy. Invasive lob- ular carcinoma (ILC) is the second most frequent subtype of breast cancer, representing around 10% of all breast cancer cases. 1 It has been demonstrated that ILC is a special disease entity that differs from the more common invasive breast carcinomas such as invasive ductal carcinoma (IDC). These differences include risk factors, histological and clinical characteristics, transcriptional signatures and genomic profiles. 2-4 ILCs usually arise in postmenopausal women, are larger in size at the time of diagnosis due to insidi- ous nature of its growth, but are typically of lower histological grade. The majority of ILC tumors ex- press hormone receptors such as estrogen recep- tors (ER), progesterone receptors (PR) but they less commonly exhibit human epidermal growth factor receptor-2 (HER-2) overexpression or ampli- fication. 5 Data about prognosis of ILC vary sub- stantially; studies have demonstrated better 6,7 , the same 8 , and worse long-term overall survival (OS) in comparison to unselected invasive breast can- cers. 9 Additionally, it has been reported that ILC is less responsive to chemotherapy (ChT). 10-12 However, these tumors respond well to endocrine therapy (ET) given their biological profile. 13 Results of an important retrospective study clearly showed an OS benefit of adjuvant ET for patients with ILC. 14 Desmedt et al. conducted the largest genomic study of ILC and found that mutations in PIK3CA (phosphatidylinositol, catalytic unit), HER and ESR1 (estrogen receptor 1) genes are more fre- quently present in ILC tumors that in other inva- sive breast cancers. Furthermore, in their study, approximately 50% of ILC tumors harboured PIK3CA-PTEN-AKT1 signaling alterations. Of note, PIK3CA mutation was associated with lower histological grade and lower Ki-67 (proliferating) index. 15 Mutations of PIK3CA gene are the most common somatic genetic alterations in ER-positive breast cancer and are associated with favorable breast cancer characteristics such as smaller tumor size, lower grade, ER positivity and increasing age. 16 Investigators of the BIG 1-98 study which com- pared letrozole monotherapy and letrozole switch- ing strategy to tamoxifen (TAM) monotherapy reported that postmenopausal patient population with ILC may derive greater benefit from letrozole (AIs) than patients with IBC, NOS (invasive breast carcinoma, no otherwise specified). 17 Furthermore, additional post-hoc analysis of the BIG 1-98 study showed that irrespective of the histological sub- type of breast cancer tumors with PIK3CA muta- tion derive greater benefit from letrozole than from tamoxifen. 18 There are no data about the exact num- ber of ILC patients with present PIK3CA mutation in the BIG 1-98 study; however in the large meta- analysis by Zardavas et al. there were 366 out of 951 (39%) patients who had PIK3CA mutated ILC. 16 In this study, we aimed to evaluate the inde- pendent prognostic role of PIK3CA activating mutations and an association between PIK3CA ac- tivating mutations and efficacy of ET in patients with operable ILC. We hypothesized that the pres- ence of PIK3CA activating mutations in primary ILC is associated with longer distant metastasis- free survival and overall survival and greater ben- efit of AIs and extended ET as compared to stand- ard 5-year treatment with tamoxifen. Patients and methods Patients After obtaining approval from the institutional review committee and ethical approval of the Ministry of Health of the Slovenian Republic (#0120-323/2019), we performed a retrospective analysis of a cohort of patients with early-stage ILC identified from a pathology database at the Institute of Oncology Ljubljana. Eligible pa- tients were treated between January 1 st 2003 and December 31 st 2008, thereby allowing at least 10 years of follow-up at the time of data analysis. Clinicopathological characteristics and determination of the PIK3CA mutation Formalin-fixed and paraffin embedded (FFPE) and hematoxylin and eosin stained tumor slides of all included patients were reviewed by an ex- perienced breast pathologist (BGK) to ensure that the diagnosis of ILC and its subtype(s) were cor- rect. We collected the following clinicopathologi- cal parameters from each patient′s from the exist- ing pathology reports: age at diagnosis, subtype of ILC, tumor size, nodal status, grade, mitotic count, ER and PR expression, HER-2 overexpression or amplification, lymphovascular invasion (LVI) and perineural invasion (PNVI). Ki-67 labeling index and tumor infiltrating lymphocytes (TILs) were additionally determined retrospectively by a sin- Radiol Oncol 2023; 57(2): 220-228. Ribnikar D et al. / PIK3CA activating mutations and outcomes in early-stage invasive lobular breast carcinoma 222 gle breast cancer pathologist as these biomarkers were not assessed routinely in years of the cohort inception. Tumor grading was classified according to the Bloom-Richardson-Elston classification. 19 ER and PR expression and HER-2 status were de- termined using the American Society of Clinical Oncology/College of American Pathologists guide- lines. 20,21 Proliferation index Ki-67 was estimated with DAKO, Glostrup antibody MIB1 accord- ing to recommendations from the International Ki-67 in Breast Cancer Working Group. 22 For the evaluation of TILs the recommendations of the 2014 International TILs Working Group were fol- lowed. 23 PIK3CA status (wild-type or mutated) was determined for each patient. DNA isolation from the FFPE tumor samples was performed by macrodissection of tumor tissue from tumor slices containing at least 70% tumor cells. Isolation was performed using the MagMAX FFPE DNA/RNA Ultra Kit (Applied Biosystems, Thermo Fisher TABLE 1. Characteristics of the included patients and their tumors Characteristic All patients (n = 365; 100%) PIK3CA mutated (n = 164; 45%) PIK3CA non-mutated (n = 201; 55%) Median age (range) (yrs) 62.8 (33−90) 63.1 62.4 Median tumor size (mm) 21 19 21 Nodal status N0 N1 N2 N3 209 (57) 84 (23) 26 (7) 46 (13) 100 (61) 37 (22) 1 (7) 16 (10) 109 (54) 47 (23) 15 (8) 30 (15) Tumor grade G1 G2 G3 52 (14) 270 (74) 43 (12) 20 (12) 134 (82) 10 (6) 32 (16) 136 (68) 33 (16) IHC* subtype ER+/PR+/HER2- ER+/PR-/HER2- HER2 + ER-/PR-/HER2- 299 (82) 39 (11) 20 (6) 7 (2) 136 (46) 14 (36) 9 (45) 2 (29) 159 (54) 25 (64) 11 (55) 5 (71) Median Ki-67 (range) (%) 3 (1−50) 2.5 (1−50) 3 (1 −40) Mitotic score M1 M2 M3 284 (78) 55 (15) 26 (7) 138 (84) 19 (12) 7 (4) 146 (73) 36 (18) 19 (9) Presence of LVI 24 (7) 4 (2) 20 (10) Median TILs (range) (%) 3 (1 −50) 3 3 Median follow-up time (range) (yrs) 10.8 (0.1 −18.6) 10.8 10.7 ER = estrogen receptor; HER = 2–human epidermal growth factor receptor-2; *IHC = immunohistochemically defined subtype; PR = progesterone receptor; LVI = lymphovascular invasion; TILs = tumor-infiltrating lymphocytes 78 patients had PIK3CA mutation on exon 9, 76 patients on exon 20; five patients had dual PIK3CA mutation on exon 9 and five patients had both, exon 9 and exon 20 PIK3CA mutations TABLE 2. Distribution of systemic therapy in all patients and according to PIK3CA mutation status Systemic therapy All pts N (%) PIK3CA mutated pts N (%) PIK3CA non-mutated pts N (%) None ET only ChT only ET and ChT 12 (3) 249 (68) 13 (4) 91 (25) 5 (3) 115 (70) 5 (3) 39 (24) 7 (3) 134 (67) 8 (4) 52 (26) None ET 25 (7) 10 (6) 15 (7) ET with TAM 106 (29) 45 (27) 61 (30) ET with AIs 127 (35) 54 (33) 73 (37) Sequence TAM-AI 107 (29) 55 (34) 52 (26) Median duration (range) of ET 5 (0.2−16.7) 5 (0.4−16.7) 5 (0.2−11.2) AIs = aromatase inhibitors; ChT = chemotherapy; ET = endocrine therapy; pts = patients, TAM = tamoxifen (TAM) Radiol Oncol 2023; 57(2): 220-228. Ribnikar D et al. / PIK3CA activating mutations and outcomes in early-stage invasive lobular breast carcinoma 223 Scientific, Austin, Tx, USA) according to the manu- facturer’s instructions. Isolated DNA was used to identify five common mutations in the PIK3CA gene using the PIK3CA Mutation Analysis Kit (EntroGen, Inc. Woodland Hills, CA, USA) ac- cording to the manufacturer’s protocol. This is a quantitative polymerase chain reaction (PCR) based assay that detects five hot spot mutations in the PIK3CA gene (NM_006218.2): c.1624G > A p.(Glu542Lys), c.1633G > A p.(Glu545Lys), c.1633G > C p.(Glu545Gln), c.3140A > G p.(His1047Arg) and c.3140A > T p.(His1047Leu). These five hot spot mu- tations represent around 80% of all PIK3CA muta- tions. 24 Systemic treatment We collected all data about systemic therapy for each patient included into the study such as whether the patient received ChT, ET and HER- 2-targeted therapy. We also collected data about the type of ChT (cyclophosphamide – metotrex- ate-5-fluorouracil regimen (CMF), anthracycline- based chemotherapy, anthracyclines and taxa- nes, taxanes without anthracyclines), type of ET (tamoxifen monotherapy, AI monotherapy or a switching approach [TAM-AI]) and duration of ET (up to 5 years or more than 5 years). Statistical analysis The outcomes of interest were distant metastasis- free survival (DMFS) and overall survival (OS) defined as time from diagnosis (date of surgery) to distant relapse or death (whichever occurred first) and as time from diagnosis to death from any cause, respectively. The last date of a follow-up was October 31 st 2021. Data were analyzed using R version 4.1.2 (Vienna, Austria). Descriptive statis- tics were used to describe patient characteristics. Age and duration of therapy were measured in years, tumor size was measured in millimeters, Ki- 67 index was measured as a proportion of positive tumor cells and logarithmised and for the number of positive axillary lymph nodes the square root was taken of in order to normalize the distribution of the variables and improve model fit. DMFS and OS were estimated using Kaplan-Meier analysis. Log-rank tests were performed to compare the survival of PIK3CA mutated and PIK3CA non-mu- tated cohorts of patients. Four Cox proportional hazard models were used to assess the impact of ET on survival of ER and/or PR-positive PIK3CA mutated patients: a simple and an advanced model for each of DMFS and OS. The simple Cox model regressed on only three variables; age at diagnosis, prior duration of tamoxifen and prior duration of AI therapy. The advanced model regressed on ad- ditional variables which are considered known or possible prognostic factors: tumor size, nodal sta- tus, histological grade, Ki-67 index, PR expression and location of PIK3CA mutation (exon 9 or exon 20), respectively. To avoid immortal time bias, any prior duration of tamoxifen and any prior duration of AI therapy were considered as time-dependent variables, while all other variables were known at a patients′ entry in the study (at the time of sur- gery). Finally, adjustment for multiple significance testing was performed using Holm’s method. An adjusted p-value ≤ 0.05 was considered statistically significant. Results Patients and their tumors Among an initial cohort of 428 patients, we ex- cluded 24 patients who were treated in other can- cer centers in Slovenia, nine patients who had de novo metastatic lobular cancer, eight patients who received neoadjuvant ChT, five patients with other histological subtypes of breast cancer (IBC, NOS, ILC and IBC, NOS), five patients with very small tumors (less than 5 mm in size) and five patients whose biopsies were from years before 2003. Two patients were lost very early in the follow up as they moved out of country, so they were excluded from the analysis. PIK3CA mutation status could 428 patients identified 365 patients eligible for analysis Excluded: 24 patients from other cancer centers 9 patients with de novo metastatic lobular carcinoma 8 patients who received neoadjuvant chemotherapy 5 patients with other histological type of breast cancer (IDC, IDC and ILC) 5 patients with a very small tumor (less than 5 mm) 5 biopsies from years before 2003 2 patients who moved abroad 5 tumors did not pass RT-PCR for determination of PIK3CA mutation status FIGURE 1. Consort diagram. Radiol Oncol 2023; 57(2): 220-228. Ribnikar D et al. / PIK3CA activating mutations and outcomes in early-stage invasive lobular breast carcinoma 224 not be determined in five cases due to technical is- sues with the samples. This resulted in an analytic cohort of 365 patients (Figure 1; Consort diagram, for patient selection schema). Characteristics of in- cluded patients and their tumors are presented in Table 1. Among all patients, 164 (45%) patients had PIK3CA mutated ILC. Of these 78 (48%) patients had PIK3CA mutations in the helical domain (exon 9) and 76 (46%) patients had the mutation present in the kinase domain (exon 20). Five patients (1%) had dual PIK3CA mutations on exon 9 and five patients (1%) had PIK3CA mutations on both, exon 9 and 20. As shown in Table 1, PIK3CA mutated and PIK3CA non-mutated cohort of patients were well balanced according to prognostic tumor characteristics. Systemic therapy exposure of all included pa- tients is shown in Table 2. As expected, the major- ity of patients received ET with either TAM (106, 29%), AIs (127, 35%) or a switch approach TAM-AI (107, 29%). Ninety-one (25%) patients received ChT and ET and only 13 (4%) patients received ChT alone. Among 25 patients who did not receive any ET, 10 patients were PIK3CA mutated. ET agents were well balanced between PIK3CA mutated and PIK3CA non-mutated patients as shown in Table 2. Median duration of ET was 5 years in both groups of patients. 191 (52%) patients had extended ET and the longest duration of extended ET was 16.7 years. An association between PIK3CA mutation status and disease outcome Median OS was 15.7 years for patients with PIK3CA mutation and 14.6 years for patients with- out PIK3CA mutation. Furthermore, median DMFS for PIK3CA mutated and for PIK3CA non-mutated ILC patients was 15.6 and 13.4 years, respectively. The presence of PIK3CA mutation in early-stage ILC patients was not associated with differential OS (p = 0.42, Figure 2) or DMFS (p = 0.36, Figure 3). Overall, 81 (22%) patients developed distant me- tastases, of these 48 (24%) were PIK3CA non-mu- tated and 33 (16%) were PIK3CA mutated. Twenty (26%) patients of those who developed distant me- tastases had PIK3CA mutation on exon 20, 12 (15%) had PIK3CA mutation on exon 9 and only one (5%) patient had both, exon 9 and 20 PIK3CA mutation. An association between type and duration of ET and disease outcome in patients with PIK3CA mutation The effect of the type of systemic ET and its duration on DMFS Patients who received AIs in comparison to those who received tamoxifen were on average 10 years older at the time of diagnosis, therefore a Cox proportional hazards model was used for adjust- ment. In a simple model, neither type of ET nor its duration and patient age at the time of diagnosis were associated with the risk of relapse (Table 3). In advanced Cox proportional hazards model only a higher number of positive axillary lymph nodes increased the risk of distant relapse (HR 1.64, adj.p < 0.001; see Table 4). The effect of the type of systemic ET and its duration on OS We found that each year of aging increased the risk of death by 5%. Each year of treatment with tamoxifen decreased the risk of death by 27% in comparison to no ET for a patient of the same age (Table 5). Similarly, each year of treatment with AIs decreased the risk of death by 21% compared to no ET for a patient of the same age (Table 5). TABLE 3. Simple Cox proportional hazards model for DMFS in the PIK3CA mutated patient cohort Characteristic HR 95% CI Adj p Age 0.99 0.96−1.02 1.0 Duration of TAM 0.92 0.75−1.14 1.0 Duration of AIs 1.07 0.90−1.28 1.0 Adj p = adjusted p value; AIs = aromatase inhibitors; CI = confidence interval, HR = hazard ratio; TAM = tamoxifen, FIGURE 2. Kaplan-Meier graph showing association between PIK3CA mutation and overall survival. Radiol Oncol 2023; 57(2): 220-228. Ribnikar D et al. / PIK3CA activating mutations and outcomes in early-stage invasive lobular breast carcinoma 225 In the advanced Cox proportional hazards mod- el age at the time of diagnosis, grade 3 and higher number of positive axillary lymph nodes were all associated with higher risk of death (HR 1.05, adj. p = 0.014, HR 5.5, adj. p = 0.040 and HR 1.58, adj. p < 0.001, respectively). We also found that longer duration of ET with either tamoxifen or an AI had a favorable impact on OS (p = 0.010 for tamoxifen and p = 0.010 for an AI, respectively). In a time-de- pendent Cox analysis each year of prior treatment with tamoxifen decreased the risk of death by 32% in comparison to no prior ET for a patient with the same characteristics. Similarly, to tamoxifen, each year of prior AI therapy decreased the risk of death by 27% compared to no prior ET for a patient with the same characteristics (Table 6). Discussion ILC of the breast is a distinct entity with unique clinical, histological and molecular characteristics that differ from more common invasive breast can- cer subtypes. The majority of older studies demon- strated that the outcome of ILC patients was better, however more recent data have suggested ILC has worse outcome compared to invasive ductal car- cinoma, especially in the long term. It seems that delay and difficulty in early diagnosis, acquired resistance to conventional therapy and risk of late relapse pose challenges in management of patients with ILC. 25 In this retrospective study our main goal was to evaluate the association between the PIK3CA mutational status and prognosis in patients with operable ILC. We found that PIK3CA mutation status has no association with either DMFS or OS. The median OS was 15.7 years for patients with PIK3CA mutation and 14.6 years for patients with PIK3CA wild type ILC. PIK3CA mutations are the most frequent somat- ic genetic alterations in ER+/HER-2– breast cancer. 18 Prior data in early-stage disease have demonstrat- ed a favorable disease outcome in patients harbor- ing PIK3CA mutations. 16 At least to some extent this could be explained by a relative mutual exclu- sion of other somatic genetic alterations which are associated with higher proliferation and increased risk of distant relapse such as TP53 mutations and amplifications of MYC (MYC proto - oncogene ) gene. 18 Furthermore, recent work has shown that ultra-low risk breast cancers have higher expres- sion scores for the PIK3CA-mutation-associated genes. 26 Previous data exploring the prognostic effect of specific somatic mutations in ER+/HER-2– early- stage breast cancer have shown that PIK3CA and MAP3K1 (mitogen-activated protein kinase 1) mu- tations co-associate and patients with tumors har- boring both mutations have a more favorable clini- cal course than those with only one gene mutation or without any of these two mutations. 27 This may suggest that PIK3CA mutation alone is not prog- nostic in general breast cancer patient population, nor is it in ILC specifically. The majority of first- and second-generation gene-expression profiles that have been used for molecular prognostication in early-stage breast cancer were developed initially in invasive ductal carcinomas and therefore their use in ILC is uncer- tain. However recent data have shown they could also be informative for prognostication in ILC pa- tients. ILC tumors are quite homogenous clinically TABLE 4. Advanced Cox proportional hazards model for DMFS in the PIK3CA mutated patient cohort Characteristic HR 95% CI Adj p Age 1.01 0.97−1.05 1.0 Tumor size 0.99 0.57−1.71 1.0 Grade G1 G2 G3 Ref 0.91 1.03 0.26−3.22 0.16−6.83 1.0 1.0 Ki-67 1.35 0.93−1.95 1.0 PR 1 0.99−1.02 1.0 PIK3CA status Other than exon 20 Only exon 20 Ref 2.19 1.01−4.80 0.582 No. of positive axillary LNs 1.64 1.30−2.06 < 0.001 Duration of TAM 1.02 0.81−1.29 1.0 Duration of AIs 1.06 0.87−1.28 1.0 Adj p = adjusted p value; AIs = aromatase inhibitors; CI = confidence interval; HR = hazard ratio; LNs = lymph nodes; No. = number; PR = progesterone receptor; TAM = tamoxifen TABLE 5. Simple Cox proportional hazards model for OS in the PIK3CA mutated patient cohort Characteristic HR 95% CI Adj p Age 1.05 1.02−1.07 < 0.001 Duration of TAM 0.73 0.61−0.88 0.002 Duration of AIs 0.79 0.67−0.93 0.005 Adj p = adjusted p value; AIs = aromatase inhibitors; CI = confidence interval; HR = hazard ratio; TAM = tamoxifen Radiol Oncol 2023; 57(2): 220-228. Ribnikar D et al. / PIK3CA activating mutations and outcomes in early-stage invasive lobular breast carcinoma 226 especially relating to classical prognostic features. The majority are grade 2, ER and PR positive, HER- 2 negative and have a relatively low proliferative activity. 28 In contrast, it has been shown recently that a 194-gene signature called LobSig may be the best in prognosticating ILC tumors. LobSig out- performed Nottingham Prognostic Index (NPI), Prosigna ROR, Oncotype Dx and Genomic Grade Index (GGI) in a multivariate Cox proportional hazards model, especially in grade 2 ILC moder- ate NPI cases. 29 The authors reported that ILCs that were associated with a high-risk score were enriched for mutations in HER-2, HER-3, TP53 (TP53 tumor suppressor gene), ROS1 (ROS proto- oncogene 1, receptor tyrosine kinase) and AKT1 (AKT serin - threonin protein kinase 1) genes. In contrast, those ILCs that had a good outcome (a low-risk score) had relatively few genomic altera- tions and interestingly, PIK3CA mutations were more common in the latter group than in the high risk ILC group. Based on all data discussed above it seems that PIK3CA mutation alone may not be an optimal genetic alteration to stratify ILC pa- tients into better and worse prognostic group but rather an integration of several concurrent genetic alterations might outperform PIK3CA mutation in an individual tumor. We also aimed to explore the predictive effect of PIK3CA mutation on benefit from ET. Efficacy of AIs as compared to tamoxifen have previously been explored and reported. 18 In the analysis from the randomized BIG 1-98 trial, the authors demon- strated a greater magnitude of benefit with an AI letrozole in comparison to tamoxifen for patients with PIK3CA mutated early-stage breast cancer (HR 0.18; 95% CI 0.06 – 0.50). Additionally, the same study demonstrated that letrozole was as- sociated with a significantly reduced risk of dis- ease - free survival event (DFS) in patients with luminal A and luminal B ILC concluding the mag- nitude of benefit from AI was greater in patients with ILC versus invasive ductal carcinomas. 30 In our study we did not find any substantial differ- ences between the effect of tamoxifen and AIs in PIK3CA mutated early-stage ILC patients when compared to patients who did not receive any ad- juvant therapy. We found that each year of treat- ment with tamoxifen decreased the risk of death by 27% in comparison to no ET for a patient of the same age. Similarly, each year of AI therapy de- creased the risk of death by 21% compared to no ET for a patient of the same age. In addition, in an advanced Cox model we found similar results. The duration of any ET, either with tamoxifen or an AI, was statistically significantly associated with OS. In patients with PIK3CA mutation each year of ta- moxifen and AIs treatment decreased the risk of death by 32% and 27% in comparison to a patient with the same tumor characteristics who did not receive any ET, respectively. However, we did not find any associations between type and duration of ET and DMFS. Only a higher number of positive axillary lymph nodes increased the risk of distant relapse of ILC in our cohort. An interesting study evaluating the relation- ship of PIK3CA mutations and response to short- term neoadjuvant AI therapy demonstrated that the presence of PIK3CA mutation did not pre- TABLE 6. Advanced Cox proportional hazards model for OS in the PIK3CA mutated patient cohort Characteristic HR 95% CI Adj p Age 1.05 1.02−1.08 0.014 Tumor size 1.38 0.90−2.10 0.84 Grade G1 G2 G3 Ref 1.41 5.52 0.59−3.38 1.67−18.18 1.0 0.04 Ki-67 0.94 0.72−1.22 1.0 PR 1 0.99−1.01 1.0 PIK3CA status Other than exon 20 Only exon 20 Ref 1.32 0.79−2.21 1.0 No. of positive axillary LNs 1.58 1.32−1.88 < 0.001 Duration of TAM 0.68 0.55−0.86 0.01 Duration of AI 0.73 0.60−0.88 0.01 Adj p = adjusted p value; AIs = aromatase inhibitors; CI = confidence interval; HR = hazard ratio; No. = number; PR = progesterone receptor; TAM = tamoxifen FIGURE 3. Kaplan-Meier graph showing association between PIK3CA mutation and distant metastasis-free survival Radiol Oncol 2023; 57(2): 220-228. Ribnikar D et al. / PIK3CA activating mutations and outcomes in early-stage invasive lobular breast carcinoma 227 clude a response to AIs. In this study suppression of Ki-67 proliferating index, which is a validated intermediate endpoint biomarker for endocrine sensitivity was used as a surrogate marker of re- sponse. 31 Furthermore, another study showed that patients with PIK3CA mutated early-stage breast cancer either treated with tamoxifen or untreated had improved outcome in comparison to PIK3CA non-mutated patients. 32 These data were concord- ant with previous findings from in vitro study showing that PIK3CA mutated cell lines were more sensitive to tamoxifen compared to PIK3CA wild type cell lines. 33 Our study has several limitations. First, due to the retrospective nature of this study our findings need to be interpreted with caution and possibly validated in the prospective setting. Second, our cohort of ILC patients is very likely molecularly heterogeneous which limits the generalizability of our findings. Third, longer duration of ET in this study is associated with significantly improved OS but not DMFS. This indicates that the impact of immortal time bias may persist despite the use of time-dependent Cox model in our analysis. Fourth, we used a PCR based assay which detects five hot spot mutations in the PIK3CA gene representing about 80% of all PIK3CA activating mutations. Results could be different if we used an alternative method for determination of all PIK3CA mutations such as next generation sequencing (NGS). Finally, the biggest limitation of our study is that despite the use of Cox models confounding by indication for treatment might still have an impact on our re- sults. In conclusion, the presence of PIK3CA mutation was not associated with differential DMFS or OS in patients with operable ILC. Adjuvant ET with either tamoxifen or an AI significantly decreased the risk of death in PIK3CA mutated ILC cohort. Furthermore, there was no significant difference in efficacy between the two classes of endocrine agents. Longer duration of ET with either tamox- ifen or AIs decreases the risk of death in ILC pa- tients with PIK3CA mutation. Future prospective studies based on molecular analyses will hopeful- ly give us more exact answer about optimal endo- crine agent for specific subpopulations of patients with early-stage ILC. Acknowledgment The authors acknowledge the financial support from the Slovenian Research agency (research core funding No P3-0321) and unrestricted donation to the Institute of Oncology Ljubljana for research. References 1. Martinez V, Azzopardi JG. Invasive lobular carcinoma of the breast: inci- dence and variants. Histopathology 1979; 3: 467-88. doi: 10.1111/j.1365- 2559.1979.tb03029.x 2. Yoder BJ, Wilkinson EJ, Massoli NA. Molecular and morphologic distinctions between infiltrating ductal and lobular carcinoma of the breast. Breast J 2007; 13: 172-79. doi: 10.1111/j.1524-4741.2007.00393.x 3. Korkola JE, DeVries S, Fridlyand J, Hwang ES, Estep AL, Chen YY, et al. Differentiation of lobular versus ductal breast carcinomas by expression microarray analysis. Cancer Res 2003; 63: 7167-75. PMID: 14612510 4. Cleton-Jansen AM. E-cadherin and loss of heterozygosity at chromosome 16 in breast carcinogenesis: different genetic pathways in ductal and lobular breast cancer? Breast Cancer Res 2002; 4: 5-8. doi: 10.1186/bcr416 5. Coradini D, Pellizzaro C, Veneroni S, Ventura L, Daidone MG. Infiltrating ductal and lobular breast carcinomas are characterised by different interre- lationships among markers related to angiogenesis and hormone depend- ence. Br J Cancer 2002; 87: 1105-11. doi: 10.1038/sj.bjc.6600556 6. Toikkanen S, Pylkkanen L, Joensuu H. Invasive lobular carcinoma of the breast has better short- and long-term survival than invasive ductal carci- noma. Br J Cancer 1997; 76: 1234-40. doi: 10.1038/bjc.1997.540 7. Silverstein MJ, Lewinsky BS, Waisman JR, Gierson ED, Colburn WJ, Senofsky GM, et al. Infiltrating lobular carcinoma. Is it differ- ent from infiltrating duct carcinoma? Cancer 1994; 73: 1673-7. doi: 10.1002/1097-0142(19940315)73:6<1673::aid-cncr2820730620>3.0.co 8. Molland JG, Donnellan M, Janu NC, Carmalt HL, Kennedy CW, Gillet DJ. Infiltrating lobular carcinoma – a comparison of diagnosis, management and outcome with infiltrating duct carcinoma. Breast 2004; 13: 389-96. doi: 10.1016/j.breast.2004.03.004 9. Ashikari R, Huvos AG, Urban JA, Robbins GF. Infiltrating lobu- lar carcinoma of the breast. Cancer 1973; 31: 110-6. doi: 10.1002/1097-0142(197301)31:1<110::aid-cncr2820310115>3.0.co;2-v 10. Tubiana-Hulin M, Stevens D, Lasry S, Guinebretiere JM, Bouita L, Cohen- Solal C, et al. Response to neoadjuvant chemotherapy in lobular and ductal breast carcinomas: a retrospective study on 860 patients from one institu- tion. Ann Oncology 2006; 17: 1228-33. doi: 10.1093/annonc/mdl114 11. Cristofanilli M, Gonzales-Angulo A, Sneige N, Kau S-W, Broglio K, Theriault RL, et al. Invasive lobular carcinoma classic type: response to primary chemotherapy and survival outcomes. J Clin Oncol 2005; 23: 41-8. doi: 10.1200/JCO.2005.03.111 12. Mathieu MC, Rouzier R, Llombart-Cussac A, Sideris L, Koscielny S, Travagli JP , et al. The poor responsiveness of infiltrating lobular breast carcinomas to neoadjuvant chemotherapy can be explained by their biological profile. Eur J Cancer 2004; 40: 342-51. doi: 10.1016/j.ejca.2003.08.015 13. Du Toit RS, Locker AP, Ellis IQ, Elston CW, Nicholson RI, Robertson JF, et al. An evaluation of differences in prognosis, recurrence patterns and recep- tor status between invasive lobular and other invasive carcinomas of the breast. Eur J Surg Oncol 1991; 17: 251-7. PMID: 1646127 14. Rakha EA, EL-Sayed ME, Powe DG, Green AR, Habashy H, Grainge MJ, et al. Invasive lobular carcinoma of the breast: response to hormonal therapy and outcomes. Eur J Cancer 2008; 44: 73-83. doi: 10.1016/j.ejca.2007.10.009 15. Desmedt C, Zoppoli G, Gundem G, Pruneri G, Larsimont D, Fornili M, et al. Genomic characterization of primary invasive lobular breast cancer . J Clinical Oncol 2016; 34: 1872-81. doi: 10.1200/JCO.2015.64.0334 16. Zardavas D, Te Marvelde L, Milne RL, Fumagalli D, Fountzilas G, Kotoula V, et al. Tumor PIK3CA genotype and prognosis in early-stage breast cancer: a pooled analysis of individual patient data. J Clin Oncol 2018; 36: 981-90. doi: 10.1200/JCO.2017.74.830 17. Metzger Filho O, Giobbie-Hurder A, Mallon E, Gusterson B, Viale G, Winer EP, et al. Relative effectiveness of letrozole compared with tamoxifen for patients with lobular carcinoma in the BIG 1-98 trial. J Clin Oncol 2015; 33: 2772-9. doi: 10.1200/JCO.2015.60.8133 Radiol Oncol 2023; 57(2): 220-228. Ribnikar D et al. / PIK3CA activating mutations and outcomes in early-stage invasive lobular breast carcinoma 228 18. Luen SJ, Asher R, Lee CK, Savas P , Kammler R, Dell”Orto P , et al. Association of somatic driver alterations with prognosis in postmenopausal, hormone receptor-positive, HER2-negative early breast cancer: a secondary analysis of the BIG 1-98 randomized clinical trial. JAMA Oncol 2018; 4: 1335-43. doi: 10.1001/jamaoncol.2018.1778 19. Elston CW, Ellis IO. Pathological prognostic factors in breast cancer. I. The value of histopathological grade in breast cancer: experience from a large study with long-term follow up. Histopathology 1991; 19: 403-10. doi: 10.1111/j.1365-2559.1991.tb00229.x 20. Hammond ME, Hayes DF, Dowsett M, Allred DC, Hagerty KL, Badve S, et al. American Society of Clinical Oncology/College of American Pathologists guideline recommendations for immunohistochemical testing of estrogen and progesterone receptors in breast cancer. Arch Pathol Lab Med 2010; 134: e48-72. doi: 10.1200/JCO.2009.25.6529 21. Wolff AC, Hammond ME, Hicks DG, Dowsett M, McShane LM, Allison KH, et al. Recommendations for human epidermal growth factor receptor-2 testing in breast cancer: American Society of Clinical Oncology/College of American Pathologists clinical practice guideline update. J Clin Oncol 2013; 31: 3997-4013. doi: 10.1200/JCO.2013.50.9984 22. Dowsett M, O Nielsen T, A”Hern R, BartletT J, Coombes RC, Cuzick J, et al. Assessment of Ki-67 in breast cancer: recommendations from the International Ki-67 in Breast Cancer Working Group. J Natl Cancer Inst 2011; 103: 1656-64. doi: 10.1093/jnci/djr393 23. Salgado R, Denkert C, Demaria S, Sirtaine N, Klauschen F, Pruneri G, et al. The evaluation of tumor-infiltrating lymphocytes (TILs) in breast cancer: recommendations by an International TILs Working Group 2014. Ann Oncol 2015; 26: 259-71. doi: 10.1093/annonc/mdu450 24. Samuels Y, Wang Z, Bardeli A, Silliman N, Ptak J, Szabo S, et al. High fre- quency of mutation of the PIK3CA gene in human cancers. Science 2004; 304: 554. doi: 10.1126/science.1096502 25. Pramod N, Nigam A, Basree M, Mawalkar R, Mehra S, Shinde N, et al. Comprehensive review of molecular mechanisms and clinical features of invasive lobular cancer. Oncologist 2021; 26: e943-53. doi: 10.1002/ onco.13734 26. Johansson A, Yu NY, Iftimi A, Tobin NP, vant Veer L, Nordenskjold B, et al. Clinical and molecular characteristics of estrogen receptor-positive ultralow risk breast cancer tumors identified by the 70-gene signature. Int J Cancer 2022; 150: 2072-82. doi: 10.1002/ijc.33969 27. Griffith OL, Spies NC, Anurag M, Griffith M, Luo J, Tu D, et al. The prognostic effects of somatic mutations in ER-positive breast cancer. Nat Commun 2018; 9: 3476. doi: 10.1038/s41467-018-05914-x 28. McCart Reed AE, Kalinowski L, Simpson PT, Lakhani SR. Invasive lobular carcinoma of the breast: the increasing importance of this special subtype. Breast Cancer Res 2021; 23: 6. doi: 10.1186/s13058-020-01384-6 29. McCart Reed AE, Lal S, Kutasovic JR, Wockner L, Robertson A, de Luca XM, et al. LobSig is a multigene predictor of outcome in invasive lobular carcinoma. NPJ Breast Cancer 2019; 5: 18. doi: 10.1038/s41523-019-0113-y 30. Metzger FO, Giobbie-Hurder A, Mallon E, Gusterson B, Viale G, Winer EP , et al. Relative effectiveness of letrozole compared with tamoxifen for patients with lobular carcinoma in the BIG 1-98 trial. J Clin Oncol 2015; 33: 2772-79. doi: 10.1200/JCO.2015.60.813 31. Lopez-Knowles E, Segal CV, Gao Q, Garcia-Murillas I, Turner NC, Smith I, et al. Relationship of PIK3CA mutation and pathway activity with antiprolifera- tive response to aromatase inhibition. Breast Cancer Res 2014; 16: R68. doi: 10.1186/bcr3683 32. Loi S, Haibe-Kains B, Majjaj S, Lallemand F, Durbeck V, Larsimont D, et al. PIK3CA mutations associated with gene signature of low mTORC1 signaling and better outcomes in estrogen receptor-positive breast cancer. Proc Natl Acad Sci USA 2010; 107: 10208-13. doi: 10.1073/pnas.0907011107 33. Whyte DB, Holbeck SL. Correlation of PIK3CA mutations with gene expres- sion and drug sensitivity in NCI-60 cell lines. Biochem Biophys Res Commun 2006; 340: 469-75. doi: 10.1016/j.bbrc.2005.12.025